PDA

View Full Version : OT: Vacuum Tubes - Shared Information



randyc
11-28-2010, 10:22 PM
I wrote a book about the design of vacuum tube amplifiers during my absence from this forum (a medical "leave of absence", so to speak, LOL). The intention was to document the procedures - as best as I could infer them - used by the old-timers to design these products.

I've "heard" there may be a few members on this forum that are interested in this subject. If you click on the link provided below, the book can be easily downloaded (it may take a few minutes - it's about 350 pages). The service and the link is provided by "Adobe" and it's free for uploading AND downloading, apparently to promote the use of their products. That's a good thing.

Although the book is free please observe the included cautionary note about re-posting and/or selling any part of the material. It is NOT in the public domain so permission to download is restricted to your personal use, OK ?

https://acrobat.com/#d=4V0KMuxAaMet6mwFQOfhiA

I constructed a set of spread sheets that simplify some of the design procedures. Their use is described in the book. They can be downloaded from a different link. (Patience is required - one must sit through 30 seconds of advertising until the link at the bottom of the page becomes available.)

http://www.webfilehost.com/?mode=viewupload&id=215029

I've received complaints that this site limits downloads to a specific number. If an error message of that nature is displayed, post a request for the spreadsheets on this thread and I'll upload them again. (Or if you know of a better file service - that is FREE - please post that too.)

The readers that I had in mind have an understanding of basic electronics and the ability to use both calculator and TOOLS. I don't know why I hadn't thought of this before but a book like this might be more interesting HERE than on the jazz forums where I sometimes hang out. The guys there - intelligent as they might be - are NOT very adept with tools other than their musical instrument. (You will note that I have slanted the simple construction details toward those that are unaccustomed to metalworking.)

Oh yeah, one last point. Since the book has never been peer-reviewed or proof-read, PM's with technical suggestions/corrections are welcome. Nitpicks about chassis construction and the like are NOT welcome, LOL.

Hope this will be of interest to some of you !
Randy C

Willy
11-28-2010, 10:55 PM
Thanks Randy,
I have downloaded the book and although I haven't yet had any time to digest any of it, it does indeed look very comprehensive judging by the table of contents.

I'll be sending a link to this page out to my brother (an EE) who will have a greater understanding of the nuances of vacuum tube design than myself.
I also have a couple of people in mind that have and advanced interest in audio equipment who would greatly appreciate any info on vacuum tube audio equipment, let alone a 350 page book.

Good to see you back here from your "leave of absence" from this site.
From your contributions of late it doesn't look like your head took a sabbatical.;)

Evan
11-28-2010, 11:19 PM
A quick check shows that I can't save it for further use. Adobe is not providing the PDF but instead a flash version of it with no provision to save. If the file isn't Too large I can host on my web server for you. How many megs is it?

macona
11-28-2010, 11:20 PM
There some good tube info on this other forum I am on:

http://www.coultersmithing.com/forums/viewforum.php?f=10

randyc
11-28-2010, 11:30 PM
A quick check shows that I can't save it for further use. Adobe is not providing the PDF but instead a flash version of it with no provision to save. If the file isn't Too large I can host on my web server for you. How many megs is it?

Hmmm, it downloaded OK for me just now. The size is 11 MB and the spreadsheet is tiny, of course. (I did note that the website requested that I accept their terms - odd because I'd already done so.)

Earlier revisions of the book are available on many tube forums and they are substantially the same as this edition - perhaps lacking a few paragraphs or a chapter ...

edited to add: Willy, I'd be happy to hear your brother's comments. Sending my personal e-mail address.

Willy
11-28-2010, 11:50 PM
Not a problem here with the download, only took a few moments.
I don't remember accepting any terms of use either.

Randy my brother is out of the country for a couple of weeks on an assignment, but I'll see to it that he gets a copy.
I'll also persuade him to send me an evaluation of his thoughts on your work.

Evan
11-29-2010, 12:42 AM
Willy,

Is it downloading as a PDF? It doesn't for me.

dp
11-29-2010, 12:48 AM
Willy,

Is it downloading as a PDF? It doesn't for me.

I hit the button in the top left corner and got a true PDF file.

That's a pretty nice read. Took me on the way-back to my first job designing radio telephone systems. The only thing different is the use of zener diodes for regulating screen voltage vs VR tubes. I prefer the tubes :)

Evan
11-29-2010, 01:05 AM
I missed that. It downloading now.

randyc
11-29-2010, 01:10 AM
...The only thing different is the use of zener diodes for regulating screen voltage vs VR tubes. I prefer the tubes :)

Oh yes, many musicians prefer gas regulators because of their non-linear performance. I've not seen them used for screen regulation (except as an incidental by-product of plate regulation).

I'd be interested in a reference to a schematic using tubes for screen regulation. If you PM me I'll provide an e-mail address that accepts attachments.

Cheers,
Randy C

Evan
11-29-2010, 01:23 AM
Randy,

Do you want me to post it on my server?

randyc
11-29-2010, 01:29 AM
Randy,

Do you want me to post it on my server?

Sure - it's free to those with no commercial interest !

Cheers,
Randy

Evan
11-29-2010, 01:41 AM
OK. I will upload it later tonight when the satellite is running faster and then I will post a link here. It will stay up just as long as it doesn't become a bandwith problem. I have a very generous bandwith allowance but the main issue is server load. It is restricted to a maximum of 50 simultaneous connections.

dp
11-29-2010, 01:52 AM
Here's a link:
http://www.bentonelectronics.com/images/leslie147.gif

Note the cathode circuit in the final stage - interesting way to get feedback (and a good jolt while jacking the speakers).

The circuit discussion is here:
http://www.bentonelectronics.com/leslieamp147.html

Evan
11-29-2010, 08:12 AM
The book is located in the Electronics section at:

http://ixian.ca/gallery/index.php

Willy
11-29-2010, 09:54 AM
Evan, I haven't been on your site lately, I know I have to get out more. It looks good.
Lots of stuff there for everybody! How do you find the time?

randyc
11-29-2010, 11:55 AM
Here's a link:
http://www.bentonelectronics.com/images/leslie147.gif

Note the cathode circuit in the final stage - interesting way to get feedback (and a good jolt while jacking the speakers).

The circuit discussion is here:
http://www.bentonelectronics.com/leslieamp147.html

Just glanced at the circuit briefly but that appears to be a partly-regulated bias circuit, the screens are biased through a dropping resistor. In other words, their voltage varies with screen current.

NOT that it's a disadvantage, it depends on one's school of thought. From the aspect of reliability, screen dropping resistors guarantee that the screen voltage can never be greater than the plate voltage. From the viewpoint of those who espouse ultra-linear operation (negligible distortion) the screen should be independantly regulated. (It may be that the circuit in the link provides an optimal compromise that could satisfy both viewpoints.)

I've managed to accumulate eight vacuum tube amplifiers and a half-dozen solid-state amplifiers over forty years (I'm a jazz guitarist) and there are situations ideal for both types, as I see it. Some of the tube amplifiers (my Fender "Twin" is one) that have the reputation of being super-clean have neither regulated screen bias nor ultra-linear output transformers.

In general, a point that I made many times in the book, manufacturing costs trumped EVERYTHING with most producers of this equipment. By todays standards, most vacuum tube amplifiers are primitive in performance. Of course this statement puts one squarely into the midst of "psycho-accoustic" debate. However, like many others, I enjoy them and the monitor system in my studio is an old 35 watt Dynaco driving Carvin Studio Monitors.

It doesn't take much to start a fierce debate in the tube forums over a few of these trivial topics, LOL. Thanks for providing the schematic, it IS an interesting one. There are aspects that remind me of the final evolution of the "Music Man" hybrid guitar amplifiers.

Cheers,
Randy C

3jaw
11-29-2010, 12:14 PM
I'm having trouble downloading the spreadsheet. It won't let me click on the file name. Am I doing something wrong?

Thanks,
Greg

randyc
11-29-2010, 12:26 PM
I'm having trouble downloading the spreadsheet. It won't let me click on the file name. Am I doing something wrong?

Thanks,
Greg

I mentioned in the OP that there is a waiting period on that site, you must sit there and stare at their ads for 30 seconds (I think). Finally, at the bottom of the screen, the download link should appear.

If you've done all this and it still doesn't allow a download, PM me with an e-mail address that will accept attachments and I'll send the spreadsheets.

If Evan has included the spreadsheets on his site, that may be an easier alternative.

