PDA

View Full Version : OT HP 5335a frequency counter - help with test equipment repair



Mcgyver
01-16-2018, 02:12 PM
In a thread a few weeks ago I posted the following, and I'm just now getting to this mini project and figured

I'd posted:

I mentioned the 5335A is busted....As per manual, I've been checking the voltages and some are not there or wrong. The next step in the diagnostics are to disconnect a bunch of stuff form the power module and see what the voltages are, then start reconnecting so as to isolate the problem.

One of the things disconnected is the keyboard and display, so the on/off switch is no longer connected. Is the unit supposed to be powered and then start disconnecting things?

here's the page (212/246) from the service manual

https://i.imgur.com/iCQjksj.jpg


Thanks

Mcgyver
01-16-2018, 02:19 PM
and Jerry replied




As for the counter repair......

The manual is starting where I would.

Power supply voltages first. The thing will need to be powered to do that. See if you can disconnect any loads....it appears you can, from the manual step 4. I never want a bad power supply hooked to potentially still good circuits if there is any way to avoid it.

You can disconnect and then power up, check, power down, disconnect something else, power up, and so on. It can be an issue if the power is not a real power switch, but some sort of an "enable" that requires another board in order to work.

If that power switch is a real one that just happens to be on that board, then use a jumper in place of the switch, and turn on and off with an external switch.

I REALLY like to have a power panel on the bench, with a variac, switches, ammeter and voltmeter for AC. If you do not, then things can be more hairy if there are serious issues that blow fuses. Much easier to turn up voltage, observe a high current, turn back down and shut off before the fuse blows. Saves a lot of fuses, and gives more info.

I do have an isolation transformer and variac setup as you mention, but I'm struggling still with HP's sequence to isolating the problem. Voltages are present but incorrect values. They want me to unplug boards from the power board to isolate what board is the problem. I don't know how to do this - if I disconnect the other boards, I can't turn the unit on. Front power switch board connects to A4 main board that connects to power board - at the power board I have no idea what I'd jump to turn it on - the switch is two boards away.

Do I just turn it on with everything connected then start disconnecting? or is that not done?

thanks

J Tiers
01-16-2018, 03:06 PM
Try other PWBs first. You may not need to go that far.

Is it a momentary keyboard type switch? If so, it is probably "on" until an "off" is pressed, or until "on" is pressed again (similar to many laptops).

In that case you may be able to pull it while power is on. If it changes nothing you pull the line cord plug and reconnect the PWB, then move on to other tests.

If the switch is a "maintained connection", then it may be possible to do a temporary switch just for testing. Again, as long as the disconnection does not identify a problem with the PWB, you hook it back up and the problem is over.

It should be possible to identify the power switch on the power supply board, from information given, Usually there are connector pinouts given, or at least a schematic which shows the pinouts and maybe the actual function.

Mcgyver
01-16-2018, 03:16 PM
k thanks, I just don't enough experience and didn't want to possibly do damage pulling things while its on. I'll report back :)

Mcgyver
01-16-2018, 04:31 PM
the good news is it was easier than I thought to jump the swith. "S" label on connector to ground and it turned on. Bad news is it isolates the problem to the main board. I don't know where to start or whether its likely in the hopeless catagory. Any suggestions?

lousy resolution scan of main board schematic: https://drive.google.com/file/d/1t44zJpB_8nhDl4kiTK0Nhanv0V6bcDIA/view?usp=sharing

thanks

BobinOK
01-16-2018, 04:46 PM
Which voltages are off and what do they read?

Mcgyver
01-16-2018, 04:52 PM
Which voltages are off and what do they read?

What should be 5V reads 2.2. There others seem ok.

J Tiers
01-16-2018, 04:53 PM
The schematic stops getting bigger with the "+" button just before it gets readable.

There are lots of techniques, starting with looking for hot things. Then using millivolt scale on DC, look for traces that seem to have a lot of current..... larger millivolt drop. THat does not always work, often the drop is fairly small.

