Fluke 5440B PSU Fault Check Guarded Power

[denimdragon]

I picked this up today and I hope someone can help me with this error. I'm going to start clearing out an area for disassembly in the meantime and check the voltages in the service manual (This thing is F#@%!! Heavy and Huge). Hopefully it's not too much of a rabbit hole. I'm good at tripping and falling down those for some reason 

Thanks in advance!

[wn1fju]

The rabbit hole shouldn't be too big!  My experience with fixing my 5440B is that everything is pretty straightforward.  Just methodically trace your way
through the circuit(s).  Do the obvious visual/simple inspections first (like the 25+ fuses in it) and look for bulging electrolytics, etc.  Also check that none
of the relays have fallen out of their sockets (two did on mine and wedged themselves into the motherboard).

Once you get the P/S straightened out, run through all the internal tests (after the ovens have all warmed up).  That should reveal any remaining problems.

The 5440B is a fantastic piece and yours looks like it has a nice bright display.  Enjoy!

[denimdragon]

The rabbit hole shouldn't be too big!  My experience with fixing my 5440B is that everything is pretty straightforward.  Just methodically trace your way
through the circuit(s).  Do the obvious visual/simple inspections first (like the 25+ fuses in it) and look for bulging electrolytics, etc.  Also check that none
of the relays have fallen out of their sockets (two did on mine and wedged themselves into the motherboard).

Once you get the P/S straightened out, run through all the internal tests (after the ovens have all warmed up).  That should reveal any remaining problems.

The 5440B is a fantastic piece and yours looks like it has a nice bright display.  Enjoy!

Thanks for the tips, it's greatly appreciated!

I've checked the voltages on the A10, A17, and A19 PCA boards. The A10 PCA has zero volts on test points 12 and 13. Also I checked the A8 PCA and there was zero output on the A/D pins. I'm going to study the schematics and make sure my hunch is correct that the TP12 and TP13 pins on the A10 which should be +30V and -15V respectively, are the source for the A/D output.

Here are the voltages I measured from the 3 PCAs:



I also took a look at some of the components and did a partial microscopic visual inspection of the A10 with about approximately 50-60% coverage. I notice that someone has done some work on this unit in the past. (I'll post those after I figure out how I can insert a photo that's clickable to larger more detailed .jpg files on my server.)

I found that the Hex inverter, Multivibrator, a couple of 1N4002 diodes, and 2 voltage regulators had flux and tamper evidence around the pads. Judging from the date codes on the ICs it looks like this was an early repair (Cerca 1987-ish). I also discovered a bad 1N4002 diode (CR5) which was conducting in both directions. I pulled it off the board for a more accurate evaluation. I also didn't have any 1N4002s in stock. I'll be looking into a closer evaluation of the failed component later today.

[denimdragon]

I have found a few things that need replacing on these boards. The A5 Output/HV Control PCA has one crispy zener diode I need to replace and another one in an area that seems suspect. My question here is can these zeners be replaced with other zener diodes with the same values? The burnt zener is UZ8706 and the other is UZ8712. I have a NOS Motorola 1N4736A I want to use to replace the UZ8706 and the specs are very similar, but there are some differences. Can someone tell me if it's ok to use the 1N4736A as a replacement?

I also found swollen caps on the A9 REF/DAC Analog PCA. I'm going to replace those as well and hope this will get rid of the fault. What I'm really concerned about is the crispy zener. I'll have to hunt down what caused the issue to see if I can fix that. I would hate for that to happen again and damage something else.

I appreciate any help. The pictures below are of the zeners and also snippets from the datasheets to see the differences in the electrical characteristics.

[denimdragon]

These 6.8uf 35V elec caps were on the A9 REF/DAC Analog board and I'm looking to replace them. The "IC" branded ones are Illinois Capacitor I'm assuming and they had some other brand I've never heard of called "marom"? I'm looking to replace them but I'm not sure if ESR (low ESR or general purpose) is a factor. Also how much would it effect the machine if I used 50V instead of 35V. Can anyone chime in with some suggestions. Thanks!

[wn1fju]

I would just go with good 105 deg rated general purpose, good quality name brand (Nichicon, Panasonic, etc.) electrolytics.  And 50VDC is
fine.  It looks like these 6.8uF caps are just used as extra filtering on the various DC lines coming onto the board.  You could probably omit
half of them and you wouldn't even notice!   And the reason I'm suggesting 105 deg caps is because there is an oven nearby on the board.