Paul Alciatore
11-29-2010, 12:35 PM
In cleaning out my office room so I can actually use it as an office, I came across a Heathkit stereo amp. - tube type. It is in a nice wood cabinet and as I recall it had a nice tonal quality. I used it for years. No idea if it still works. I may just set it up and use it if it does. Just a bit of related info.

Oh, I also have a tube tester. If anyone has any tubes that they would like to have tested, arrangements can be made. For a dozen or less, no charge: just pay postage both ways. It can provide comparitive readings so tubes can be matched.

Evan
11-29-2010, 12:51 PM
I forgot about the spreadsheets. They are now in the same album as the book.

dp
11-29-2010, 02:22 PM
Just glanced at the circuit briefly but that appears to be a partly-regulated bias circuit, the screens are biased through a dropping resistor. In other words, their voltage varies with screen current.

That is pretty much the case with zener diodes in clamping mode, too. The voltage at the screen can never be higher than the zener voltage, but can be less, depending on how much screen current there is through the series resistor. The need then is to ensure the series resistance is chosen to handle worst-case screen grid current (which is never really very high).

The VR tubes are also commonly used to provide negative voltage for control grids, of course. I always liked the glow in those old power supplies.

The Leslie circuit sent me off across the internet looking at Leslie effect speakers - a hobby interest I had in high school. There's still a lot of interest in that. I built an electronic Leslie in the 1970s but it never had that real wet sound of a good Leslie. I'd started working on a bucket brigade delay to make the phase relationship of the amplitude shift and the frequency shift variable, and to drive the mono signal in a voltage controlled divider to a stereo amp to simulate the Leslie horn displacement. That got sidelined by a pending divorce :mad:

randyc
11-29-2010, 02:58 PM
That is pretty much the case with zener diodes in clamping mode, too. The voltage at the screen can never be higher than the zener voltage, but can be less, depending on how much screen current there is through the series resistor. The need then is to ensure the series resistance is chosen to handle worst-case screen grid current (which is never really very high).

The VR tubes are also commonly used to provide negative voltage for control grids, of course. I always liked the glow in those old power supplies.

The Leslie circuit sent me off across the internet looking at Leslie effect speakers - a hobby interest I had in high school. There's still a lot of interest in that. I built an electronic Leslie in the 1970s but it never had that real wet sound of a good Leslie. I'd started working on a bucket brigade delay to make the phase relationship of the amplitude shift and the frequency shift variable, and to drive the mono signal in a voltage controlled divider to a stereo amp to simulate the Leslie horn displacement. That got sidelined by a pending divorce :mad:

You are exactly correct - configuring zeners as a constant voltage DROP rather than as a voltage REGULATOR prevents the plate signal voltage swing from dropping below the screen voltage. A true voltage source at the screen grid would satisfy many linear purists but wouldn't be good for tube life !

There is a trivial concern also associated with any form of regulation (gas tube or solid state) at that point, in that the screen voltage is modulated by the noise of the regulator circuit. At the signal levels present in the output of the chain this is of little importance - especially since vacuum tube amplifier noise is almost always dominated by power supply ripple rather than thermal noise.

But it CAN be significant in power stages like cathode-follower circuits, where the highest gain input terminal in the circuit is the screen grid ! I think that I mentioned in the book that if the external impedance at the screen grid is kept below about 2k (or about 1/10 the screen grid impedance), problems like this won't occur. (Many, many amplifier circuits don't bother bypassing the screen grid but I've always thought that A.C. coupling it to the cathode was a fine idea.)

I played with the bucket-brigade phase shifters too, back in the early eighties. (I don't think those ICs are even available any more.) Another interesting concept (my background is microwave/RF) is upconverting the baseband signal which would allow using a high-Q filter as a phase shifter or some of the many SAW phase shift networks (maybe even a crystal) and then downconverting back to baseband.

At one time this would have been bulky and expensive. Given the huge advances in miniature RF circuits, driven by modem and cellular telephone technologies, it would be a piece of cake now. Everthing needed can be ordered from Digi-Key for a couple of dollars.

In fact, the class "D" power amplifiers almost approach the upconversion concept except they are configured as a switching circuit (as are DSP-configured upconverters).

Thanks for your comments !
Randy C

squirrel
11-29-2010, 03:16 PM
WOW! I don't feel like a odd ball anymore. My winter project is a EMP proof receiver built from Nuvistors. I did find an original copy of Radio Reciever Design by Sturley on eBay, it was almost ready to fall apart but the content is well worth the cost we had in getting it rebound. Thank you very much for posting your book, if you need something CNC machined or modeled in Solidworks call my office so I can return the favor.

3jaw
11-29-2010, 03:58 PM
Got the spreadsheet. Thanks Evan!

randyc
11-29-2010, 04:00 PM
WOW! I don't feel like a odd ball anymore. My winter project is a EMP proof receiver built from Nuvistors. I did find an original copy of Radio Reciever Design by Sturley on eBay, it was almost ready to fall apart but the content is well worth the cost we had in getting it rebound. Thank you very much for posting your book, if you need something CNC machined or modeled in Solidworks call my office so I can return the favor.

And thank YOU for the incredibly generous offer, that is VERY kind ! I think that may be more of "us" than we imagine. I say this because about once a week, a message is forwarded from one of the two jazz guitar forums where the book was originally rolled-out.

The messages are always gracious and frequently contain details of a project that the sender is working on. Sometimes the authors encountered references to the book in odd places and took the time to track down the forums where the book originated. Almost none of these guys are guitarists but became members of the jazz guitar forums so that they could send a message and ask the occasional question !

Counteracting these flattering messages were some of the dung-throwers on several forums devoted exclusively to vaccuum tubes. I'm sure most of you can infer the type of responses these self-proclaimed experts wrote, ha-ha !

FWIW, during my convalescence, there was nothing to do but watch daytime television, sigh. My wife bought this little laptop for me to use when sitting up in bed. It was her idea that I consolidate the lengthy posts that I made on various forums into a single document.

The disjointed style of the thing is a direct result of cutting/pasting many forum posts that were originally written over a period of several years. Many were left out of the book (for example a very long, very boring, very nerdy dissertation on loudspeaker mechanical design and performance). I haven't the patience to be an editor so the book is what it is, LOL.

Thanks again, both for your comments and for your offer - you're a generous man !
Randy C

gary350
11-29-2010, 04:06 PM
I love Tube Amps I have built several amps from scratch and I built several from kits. I have also built transmitters and receivers and a few other things. Tubes are fun.

I am building a home movie theater system with a 12 ft wide screen. I intend to use new technology, no tubes. I want the best quality sound system I can get.

http://en.wikipedia.org/wiki/Loudspeaker_enclosure

http://ibsubwoofers.proboards.com/index.cgi

gwilson
11-29-2010, 04:06 PM
I hope they aren't primitive. I bought myself a Standel reissue last year. Great,clear tone. they only made 50 originals,but anyone who was anyone played one in the 1950's.

Evan
11-29-2010, 04:21 PM
I took three years of vocational electronics in High school and one of the first real projects we built was a five tube superhet radio. Then we got to put bugs in each other's radios like a loose solder connection or swap in a bad tube and then troubleshoot the problems with a concise writeup.

One of the first Jobs I had that was in part responsible for me dropping out of UC Berkeley was working for Pacific Stereo and Hi Fi in Berkeley. At the time Marantz and Macintosh were the big names in Hi Fi audio and it was my job to fix them. Mac was all tubes of course but Marantz had just started making transistor amps. The reason I dropped out was because I ended up making about $800 per month after just three months on the job. Back in the sixties that was a very handsome wage. Troubleshooting has always been my strong point.

I just am now going in to town to pick up an order from DX. Included in that order is a tiny circuit board slightly larger than a postage stamp. On that board is a power supply filter capacitor and a chip with connectors. That chip is a class D audio amp that puts out 15 rms watts per channel with no heat sinking and no external components. The output is flat from near DC to 40khz and the distortion is nearly unmeasurable. It cost under 20 bucks.

Things have changed.

RTPBurnsville
11-29-2010, 06:47 PM
Randy,

Very nice job..... Thanks!

Lew Hartswick
11-29-2010, 06:58 PM
That schematic: What no type 80 rectifier tube??? :-)
What's with that "solid state" bridge? :-)
...lew...