Also, just because the main board SEEMS to be the problem, that does not prove it is. Might be that a particular supply just cannot hold up a normal load for some reason. That may be easier to check first, presumably there is a procedure for the check of the PS board.

If there are a lot of the small epoxy dipped tantalum capacitors, they are a known issue. In fact, many tantalum electrolytics are a problem in older equipment. They like to fail as nearly a dead short, but may not be quite that dead.

Follow the 5V, which supplies most of the logic parts, and see what you find.

If you happen to have a FLIR camera, they do the job of finding hotter stuff quite well.

BobinOK
01-16-2018, 05:19 PM
J Tiers has posted allot of good info. If your volt meter is good enough with the power off check the 5 volt supply pins right at each chip with an ohm meter, might uncover an area to look at. Feel each chip with your finger and see if any one seems hotter than the others. You can also measure the 5 volt supply pins at each chip with it hot and look for the lowest reading.

tomato coupe
01-16-2018, 05:31 PM
A working HP counter of that vintage can be purchased for $50-$200 on eBay. Are you sure it's worth your time to try to fix this one?

P.S. If it were me, I would just start snipping the +5V decoupling caps off the main board. If one or more are shorted, the counter will start working once they are removed. (Assuming, of course, that is the only problem.) Replace the caps later if the counter looks salvageable.

J Tiers
01-16-2018, 05:32 PM
J Tiers has posted allot of good info. If your volt meter is good enough with the power off check the 5 volt supply pins right at each chip with an ohm meter, might uncover an area to look at. Feel each chip with your finger and see if any one seems hotter than the others. You can also measure the 5 volt supply pins at each chip with it hot and look for the lowest reading.

Yes, that is equivalent to checking the drop in the supply line.

The one problem is that you can get confused by drops that occur in the GROUND lines. If you use one ground point, then you may end up missing some drops because the ground-side drop does not show up (the drop measures as part of the actual load voltage). If you check + and ground at each chip, you may get confused if there is a common ground line which serves both the area with a problem, and the area you are measuring. The drop in the common ground may fool you into thinking there is a problem when it is actually in a different place.

polaraligned
01-16-2018, 08:03 PM
I'd look for any tantalum capacitors and check them first.

JoeFin
01-16-2018, 08:28 PM
A working HP counter of that vintage can be purchased for $50-$200 on eBay. Are you sure it's worth your time to try to fix this one?

.



Sorry to say it guys but tomato coupe is right

I got a AS degree and worked in electronics back in the dark ages 75 to mid 80s. I'm not sure as I haven't checked yet but it is more then likely even if you chase down the shorted chip its is no longer available

Go ahead if it makes you feel better, find the shorted chip - but don't be surprised when you find out it hasn't been manufactured for quite some time and not available at any price

Mcgyver
01-16-2018, 08:45 PM
Thanks for all the ideas guys. I'll see what I can do.


Sorry to say it guys but tomato coupe is righte

I hear you, but really it depends. If its a cap I might have a fighting chance and as I said, being on the main board it may be scrap and above my abilities. otoh all this home shop stuff is for kicks, I'm learning and having fun for small to no dollars so I'll go a bit more.... when that stops i'll buy one. :)

Keep in mind that whats $200 on ebay might be 450-500 by the time its in my shop. Add in international shipping, exchange rate, HST, pain the ass carriers service charge for clearing customs and collecting hst. So for 500 its worth asking a few questions and poking about (without getting too carried away)

Mcgyver
01-16-2018, 08:50 PM
J Tiers has posted allot of good info. If your volt meter is good enough with the power off check the 5 volt supply pins right at each chip with an ohm meter, might uncover an area to look at.

I'm not 100% understanding what to do or look for. Isn't the 5V supply on the same net while ground is ground - i.e. would i get either the same resistance or voltage anywhere in the circuit from a 5V supply pin to ground?