[denimdragon]

I would just go with good 105 deg rated general purpose, good quality name brand (Nichicon, Panasonic, etc.) electrolytics.  And 50VDC is
fine.  It looks like these 6.8uF caps are just used as extra filtering on the various DC lines coming onto the board.  You could probably omit
half of them and you wouldn't even notice!   And the reason I'm suggesting 105 deg caps is because there is an oven nearby on the board.

Good to know I'm thinking correctly. I had decided on 105°C caps when I started removing them from the board due to the closeness to the oven. Thanks for your help. Now I can put the order in to Mouser. Over time, I'm going to recap the entire unit, but one board at a time. That way I can see the end result.

Another question, other than a 3458A and a 732B, what can I use to verify these beast with? 

[TheSteve]

I'm also working on a 5440B and that same 6.8 volt zener is crispy(along with the pcb around it). I suspect it is a design flaw more then anything. I don't think we're the first to experience the problem. I am upgrading mine to a 1 watt zener, but have been running the unit a full week with the crispy one in place. The burn occurred so long ago that the PCB doesn't even have a burnt smell anymore.
These are amazing units to work on, a big heavy box of fun.

[wn1fju]

You asked, "other than a 3458A and a 732B, what can I use to verify this beast with?"

That question is better addressed to the metrology forum where you are sure to get N opinions from N replies!  It is far from trivial, since
the word "verify" has all types of meanings, e.g., accuracy, linearity, drift, repeatability, etc. 

I don't have either a 3458A or 732B, but I have a slew of 6-1/2 digit DMMs and it was very easy with them to see that my 5440B was at first not
acting up to par as it would drift almost 1 mV when set to zero and acted "differently" every time I turned it on.  The problem was traced
to dirty relays and I eventually ended up replacing all of them! 

[denimdragon]

I'm also working on a 5440B and that same 6.8 volt zener is crispy(along with the pcb around it). I suspect it is a design flaw more then anything. I don't think we're the first to experience the problem. I am upgrading mine to a 1 watt zener, but have been running the unit a full week with the crispy one in place. The burn occurred so long ago that the PCB doesn't even have a burnt smell anymore.
These are amazing units to work on, a big heavy box of fun.

And it still works as designed with the crispy zener? Wow these things are amazing! Also, in the service manual it says that these are 1 watt zeners. Maybe we need something bigger. Keep me posted on the progress or else I'll bug you when I'm on TiNs site (if it's the same "The Steve")lol.

You asked, "other than a 3458A and a 732B, what can I use to verify this beast with?"

That question is better addressed to the metrology forum where you are sure to get N opinions from N replies!  It is far from trivial, since
the word "verify" has all types of meanings, e.g., accuracy, linearity, drift, repeatability, etc. 

I don't have either a 3458A or 732B, but I have a slew of 6-1/2 digit DMMs and it was very easy with them to see that my 5440B was at first not
acting up to par as it would drift almost 1 mV when set to zero and acted "differently" every time I turned it on.  The problem was traced
to dirty relays and I eventually ended up replacing all of them!

I'm glad you mentioned the Relays. I need to make sure to clean them once I replace any more bad components. I have a 3456A, and a 3457A meter that I can use for now. In particular, the 3456A (I love that machine). In the next month or so I hope to pick up another meter that's better than my 3456A. I'm currently waiting for my Mouser order to come in so I can get this show on the road. 

Thanks guys for all the input.

[TheSteve]

I have my 5440B working well now. Now that I've said that something else in it will probably fail.

Anyway the 6.8 volt zener that has gone all crispy on both of our A6 boards should not have failed. It is across the 5 volt supply that powers the relays. There is little to no chance that the voltage was too high as the supply has a crowbar circuit that will fire and blow the fuse long before it gets to 6.8 volts. That leads me to believe the zener's have been weakened by the back EMF from the relays and eventually started failing or perhaps Fluke installed the wrong value from the factory. So after it burns and eventually fails open that is why the unit will keep operating just fine. I suspect mine has been that way for many years as the burnt board no longer had any smell too it. I have since cut out all of the burnt portions, ran 3 jumpers and added a new diode.
The biggest pain has the been the relays. I opened each one and cleaned the contacts with 99% alcohol and then Deoxit D5. I also cleaned the sockets/pins with D5. After that the 5440B had a collection of faults. I ended up finding a few relays that had less then perfect contact(tested resistance of every single one, opened/closed etc). However the biggest problem was that one relay had stopped cycling properly, it was stuck internally. I suspect my problem was never dirty relays as the unit was so clean inside you could eat off of it. It was most likely a stuck relay all along. I got that relay unstuck and had been running the 5440B for a few days without any faults. Then tonight I had another fault - turned out a different relay had stuck in the exact same way. I am going to have nightmares about American Zetler relays!