Evan
11-29-2010, 07:59 PM
The 80s burn out pretty quick. I haven't checked but I suspect they are hard to find.

dp
11-29-2010, 08:59 PM
Real tubers use selenium rectifiers.

squirrel
11-29-2010, 09:06 PM
Real tubers use selenium rectifiers.
Do you where any can be found?

dp
11-29-2010, 09:10 PM
These people advertise some:

http://www.lphoying.com/selenium.html

randyc
11-29-2010, 09:20 PM
Real tubers use selenium rectifiers.

LOL, there is a LOT of truth in that statement ! But blues guitarists have always loved the "sag" caused by tube rectifiers. That preference transferred to mainstream musicians eventually (our ears seem to become "conditioned" to certain sounds).

FWIW, looks like more vacuum tube guys might hang out HERE than on the two jazz forums where I posted the original download link about a year ago. Most guitarists profess to love tube amplifiers but there has been more response here in 24 hours than I received during the first three days after posting on the music forums.

(That probably is because most musicians aren't tool-oriented.)

Cheers,
Randy C

dp
11-30-2010, 02:03 AM
LOL, there is a LOT of truth in that statement ! But blues guitarists have always loved the "sag" caused by tube rectifiers. That preference transferred to mainstream musicians eventually (our ears seem to become "conditioned" to certain sounds).

I have to agree. I love playing blues but my amp is a 100 watt Fender transistor rig and it has no soul. In the 60's I had a 25 watt all-tube amp - damn - drawing a blank on the maker (getting old's a bitch). Anyway, it had that sag, even for a low-power amp (not all that low for the time).



FWIW, looks like more vacuum tube guys might hang out HERE than on the two jazz forums where I posted the original download link about a year ago.

Yep - I'm always amazed at what these people do when not making chips or arguing here about not-making-chip stuff :)

I started out in the 60's as a EE working in SoCal on SSB radio-telephone design using tubes and/or TRW transistors, wireless mic's and receivers, nearly got into microwave stuff but didn't care for working with tiny machined parts, so went into field engineering with Raytheon on ship-board radar systems (I liked to travel and saw the world), then sole proprietorship in marine electronics working primarily in Alaska. After a couple winters working on radar masts in white-out conditions with a spare set of gloves in the ship's clothes dryer between trips below to warm up, I came home for good and got into computers and that's what I still do.

So here I am in the Seattle area about to get dumped on by a snow/wind storm. :D

RTPBurnsville
11-30-2010, 07:47 AM
FWIW, looks like more vacuum tube guys might hang out HERE than on the two jazz forums where I posted the original download link about a year ago. Most guitarists profess to love tube amplifiers but there has been more response here in 24 hours than I received during the first three days after posting on the music forums.

(That probably is because most musicians aren't tool-oriented.)

Cheers,
Randy C


Could be because lots of engineering types are also musicians. I for one have played pedal steel guitar from age 14 and have several Music Man and Fender tube amps. Music, electronics, metalwork and even some woodworking all sorta blend together.

Robert

topct
11-30-2010, 08:15 AM
I am a tube fan. I am waiting for my upgrade in connection speed to start next week so that I can download your book.

I have a Super reverb that I play through but it's a bit much for harmonica. I also have an old Princeton that has been around the block a million times. I have recapped them both once entirely and now just to keep them working.

One of my favorite amps was another Princeton that I gutted and using the original chassis with some extra holes punched in it, built a Bassman circuit in it. At the time the local Fender dealer was actually scrapping amps and would sell the transformers for cheap.

To me Leo Fender was God. His approach to building amps was so easy to copy that even someone like myself with no training could build one of his amps. You could just remove the back cover and everything was right there. Most repairs could be made without having to remove the chassis from the case. Fish paper and brass eyelets can make for a very tidy circuit board. I recently had to replace a pot on a newer solid state something or other amp. It took me an hour just to wade through the layers and layers of boards just to get to the thing.

I have also tried the rectifier tube replacement thing and it really does change the sound. I don't like it. Just to darn hard sounding

Evan
11-30-2010, 09:27 AM
Could be because lots of engineering types are also musicians.

:D

http://www.youtube.com/v/tmn4o8Vs-H8?fs=1&hl=en_US

dp
11-30-2010, 03:43 PM
I played with the bucket-brigade phase shifters too, back in the early eighties. (I don't think those ICs are even available any more.)

So this lit off another trip across the internet as those were fun devices to play with. And yep - nobody makes them any longer. There are plenty on Ebay (SAD1024 BBD and PANASONIC MN3007 BBD). Prices are enough to choke on!

The current method is to use Arduino technology and A/D to RAM, let a few milliseconds go by, the D/A out of RAM to get the delay. It also solves the problem of switching noise the BBD chips had. It's not cheap and it's more complex but it does a good enough job.

In the mid 1970's I got a reverb cab from my neighbor. It was about 4' tall, 30" wide, probably, had a tube amp with 6L6 finals, and multiple wet spring reverb columns in it. As I recall it also had 11 speakers in the box. Can't recall if it was a Hammond or Wurlitzer, but it had quite a sound.

Nice thread with lots of memories!

randyc
11-30-2010, 04:30 PM
Digital signal processing has probably come further, faster than any branch of electronics in my lifetime. Most of my career was spent in defense and aerospace and one fascinating job was for an ELINT company (electronic intelligence). Their claim to fame was state-of-the-art TDOA direction finding equipment (mostly on aircraft).

When DSP was introduced to conventional analog technology, it opened up a whole new area of electronic surveillance and countermeasures. For example, a common technique at the time was to digitally record Soviet radars, storing their pulse/chirp characteristics. This had several useful purposes.

One was the fact that EVERY radar has a distinct electronic signature and a digital lookup table onboard the aircraft could quickly tell the operator that he was just out of detection range of a "Chernov" class guided-missile-cruiser (using an imaginary example).

Even more useful was the ability to "parrot" back the pulse and chirp, using DSP modulators ... by using a little sleight-of-hand in the timing, it was completely practical to fool the transmitting radar into thinking the digitally reproduced signal was his own echo. This made his detection and processing circuits place the "echo" hundreds of miles away from the actual surveillance aircraft.

At the current time, microwave and RF engineers needn't worry about digital converters and complete digital radios replacing hard-won design skills. The processing speeds aren't anywhere near what they need to be. Analog designs, like the ones with which I was associated at Space Systems were easily reaching above 40 GHz using GaAs and even silicon MMICs. Newer materials have pushed the performance edge even further beyond those milestones !

One of the main reasons for my interest in microwave design is that the "circuits", more often than not, are machined parts rather than circuit boards. My consulting business was primarily oriented toward producing DESIGNS - not hardware. But often, in order to satisfy myself or a customer, it was necessary to build a prototype "circuit". My choice of machinery (8 x 18 lathe and 8 x 30 vertical mill) were specifically selected to accomodate these small projects.

Here's an example of a synthesized receiver module that I designed for a commercial digital radio company (point to point microwave telephone links):

http://i97.photobucket.com/albums/l219/randy9944/P1010594-1.jpg

(This particular device is low enough in frequency that conventional microstrip circuits could be used - the only things that I was required to machine were the shielded housing, the covers and several heat exchangers on the far side of the assembly.)

The slightly larger Leitz horizontal mill satisfies the need for fast, accurate metal removal and large boring chores. Not to mention the fact that I just LIKE it, LOL !

But I digress - nothing could be further from the point of this thread (vacuum tubes) than this type of technology. Vacuum tubes are so much fun probably because they ARE so primitive by today's standards. Nevertheless, I own a lot of equipment (sensitive receivers and low phase-noise signal generators) that are completely vacuum-tube based. Some can give a solid-state generator like my HP8640B a little competition when allowed to stabilize for a day or so, LOL. (Well, not really, but they ARE pretty fine examples of technology for their time.)

Thanks for your comments,
Randy C

rohart
11-30-2010, 05:17 PM
Your book looks great, randy - for when I'm not commissioning new/old bits of mechanical kit. My old Vox is still in operation, so I'll read your book in conjunction with the schematic for that.

On a trivial note, does any one happen to know the earliest Adobe Reader that will load the downloaded PDF ? I have a thing against unnecessary updates, and I'm a bit out of date with this.

randyc
11-30-2010, 06:04 PM
Your book looks great, randy - for when I'm not commissioning new/old bits of mechanical kit. My old Vox is still in operation, so I'll read your book in conjunction with the schematic for that...

