Fluke 5440B repair

[Rax]

I landed a 5440B at my bench that's exhibiting some confusing symptoms.

The unit came to me in pristine form - Fluke original stickers everywhere, original manuals and cables. Even the registration form was enclosed.

The unit exhibits no signs of life, other than the green power LED lighting (next to the turn on knob). Nothing else seems to be alive, including the fan. I opened the unit (tearing all those cal stickers...) and checked the rails, both unreg and reg on A10 and A12, and they seem reasonably close. Definitely "alive."

Upon checking A19 though, there's no DCV present on it. Which makes some sense, given the fan is also off (they feed off of the same 120V VAC line). Upon checking for continuity between J2 (13/26; 14/28) and J8 (12/24; 11/23), I see that the lines on the motherboard seem to be OK.

Other checks:

• I see 116VAC between 9F and 9S on A11
• The line switches (S1, S2, S3) seem to show correct continuity between their poles
• I measured all "full winding" taps of the GUARDED XFMR and they all show reasonable ohmic values

I'm not sure how to explain the presence of the line VAC between 9F and 9S, and no VAC at the fan and UNGUARDED XFMR... Particularly as CR2 is lighting. Not sure if I should suspect the E1 spark arrester or what else could be going on here.

Thanks for all input.

[picburner]

with the 120V mains one reason for the failure of the instrument and the fan to turn on could be one of the switches s1-s3 which has/have false contacts. Although you have measured them (in circuit) the presence of the transformer windings may have distorted the measurement. Try exercising the three switches several times.
The green LED CR2, connected to a secondary winding of the transformer (by cr1/r1), could light up dimly even with a small flow of current and this contributes to confusion.
Check that s4 is also working properly of course.

Also, to be safe, I would also check the fan itself by unplugging it from the instrument.

[leighcorrigall]

You probably want to start by reviewing the following documentation from the service manual:

Overview:
Figure 3-1 module locations

Fan-related diagrams:
Figure 2-7 power supply simplified schematic
Figure 7-6 A3 (motherboard)
Figure 7-14 A11 (guarded transformer termination)

Specific to the outside regulation that is connected with the fan:
Figure 7-20 A17 (outside guard regulator)
Figure 7-22 A19 (outside guard terminator)

Assuming the fan is working and unable to receive power, you should look at the switch (Figure 5-25), the mains input and guarded transformer (Figures 7-6 and -14), and all of the fuses (Figure 7-14).

EDIT: Believe it or not, you should also remove the fan filter. Sometimes it gets stuck in the fan and prevents free movement. I have seen a video somewhere on YouTube with this problem. Might as well rule out the really simple things first.

EDIT: If you are to remove the front panel to access the switch, pay extra attention to the DIP plug. It is extremely fragile and should be handled with great care. The front panel can be removed after both clam-shell covers are removed and the rack handles.

EDIT: To remove A19, the fan must be removed and both of the large mounted capacitors. The large filter capacitor should eventually be replaced as it can cause overheating of A17 and A19 boards, resulting in a 'check guarded power supplies' fault. C7 36DX343G015AF2B (CAP, AL, 34000uF, +/-10%, 15V) should be replaced by CGO343M020L or MAL210116333E3. I use the Vishay capacitor even though it is much shorter than the original.

[leighcorrigall]

Attached is a picture of the guarded transformer (A11) with a lot of the fuses.

EDIT: This module is connected to the frame from the adjacent side with a set of screws. You will have to remove the clam-shell covers to access them.

[Rax]

You probably want to start by reviewing the following documentation from the service manual:

Overview:
Figure 3-1 module locations

Fan-related diagrams:
Figure 2-7 power supply simplified schematic
Figure 7-6 A3 (motherboard)
Figure 7-14 A11 (guarded transformer termination)

Specific to the outside regulation that is connected with the fan:
Figure 7-20 A17 (outside guard regulator)
Figure 7-22 A19 (outside guard terminator)

Assuming the fan is working and unable to receive power, you should look at the switch (Figure 5-25), the mains input and guarded transformer (Figures 7-6 and -14), and all of the fuses (Figure 7-14).

EDIT: Believe it or not, you should also remove the fan filter. Sometimes it gets stuck in the fan and prevents free movement. I have seen a video somewhere on YouTube with this problem. Might as well rule out the really simple things first.

EDIT: If you are to remove the front panel to access the switch, pay extra attention to the DIP plug. It is extremely fragile and should be handled with great care. The front panel can be removed after both clam-shell covers are removed and the rack handles.