Anyway, if you think I can help at all feel free to drop by TiN's chat and we can work through the problems. There are others there that have also fixed a 5440B before so we'll have you covered pretty good.

[denimdragon]

I have my 5440B working well now. Now that I've said that something else in it will probably fail.

Anyway the 6.8 volt zener that has gone all crispy on both of our A6 boards should not have failed. It is across the 5 volt supply that powers the relays. There is little to no chance that the voltage was too high as the supply has a crowbar circuit that will fire and blow the fuse long before it gets to 6.8 volts. That leads me to believe the zener's have been weakened by the back EMF from the relays and eventually started failing or perhaps Fluke installed the wrong value from the factory. So after it burns and eventually fails open that is why the unit will keep operating just fine. I suspect mine has been that way for many years as the burnt board no longer had any smell too it. I have since cut out all of the burnt portions, ran 3 jumpers and added a new diode.
The biggest pain has the been the relays. I opened each one and cleaned the contacts with 99% alcohol and then Deoxit D5. I also cleaned the sockets/pins with D5. After that the 5440B had a collection of faults. I ended up finding a few relays that had less then perfect contact(tested resistance of every single one, opened/closed etc). However the biggest problem was that one relay had stopped cycling properly, it was stuck internally. I suspect my problem was never dirty relays as the unit was so clean inside you could eat off of it. It was most likely a stuck relay all along. I got that relay unstuck and had been running the 5440B for a few days without any faults. Then tonight I had another fault - turned out a different relay had stuck in the exact same way. I am going to have nightmares about American Zetler relays!

Anyway, if you think I can help at all feel free to drop by TiN's chat and we can work through the problems. There are others there that have also fixed a 5440B before so we'll have you covered pretty good.

That is awesome information! I'll be sure to focus on getting those relays spotless and making sure I clean the boards good. I'm taking my time with it and hope to have it done before the end of the month. It's amazing that these things are the tanks that they are. It sucks that most companies don't build with this type of quality anymore, but I guess for the price of the 5440B new, the quality is to be expected. I love the way the system is laid out as well. Very easy to understand for a rookie like me, once I got over the overwhelming amount of boards in it. I'm learning a lot about circuitry in general from this unit.

[99tito99]

Hi denimdragon:

Your table as does the Fluke 'Table 3-4' shows the TP4 at +28.0, but the Fluke schematic 'Figure 7-18. A17 Outside Guard Regulator PCA' shows the TP4 on an UNREG +24 volt rail.  Is this a typo in the table?  This came to my attention when I tried to adjust R18 to get +28, but the pot bottomed at only +27.
Cheers,
Mark
***********************
PS: The TP4 on the board is labled +28.8.  +28.0 or +24.0, what am I missing?

[Scopetechniques]

I was changing the 0.150 ohm resistor on A17 which had crept up to over 0.3 ohms and was causing a slew of problems.
I replaced it with a new resistor from Digikey (P/N ALSR5J-.15-ND).

I also noticed the same issue with TP4 on A17.  The schematic shows that it should be 24V at TP4.
However the checkout section says to adjust to 28V, but it will not go that high.

I looked at the LM317HVK circuit and datasheet, with the two resistors and trimmer in that circuit the LM317HVK will not go to 28V.
So I assume the 28V note should say 24V.

--Victor

[denimdragon]

Just for the record, I never had time to finish this project  School has been consuming all me time. I decided to let this one go as bad as I wanted to repair it. If anyone is interested PM. I just listed it on ebay a couple of hours ago.

[TheSteve]

Just for the record, I never had time to finish this project  School has been consuming all me time. I decided to let this one go as bad as I wanted to repair it. If anyone is interested PM. I just listed it on ebay a couple of hours ago.

So what is the "little bit better" item you've replaced it with?

[denimdragon]

Just for the record, I never had time to finish this project  School has been consuming all me time. I decided to let this one go as bad as I wanted to repair it. If anyone is interested PM. I just listed it on ebay a couple of hours ago.

So what is the "little bit better" item you've replaced it with?

