PDA

View Full Version : Magnetic clock.



Bob Fisher
04-23-2012, 05:06 PM
Did any of you see Wes Bye's magnetic clock at the NAMES show that past weekend? I forgot to take a pic of it or would post it.Maybe Wes will do us a favor and post a few shots. Bob.

jdunmyer
04-23-2012, 07:35 PM
Well, I hope Wes is getting a commission on magazine subscriptions, I bought one because of his clock. Probably won't get around to actually building it, but will enjoy reading about it.

Neat device!!

Weston Bye
04-23-2012, 08:08 PM
I can post pictures, but I'd rather see other people's pictures. I am ecstatic with the reception the clock received, and indeed, there were quite a few subscriptions to DM sold because of it.

The reception has encouraged me to explore having printed circuit boards made for the electronic portion of the project. I find that I will be able to do it, and at a reasonable cost, so people who might be reluctant to hand wire a board will find it easier to build the clock. The article will still give instructions for hand wiring, but a board will also be offered.

I was surprised that some of my older projects that I exhibeted also received a lot of attention and intreats. Some of those may have sold some back issues for the magazine.

I enjoyed seeing and talking with Bob and Jim again as well as others, and meeting some of the other members from the BBS for the first time.

All in all, NAMES was a very good experience for me this year.

George_Race
04-24-2012, 08:09 PM
Hi Weston:
I talked with you for some time just before noon. You may remember that I was wearing a HMEM name tag. Yes, i went and got the last of the Spring edition, along with the Winter. I also subscribed for the next 4 issues.
I am really impressed with your Magnetic Clock design. So impressed that I have to make one! Yesterday I drafted up almost all of the wheels, described in the Spring edition. I am doing them in CamBam and it is really easy using the detail that you have provided, and of course, the GCode is generated immediately for my home made CNC system.
I have noticed a couple of idiosyncrasies in the drawing dimensions between the downloaded DXF files and the magazine.
Is it OK to discuss and ask questions about the build here on the forum, or do you have an email address that I can use to contact you directly?
Boy, I can hardly wait until the next edition to get the next group of pieces to draw and build. What a great project you have created!
And yes, I will post pictures as soon as I start cutting metal.

George

Weston Bye
04-25-2012, 10:32 PM
George,
I do indeed remember you. Thank you for your comments and interest in the clock. I would be happy to discuss the clock here, or in Personal messages or by e-mail, and if necessary, by phone. A PM to me or an e-mail to Village Press will make contact with me for direct e-mail or phone.

I can discuss nearly anything about the clock as long as I don't get too far ahead of the published articles.

winchman
04-26-2012, 02:47 AM
Is this the clock?

http://neme-s.org/NAMES_2012/2012-04-24/Section_F/DSC03477.jpg

There are more pics here: http://neme-s.org/NAMES_2012/north_american_model_engineering_page_2.htm

Black_Moons
04-26-2012, 04:17 AM
Weston Bye: how are the magnets 'toothed'?
ie, do you have the two disks attracting with north then south magnets interleaved?

I thought of an interesting alternative of having all north magnets where the magnets from each disk would repel each other, producing magnetic 'teeth' that refuse to overlap and push each other apart.

Weston Bye
04-26-2012, 05:50 AM
Winchman - That's the one.

Black Moons - The magnets on each wheel alternate North - South. I tried various configurations and found the N-S engagement to be the most reliable. Such an arrangement dictated even numbers of magnets on all wheels, but works for defining time as we know it. I haven't done the math, but using magnetic gears for something like an orrery might prove complicated.

To give an idea of scale, the chapter ring is just a little over 10 inches in diameter.

dp
04-26-2012, 12:17 PM
Evan has some fascinating posts here about magnetic coupled drives he made. Kinda hard to call them gears, but power transfer was quite remarkable for an airgap system.

Weston Bye
04-26-2012, 12:54 PM
Evan has some fascinating posts here about magnetic coupled drives...

Yep, saw them. They might have been the seed of knowledge that inspired the clock.

Errol Groff
04-26-2012, 04:50 PM
Winchman: Thanks for referencing the NEMES web page!

