Two random ebay finds, two easy repairs!

[slburris]

First up I got a Racal/Dana 5002 wideband level meter with no display on power up.  Doing some sleuthing on the net, I find a repair story from the emperor of test equipment http://emperoroftestequipment.weebly.com describing a similar problem.  He suggested a faulty lithium battery was at fault.

Cracked mine open, measured the battery and it was dead as a doornail.  Clipped it out and applied 3.6v from an external power supply and the meter started right up!  Tossed a replacement battery into this week's Digikey order.

Next up is an HP3456A voltmeter.  It worked on power up, sort of, but the readings were always at least 10% wrong. Plus the readings jumped around, not the usual last digit rolling +- 1 or 2 numbers. Ran the diagnostic tests and got an error -4.  Sigh.  I opened it up and measured the power supplies.  All ok except for the +34v which measured about 25v and it wasnt very stable.

Hmm, so I started looking the power supply over and noticed a bulging cap, C8.  Checked the schematic and it's on the output of the failed supply, bingo!  Bodged in a cap I had on hand and the meter works perfectly now and still seems well calibrated.  Tossed a 105C replacement cap into the Digikey order as well.

I love it when people toss equipment due to a failed cap or battery!  I guess I was due for some easy ones, because I've been having a run of equipment with difficult to diagnose problems.  I'll take it!

Scott

[radhaz]

There are certainly deals to be had if you are handy/lucky. I found some nice amps that only needed the output section replaced.

[saturation]

Nice.  On the 3456a what is the first 4 digits of the serial number, just before the letter "A"?

[HiTech]

That HP 3456A is a very respectable meter. If you paid very little for it and the repair was a cinch, then you got yourself a bargain. You could flip that for a decent profit.

[MarkPalmer]

I scored well on two of the 3456A meters recently from equipment recyclers.  One just needed an out of tolerance resistor replaced in a relay control, and the other only needed a good cleaning and a couple of missing button springs replaced.  Measurements on both of the HP’s compare very favorably to one another and to my Fluke 8505A meters.  The world of recycling can bring a lot of good, still serviceable equipment to the ambitious hobbyist that would have otherwise hit the bottom of dumpsters. 

-Mark-

[slburris]

The serial number prefix on the 3456A is 2512.  What does that tell you about the meter?  It was $75 including shipping.  I also have a Fluke 8505A -- I should go compare the two.  As I recall, the Fluke holds its calibration extremely well.

The Racal/Dana 5002 was $65 with shipping.

I rarely flip test equipment -- I'm still in the accumulation phase of the disease :-)

Scott

[slburris]

Here's the bodged cap.  The original was a 100uf 50v.  Most of my parts bin caps are 16v or 25v, but I found a 250v axial, and temporarily put it in.  I ordered a Nichicon 105C 100uf 100v cap as a replacement, as I prefer to be more conservative in voltage derating.

[MarkPalmer]

On much of the 70’s through ‘90’s HP test equipment the first part of the serial number gives an approximate build date.  The first two numbers are years past 1960, the second two the month.  So 2512 would be December of 1985. 

I'm not a big flipper of equipment either.  I usually just keep what I fix, someone else can sell it after I die 

-Mark-

[saturation]

https://www.febo.com/pipermail/volt-nuts/2010-November/000626.html

Its a good manufacture date.

The serial number prefix on the 3456A is 2512.  What does that tell you about the meter?  It was $75 including shipping.  I also have a Fluke 8505A -- I should go compare the two.  As I recall, the Fluke holds its calibration extremely well.

The Racal/Dana 5002 was $65 with shipping.

I rarely flip test equipment -- I'm still in the accumulation phase of the disease :-)

Scott

[MarkPalmer]

The newer 3456's got the soft rubber panel buttons that aren’t prone to becoming sticky or losing/bending the springs.  But they also lost the traditional HP quality look and feel to them (when they work right) at that point.  A trade-off I guess. 

HP did learn with the 3455A and early 3456A's that a cooling fan is not something you want to have on a lab grade multimeter.
 
-Mark-
 

[macboy]

Good job on the repairs. Always check the power supplies first.

I still maintain I have the easiest fix ever for an ebay "broken/not working" instrument. Keithley 199 5.5 digit multimeter, would not take readings on any function, just displays "------". Scored it for $25. Upon receipt, it is in excellent physical condition. Clean, no nasty stickers, permanent marker, etc. Plugged it in, and it powers on fine, but displays "------". I know exactly how to fix it because I have another 199 in my lab.

I pressed the Trigger button. Then it started to take readings. And it has done so perfectly for the past 3 years.

[MarkPalmer]

I'm sure teardowns of the 3456A have been on here before, but I had these pictures in my archive.  These meters are packed full of all kinds of goodies:

Top:


Bottom:


-Mark-

[alex.forencich]

Good job on the repairs. Always check the power supplies first.