Hopefully a 732A that works but it's not in stone yet. One of the local shops I do repairs for on the side are looking to upgrade to a new Fluke 732C? Not sure if this the right nomenclature but it's the newest standard/s and a new Calibrator. I usually get there old stuff pretty reasonably priced. However, now you are making me second guess myself.

I wanted to keep my 5440 but school has been mauling me latelyand I can't stand to see it sit up and go to waste. I know I shouldn't be asking this, but what do you think? The 732A is better right??

[TheSteve]

As a 10 volt (or 1.018) volt reference the 732A is better, but it does a lousy job of delivering other voltages
The 732A should sit quietly and run 247. The 5440 is a very nice DC volts calibrator, not something you likely want to leave running 247. They both technically have similar internal references. The 5440B uses two of them in series. The 732A has one ref and a boost network to get to 10 volts.
So in the end I guess it depends on what you're looking for.

[Tony_G]

The problem was traced to dirty relays and I eventually ended up replacing all of them!

Out of interest where did you get the replacements and, if I may, what did they cost you?

I have 2 5440's (B & AF versions) and I need to replace the relays in them for the same reason you did.

Thanks,

TonyG

[picburner]

I replaced only one relay in my F5440B with the original type (the AZ431-17-205) and it cost me a fortune: $95 + $40 shipping + customs fee.
Replacing them all would be a very crazy expense!

[TiN]

Who am I to say, but since you mentioned it, I would never let boxes like 5440B go, unless you really have an emergency. There are not much alternative (reasonably priced) options to get precise up to 1kV DC programmable source with good linearity. 732A is a support not, not a replacement for the calibrator. You cannot use 732A to calibrate DC function on any meter, except maybe 3458A.

Another benefit of oldie moldie 5440B is that many volt-nuts here have them and can help with repairs/support if you need so. For example I even bought few 5440B boards (don't have 5440B) just for sake of taking magic parts out of them to do experiments. Relays that are on them (DC ref board assembly and reference string) are no use for me, so there are ways to get them if your unit need parts .

And 732A can be a lengthy and deep rabbit hole on it's own like unit here, that had popcorn noise spikes on the output, which I don't like. Already spent dozen of evenings checking resistors, cleaning parts (oven assembly is PITA to get inside!) and ended up replacing SZA263 zener chip. Now it sits silly uncalibrated, as few hundreds more hours need to be spent to get it into shape...   

[denimdragon]

As a 10 volt (or 1.018) volt reference the 732A is better, but it does a lousy job of delivering other voltages
The 732A should sit quietly and run 247. The 5440 is a very nice DC volts calibrator, not something you likely want to leave running 247. They both technically have similar internal references. The 5440B uses two of them in series. The 732A has one ref and a boost network to get to 10 volts.
So in the end I guess it depends on what you're looking for.

I knew I shouldn't have asked. It was already hard enough to list it 

Who am I to say, but since you mentioned it, I would never let boxes like 5440B go, unless you really have an emergency. There are not much alternative (reasonably priced) options to get precise up to 1kV DC programmable source with good linearity. 732A is a support not, not a replacement for the calibrator. You cannot use 732A to calibrate DC function on any meter, except maybe 3458A.

Another benefit of oldie moldie 5440B is that many volt-nuts here have them and can help with repairs/support if you need so. For example I even bought few 5440B boards (don't have 5440B) just for sake of taking magic parts out of them to do experiments. Relays that are on them (DC ref board assembly and reference string) are no use for me, so there are ways to get them if your unit need parts .

And 732A can be a lengthy and deep rabbit hole on it's own like unit here, that had popcorn noise spikes on the output, which I don't like. Already spent dozen of evenings checking resistors, cleaning parts (oven assembly is PITA to get inside!) and ended up replacing SZA263 zener chip. Now it sits silly uncalibrated, as few hundreds more hours need to be spent to get it into shape...   

Well it looks like it will have to sit until I make time for it. Thanks guys for wisdom in this area. I'm glad no one bought it 

In fact, I'm going to start working on it again at the beginning of June. I'll have a few days I can free up then.

[denimdragon]

Here we go again  ...