Wes: I must apologize for the so so quality of the photos. I wish I had done better justice to your work. Maybe next year will be better!


To all: The first half of the video I took at the show is rendering now, a 5 hour + trudge on my computer and will be uploade to FB as soon as it is done. I will do the rest of the video tomorrow and post it as soon as possible.

Rich Carlstedt
04-26-2012, 05:35 PM
It's not only a very beautiful piece of work, but the guy standing there was about as nice as they come !


Congrats on a very special time piece, Wes !
enjoyed our talk

Rich

boslab
04-26-2012, 05:40 PM
Yep, saw them. They might have been the seed of knowledge that inspired the clock.
irrispective of origin a nice clock, reminded me of a harrison h series i saw in london, well done
regards
mark

George_Race
04-29-2012, 08:05 AM
Weston, just want to clarify what I think is a difference in the magazine article and the .DXF files you have for download.

The Great Wheel on page 49, shows an outside radius of 3.381 inches.
The Great Wheel DXF file, det2.DXF, is drawn actual size with a 3.381 diameter.
I don't see a note anywhere that the DXF file is drawn 1/2 scale.

I assume that the magazine article size of 3.381 radius is correct, as the Great Wheel shows material size to be 6 13/16 in the bill of materials on page 55.

Clarification would be appreciated,
George

Weston Bye
04-29-2012, 08:27 AM
The 3.381 radius is correct.

I don't know if or how the DXF came to be in error, but I have occasionally seen DXF file transfers experience scaling errors when transferring between different CAD programs.

George_Race
04-29-2012, 08:42 AM
Weston, more info needed to confirm dimensions.

The drawing, det5.DXF appears to be drawn twice actual size of the magazine drawing on page 52. (Wheel J)

Based on what I can see, I believe that the magazine drawing dimensions are correct.

Please confirm.

Thanks,
George

Weston Bye
04-29-2012, 09:31 AM
The magazine drawing dimensions are correct.

George_Race
04-29-2012, 11:05 AM
Hi Weston, back for more information.
Incidentally, I use AutoCAD version 9 for importing your DXF files. Makes no sense that some would open properly, and some double size.

Thanks for confirming that Wheel J is correct size in magazine.

Now, my next confusion. det6.DXF shows an outer diameter of 1.62, but the actual diameter on the drawing is 3.24. The drawing on page 52 shows a diameter of 1.62, and the material list on page 55 shows that the Wheel K material is 1 5/8 in diameter. Those two seem to correspond, but, when I look at the Full Assembly drawing on page 48, Wheel K is obviously much larger than Wheel J.

By the Assembly drawing I would conclude that Wheel K is correct in the DXF file at 3.24, and the magazine drawing and material size in the list are incorrect.

If the magazine drawing changed to show the outer dimension as R1.620 than the magazine drawing would be correct. Also the magazine drawing does not show the "Radius" of the Magnet Hole's center line. Of course then the material list size would need to be up-sized as well to bring everything into the same size.

To even add more to this overall confusion, on my end, when I look at the full assembly drawing on page 48, the Great Wheel E, looks about 4 times the diameter of Wheel K. That would make the Great Wheel E, close to 12 inches in diameter, not 6 as reflected in the drawings and materials list. Or, is the Great Wheel really about 6 inches and the K Wheel about 1.5?

I am really feeling like I am being a really big bother here Weston, but am a stickler for detail, especially when I have conflicting information and am not sure of which is correct.

Thanks for the repeated help, really appreciated.

Weston Bye
04-29-2012, 04:51 PM
No bother George, I checked the original CAD file and the magnet circle radius for Wheel K is .695 - looks like we have an omission in the magazine.

The wheel would work even if a reader "eyeballed" the radius. The magnetic coupling will tolerate considerable misalignment.

George Bulliss
05-01-2012, 09:35 AM
Iím sorry for any confusion the dxf files may have caused. In the case of Detail 6, the dimension of the magnet pocket circle radius was not included in the magazine. The correct drawing is displayed below. This was simply an omission caused by me when redrawing the detail.

http://i358.photobucket.com/albums/oo28/gbull3/det6.jpg

As to the scale problem; the details are drawn to the correct size, but I do scale the drawings differently to fit into the magazine. Detail 6 appears 1:1 in the magazine, but we often scale a drawing differently, with small drawings sometimes being scaled up and larger drawings scaled down.