EDIT: To remove A19, the fan must be removed and both of the large mounted capacitors. The large filter capacitor should eventually be replaced as it can cause overheating of A17 and A19 boards, resulting in a 'check guarded power supplies' fault. C7 36DX343G015AF2B (CAP, AL, 34000uF, +/-10%, 15V) should be replaced by CGO343M020L or MAL210116333E3. I use the Vishay capacitor even though it is much shorter than the original.

Thank you, Leigh. Very thoughtful, insightful input.

I am pretty intimately knowledgeable of the 5440 series: https://www.eevblog.com/forum/metrology/fluke-5440a-datasheet/, which I think you've seen.

I am currently suspicious of the card edge connectors (bad contact?...) and maybe the E1 spark arrestor. Also, picburner's recommendation to closer examine the line switches. I may either condition them and try again on board, or pull them off the board and properly test them.

I suspect once I figure out why the 120VAC line going to the outboard supply and fan is fixed, the unit may just go back to operation. But why that's offline is currently a mystery to me.

[leighcorrigall]

Hi Rax,

I just thought I would give you some ideas to rule out the trivial solutions and perhaps provide advice for others who have not had much experience tackling the Fluke 5440B before. Your introduction post sounded like you just opened one of these devices for the first time. Sorry about that.

If you suspect that the card slots are dirty, then try the following:
- Use a paintbrush to loosen the dirt, then apply a blower (not compressed air as it has oils)
- Find a stick of the same width of the card
+ I have heard people using erasers, but I have not tried this approach before
+ If a stick is not easy to find, perhaps just fold a piece of paper until it becomes the correct thickness
- Fold some printer paper around it
- Apply a solvent like isopropyl alcohol or a contact cleaner
+ I use a spray bottle to reduce waste and contamination
- Run it in and out of the affected slots until no more dirt can be seen, then keep going for a while just for reassurance

- Make sure to clean every contact on the cards with a Q-tip or paper towel before reinsertion
+ The fibres and hairs that may get caught along the trace corners can be removed with a brush or tweezers

I know that the instrument will not respond or display anything without A11, A17 and A19, which seems to be your problem. I would start there for the slot cleaning. I do not recall if the green light turns on when the cards are pulled. My instrument was really dusty when I first got it, especially on the bottom where the motherboard is mounted. While you are there, there are some shielded cables that you could examine. I think one is for the power button, but I don't remember for sure. Perhaps there is a bad solder joint, or something got torn -- hard to believe, though, given that no one ever wants to go back there.

The green light being on seems odd to me. Maybe there is something to that.

If you have another 5440B that you do not mind opening, I would suggest swapping those cards, including the transformer, to see if you can isolate the problem.

In any case, I would pull every single card from the 5440B to give the mainboard a thorough examination. I know that some people with 5700As have had mainboard burnout traces before. That would be an ugly repair, but it is possible.

Regards.

[Rax]

Hi Rax,

I just thought I would give you some ideas to rule out the trivial solutions and perhaps provide advice for others who have not had much experience tackling the Fluke 5440B before. Your introduction post sounded like you just opened one of these devices for the first time. Sorry about that.

Absolutely no problem, Leigh! I meant to manifest appreciation for your very detailed, 100% comprehensive, and extremely thoughtful post, while stating my prior exposure to this unit group. I am also always striving to close the loop on any one of my repairs, so anyone coming behind me would have a clear depiction of the symptoms, repair endeavor and hopefully solution.

If you suspect that the card slots are dirty, then try the following:
- Use a paintbrush to loosen the dirt, then apply a blower (not compressed air as it has oils)
- Find a stick of the same width of the card
+ I have heard people using erasers, but I have not tried this approach before
+ If a stick is not easy to find, perhaps just fold a piece of paper until it becomes the correct thickness
- Fold some printer paper around it
- Apply a solvent like isopropyl alcohol or a contact cleaner
+ I use a spray bottle to reduce waste and contamination
- Run it in and out of the affected slots until no more dirt can be seen, then keep going for a while just for reassurance

- Make sure to clean every contact on the cards with a Q-tip or paper towel before reinsertion
+ The fibres and hairs that may get caught along the trace corners can be removed with a brush or tweezers

I know that the instrument will not respond or display anything without A11, A17 and A19, which seems to be your problem. I would start there for the slot cleaning. I do not recall if the green light turns on when the cards are pulled. My instrument was really dusty when I first got it, especially on the bottom where the motherboard is mounted. While you are there, there are some shielded cables that you could examine. I think one is for the power button, but I don't remember for sure. Perhaps there is a bad solder joint, or something got torn -- hard to believe, though, given that no one ever wants to go back there.