I'm wishing I have heat sensitive camera - that makes a lot of sense

polaraligned
01-16-2018, 09:49 PM
Sorry to say it guys but tomato coupe is right

I got a AS degree and worked in electronics back in the dark ages 75 to mid 80s. I'm not sure as I haven't checked yet but it is more then likely even if you chase down the shorted chip its is no longer available

Go ahead if it makes you feel better, find the shorted chip - but don't be surprised when you find out it hasn't been manufactured for quite some time and not available at any price

You are assuming it is a shorted chip. Certainly may not be. Might even be the power supply not being able to supply the current needed to power the board. I have repaired a number of HP equipment over the years and the only one that I ever need a chip for was in a HP 8601A sweep generator. Thankfully, the output hybrid chip was notorious for failing in it so a guy had reverse engineered the original and was selling replacements. The replacement worked perfectly. Often you can find people parting out HP equipment. I have an HP 3456A which will measure down to milliohms and it is perfect for finding shorts- if that is what you have, but at this point I would not make that assumption.

J Tiers
01-16-2018, 10:20 PM
I'm not 100% understanding what to do or look for. Isn't the 5V supply on the same net while ground is ground - i.e. would i get either the same resistance or voltage anywhere in the circuit from a 5V supply pin to ground?

I'm wishing I have heat sensitive camera - that makes a lot of sense

When there is current flow, there is voltage drop. maybe not very much but it is there. The more current the more drop. You presumably have lots of current going somewhere (or a bad P,S.), so you should find a voltage drop that drops more some places than others.

That is one way of finding the bad part, measure the drop in the 5V line, from the 5V input to the board, to any other place where 5V is on the board. + on the input pin, negative on the other that you check with, set to read DC millivolts. Odds are some area will show up noticeably lower, and likely the problem is there.

PStechPaul
01-16-2018, 11:07 PM
I was able to download the PDF and use Adobe Reader to 300% and still read most of the text. But it looks like it uses some rather old school microprocessor with two 16k x 8 EPROMS (or maybe mask ROMs), and two 128 x8 RAM chips. The manual is from 1980, and the processor is likely to be an Intel 8080 which was introduced in 1974:
https://en.wikipedia.org/wiki/Intel_8080

You might look at a counter with similar specs but using old school simple logic components. I have a HP 5316 that I bought for $55, and it does the job.
https://www.ebay.com/itm/Hewlett-Packard-HP-5316A-Universal-Counter/372183696602?epid=1303899486&hash=item56a7e084da:g:kgYAAOSwVNxaTT4Y

https://i.ebayimg.com/thumbs/images/g/kgYAAOSwVNxaTT4Y/s-l225.jpg

lakeside53
01-16-2018, 11:18 PM
If it has eproms.. good luck on them retaining their contents for 30+ years.

darryl
01-17-2018, 02:24 AM
Not sure if it's been mentioned, but have you checked out the regulators- if they even look browned at all, they are suspect along with their associated capacitors. There may not be a failure in the counter circuitry except for a burned regulator that supplies it. I've found several times where a part- often a regulator- is not on a good enough heat sink. While they supposedly have over current and over heating protection, that's no guarantee that the high heat abuse will not damage it. I'm not sure how much current would be drawn by a typical frequency counter circuit, but perhaps it is substantial. And in turn there could be another part that normally runs hot, perhaps one of the chips. My guess is that you've already found which parts do run hot, and need to know if that's an error or if it's normal.

ptjw7uk
01-17-2018, 04:08 AM
Could check this out :- https://www.eevblog.com/forum/testgear/hp-5335a-timer-counter-anything-i-should-know/

They say about shorted tants in the power supply ( post 13)

Peter

Mcgyver
01-17-2018, 10:30 AM
Could check this out :- https://www.eevblog.com/forum/testgear/hp-5335a-timer-counter-anything-i-should-know/

They say about shorted tants in the power supply ( post 13)

Peter

that is an awesome thread, thanks......later on in it I found a link to a hi res manual

RichR
01-17-2018, 10:43 AM
What should be 5V reads 2.2. There others seem ok.

The +5V output also supplies the +3V regulator. Did you make your measurements at the power supply board or elsewhere?