The fact that the dxf file is set to a scale that is twice that of what appears in the magazine is simply due to the art department sticking with a policy of reducing every drawing going into the magazine by 50%. Iím sure they had a reason for doing so in the past, but I canít figure it out. After months of trying to change things, I gave up and simply submit the drafting files in double the scale they will appear in the magazine.

Thatís the reason for the scales being different between the magazine and the dxf files, but it should have no bearing on your use of the files. If I import one of the dxf files and check a dimension, I get the correct information. The geometry was drawn at the correct size and the scaling only comes into play when printing or when I export a tif file for the art department.

Your import settings in your CAD program should have the ability to set the scale of the incoming drawing. Mine defaults to the scale at which I drew it, but yours may not be defaulting to the correct settings. Try importing Detail 6 with the scale setting at 2:1 and see what happens.

The bottom line is the dimensions depicted in the magazine and on the drawing files are correct (to the best of our knowledge) and if there is a discrepancy between the dimensions detailed and your measurements, I would look at the scale settings used when importing.

Hope this long-winded ramble helps.

George

George_Race
05-01-2012, 08:18 PM
George, thanks for the explanation, it really helps a lot.
Knowing that the published dimensions are correct are all that I needed to move ahead with my drawings.
Thanks again for the explanation,

Weston Bye
05-02-2012, 06:23 AM
From another thread. I am copying it here because it relates to the clock:



Originally Posted by topct
Another example of socket head screw abuse.

Ah, how easy it is to offend some people's sensibilities.

Well, to explain myself, the clock uses a multitude of non-traditional materials, methods and principles, and was indeed designed to use socket head screws as another departure from tradition. Indeed, in most places, nothing else will fit. The stainless steel screws have a bright enough finish that they don't seem out of place with the aluminum, and there will be no rusting or tarnish problems in the future.

As to tradition, one need only pick up a magazine about high-end Swiss and German watches these days to behold all manner of non-traditional fasteners.

The black oxide button head screws used on the chapter ring are just place holders. The clock was not completed by the time NAMES came along. I used black oxide finish to contrast with the aluminum but approximating the color of the hands. They were only there to fill the indexing holes that I will be using when I engrave the numerals.

George_Race
05-04-2012, 07:43 PM
Weston:
I would like to mount the magnets in the wheels. The article states that each adjacent magnet has its pole reversed from the one next to it.
By looking at figure two on page 49 of the Winter 2011 edition, your example makes it very clear.
I have the magnets and would like to proceed and put them in place.
I am considering super glue to hold them in place. Is that a good solution?
Any suggestions would be appreciated.

When I broached the three wheels, I did not pay any attention to where the splines were pointing, in reference to magnets. It that going to be a problem, or does it not matter?

I think that is all that I can do until the next issue of the magazine.

This is a real fun project, thanks for putting it out there.
George

Weston Bye
05-04-2012, 07:59 PM
Super glue. More specifically, I used Gorilla Glue's version of super glue, although any version should work. I placed the magnet on the end of a steel rod to place it and used used a brass pick to get the magnet off the rod.

If you get a magnet in backwards you will have to use a propane torch to get it out. This will destroy (demagnetize) the magnet in the process.

Sage5902
05-21-2012, 08:36 AM
Weston:

George asked how to tell (or if it matters) the orientation of the hex holes with relation to the magnets. It's difficult to be sure from the drawings.
I'm about to make chips and curious minds would like to know.

Thanks

Sage

Weston Bye
05-21-2012, 08:45 AM
The hex broach in the bore of the wheel makes no difference - there are other features built into the clock train to accurately align the hands.

Sage5902
05-24-2012, 10:50 PM
Thanks Weston:

Would it be allowed to get a shopping list of the bearings required so I can get a head start on ordering them?
(If they were in the first two articles then sorry I didn't see it)

Sage

George_Race
05-25-2012, 06:50 AM
Thanks Sage for re-asking the question.
I have been sitting back worrying about it.