The green light being on seems odd to me. Maybe there is something to that.

If you have another 5440B that you do not mind opening, I would suggest swapping those cards, including the transformer, to see if you can isolate the problem.

In any case, I would pull every single card from the 5440B to give the mainboard a thorough examination. I know that some people with 5700As have had mainboard burnout traces before. That would be an ugly repair, but it is possible.

Regards.

Thank you for the above also (me and all others taking on a similar project!). I'm actually done with the project... (or 99%! More on this below).

I think the issue was the line voltage switches - which needed a bit of a wash with Deoxit (D5). I also cleaned the card edges of A11 and A19 (only IPA, didn't condition those edge pads). I've inserted an IPA soaked swab (small one) also on the last couple of connectors of A11, where the 120VAC is sent to A19. So any one of these may have corrected the issue. So just a tiny bit anticlimactic, but the satisfaction is there in strides. I have two 5440s operating at my bench!

The 99% statement above comes from some residual flakiness of the fix. The unit came with A4 fit for rear output, and after seeing this works, I turned it off and switched A4 to the other position. Upon turning it on, it flaked on me again, but reseating the cards a bit and a bit more fiddling around to make sure everything sits well into their slots the unit came back online without aggravation. I probably have to sit everything - and while at that, clean all card edge connectors, like you're suggesting, Leigh, and like I did with my 5440A - and close the shield and case and it'll work reliably.

Thank you very much, Leigh and picburner! Happy camper here.

[Rax]

I may get a lot of flack for this, but upon warming up the unit up the about 4hrs it typically needs to be in spec, I am seeing it around 2.2ppm low of 10V, and about as much away from 1V too, as per my Prema 6048's 10V/1V, so to speak. I have some level of trust on the Prema's volt, so I take this as good news.

The only downside is that, as much as the unit is just about pristine, the display is really dim (and some digits are brighter than others, which I think I've seen before). A bit of a disappointment, as one thing I was looking forward relative to my 5440A is a legible display, as the 5440A's is very inconvenient to use in practical terms. The 5440B was probably on for years and years in a system environment where it was probably never turned off.

[Rax]

Stop press!! It looks like there's a bunch of work ahead.

As soon as I try to go over about 40V (>= 50V), the unit fires an "OUTPUT LIMIT FAULT. OUTPUT UNDER VOLTAGE" and resets itself. Per SM (p. 4-14), I'll need to do some A8 (possibly A9 and A4) investigations... Back to pulling up the sleeves!

[TheDefpom]

hmmm, just did a quick search and found this....

https://www.adret-electronique.fr/fluke_5440B/fluke_5440B_display.html

[Rax]

hmmm, just did a quick search and found this....

https://www.adret-electronique.fr/fluke_5440B/fluke_5440B_display.html

Very interesting, but I'm not sure I get it - is there a board and BOM available to do this upgrade? They say this much, but I couldn't find a link to a storefront. I'm reading this with google translate, my French speaking days in school are long gone.

The other thing interesting to find out is whether this upgrade can be applied to the 5440A. I know the display module is completely different between the two.

[picburner]

The person who made that board, if I remember correctly, is a user of this forum: jfphp
The board in the link is a prototype, a long time ago he sold the assembled board in his eBay shop but now I can't find it.

[leighcorrigall]

...

The only downside is that, as much as the unit is just about pristine, the display is really dim (and some digits are brighter than others, which I think I've seen before). A bit of a disappointment, as one thing I was looking forward relative to my 5440A is a legible display, as the 5440A's is very inconvenient to use in practical terms. The 5440B was probably on for years and years in a system environment where it was probably never turned off.

You reminded me of something very important. The Outside Guard Regulator (A17) should be operating TP4 at 24.0 V, not the specified value in Tables 3-4. This supply is for the VFD. R18 on A17 can be adjusted to lower this voltage.

In your case, the VFD is dimming because of age, or, if you are really lucky, the voltage is too low.

[leighcorrigall]

The person who made that board, if I remember correctly, is a user of this forum: jfphp
The board in the link is a prototype, a long time ago he sold the assembled board in his eBay shop but now I can't find it.