Mcgyver
01-18-2018, 03:19 PM
The +5V output also supplies the +3V regulator. Did you make your measurements at the power supply board or elsewhere?

Here's the voltages. Sorry for all periods, spaces didn't seem to work in laying it out as a table

listed..................on PS with..................on main board
.........................main board..................connected
.......................disconnected

15.7 ....................15.4...................14.86
-15 .....................- 14.89................-14.8
5.......................... 5.05....................1.2
-5.2......................-5.21................... -5.2
3.1........................3.14................... ..0.2
24.5.......................26

J Tiers
01-18-2018, 04:37 PM
Problem is likely with the 5V as you mentioned.

3V may come from it, fix 5V and see if 3V comes up.

Tantalums suspected, other sources suggest it has plenty of them.

Mcgyver
01-18-2018, 05:18 PM
Here's a shot of three caps. i know enough that the bulbous looking middle one is a tantalum cap, could the smaller ones be as well?

https://i.imgur.com/TSQbr9J.jpg

J Tiers
01-18-2018, 05:35 PM
The large blue is tantalum.

The small blue above should not be.

The greenish axial is not.

Board looks old enough to have some bad solder joints on it.

polaraligned
01-18-2018, 05:35 PM
The top one probably isn't, but you really need to look at the schematic and parts list and make sure you to go thru this board properly and replace with same components.

polaraligned
01-18-2018, 05:38 PM
I hope you know how to desolder and solder properly as it is easy to destroy a board with bad skills.

Mcgyver
01-18-2018, 05:40 PM
The top one probably isn't, but you really need to look at the schematic and parts list and make sure you to go thru this board properly and replace with same components.

I hear you....not sure if my brain is up for the level of tedium with component #'s not marked on the board :( I was going to follow Tomatoes idea, clip that tats and see if the voltage comes back up but I wanted to restrict the clipping to the most likely suspects. If it works, then I agree, I have to go through the tedium.....


I hope you know how to desolder and solder properly as it is easy to destroy a board with bad skills.

Reasonably well, and I know the traces on old boards sometimes aren't up a de/re soldering, but whats the alternative?

J Tiers
01-18-2018, 05:53 PM
99% or more of items like the top one will be a radial ceramic (possibly film) part. I've used a ton of them in products. We bought from a company whose parts were tan, which is the only reason I'd wonder, this is a different company.

Focus on the ones that look like the lower blue one. If they do not solve it, then think about it some more.

Tab=ntalum capacitors are a form of electrolytic capacitor. One thing about ANY electrolytic capacitor is that it will have a mark to indicate which lead is + and which is -.

No mark means it is either not electrolytic, or it is an aluminum non-polar electrolytic (internally two opposite polarity parts). Either way it is not going to be between + and ground.

PStechPaul
01-18-2018, 07:14 PM
You might try putting a 1 ohm resistor in series with the 5V supply to the board. That will tell you if the 5V supply is shutting down or if the board has a short or high current draw, and you can read the current. It might have an overvoltage crowbar (SCR) that is triggered. Since you read some voltage on the 5V bus at some point, you may be able to trace the fault to where it shows much lower voltage. You may need to find places on the board where GND and +5V tracks are located, and mark them.

[PS] For tables, it helps to use a monospace font like Courier, and underlines instead of periods:

listed________on PS with__________on main board
______________main board__________connected
______________disconnected

+15.70________+15.40______________+14.86
-15.00________-14.89______________-14.80
+5.00_________+5.05_______________+1.20
-5.20_________-5.21_______________-5.20
+3.10_________+3.14_______________+0.20
+24.50________+26.00

bob_s
01-18-2018, 08:08 PM
The production date code on the chip shown 40th week of 1980 would seem to indicate that you won't have the problem of dealing with Rohs solder.

You can use an Exacto knife to carve off most of the solder around the exposed component leads, then use heat and solder sucker remove - that way minimal heating of the circuit board traces.

As Paul indicated it would be easiest to first load test each of the PS sections.