Like you, I am anxious to keep building. Can hardly wait for the next issue.
George

Weston Bye
05-25-2012, 04:44 PM
Bearings. Two different sizes and styles are needed.

Flanged, shielded ball bearing- .375 od x .1875 id x .125 wide
McMaster-Carr #57155K319

Shielded ball bearing- .3125 od x .125 id x .1406 wide
McMaster-Carr #57155K366

6 of each style are required.

I purchased mine from McMaster-Carr, but they are standard size and can be had elsewhere. Don't get sealed bearings, just shielded. The seals will cause too much friction.

There are a couple of O-rings used also.

Type AS568A-001-1/2
1/16 id x 1/8 od x 1/32 wide
McMaster-Carr #9452K311

Type AS568A-005
7/64 id x 15/64 od x 1/16 wide
McMaster-Carr #9452K13


The O-rings are sold in bags of 100 from McMaster, sad to say, but they aren't costly.

I will post the electronics stock list later. I have looked into printed circuit boards and find it feasable so I will be offering them for sale, and possibly even assembled and tested board assemblies. This little bit of commercialism has been approved by George as it will be of benefit to readers and the magazine. More about this later.

Paul Alciatore
05-25-2012, 07:39 PM
Winchman - That's the one.

Black Moons - The magnets on each wheel alternate North - South. I tried various configurations and found the N-S engagement to be the most reliable. Such an arrangement dictated even numbers of magnets on all wheels, but works for defining time as we know it. I haven't done the math, but using magnetic gears for something like an orrery might prove complicated.

To give an idea of scale, the chapter ring is just a little over 10 inches in diameter.

Weston,

I loved your clock article. It is a great idea.

Your remark about the difficulty of doing an odd number of "teeth" in a magnetic gear started me thinking. Here is an idea for doing that with an "idler" magnetic wheel between two others. One or both of the outer wheels could have an odd number of magnets.

http://img.photobucket.com/albums/v55/EPAIII/MagnetGearsOdds.jpg

The outer wheels have all magnets oriented the same way, all the North poles are up in the illustration (I showed Souths on the BOTTOM of the right wheel for the illustration). The intermediate, "idler" wheel would transfer the motion from one to the other outer wheel. It has alternate pole orientations but at half the spacing so an opposing magnet on it will always be between two attracting pairs on the outer wheels. This will allow the outer wheels to have any number of magnets, even or odd, in any combination that physically fits.

So, go ahead and make your magnetic wheel orrery. I look forward to the article and photos.

Weston Bye
05-26-2012, 02:35 PM
Paul,
Excellent idea. However, supporting the idler wheel might be problematic, sandwiched between the two larger wheels. I think it could still be done, but My first thought was to Move the 17-magnet wheel to the right to engage magnets on the other side of the idler.

Interesting concept - I like it. If I do incorporate it in future projects I will be sure to attribute it to you when writing the article.

Paul Alciatore
05-26-2012, 04:37 PM
Weston,

My drawing was just a study of the concept. Yes, positioning the idler wheel as I have drawn it may be problematic with some combinations of numbers. While with others it would work as I have shown. As I have drawn, it would certainly need to be on a shaft that only protrudes from one side of the idler. I debated about adding shafts to the drawing but omitted them for clarity. I did make sure the idler does not interfere with the shafts of either of the other wheels. Your idea of putting the 17 magnet wheel on the other side or even top or bottom or any other angle that does not interfere would definitely work.

Again, remember that not all combinations of gears with teeth can be physically assembled in a world that is limited to only three dimensions. I have run into this on my lathe's gearing: the numbers may work but the gears just can not be assembled that way.

PS: Credit is optional. I would just love to see it in use and I doubt I would have time for such a project unless I were going to sell it.

Edit: I was just looking at the drawing and thinking about Weston's idea of putting the 17 magnet wheel on the far side. If that was done, then the magnets in it could be oriented either way, either all N or all S up. It would work both ways if you are free of the field from the 16 magnet wheel. It may even work in my original configuration, I/you would need to experiment on that one. Actually, you could even go to a alternate orientation at that point if you spaced them at the same 1/2 interval of the "idler" wheel. This would facilitate adding more wheels to pick off the 17 magnet wheel which would then become a 34 magnet wheel. But this would still preserve the 16:17 ratio of these two wheels.