Vox are fine amplifiers, particular the AC-30 which is highly regarded for jazz. Unfortunately, I've never owned one but I do have a 1961 Ampeg Reverberocket - same size and weight and which produces similar tonal qualities at about the same power level.

Thanks for your interest,
Randy C

gary350
11-30-2010, 06:16 PM
Real tubers use selenium rectifiers.
__________________
Dennis


Real tubes?????????? I have never seen a fake tube.

Use a 5U4 rectifier tube.

dp
11-30-2010, 06:32 PM
Real tubes?????????? I have never seen a fake tube.

Use a 5U4 rectifier tube.


It sez "tubers" as in manly tube users :)

Those 5U4's were some beefy rectifiers. They would survive red hot plates where a 5Y3 would gas the getter white.

Evan
11-30-2010, 06:33 PM
At the current time, microwave and RF engineers needn't worry about digital converters and complete digital radios replacing hard-won design skills. The processing speeds aren't anywhere near what they need to be.

I have heard of 200 ghz DSP chips and that was a few years ago.

As for complete digital radios, I am holding in my hand right now a complete FM transmitter on a single chip. It requires no external parts except a power supply and is stereo broadcast quality complete with an RS232 interface for frequency control from 88 to 108 mhz. It's under 20 dollars. It will even do frequency hopping.

CCWKen
11-30-2010, 06:55 PM
:( Boo Hoo. All I get on the original link is Acrobat's home page wanting me to download software. I did notice that "DoubleClick" was trying to track activity so that's probably the problem.

randyc
11-30-2010, 07:32 PM
I have heard of 200 ghz DSP chips and that was a few years ago.

As for complete digital radios, I am holding in my hand right now a complete FM transmitter on a single chip. It requires no external parts except a power supply and is stereo broadcast quality complete with an RS232 interface for frequency control from 88 to 108 mhz. It's under 20 dollars. It will even do frequency hopping.

The definitions used to describe frequency capability don't imply that the devices can actually be operated at that frequency. The classic definition for frequency capability of a transistor, for example, is Ft, cutoff frequency or the frequency at which the gain is 0 dB (transistor can perform no useful function). These terms are useful mainly because one must have some kind of yardstick as a basis of comparison between various devices. To my knowledge there are zero PRACTICAL analog devices that can operate at 200 GHz except perhaps "bulk-effect" devices, which are single-terminal negative resistance semiconductors. They are generally used as high frequency oscillators.

As we know, all digital circuits are composed of analog building blocks so it follows that it's not possible to construct a digital circuit that is capable of operating at a higher frequency than the individual components (except in the esoteric cases of frequency multiplication or up-conversion, which rely not on gain but on non-linearity for proper function).

The transmitter that you describe is not the "radio" to which I alluded. The part shown in my hand (the equivalent of thousands of billions of individual transistors - what look like individual transistors on the circuit are actually MMICs) in the photo is simply one plug-in of six to ten in a microwave digital radio, not to mention the high-frequency head on the antenna (anywhere from 13 to 40 GHz upconverters/downconverters and LNAs). The cable feed only carries the I.F. and power (I.F. is usually around 4 to 6 GHz for commercial digital radios).

Definitions and descriptions are sometimes unclear. For example what MOST people call a radio is an AM or FM receiver. The little FM transmitter that you describe has been around for decades in various configurations - I have two designs in one of my lab books from the early nineties that use Motorola single-chip FM transmitters as frequency multipliers. Very useful IC !

To the military, aerospace and telecommunications people, the term "radio" means what most would call a "transceiver", I suppose, but with all of the DSP circuits, telecommunications link to the hub, peripheral support circuits and the microprocessor that makes everything work. They are full duplex - simultaneous transmit/receive on as many as 300 channels (for "little" radios) - and remain in full-duplex digital communication with the command hub for fault-reporting, channel switching, automatic level control, transmit power adjustment, switchover to "hot" standbys and so forth. The radios have their own mass storage media and can be re-programmed on the fly from the systems hub. In the U.S. (where E911 requirements exist), they may also include TDOA direction finding capability for the purpose of locating emergency calls from a cellular telephone.

The modern digital radio is as different from the popular conception of a radio as a space shuttle is to a Piper "Cub". The technology and current architecture is less than two decades old. In fact even some of the modulation techniques were unknown two decades ago and the RF and microwave circuits were still dreams of IC and MMIC designers.

Cheers,
Randy C

randyc
11-30-2010, 07:36 PM
:( Boo Hoo. All I get on the original link is Acrobat's home page wanting me to download software. I did notice that "DoubleClick" was trying to track activity so that's probably the problem.

Try Evan's site -

Forestgnome
11-30-2010, 07:46 PM
Nice document. My guitar-playing, tube amp fanatic friend loved it! Hey Randy, was wondering what elint systems you worked on (if not secret). I used to work with Gaurdrail/Common Sensor , Trailblazer, and Quicklook/Quickfix. All pretty much ESL/TRW stuff. Fairly dated now. Interesting thing is, the company I work for now is just getting into building some DF equipment and FFT analyzers for elint.

randyc
11-30-2010, 08:09 PM
Nice document. My guitar-playing, tube amp fanatic friend loved it! Hey Randy, was wondering what elint systems you worked on (if not secret). I used to work with Gaurdrail/Common Sensor , Trailblazer, and Quicklook/Quickfix. All pretty much ESL/TRW stuff. Fairly dated now. Interesting thing is, the company I work for now is just getting into building some DF equipment and FFT analyzers for elint.

Thanks for your interest. My ELINT career was mainly as an employee of "Condor Systems" in San Jose (they've since moved to Morgan Hill). I worked as one of several RF designers on two Army programs - both were portable detection/location systems but the program name now eludes me (age and chemo-brain, as my wife calls it).

The most fun program, which IS classified, was a complete digital surveillance receiver for the RA-2G aircraft (what used to be called the U-2). My participation was also limited to the RF parts of the receiver; a very bright young engineer did the DSP design and programming. The program required trips to Texas (the prime was E-systems, Garland, and the airplane was parked nearby) and involved a lot of beer-drinking - oops, I mean late-night meetings - with the E-systems engineers and the Air Force guys. Tough work but someone had to do it ! (The job also required periodic reviews in Washington with the NSA guys which was NOT fun !)

The work was interesting but greener pastures called so I moved to "Space Systems Loral" in Palo Alto as senior staff engineer/project manager for various satellite programs. I was just starting to hit my stride with the space hardware when the doc suggested I retire (I was 55 at the time and had developed high blood pressure). So I left the bay area, moved up here and started a consulting business which was great fun until further health problems caused me to shut it down a few years ago. One of the great things about the consulting business was the ability to write off part of the metalworking machinery cost as capital equipment, against my self-employment taxes.

I did a lot of proposal work on Guardrail/Common Sensor when I was with Zeta Laboratories in the early nineties but none of it ever materialized as real hardware orders. It's a fascinating field and I'm sure you'll enjoy it. Based on the programs with which you've been associated, I'm inferring that you're an experienced guy and been around a while, LOL.

Cheers and thank your friend for me !
Randy C

sprkt01
11-30-2010, 08:56 PM
Book looks good. Always enjoy some tube reading. Thanks

Evan
12-01-2010, 01:49 AM
The definitions used to describe frequency capability don't imply that the devices can actually be operated at that frequency. The classic definition for frequency capability of a transistor, for example, is Ft, cutoff frequency or the frequency at which the gain is 0 dB (transistor can perform no useful function).

It's already here.




Speaker:
Dr. Sorin P. Voinigescu
Electrical and Computer Engineering, University of Toronto

Abstract:
The most advanced SiGe BiCMOS and 32-nm CMOS technologies with f_T and f_MAX simultaneously exceeding 300 GHz have opened up the 100-200 GHz spectrum for massive integration of transceivers, antennas, and digital signal processors into Systems on Chip (SoCs) that would have been unthinkable even 5 years ago. What are the applications that we can address with such systems? What are the design techniques and circuit topologies that can be employed at 160 GHz in silicon? Can we test such systems in production at less than 10 cents per part?


http://ece.uwaterloo.ca/Seminars/Invited/2010/voinigescu_apr_30.html

randyc
12-01-2010, 02:53 AM
It's already here.
http://ece.uwaterloo.ca/Seminars/Invited/2010/voinigescu_apr_30.html

I can't think of ANY practical applications for that technology and I've spent my career in that "neighborhood". Consider the tolerances required to produce tight-beam antennas with GAIN at those wavelengths (1/64 inch for quarter wavelengths !!!!) Power levels available above 50 GHz or so can be measured in microwatts or MAYBE a few milliwatts, LOL !