Go through the main board looking for ECL (high speed emitter coupled devices) because they typically have much higher current draw than the LS...ttl devices. Look for devices with prefix numbers 101 - 116. Prescaler, time base, A-D conversion sections would be highly stressed.

I recall that HP had the nasty habit of fusing BNC connectors in scopes and other test equipment

PStechPaul
01-18-2018, 11:32 PM
The image of the IC on the PCB shows that the leads may be corroded. Perhaps the board was exposed to dampness and salt or other corrosive atmosphere. It might not hurt to scrub the board with alcohol, followed by detergent, hot water, and hot air drying. The 1.20 volts on the 5V supply indicates something less than a solid short, perhaps a leaky capacitor, so it should be easy to track down. But it's also possible that the 5V regulator is not providing the required current.

If you determine that an IC is defective, you can snip the leads and then remove each one carefully so as not to damage the board, particularly plated thru holes.

J Tiers
01-19-2018, 12:40 AM
The voltage may also be due to just plain resistance of leads and traces. PC trace can have a fair bit of resistance. And not every tantalum forms a "bolted short". Usually they are just so crummy that nothing works.

The combo of resistances of the bad part(s) and traces etc can easily cause that. If the 5V can supply a half amp, then 2.5 ohms would make about the observed voltage. And that's a pretty good short.

+1 on cutting leads of bad ICs to remove. If leads are straight and not crimped over on the solder side, cut closer to the IC body, then grab with small tweezers. Bring over the iron and melt the solder (use extra flux if you have it), then pull out the lead from the component side.

If the ends of the leads are bent/crimped over, you may have to change approaches. Cut the lead near the board, at a point it is necked down so as to fit in the hole. Then pull out from the solder side. This works in either case, actually, but when the leads are straight, it is easier to grab from above and pull out that way. And it is easier to cut them at the IC body without damaging the board or other traces.

RichR
01-19-2018, 01:35 AM
Based on the value of R7 I'd estimate that current limiting starts kicking in a little over 3 Amps.
https://s13.postimg.org/k5j4v3i1j/grabber2018_Jan19-005243.png

With that kind of current I would expect to find a little bit of heat. Feel around for a hot spot. Any ECL chips will be quite warm but run on -5.2 Volts
not 5 Volts. Exceptions to this would be any ECL to TTL or TTL to ECL translators (MC10125 and MC10124 respectively), they run on both +5 Volts
and -5.2 Volts.

Also, any chips in sockets should be carefully reseated to make sure excessive current draw isn't being caused by logic misbehaving.

polaraligned
01-19-2018, 07:03 AM
Reasonably well, and I know the traces on old boards sometimes aren't up a de/re soldering, but whats the alternative?

A desolder iron like a Hakko 808 is an excellent choice and has worked for me perfectly on many HP boards. You get the solder melted fast, and then sucked right up. Using a separate solder sucker is not quite as good as you have to heat the solder a little more to account for the time it takes to put your soldering iron down, pick up the solder sucker and get the solder sucked.

It is a given that all tantalums of this age should be replaced regardless of whether they are working currently or not. There is no other component on HP equipment that I would replace as a matter of course except tantalums. Even their regular electrolytic capacitors going back into the 60's are highly reliable and don't need replacing unless they have failed.

MattiJ
01-19-2018, 07:14 AM
..and I know the traces on old boards sometimes aren't up a de/re soldering, but whats the alternative?

Looks like glass fiber(FR-4) board with plated trough holes. About as durable as it ever gets. Unless you seriously cook the board with too hot iron or the board is already toasted next to hot components I would not expect much drama.
Phenolic paper (FR-2) is the fugly stuff I rather stay away from.

J Tiers
01-19-2018, 11:48 AM
A desolder iron like a Hakko 808 is an excellent choice and has worked for me perfectly on many HP boards. You get the solder melted fast, and then sucked right up. Using a separate solder sucker is not quite as good as you have to heat the solder a little more to account for the time it takes to put your soldering iron down, pick up the solder sucker and get the solder sucked.

.....