Oh, the wheels are spinning and I love it.

Paul Alciatore
05-26-2012, 04:56 PM
Taking my last thought in my previous post just one step further, it seems that you could also double up on the magnets in both the 16 and the 17 magnet wheels to match the half spacing of the magnets in the idler. If you do this, it suddenly becomes obvious that you do not need the idler in between them. They can then mesh with other and the 16:17 ration is still maintained.

Thus, the easiest way to get such an odd ratio or a ratio with an odd number of teeth/magnets would be to simply double the tooth/magnet count on each wheel and presto, you now have an even number of teeth/magnets on each one. So the original scheme of alternating N and S poles will still work. This method can be generalized to work with all possible ratios. My "brilliant" idler is not really necessary after all. Just need more magnets.

Of course the idler may be more interesting in a visual sense. Or it may allow a more compact design.

dp
05-26-2012, 05:00 PM
Don't know if it's been suggested, but it might be worth while to drill a brass or aluminum dowel and put in it a "master" magnet to test against to ensure alignment of magnets prior to gluing them in place.

Sage5902
05-26-2012, 09:18 PM
Weston:

Sorry I didn't ask you these questions when I saw you at NAMES but what's the basic design of the electronics?
Does it use a crystal as a time base or does it count AC line cycles?
Is it 555 timer / microprocessor based or TTL / CMOS logic.
battery or ac line powered.

and other such clues would be interesting.

Thanks

Sage

Weston Bye
05-27-2012, 08:10 PM
The clock is quartz crystal controlled, qualifying it as a "Quartz Movement", using the crystal and a CMOS oscillator/divider chip. A second chip further divides and decodes the pulses along with some diodes to drive three power FETs to drive the three coils, forming a three-phase variable reluctance stepper motor.

Here is the electronics stocklist, all pretty common stuff:
1 32,768 Hz tuning fork crystal
1 CD4060BC CMOS 14-stage ripple-carry binary counter
1 CD4022BC CMOS Divide-by-8 counter/divider with 8 decoded outputs
3 IRLU024N N-channel Power Field Effect Transistor (many others will also work)
4 1N4004 diode
6 1N914 diode
Note: all resistors are ľ Watt
1 330K Ohm resistor
1 15M Ohm resistor (or use a 10M and a 5M in series)
3 100 Ohm resistor
3 10K Ohm resistor
1 10 pF capacitor
1 22 pF capacitor
3 .1 uF capacitor
1 100 - 500 uF, 25 Volt electrolytic capacitor

The clock, as built, draws about, 140mA - too much for extended battery use, so I run it on a wall wart. I will be doing some experimenting before the electronics article comes out to try to reduce the current consumption by changing the coil windings.

Weston Bye
06-24-2012, 04:27 PM
Due to popular demand, I have designed and contracted to have printed circuit boards made for the electronic portion of the clock and will be offering them for $15 each for the bare board.

I have had requests for fully assembled and tested boards and even the wound coils. I am considering these but have made no decisions.

This bit of obvious commercialism has George's approval as it may aid some readers in attempting to build the clock and is helping to sell new subscriptions of Digital Machinist magazine.

dp
06-24-2012, 05:28 PM
I'd forgotten just how long this project is been in the works. Here's a post from the just beyond the napkin drawings phase:

http://bbs.homeshopmachinist.net/showthread.php?t=36695

lvanduyn
06-26-2012, 04:03 PM
Hi Weston..
On looking through the parts list, I note a couple of 4000 series CMOS parts. These have been obsolete for years. I did a search of all the usual suppliers and did find the CD4060 at Jameco. I have not found the CD4022 anywhere. Most don't even have it listed and the few who do show zero stock. Perhaps you might revisit the design to see if some other chip would work before commiting to a board design.
Thanks, Larry

George_Race
06-26-2012, 05:58 PM
Weston, I am a lifetime electronics buff, would it be possible for you to post a schematic of your Magnetic Clock electronics circuit? Retired, I do have a lot of time and it would be fun to see what I could come up with that uses today's technology for the clock timing circuit.