"Lab results" and similar are often reported. They aren't useful except as a stimulus for funding. (I interpret that summary as a request for research funds by holding out the carrot of low-cost MMICs - NOT going to happen.) Waterloo has a couple of distinguished academics (I have several excellent papers on thermodynamics from Waterloo) but NOT in the field under discussion.

I'd be MIGHTILY impressed to see a REAL 100 GHz radio - the frequencies claimed won't happen commercially in OUR lifetime, regardless of that summary. Anyway, I'm not getting your point - why would you spend time looking up that obscure reference - has this become an "argument"?

If I've pushed a button, it was unintentional. If my resume has offended you or anyone else, then let's forget it and simply accept that I'm another home machinist that wrote a free book about vacuum tubes, OK ?

Evan
12-01-2010, 05:41 AM
Not an argument at all. You mentioned that there wouldn't be any likelyhood of systems on a chip replacing "hard won skills".

The freqencies claimed are already being exploited.


Consider the tolerances required to produce tight-beam antennas with GAIN at those wavelengths (1/64 inch for quarter wavelengths !!!!) Power levels available above 50 GHz or so can be measured in microwatts or MAYBE a few milliwatts, LOL !


That simply isn't a problem. Planar arrays of elements are perhaps the easiest part of the problem to solve when you are working with feature sizes measured in nanometres. Many companies have been announcing 200 ghz parts and higher. Incidentally, the researcher at Waterloo has produced a sample and hold IC with excellent performance at 30 ghz. NEC is making parts that run at 200 ghz.


"We have achieved excellent RF performance suitable for next-generation wireless communications technologies such as long-term evolution (LTE) and WiMAX specification," said Hayashi.

NEC researchers have been able to boost fmax to 200 GHz while decreasing gate resistance to 17.1 ohms. This compares to last year's results of a preceeding tweak to the CMOS porcess that had a an fmax of 170 GHz and gate resistance of 22.7 ohms.



http://www.eetimes.com/electronics-news/4086298/IEDM-NEC-tweaks-CMOS-device-process-to-200-GHz

I have very wide ranging interests and I try to keep up with what is going on in a variety of fields. DSP is one of those as it has direct applications in astronomy which is one of my primary interests. These sorts of parts are also required to enable large bandwidth increases in optical fibre systems, much of which can be applied to in place single mode fibre optics. Limitations on bandwidth are in part caused by the dispersion of ultrashort pulses of laser light that become "smeared" out in time. Current systems are able to produce pulses with durations measured in femtoseconds. The NRC Canada is now exploring attosecond systems and 200 ghz parts will enable the production of comm systems with orders of magnitude higher bandwidth. The limits cause by dispersion can be dealt with by reshaping the pulses at intermediate DSP amplifiers in the optical systems.

Incidentally, one technology that I haven't looked into recently was the development of "tubes on a chip". I recall that there were efforts to build actual free electron switching devices using micromachining to produce micron scale vacuum devices that operate on the same principles as tubes.

dp
12-01-2010, 10:41 AM
The notion was floated some time back to use such microwave transmitters to communicate across microchips where densities don't allow bussing data between layers. I don't know how far things have gotten with that but as an idea it's pretty interesting.

randyc
12-01-2010, 11:29 AM
Oh boy, if the readership hasn't already been lost, THIS should do it, LOL ! "Exploited" is a tricky word since it doesn't necessarily imply that useful work is being performed. These things are mostly lab curiosities that may someday have an application but are not practical in the sense of my background - I would define "practical" as the ability to incorporate the devices in working systems with a real purpose and a real price (not the quotation from a researcher who may have no concept of manufacturing costs).

Planar arrays address only a very small part of housekeeping at those frequencies. Arrays are necessarily lossy and when power levels are so incredibly small, one struggles for every fraction of a dB - not just to make use of the power but to maintain reasonable noise figures. Additionally, the arrays must be phased properly and when wavelength is measured in tiny fractions of an inch and the arrays must be phase-matched within 5 degrees, the implication of the transmission lines that connect them becomes of major importance. Which leads to the problem of interconnections - circuits on a single substrate (MMICs) are no problem but they must be cascaded to form practical systems. There are only three, maybe four ways of implementing interconnections at system level - the lowest loss being waveguides. Considering the wavelengths involved, the manufacture of low-loss waveguide is a challenge in itself.

(You note that you are interested in astronomy so the serious consequences of loss as related to noise figure, are familiar. Planar arrays are NOT a good sensor element for radio astronomy at millimeter wave frequencies; in my opinion their usefulness is confined to high frequency radar where noise figure is not so serious. I think the various forms of dish antennas are far better - less loss and higher gain. But then we're right back to dimensional tolerances and perhaps more importantly: the antenna feed which returns us to waveguide manufacturing issues.)

This subject is a LONG way from vacuum tubes and in fact would probably not interest most RF engineers, actually. I found that many of my colleagues were disinterested simply because of size (the tiny wavelengths) and the inability to manipulate ("tune" if you will) the circuits for best performance. Even at Space Systems, which had the best computing systems for circuit analysis that money could buy, our 40 GHz work was frequently trial and error. At that time only governments, universities and a few companies like Space Systems and TRW (between them they controlled the commercial satellite market) could afford to dabble in high frequency research.

Computing tools are now better, particularly those that include field analysis, and foundries are much better than when I worked daily on these problems. Nevertheless, most of the reporting is, as you quoted above "next generation" ... we simply can't handle the technology that we have now, for "housekeeping" reasons like the interconnect problem. Even more serious is that the reported data is supported by what measurements ? There are no measurement systems that operate at those frequencies so the labs rely on their home-built instrumentation which has no national traceability.

I once watched a technician testing a Gunn oscillator at a company in the midwest part of the U.S. The specification was that the oscillator must produce 1 milliwatt of power at 35 GHz (I was the customer - product manager at Digital Microwave for 23 and 40 GHz radios). I asked him how he intended to measure the output power and his response went something like this: "Well, we have to downconvert the oscillator so that the frequency can be measured and we "know" that it takes one milliwatt of power to get a reading through the downconverter to the frequency counter".

My next and very obvious question was what drives the downconverter so that you can make the frequency measurement. He responded that another home-built Gunn oscillator was used as the L.O. Now anyone the least bit familiar with these things will see several major problems with this jury-rigged setup. Frequency measurement is ludicrous since it was not based on a stable L.O. but instead based on the same type of device that was to be measured.

As far as output power determination, that was simply a laugh so I asked him to summon his product engineer so that we could resolve the issue. It didn't get resolved, they simply had no capability to make either measurement so I ended up bringing the oscillator back to San Jose. Of course WE didn't have that measurement capability either so I simply installed the device in the 40 GHz radio to see if the radio worked with it.

This occurred about fifteen years ago and the instrument situation at 40 GHz is now stable and traceable. But higher frequency measurements are still a problem - the measurement of noise figures is a good example. There are three main areas of uncertainty in the measurement and just one - the interconnect VSWRs - can alter the measured data by as much as 2 dB at these frequncies. So when I see that so-and-so University "reports a 4 d B noise figure at 95 GHz" I have to take that with two pounds of salt.

This long-winded story (and I could go on for hours because I love this stuff - it was my career for almost forty years) is an illustration of just a few problems that are encountered every day by people who make their career in this field - with far more practical experience than those who perform funded research at a university. As they say, those who CAN perform, "DO" and the rest "teach". OK, I can hear the snores so I'll leave the topic alone simply noting that staying current with the literature has no real relationship to practical application.

Oh yes, apologies to all readers (those who remain) for the lengthy diversion. I rarely get to discuss my work so I tend to get caught up in the current, so to speak.

Cheers,
Randy C

dp
12-01-2010, 11:43 AM
Oh yes, apologies to all readers (those who remain) for the lengthy diversion. I rarely get to discuss my work so I tend to get caught up in the current, so to speak.

Cheers,
Randy C

There are on-topic, topic creep, topic expansion, off-topic, and topic hijacking variants of thread trends. The first three tend to remain interesting to a good many readers particularly when the OP contributes, and the original topic has been well covered.

There are also abrupt interjections such as this one that address a sub-topic of a specific post and which tend to lead to topic hijacking, so I won't comment further than to say I've enjoyed this thread more than most !!, and never fear topic expansion :)

randyc
12-01-2010, 12:36 PM
There are on-topic, topic creep, topic expansion, off-topic, and topic hijacking variants of thread trends. The first three tend to remain interesting to a good many readers particularly when the OP contributes, and the original topic has been well covered.