You do it two-handed. Iron in one, solder sucker in the other (if you do not have a PACE type workstation with both in one). heat the solder, just a little past what seems melted, to make sure to melt through the plate-thru hole, then trigger the sucker as you remove the iron.

I actually DO NOT LIKE soldersuckers to remove components. The problem is that the leads usually end up still stuck to the side of the hole by a little solder, and that can be a real pain to melt so the lead can come out.

I prefer to cut the leads, pull the leads out with solder melted, then suck the hole clean. It's a little more heating and melting, but it ends up with a LOT less pulling on the copper. If you cut and pull the leads one by one, and then go back to clear the holes, you let the copper cool off before clearing the hole, which helps it not get damaged.

As with all PCB soldering, a larger, hotter iron actually does less damage, because you can get in, melt solder, and get out much faster. Small irons take longer to heat the solder, and tend to heat and damage a larger area of the board than larger irons do.

For soldering practice, you can "build a solder tower". Solder a blob to a piece of scrap PC board. Then get another one soldered to the top of the first, without fully melting the first. Solder a third , fourth and fifth on top of those, making sure to keep the "tower" the same width as you build it upwards.

You will soon get the hang of how solder melts, and start to see the benefits of a hot iron that has the mass to heat a part of the solder to melting before the rest gets there. That is how it works for PC soldering, you can melt what you need to and be done before the rest of the trace heats up.

There are cases where a small iron is better, but used carefully, a bigger, hotter iron does a better job in a lot of situations..

MattiJ
01-19-2018, 12:14 PM
You do it two-handed. Iron in one, solder sucker in the other (if you do not have a PACE type workstation with both in one). heat the solder, just a little past what seems melted, to make sure to melt through the plate-thru hole, then trigger the sucker as you remove the iron.

I actually DO NOT LIKE soldersuckers to remove components. The problem is that the leads usually end up still stuck to the side of the hole by a little solder, and that can be a real pain to melt so the lead can come out.


Some boards are total nightmare if the holes are on small side of tolerance. But usually I get 9 out of 10 component leads completely free on first try.
One trick (besides good desoldering gun like Hakko 470+ Hakko 802) is to add plenty of fresh solder to the joint before any attempts. Flux and fresh solder helps with oxidized or otherwise nasty joints
I usually give the desoldering gun a little "wiggle" after sucking the solder out, this often helps if the leads have habit of getting stuck to walls.
And when I say "little" wiggle I mean it, purpose is not to peel of the heated trace from pcb.

J Tiers
01-19-2018, 12:55 PM
....
I usually give the desoldering gun a little "wiggle" after sucking the solder out, this often helps if the leads have habit of getting stuck to walls.
And when I say "little" wiggle I mean it, purpose is not to peel of the heated trace from pcb.

That's the best argument for clipping leads. Pull them out clean while the solder is melted, THEN clear the hole. Do what the professional techs do, they do it all day, and that has been the technique of nearly every one.

OK, if you are going to sent a part back for a failure report, you may have to do it the hard way. Otherwise, make it easy and as foolproof as possible.

It's that "little wiggle" that breaks the "through hole barrel" away from the pads, etc. Plus you never now if the "barrel" is perhaps already cracked, held in contact only by solder. Try to NEVER stress the pads and/or thru hole barrel.

MattiJ
01-19-2018, 02:36 PM
That's the best argument for clipping leads. Pull them out clean while the solder is melted, THEN clear the hole. Do what the professional techs do, they do it all day, and that has been the technique of nearly every one.

Well not for me. Too slow. :rolleyes: (If it counts I have been doing that for living for a year or two at some point )
And I have seen more non-professional "techs" that are unable to maintain their desoldering gun than I care to remember :p

But yeah, I agree that if you are working on unobtanium slightly pre-toasted circuit board its safer to clip the leads.
Done that also, sometimes it was most cost efficient to replace every logic ic on the board so you end up cutting 500 leads per board.