I still maintain I have the easiest fix ever for an ebay "broken/not working" instrument. Keithley 199 5.5 digit multimeter, would not take readings on any function, just displays "------". Scored it for $25. Upon receipt, it is in excellent physical condition. Clean, no nasty stickers, permanent marker, etc. Plugged it in, and it powers on fine, but displays "------". I know exactly how to fix it because I have another 199 in my lab.

I pressed the Trigger button. Then it started to take readings. And it has done so perfectly for the past 3 years.

Oh, I have another good one.  I bought an AM503 current probe amp in a one-slot power supply frame that 'does not power on'.  Can't remember how much I paid for it, but it wasn't very much.  It came in.  I plug it in.  I go to turn it on.  Hmm, power switch feels a bit stiff.  I pull harder on the power switch.  POP.  The switch comes un-stuck, and it turns on.  A few cycles of the switch later, and it works just fine. 

Definitely a good idea to check the power supply circuitry first.  I have a whole list of stuff I picked up with power supply issues that required only one or two swapped components in the power supply.  Like 3 Sorensen 40V 13A power supplies that I picked up for $50 each.  Wouldn't go to full range.  After poking around on the control board, I found that the power suppy rail was oscillating at 120 Hz.  Hmm.  Blown cap, anyone?  Electrolytic cap had failed open.  After replacing it, the unit worked fine.  I also picked up an 85645A 300 kHz to 26.5 GHz tracking source not too long ago for $1k that was failing an ADC self test.  It came in, I openend it up, 32.5V was reading zip.  Figured it was the regulator, so I popped over to radioschmuck for some LM317s.  Replaced it, plugged it in, POP.  Whoops, forgot to check if the rail was shorted.  It was.  Removed shorted cap, replaced regulator again, and hey, it now passes the self test. 

I was given an old HP 8590L spectrum analyzer that did not power on.  Opened it up, removed the power supply, looked at it, didn't see anyting obvious, reinstalled power supply, and tried again.  It powered up.  Really??

I also picked up an HP 8340A H02 10 MHz to 20 GHz sweep generator that did not power on for about $800.  Everything depends on the 20V supply, which was not starting.  Traced that to a blown fuse caused by a shorted regulator transistor.  After replacing that, the 5v rail read zip.  Noting obviously wrong with the power supply, so I started removing boards.  I think I just pulled out four or five of the boards and then turned it on.  The rail came up.  I turned it off, reinstalled one board, and turned it on.  It powered on correctly.  I repeated that until I had reinstalled all of the boards, and it still powered on correctly.  Still not sure what was going on there.  However, it failed the self test and the level could not be controlled.  Eventually traced that back to a bad op amp in the power level reference generation circuitry.  After replacing that and making a handful of adjustments, it works perfectly. 

I also picked up an Agilent 54854 Infiniium 4 GHz 20 GSa/s scope that did not want to boot up.  Eventually traced that one to a blown fuse.  In the removable hard drive caddy.  Shorted across that one with a wire as I didn't have a replacement fuse in the correct form factor.  Ended up salvaging the calibration data off of the drive and then performing a system restore due to some windows update issues. 

Moral of the story: buy stuff that doesn't power on as it is (relatively) easy to fix power supplies. 

[dr.diesel]

HP did learn with the 3455A and early 3456A's that a cooling fan is not something you want to have on a lab grade multimeter.

I'd been thinking of clipping the fan on mine, did the later model 3456s ditch the cooling fan?

[lowimpedance]

I'm not a big flipper of equipment either.  I usually just keep what I fix, someone else can sell it after I die 
-Mark-

Good man, same here. Generally past caring when dead

I'd been thinking of clipping the fan on mine, did the later model 3456s ditch the cooling fan?

Yes they did, I have one of each but cannot remember now what component/engineering changes were done between each version.
Not sure removing your fan would be in the best interest of the meter though.
 

[dr.diesel]

Not sure removing your fan would be in the best interest of the meter though.

I will thermally profile, need to re-cap it anyhow.  Nothing a few extra heat sinks and a 4" hole saw won't fix, 

[MarkPalmer]

When the fan was eliminated the LM323K 5 volt power supply regulator was moved on to the back panel which is used as its heat sink.  Other than that there were no significant changes made by eliminating the fan from the model.  The main reason they got rid of it is the meter warms up to maximum thermal stability faster and remains stable without the fan.

-Mark-
 

[slburris]

For the HP 3456A, I finally installed a decent replacement cap, the Nichicon next to the big blue cap.



And now the repair is done!  Just have to clean off the sticker mess the previous owner left.  What's with companies plastering everything with stickers?  How about just an unobtrusive property tag on the rear?  At least this didn't have permanent marker all over it, like some equipment I've received.



Scott

[slburris]

The Racal/Dana 5002 now has a new lithium battery.  Oddly, there's no
+ or - on the silkscreen.  On my unit, there's some pencil scrawled on
the PCB, but I know better than to trust it.  The service manual has a procedure
for replacing the battery and says to put the + to the front.



Success!  It's alive again.  Unfortunately, this unit has both sticker mess and
marker mess on it.  I really hate the markers because they can be difficult to
remove without causing damage.  Still, the meter now works.



Scott