There are also abrupt interjections such as this one that address a sub-topic of a specific post and which tend to lead to topic hijacking, so I won't comment further than to say I've enjoyed this thread more than most !!, and never fear topic expansion :)

Great !

But I fear that you are one of the few that did enjoy it, your background being similar to mine. It is VERY difficult, as you know, to make a point in an esoteric field without having to use esoteric terms - and unhappily I think that's where most readers throw up their hands and find something more entertaining, LOL.

Thanks for your comments, they are welcome. As noted I tend to babble when I talk about these things. Other than the occasional chat with an old friend in the San Francisco Bay Area, I don't know of anyone in my current location that has the slightest interest in this material - except perhaps the amateur radio operators ! (One lives down the road and visited me once because he wanted to see a "real" microwave spectrum analyzer.)

http://i97.photobucket.com/albums/l219/randy9944/P1010970.jpg

This is one of the last analog microwave analyzers before H-P (and then Agilent) introduced the DSP-based analyzers. It's slow, of course, and the resolution bandwidth is limited to 100 Hz but for microwave use, it's great and I'm lucky to have obtained it at a good price (eBay, of course).

BTW, the little plug-in at the bottom of the photo is a home-built downconverter to extend the range of the instrument below 10 MHz. The frequency range is now 1 kHz to 23 GHz :)

Cheers,
Randy C

Evan
12-01-2010, 12:58 PM
This long-winded story (and I could go on for hours because I love this stuff - it was my career for almost forty years)

Well, I'm not bored.
:D

My first "exposure" to microwaves was at the Lawrence National Lab. My father was using high power klystrons to pump the primitive coaxial fusion machine with which they were exploring the instability modes of superheated plasmas.

I am quite familiar with Gunn oscillators too which use a negative resistance Gunn Diode as the active element. Despite the various problems including dimensional tolerances a little bit of power goes a very long way. In the millimetre wavelengths the problems resemble optical design rather than "radio". Once you begin thinking in terms of optics the notion of dimensional tolerances takes a very large shift to a much smaller realm.

I am accustomed to working to very tight tolerances. This is an example of an optical grating ruled with diamond tooling on my shop built CNC milling machine. The grating is ruled at thirty six thousand eight hundred lines per inch.

http://ixian.ca/pics8/grating.jpg

Back on the subject of real vacuum tubes, I have a small supply of shielded peanut tubes from old aircraft radios. I don't know what they are as I haven't removed them from the equipment. If anybody is interested in a few I will have a look and see if they have any designations on them. They are the type with no sockets, just solder in leads.

Seastar
12-01-2010, 02:36 PM
Since I graduated from Purdue with my BSEE the same year as the guys at Bell Labs invented the transistor I spent about half of my professional life in a "mixed" design environment. Lots of tubes at first and mostly semiconductors in the second half.
I worked for the Navy designing radar countermeasure equipment for naval aircraft. Traveling wave tubes and the like at the gigahertz freguencies and "peanut" tubes for signal processing.
I then worked for Hazeltine Corp and was the lead engineer on the first civilian air traffic control transponder.
I founded a company that produced thousands of panel mount ATC transponders for general aviation as well as all kinds of nav and comm gear for aircraft.
We were still using tubes at 1 GHz at that time.

I also founded this company 35 years ago:
www.ritron.com

We have products in space, backhauling data on the Mexican border, monitoring geologic faults on the pacific rim and on most railroad trains in North America to name just a few markets.

As you can see the company is still in business and doing well. I designed a few of the products on the website.
I partially retired 10 years ago at age 68 and my son now runs the company.

I have several tube audio amps in my collection and some tube test equipment.
One of my hobbies is restoring tube communications receivers/transceivers.
My latest is a Hallicrafters S-85.

I relate this uninteresting history so that you will know that I am sincere when I say this is one of the most interesting posts I have ever read on this forum.
Thank you!
You have made my day!
Bill

DICKEYBIRD
12-01-2010, 03:00 PM
OK, this thread, without a shadow of a doubt confirms my suspicions. You guys that understand and write about this stuff are all from another planet or were taught by THE VISITORS at Area 51.

Go ahead, admit it; you can't fool me!:D

ps: Even though I can't understand it, I do enjoy reading about it.:)

randyc
12-01-2010, 03:37 PM
Since I graduated from Purdue with my BSEE the same year as the guys at Bell Labs invented the transistor I spent about half of my professional life in a "mixed" design environment. Lots of tubes at first and mostly semiconductors in the second half......

I relate this uninteresting history so that you will know that I am sincere when I say this is one of the most interesting posts I have ever read on this forum.
Thank you!
You have made my day!
Bill

Hi Bill,

You are clearly the most senior and experienced reader of the post - if you enjoyed the experience then the time expended has been worthwhile ! I am only 12 years younger than you but not a single tube class was listed in the curriculum by the time my E.E. education commenced. (I did take a brief detour though, by first obtaining my BSME before changing careers, so I was out of step with the rest of my class by some six years or so.)

If you ever get around to reading my modest vacuum tube document, your experience would make you the ideal peer-reviewer and I would value any comments you'd care to make. If interested, I can PM my private e-mail address for that purpose.

Sincerely,
Randy C

edited to add: that grating is nice work, Evan

Forestgnome
12-01-2010, 04:48 PM
Randy, nice antique! Can't say I worked on that one. The older equipment I've worked on are mostly microwave sources, but go back to the 8640B's. Actually I was trained in the Navy on 50's vintage high power HF gear. Now I work on PSA series analyzers and PSG series sources (I work at Agilent). If you're curious about the state of the art in high freq measurement, check out our new PNA-X's to 1.05thz. They're pretty amazing boxes.

randyc
12-01-2010, 05:36 PM
Randy, nice antique! Can't say I worked on that one. The older equipment I've worked on are mostly microwave sources, but go back to the 8640B's. Actually I was trained in the Navy on 50's vintage high power HF gear. Now I work on PSA series analyzers and PSG series sources (I work at Agilent). If you're curious about the state of the art in high freq measurement, check out our new PNA-X's to 1.05thz. They're pretty amazing boxes.

You guys (H-P, then Agilent) have ALWAYS been the best ! Every ATP (acceptance test procedure) I ever wrote universally specified your gear. FWIW, I bought one of the first automated phase-noise systems from Santa Rosa Division when I was at Zeta Labs. This preceded the nice little single chassis unit, it was based on the 11729B carrier test set and the 8662 ultra-low-noise synthesizer - sounds simple enough but the thing was in two six foot high panel-rack cabinets and had a 3585A baseband analyzer, a 8577B microwave analyzer, a microwave synthesizer (can't recall model number) and the entire system was controlled with a 9816 computer.

That ran around a quarter of a million dollars at the time - considerably more than the tiny noise test set that was introduced a few years later with BETTER performance. Since I also had a to buy a nice screen room for the phase noise test set, I went ahead and bought the 8970T microwave noise figure test set, too ! We had - outside of Hewlett Packard - probably the best noise measurement system in northern California at that time ! People would come to use our calibration lab to calibrate their noise tubes and solid-state noise diodes, LOL !

Since you started out young in the Navy, I'm sure you've aligned one or two URM-25 generators - my "D" model is still going strong and the tubes haven't been replaced since I got it in the early seventies - I still use it occasionally along with an 8640 for two-tone measurements. And you mentioned the 8640B - here's my workhorse for modulation and noise work in perfect operating condition:

http://i97.photobucket.com/albums/l219/randy9944/P1010971.jpg

A wonderful generator and I built a frequency doubler for it so that I can get 1080 MHz coverage. I also have an external power amplifier/doubler that allows up to 2160 MHz low-noise coverage. You may have noted that I added a converter to my old 8570 spectrum analyzer (same as the 8669B) to obtain baseband coverage. The reason for that was for making FM noise measurements using the 8640B as a downconverter L.O.

The spectrum analyzer YIG L.O. noise doesn't allow close-in measurements so I also made this little gadget (which will eventually end up in a shielded enclosure):

http://i97.photobucket.com/albums/l219/randy9944/P1010973.jpg

I don't think you can see the block diagram in my lab book but what this little circuit does is phase-lock the 8640B to an onboard xtal oscillator that is the exact same frequency as a xtal filter on the same board. A sampling phase detector compares the signals and provides an error correction voltage to the 8640B DCFM input port.