The Doctor
01-19-2018, 03:06 PM
If you're at the point of giving up and putting it in the bin, there is 1 last, rude & crude way to try finding the shorted component. Apply enough power to clear the short! Grab a heavy duty 5V supply, 10A or so should work. Place it's output in parallel with the onboard supply, then turn on the counter and the supplemental supply. Problem will probably quickly show itself. This is only to be done when you've given up, as you can burn traces off a board or worse with this trick. Also don't do this in your house, the stink can last for days.

Ed

MattiJ
01-19-2018, 03:09 PM
If you're at the point of giving up and putting it in the bin, there is 1 last, rude & crude way to try finding the shorted component. Apply enough power to clear the short! Grab a heavy duty 5V supply, 10A or so should work. Place it's output in parallel with the onboard supply, then turn on the counter and the supplemental supply. Problem will probably quickly show itself. This is only to be done when you've given up, as you can burn traces off a board or worse with this trick. Also don't do this in your house, the stink can last for days.

Ed

Thermal camera works miracles with this method as you can even see the current carrying tracks. Saved my day more than once when there was randomly shorted ic on board.

J Tiers
01-19-2018, 03:16 PM
If you do not want to do the clip technique, then you can do the "wiggle".

I usually desolder the hole, and then use the tip of the soldering iron to bend the lead toward the middle of the hole. If the soldering iron is a desoldering type, with the suction hole in the middle, then you can trap the end of the lead in the hole, and push it toward the center.

With luck, the solder in the hole will solidify as you do that, and the lead will not resolder itself to the thru-hole as it springs back, so the chip plus lead will pull out after you have them all done. Sometimes there is excess solder on the hole and on the lead, so that the lead gets locked-in, needing a remelting to come out.

Whoever suggested adding a bit of solder was correct, although if I have a solder flux "pen", a dab of flux will also do nearly the same thing with less mess. The "no residue" or "no clean" type flux is the best.

JRouche
01-19-2018, 03:18 PM
If you're at the point of giving up and putting it in the bin, there is 1 last, rude & crude way to try finding the shorted component. Apply enough power to clear the short!

Right... When I was working on electronics for money that was one saying that was passed around. No such thing as a short. Shorts usually become opens :) JR

J Tiers
01-19-2018, 03:22 PM
It's the "weak link" that opens, and that may be a trace, not the real problem.

I suppose if you no longer care, it makes no difference.

Mcgyver
01-19-2018, 04:55 PM
great input...i can't keep up with you guys. Yes, having a couple of the pins stick a bit has been the problem I've had with desoldering IC's using a Pace system. frustrating, so you put a bit of pressure on as you move the heat around hoping to free the pins and you end up pulling up a pad....no more! its lead clipping from now on

Let me get a block of time to work through the posts and do some tests

thanks again!

J Tiers
01-19-2018, 05:21 PM
I realize that I forgot to emphasize that when you do the "wiggle", especially with something like a Pace system, you LIFT the end of the iron a little so you are NOT scrubbing around on the pad as you move the lead around.

It's probably obvious, and maybe it was even mentioned, but it bears repeating. Failing to lift is a great way to pull off pads.

polaraligned
01-19-2018, 06:02 PM
You do it two-handed. Iron in one, solder sucker in the other (if you do not have a PACE type workstation with both in one). heat the solder, just a little past what seems melted, to make sure to melt through the plate-thru hole, then trigger the sucker as you remove the iron.

I actually DO NOT LIKE soldersuckers to remove components. The problem is that the leads usually end up still stuck to the side of the hole by a little solder, and that can be a real pain to melt so the lead can come out.

I prefer to cut the leads, pull the leads out with solder melted, then suck the hole clean. It's a little more heating and melting, but it ends up with a LOT less pulling on the copper. If you cut and pull the leads one by one, and then go back to clear the holes, you let the copper cool off before clearing the hole, which helps it not get damaged.

As with all PCB soldering, a larger, hotter iron actually does less damage, because you can get in, melt solder, and get out much faster. Small irons take longer to heat the solder, and tend to heat and damage a larger area of the board than larger irons do.