George

Weston Bye
06-27-2012, 06:18 AM
Hi Weston..
On looking through the parts list, I note a couple of 4000 series CMOS parts. These have been obsolete for years. I did a search of all the usual suppliers and did find the CD4060 at Jameco. I have not found the CD4022 anywhere. Most don't even have it listed and the few who do show zero stock. Perhaps you might revisit the design to see if some other chip would work before commiting to a board design.
Thanks, Larry

Just checked Digi-Key, 1447 available.
http://www.digikey.com/product-detail/en/CD4022BE/296-2161-5-ND/67263

Also Newark, 920 available.
http://www.newark.com/texas-instruments/cd4022be/logic-counter-divider-oct-16dip/dp/50R5027?in_merch=Popular%20Logic%20Products

Also Mouser, 740 available.
http://www.mouser.com/ProductDetail/Texas-Instruments/CD4022BE/?qs=LU5rZWrBGo3lK92hebnP3z7a1YAFeEZCEmgZn2IzUFU%3d

Weston Bye
06-27-2012, 07:58 PM
Here is the board:
http://img.photobucket.com/albums/0803/Weston/DSCN4416.jpg

...and populated for a three-phase clock. The board is also useful for a four-phase clock with the addition of another transistor and a couple of resistors and diodes and moving of a jumper.
http://img.photobucket.com/albums/0803/Weston/DSCN4415.jpg

jdunmyer
06-27-2012, 08:05 PM
Wes,
What's the chances of your offering a kit for the electronics?

Weston Bye
06-27-2012, 08:15 PM
Here is the schematic for the previous clock from the Winter 2006 issue of Digital Machinist. It is the same basic circuit as the current project except that it provided four half-second pulses to the coils. The current project requires three one-second pulses to the three coils so I had to add some circuitry. I will be holding off on posting the new circuit until after the article goes to press.

the electronics gurus among you are welcome to improve upon the circuit, but bear in mind that many (most?) readers are not adept at using programmable devices like PIC, Basic Stamp, Arduino, etc. I have to write so that the average normal reader has a chance at reproducing the circuit with only a soldering iron.

http://img.photobucket.com/albums/0803/Weston/TimeMachinerySchematic.jpg

Weston Bye
06-27-2012, 08:19 PM
Wes,
What's the chances of your offering a kit for the electronics?

I am exploring the possibility. Also, I have been asked about providing wound coils, an assembled board, and a ball bearing kit.

No firm plans yet.

Paul Alciatore
06-27-2012, 08:23 PM
Hi Weston..
On looking through the parts list, I note a couple of 4000 series CMOS parts. These have been obsolete for years. I did a search of all the usual suppliers and did find the CD4060 at Jameco. I have not found the CD4022 anywhere. Most don't even have it listed and the few who do show zero stock. Perhaps you might revisit the design to see if some other chip would work before commiting to a board design.
Thanks, Larry

40xx CMOS not available? Oh, come on.

Digi-Key CD4066BE, good old 14 pin DIP package, Digi-Key part number 296-2061-5-ND showing 6984 in stock, price $0.48.

Digi-Key CD4022BE, also good old 14 pin DIP package, Digi-Key part number 296-2161-5-ND showing 1447 in stock, price $0.52.

Both can be ordered in quantity of one. Of course, shipping would be more than the cost of the ICs but you can get almost all of the electronic parts there so just make one overall order.

For those who are less familiar with electronic specs, the DIP style package is the type for through-hole mounting in the circuit boards. Much easier to solder than the surface mount components.

As for lowering the current draw, I doubt that changing to a "more modern" chip, like a PIC device, would make much of a difference. Quiescent current in a 4066 is about 1 uAmp. and about 10 or 20 UAmp. for the 4022. Not much room for improvement there. Most of your current is undoubtedly going to the coils. You may save one or two milli-Amps in the rest of it, but even that much I doubt. Perhaps the oscillator section could be improved, but I don't know how. I mean you are already going through a 330K series resistor so not much current there either.

You can program a PIC to do this but frankly, it is gross overkill.

George_Race
06-27-2012, 08:53 PM
Excellent circuit Weston, could not be any easier. Very straight forward and efficient. Design current is probably at it's minimum.
George