The idea is that the conversion is adequately removed from the spectrum analyzer L.O. so as to reduce the noise floor of the analyzer below the measurement range. The entire rig basically notches out the L.O. contribution of the spectrum analyzer by 40 dB and allows me to make close-to-the-carrier noise measurements that aren't dominated by the YIG oscillator. It needs to go inside an enclosure so that I can ovenize the two xtal circuits - they drift out of synchronization every ten minutes of so, requiring retuning either the oscillator or the filter, LOL.

When I started my consulting business, it was a shoestring operation and clearly I couldn't afford modern equipment - so I designed and made various items like this one to make use of the equipment I COULD afford while improving their performance. There are a number of home-built, low-noise synthesizers, for example, that I use for intermodulation measurements and the like (two synthesizers are visible underneath the 8640B in unpainted alluminum chassis). I also built a 0.4 dB baseband LNA that precedes the spectrum analyzer converter so I can make measurements down to about -170 dBm (great for noise measuremetns if all the power levels are "just right").

I dig this stuff and one of the saddest things that ever happened to me was to be told that I had to retire. But once the pressure of project management was off, I could go back to being an engineer and the consulting business was doing great until I was diagnosed with cancer. I STILL tinker with this stuff and on the "drawing board" for the past couple of years is an ultra-lo-noise baseband swept generator based on a noise-cancellation scheme use in the ALQ-156 system, which I worked on not long before retirement. I have a couple of those DSP-based audio generators but their noise floors are completely unacceptable for precision work (e.g. modulating the 8970B).

Thanks much for your response - as I say, I'm a long-time admirer of the equipment and the culture at your place. I couldn't have fit in there - too much of a maverick, used to come in at 1:00 AM and maybe leave at noon when I felt like it - but I've always enjoyed interacting with your applications engineering people, both in the equipment division and the semiconductor division. I had lots of friends in those areas and if you've been there for a while, you'd know them, I'm sure (Jack Leipoff, Neil Corchrane - I wonder if Suzanne Frick is still around?).

Cheers,
Randy C

edited to add: here's a little noise source I built and calibrated for my noise figure testing:

http://i97.photobucket.com/albums/l219/randy9944/P1010974.jpg

Optics Curmudgeon
12-01-2010, 06:10 PM
I can't think of ANY practical applications for that technology and I've spent my career in that "neighborhood". Consider the tolerances required to produce tight-beam antennas with GAIN at those wavelengths (1/64 inch for quarter wavelengths !!!!) Power levels available above 50 GHz or so can be measured in microwatts or MAYBE a few milliwatts, LOL !

There is a tube called a gyrotron that works on the electron cyclotron principle and is usable as an oscillator or amplifier in the millimeter wave region. I worked on a plasma heating experiment in 1990 that used Russian tubes that put out 350Kw at 75-100Ghz (11 of them!), the Japanese subsequently developed a tube that puts out 1Mw at 150Ghz (2005). The coupling system used multimode round waveguides, horn antennas, mirrors and plastic lenses. Just like any other optics, as Evan said. For scattering experiments receivers were used that looked just like any other superheterodyne receiver, with Schottky diode mixers.

Joe

randyc
12-01-2010, 06:18 PM
There is a tube called a gyrotron that works on the electron cyclotron principle and is usable as an oscillator or amplifier in the millimeter wave region. I worked on a plasma heating experiment in 1990 that used Russian tubes that put out 350Kw at 75-100Ghz (11 of them!), the Japanese subsequently developed a tube that puts out 1Mw at 150Ghz (2005). The coupling system used multimode round waveguides, horn antennas, mirrors and plastic lenses. Just like any other optics, as Evan said. For scattering experiments receivers were used that looked just like any other superheterodyne receiver, with Schottky diode mixers.

Joe

Hi Joe,

That is a completely different technology - the discussion was focused at state-of-the-art MMIC oriented semiconductors. None of that technology lends itself to miniaturized, high-level integration, it's just more lab experiments with equipment spread out over an optical bench.

Repeating my point about practicality: practical devices should be able to be inserted into working systems that have a real purpose. They should be performance and cost effective and NEVER be a sole-source item. Been there, done that - bad news to design ANY sole-source item into a system that is required to have a finite lifetime and the capability of being repairable.

Cheers,
Randy C

Optics Curmudgeon
12-01-2010, 06:29 PM
Oops, shoulda read back farther.

Joe

randyc
12-01-2010, 06:35 PM
Oops, shoulda read back farther.

Joe

Not a problem - the ORIGINAL topic is vacuum tubes and your post fits that subject perfectly :)

The Artful Bodger
12-01-2010, 06:50 PM
Since this is a topic about valves (vacuum tubes):--

One of my trips to Afghanistan included a visit to a rather sadly damaged and thoroughly looted MF and HF transmitter site associated with Kabul airport.

Right at the back of the building stood/stand? two venerable Wilcox NDB tranmitters each taking up 3 or so floor standing cabinets. It would have been 'nice' to take them home but that was impractical so I satisfied myself by grabbing one of the spare triode tubes, I dont recall the type number but I am sure some would be familiar with them, a bulbous envelope with grid terminal on the side, I know that is far from a unique description.

I only carried one very small knapsack so I threw out the dirty laundry and put the tube in there, it filled the knapsack and gave me several moments of enjoyment watching airport x-ray operators trying to figure out what was in this apparently almost empty knapsack.

Maybe one day I will make it into a one valve receiver!

Forestgnome
12-01-2010, 07:42 PM
[QUOTE=randyc]You guys (H-P, then Agilent) have ALWAYS been the best ! Every ATP (acceptance test procedure) I ever wrote universally specified your gear. FWIW, I bought one of the first automated phase-noise systems from Santa Rosa Division when I was at Zeta Labs. This preceded the nice little single chassis unit, it was based on the 11729B carrier test set and the 8662 ultra-low-noise synthesizer - sounds simple enough but the thing was in two six foot high panel-rack cabinets and had a 3585A baseband analyzer, a 8577B microwave analyzer, a microwave synthesizer (can't recall model number) and the entire system was controlled with a 9816 computer.[QUOTE]
You're describing the 3048A system. I used to work on that too. Then we replaced it with the E5500 system. I hope we continue to lead in the phase noise market, as keeping the lead in technology is what's made us successful. As you probably know we got out of the frequency standard business, in which we were the only player. We've been recently pursuing lower phase noise in our sources to the point that where there's only one type of reference source in the world we can use to measure it. Way lower than the 8662A. I don't know the people you listed, as I've only been with Agilent for about twelve years, but I'm sure there are a few around here that do know them. Nice collection of gear you have there. You've been busy!

randyc
12-01-2010, 08:24 PM
Hey Forestgnome,

Thanks for the post and keep up your fine work ! If something can be measured practically, you people have always determined the best way to measure it. Yes, the 3048A was the system I was recalling - what a cool little machine compared to the monster in the screen room at Zeta Labs !

I did NOT know that you no longer made frequency standards and I'm shocked to learn that. (I haven't kept up with your company like I used to before my surgery.) If you guys have any old 10 MHz STALOs lying around gathering dust, I could use a nice system standard to lock everything to, LOL ! (OR any old rusty, dusty YIGs, HA-HA !!!)

The people that I mentioned are probably all retired now (except for Suzanne - if she is still there, she's probably Sales V-P - but she may have remained with the H-P "name" organization). Those other guys were major talents in the day (I think both came out of MIT), I still have app notes written by them that are useful.

Incidentally and for what it is worth, most microwave engineers regard the fact that the computer division keeping the name (rather than the measurement division) to be a travesty. The Hewlett-Packard name has no tradition in computer technology and everything to do with measurement instrumentation - particularly microwave systems. I wonder if the old-timers feel any resentment about that ...

If my old 8640B ever breaks, I NOW know who to turn to for trouble-shooting advice ! By the way, I sincerely appreciate the service you people provide by continuing to make some of the old H-P documentation available to us on your website.

I would REALLY be interested in learning about the current state-of-the-art of phase noise at Agilent. Can you recommend any literature that describes the current technology (if you PM me, I'll provide you with my personal e-mail address that will accept attachments) ?

Cheers,
Randy C

Forestgnome
12-01-2010, 08:40 PM
I would REALLY be interested in learning about the current state-of-the-art of phase noise at Agilent. Can you recommend any literature that describes the current technology (if you PM me, I'll provide you with my personal e-mail address that will accept attachments) ?