So, my 3456A started acting a bit weird with voltage readings off about -.150ish mV and I thought it might have been something I pressed in Math that did this, but I was unable to resolve the issue which means, either use my 3457A meter which is as slow as a snail, or fix the 3456A. I have exactly one week before I'm back in class full time so that means this week has to be productive in getting my lab together before school starts. Now that I have to repair "Big Shirley" (my pet name for my 3456A), I figured it's a good time to finally commit to fixing the 5440B for good. I can't afford to put it off any longer, as this situation has proved the need of the 5440B if I'm going to have any kind of semi-decent electronics lab and metrology lab. Procrastination bites me in the @$$ yet once again 

I have been working on the 5440B off and on so I  have resolve the original "Check Guarded Power". What's crazy about this is that on the A12 Filter A Board CR10 and CR11 Rectifiers were missing! I replaced those then I got my -15V and +30V rails back. After that, I powered it up and a new error surfaced.

"Guard Communication - Fault Check Garbled Data"

This error popped up so I went to the A14 Guard Crossing Board. I checked all the Optos and they were good, so I replaced the 4 electrolytic caps with new ones, replaced the board and fired it up. Guess what... Yet another error, but at least I'm making progress right?!

"DAC Digital - Fault Check A to D"

When this came up I got a bit worried, hoping that the DAC/ADC are ok. I removed the A8 DAC/REF Digital board and replaced the single electrolytic cap with a new one. Inserted the board back into the 5440B and received another error. I felt a little better since the DAC/ADC seemed not to be the problem.

"Inside Guard - Fault Check Guard CTRL Bus"

And this is the current state of affairs. I removed the A8 DAC... Digital board again because this time I used the manual method of removing each board until the error no longer occurs as instructed by the service manual. I'm starting to think that I need to replace all the electrolytic caps in order to get it up an running. I was going to wait until a later date but it seems I don't have a choice if I want to rule that out as being the case. I've already cleaned the relays with card stock and deoxit as suggested here.

Another interesting thing is when the DAC-Digital board is in, I lose my -15V and +30V on A10 board and actual the +30V goes to 8ish Volts and the -15V goes to -.284mV. When the DAC board is out, those voltage return to there normal value -15.011V and +28.9V. I assume this means that the problem is definitely on the DAC board but I'm not sure, and the issue with that is if a Logic IC is bad, how do I check it?

Another point... When I had all the boards in and tested the +5LH voltage, it's a small - voltage (-100-200ish mV). I know this rail goes to the logic on the A8 board, but I'm wondering if there something along the path that shorted to ground. I'm just not sure what would be the next move other than replacing all the caps, which I will do anyway. Any help anyone?

[dj831]

Hi denimdragon, not sure if you had success in repairing your 5440B. I'm ending in repairing mine, bought 2.5 years ago - no time, but this is something I would have never let go  . Issue of mine was just that .15R resistor on guard that was ... desoldered!

BTW, I had weird issues like yours, after repairing my unit. When it was cold, I had all sorts of errors. There was even a weird noise like a quiet sounding buzzer  . My 5440B had a bunch of these yellow Nichicon capacitors, but two had turned to brown on regulator board (I think that this calibrator ran for week/months or even years without being powered off...).

I've replaced all these capacitors by +125°C on all boards (c.a. the same price than +105°C), and I even increased values (most if not all 22uF were replaced by 47uF - exact same footprint). Since, no more issues, calibrator runs perfectly even if cold (apart stability, of course). I've checked stability after several hours against a 8505 from eBay ($50) and @ +10V, there is a difference of 10ppm between both. And both are rock stable  (of course, this is not a proof in itself).

Note about relays: I had issues too with relays, especially with smallest ones (2 form C). I cleaned contacts without any product, apart with very soft sandpaper (grain of 5000). For other relays, including sanded ones (total of 25 IIRC), I made a small board with an NE556, one half operating @ c.a. 7Hz and the other half being a monostable for c.a. 2 minutes. A kind of self-cleaning. That trick worked like a charm (same trick used with attenuator boards of Tektro TDS 700 series).

I've since dumped nearly all devices - PALs on display controller PCA are readable - and an interesting thing, my calibrator has software v2.1 (most 5440B I've seen are fitted with software v2.0). As I'm in Europe and have the habit of 6 and 9 with tails, I edited EPROM U61 to change display font (reason why I've always hated 4511's ) I will make a separate post about repair and mods of my calibrator when everything is solved (the four remaining things: being able to read 68701, solving 2x20 VFD issue - mine is readable, but not new, replace both leaky output 1uF capacitors, and add four feet).

[picburner]

Here is the dump of U3 (MC68701L).
If you can, post your firmware version 2.1 and the PAL dump on the KO4BB website (http://www.ko4bb.com/getsimple/index.php?id=manuals).
I did it time ago but my 5440B firmware was the old 2.0 version.