For soldering practice, you can "build a solder tower". Solder a blob to a piece of scrap PC board. Then get another one soldered to the top of the first, without fully melting the first. Solder a third , fourth and fifth on top of those, making sure to keep the "tower" the same width as you build it upwards.

You will soon get the hang of how solder melts, and start to see the benefits of a hot iron that has the mass to heat a part of the solder to melting before the rest gets there. That is how it works for PC soldering, you can melt what you need to and be done before the rest of the trace heats up.

There are cases where a small iron is better, but used carefully, a bigger, hotter iron does a better job in a lot of situations..

Clipping the leads is great if you know you are replacing the part. In the case of tantalums, that would be the case. But for troubleshooting you so often need to lift a lead of a component to test it that cutting leads is certainly not an option. A desolder iron like the Hakko 808 makes easy work of it IF you know how to use it. I think the Hakko is 60 watts, and you get in and out fast.
I use the desoldering iron with one hand and very gently pull and wiggle the component with needle nose pliers with the other hand. You can't do that with a separate solder sucker. A little practice and you will not be lifting traces. Of course, this will not work for things like IC's, but I have no problems with transistors. I have a habit of buying Tek and HP gear on E-bay and repairing them. My technique works well for me.

polaraligned
01-19-2018, 06:07 PM
great input...i can't keep up with you guys. Yes, having a couple of the pins stick a bit has been the problem I've had with desoldering IC's using a Pace system. frustrating, so you put a bit of pressure on as you move the heat around hoping to free the pins and you end up pulling up a pad....no more! its lead clipping from now on

Let me get a block of time to work through the posts and do some tests

thanks again!

Listen, if you are serious about repairing this, just remove all the tantalums and see if the short still exists. Even if it is another component causing the short, the tantalums need to be replaced anyway so you would not be wasting time.
I have a Leader LCG-400 that I need to restore, it has over 100 tantalums in it.....needless to say, that project is in the back burner....

MattiJ
01-19-2018, 06:47 PM
I have a Leader LCG-400 that I need to restore, it has over 100 tantalums in it.....needless to say, that project is in the back burner....

Fastest way to sort that is ...connect 230v ac to dc rails. Just ask how do I know :rolleyes:

Spectacular fireworks when I once accidentally mixed up the 230v connection and 0-1v input connector..
Exploding tantalums spew white hot burning fragments around. Had to put out multiple fires on my work desk, some of the fragments burnt trough FR4 and probably good 5 feet of traces vapourized :rolleyes:

Large smd tantals also explode with normal 9v battery, again fragmenting like incendiary grenade.

PStechPaul
01-19-2018, 06:51 PM
If you need to lift one leg of an IC, or an axial resistor or capacitor, you can snip it, leaving enough left on both sides to add a dab of solder to restore the connection. You can also lift individual leads of an SOIC component by heating the pad with an iron while carefully prying the lead off the pad, using a fine point stainless steel tweezer. You might practice a bit on an old computer motherboard or other junk.

You can also cut a track to isolate part of a circuit. Just use an Ex-Acto knife and cut enough to see a thin space of separation. Scratch off a little bit of the solder mask to expose raw copper, and verify no continuity across the cut. Once you have fixed the problem, you can bridge the gap with a dab of solder.

Here is an example of an ugly modification on a prototype PCB - rough, but it works. It can be cleaned up and made more presentable, but good enough for now:

http://enginuitysystems.com/pix/electronics/BMS_PCB_3720.jpg

JRouche
01-20-2018, 02:07 AM
I have a Leader LCG-400 that I need to restore, it has over 100 tantalums in it.....needless to say, that project is in the back burner....

Now I know you are a lil off. Because that sounds like a project I have and shelved it for a moment till I get some free time. And I know I am a lil off. JR

Electronics are fun when they are in your machine tools. my 10" lathe still fires on all six tubes. (is it six or four guys?). 10" tube fired lathe.

macona
01-20-2018, 03:26 AM
You can also get soldering needles to get through hole components out. They are pretty neat:

https://hackaday.com/2017/04/03/have-you-ever-tried-desoldering-needles/