Cheers,
Randy C

Yeah, I did feel a little miffed about losing the HP name to the computer business. Just didn't make sense to me, but a lot of consumers do associate HP only with computer stuff.

If you ever need any help or info I can provide feel free to contact me.

How about app notes?
http://www.home.agilent.com/agilent/facet.jspx?t=80030.k.1&ki[0]=phase+noise&cc=US&lc=eng&sm=g&pselect=SR.Facet.By%20Type%20of%20Content

Evan
12-01-2010, 09:05 PM
I have several tube powered devices that are still in good working order.

A Hallicrafters S-40A Receiver:

http://ixian.ca/pics8/halli.jpg

A first generation digital rate meter/scaler (fancy Geiger counter) complete with vacuum tube flip flops for the binary counter.

http://ixian.ca/pics8/scaler.jpg

And one that needs a new tube that I don't have. Not much point in fixing it since it is intended to warn you that you are dead meat and would you please turn off whatever made you that way.

http://ixian.ca/pics8/gamma1.jpg

I also collect anything in a sealed glass envelope. Here is a small sampling.

http://ixian.ca/pics8/tubes.jpg

randyc
12-01-2010, 09:35 PM
Yeah, I did feel a little miffed about losing the HP name to the computer business. Just didn't make sense to me, but a lot of consumers do associate HP only with computer stuff.

If you ever need any help or info I can provide feel free to contact me.

How about app notes?


Thanks, that site is a long-time bookmark on my desktop computer and I've added it to this laptop. I'm going to read through the app notes on the E5XXX system tonight.

Not only was H-P noted for microwave instrumentation but at one time no hospital in the U.S. was considered well-equipped without H-P medical monitoring equipment ! The list goes on - especially in the calculator and computer field (many of the major advances have been forgotten but the scientific calculator was born there).

In semiconductors, Hewlett Packard also had a fine reputation - the step-recovery diode was invented there and the Schottky-barrier diode was developed to it's final configuration at HPA (Hewlett-Packard Associates, which was a different organization from H-P in the late sixties).

The guys that designed measuring hardware were always accessible (it helped to have worked just a few blocks away, ha-ha) and a visit to the Applications Lab was useful to resolve tricky measurement problems. (One of the coolest - and most esoteric - measurements I learned came about from one of those visits and I still use the technique.)

The sampling phase-detector, which is the key to all high-stability microwave and millimeter wave low-noise oscillator applications, was developed at H-P. It resulted as a stepchild of the sampling oscilloscope, the first instrument developed that could display 1 GHz waveforms in real time.

Computer Aided Design was born at Hewlett-Packard, invented by my boss (Les Besser) when I worked at Fairchild Microwave, where he finally completed the first practical analysis/optimization program. It was called "Speedy" and ran from a rented mainframe on a time-sharing basis. Les went on to form Compact Engineering and marketed the program for many years before becoming a respected lecturer/consultant.

I could continue listing H-P achievements for days. Silicon Valley was a hotbed of development (especially in the microwave frequencies) and a major reason for that were the H-P facilities in Palo Alto and Mountain View. I was fortunate enough to have been associated with the technology almost from the beginning - due to location and LOTS of luck, knowing the right people at the appropriate time. By association, I was convinced that this technology was the future and that prompted me to change my career from mechanical engineering to electrical engineering, specializing in RF/Microwave/Millimeter Wave technologies.

I have many memories of MAJOR technical breakthroughs that were measured and documented by Hewlett-Packard measurement systems ! One was the characterization (at Fairchild R & D) of the first commercial GaAs FET using a custom S-parameter test set that H-P built for us. It was one-of-a-kind and VERY expensive since S-parameter characterization was virtually unknown in the early seventies !!! However, we sold those first GaAs FETs for $500 each in 1973 (IIRC) and couldn't make them fast enough, LOL. (Incidentally, H-P introduced their own GaAs FET products not too many years later and they were excellent devices but had reliability problems.

The entire design and production of the H-P GaAs FET line was sold to "Dexcel" (Santa Clara, CA) which was owned by a Japanese entrepeneur at the time. They were located exactly next-door in the small industrial complex as my employer at the time (Solid State Technology). Their application engineer was my colleague from Fairchild, George Bechtel (who also worked for Les Besser and took over some of the research projects that Les left behind when he moved to "Farinon Electric" just prior to getting his own business started.). (I believe George went on to publish almost as prolifically as Les, LOL.)

Now I'm so far off-topic that this thread looks like the ones on the engineering forums where I hang out sometimes, LOL !

Cheers,
Randy C

edited to add a couple more memories -

randyc
12-01-2010, 11:16 PM
Although I haven't spoken with either of the guys mentioned in the previous post for YEARS, I just did a quick google and found both of them:

Les is listed in the "Microwave Hall of Fame" site:
http://www.microwaves101.com/encyclopedia/halloffame3.cfm

George can be found on the National Institute of Standards site:
http://www.atp.nist.gov/1999national/bechtel.htm

Great guys and interesting times, I was fortunate to have been there !

edited to add: Les is listed as Hungarian/Canadian which isn't strictly correct. According to what he told me as we were driving from a meeting one morning, he slipped out of Hungary (this was during Soviet domination) and spent two years in Toronto before finding his way to the Santa Clara Valley. He described his amazement at the fact that he could earn enough money (in Toronto) to buy a car. After moving to the San Francisco area he was RICHLY rewarded for his abilities. This was a man with the right idea in the right profession at the right time :)

randyc
12-03-2010, 12:16 AM
The topic of the original post has run its course, LOL.

BUT - I wanted to thank forum member forestgnome, who (this morning) steered me to an online reference describing the architecture of a very interesting, state-of-the-art measurement instrument. Equipment manufactured by Agilent (formerly Hewlett-Packard Measurement Division) made them the most highly-respected manufacturer of microwave/millimeter-wave measurement systems for the past sixty years.

Forestgnome is one of the guys that keep Agilent at the cutting edge of high-frequency measurement systems. Thank you, Eric, you're a great guy and keep up the good work !

Randy C

bob_s
12-03-2010, 04:50 PM
Surely you jest when you say that HP doesn't have a background in computers!

9800 series (up to 9840) about 1970-72 only predates IBM by a decade.

Of course HP1 (gazinta language) was not nearly as friendly as BASICA, no disc drive, but had a reasonably powerful cassette interface - load program, serve as cache memory, or intermediate results, or swap space. Of course it was typical engineering number cruncher not one of these useless game playing, cruise for porn boxes.

randyc
12-03-2010, 05:33 PM
Surely you jest when you say that HP doesn't have a background in computers!

9800 series (up to 9840) about 1970-72 only predates IBM by a decade.

Of course HP1 (gazinta language) was not nearly as friendly as BASICA, no disc drive, but had a reasonably powerful cassette interface - load program, serve as cache memory, or intermediate results, or swap space. Of course it was typical engineering number cruncher not one of these useless game playing, cruise for porn boxes.

Actually, I didn't say that. I said they didn't have a "tradition" in computers (like IBM, for example).

Oh sure, I had a number of those old H-P computers, a 9845, two 9816s, HP-85A one for personal design/analysis and three for instrument controllers) but the introduction of computers to the H-P product line in the seventies post-dated the metrological business that was founded decades earlier and which made the company the world leader in that business.

When HP-BASIC was introduced (may have been simultaneous with the roll-out of the HP-85A), good things started to happen for the working design engineer and test engineer. For the first time, the average joe could interface an inexpensive computer with H-P measurement equipment and make a dedicated automated test station with previously unheard of cost-effectiveness.

I first saw this when a buddy spent a couple of days hooking his HP-85 to an old HP8410 network analyzer. Not only was he able to control every aspect of the operation but using the computer, his results made a LOT more sense than visual screen display interpretation because he was able to use statistical tools to improve measurement accuracy. All later network models used variations of what Tim Healey came up with in 1981 ...

Also, since HP-BASIC was a superset of earlier BASIC languages, it was fairly easy to convert heritage programs (the very few that existed) to the newer platforms.

Younger engineers and technologists will associate the company with computers as forestgnome (an employee of Agilent) suggested.

Us old-timers will always think of test and measurement instrumentation when we hear the name "Hewlett-Packard". (With the significant exception of the scientific calculator which will always be associated with Hewlett Packard.)

Thanks for your comments,
Randy C