I remember this being true. The user might be willing to help out.

[leighcorrigall]

...

As soon as I try to go over about 40V (>= 50V), the unit fires an "OUTPUT LIMIT FAULT. OUTPUT UNDER VOLTAGE" and resets itself. Per SM (p. 4-14), I'll need to do some A8 (possibly A9 and A4) investigations... Back to pulling up the sleeves!

Glad that you have fixed the line voltage issue. Now, it is onto more problems!

Although I have not experienced this fault before, I highly recommend that while you are removing boards, you replace all of the thermal paste on each of the heat sinks. A common problem with overheated regulators is that they tend to produce undervoltage. These calibrators run very hot, and this type of maintenance will greatly improve the performance and lifespan of the unit.

You should also continue to clean the card slots as well. A4, A8, and A9, according to Table 4-1.

[Rax]

These will come in handy (see below) - extender boards I designed and fabricate for the 5440A/B. They were instrumental in fixing my 5440A, and they helped some others that used them in their repairs.
They're keyed for the compounded combination of slots for all modules in the main compartments.

[Rax]

Some progress today with this.

First thing I had to acknowledge today was that my board needs an accessory little pad to serve A4. I enclose pics on how I built it. Having the custom EDAC card edge connector handy helped. I also had to trim a "two contact" segment, and the same for the PCB part. It works perfectly as built.

I'm still digesting through the data I collected, but here's what I'm seeing this far:

• setting a non-faulty condition output (say, 40V) sets K1 off and K2 on (about 4.5V over it)
• going into a fault-triggering condition (say, 100V), outputs a non-compliant output (about 60V for a setting of 100V), which, upon the unit actualizing the fault, turns K2 off (0V over it from the 4.5V just prior to that) and resets the unit and puts the output in STBY
• the other thing the manual says pay attention to is the reading over C9. I am seeing about 26.6V over the output over it unless the fault occurs and then the drop over it goes to zero.

[Rax]

Pics

[TheDefpom]

Something to consider is that those old relays often give trouble due to oxidisation on the contacts, I have come across this issue in different pieces of equiment.

it is probably worth opening each up and running a piece of paper soaked with deoxit through each one to clean them up.

[Rax]

I've collected more info per SM troubleshooting steps, none of which makes a ton of sense yet to me in terms of origin of the fault(s).

For instance, 4-23 at p.4-29. I'm measuring the below with a DC voltmeter, though maybe I should look at these with a scope.

• TP2-TP1 20V out: 0V (expected, this doesn't engage yet the 23-275V HV path). 25V out: 14.5V (ramping up slowly, which doesn't look right; but the output is solid and correct). 35V out: 13.4V (sloping down now). 60V: 7.3V. 100V: 2.25V, which goes to 0V upon fault condition actuating the monitoring system. 300V: 2.5V (same as 100V)
• TP2-TP3 20V out: 0V. 25V out: -16V (slowly ramping up in absolute value, as above). 35V out: -14.35V (same fluctuation as above). 60V: -7.9V. 100V: -2.4V (zeroes at fault). 300V: -2.65V (zeroes at fault)

Noone of the above is anywhere near the +-40V indicated by Table 4-7. Additionally, they vary widely with the set voltage, which I'd have expected to rather stay close to those +40V or -40V turning the state of the triacs on or off (if I understand how they're supposed to work).

Generally speaking, setting 100V out outputs about 57V (though sometimes seems to go to 100V, then dropping at this 57V), which triggers the fault condition, fires the error code and puts the unit in STBY. If I request 300V out, the same happens, though I don't this I've ever seen it outputting more than the 65V it settles at, right before the unit gets switched to STBY.

Further on the same procedure (step 5), when I try to measure state at pin 8 of P21 (TS1+) relative to the PCOM (23/13), for:

• 20V out: 4.46VDC
• 30V out: 3.8V
• 70V out: 3.8V which goes to 4.46V at fault triggering

[TheDefpom]

Sounds similar to an issue I had on my 5200a, one of the output transistors were breaking down under high voltage, causing it to short out and drag the output down.

[Rax]

Sounds similar to an issue I had on my 5200a, one of the output transistors were breaking down under high voltage, causing it to short out and drag the output down.

Trying to see trough the weeds of the troubleshooting steps and the eclectic results they return, one thing stands out from my symptoms, namely the fact that the output plateaus at about 60V regardless of setting, which I think seems to point to a hard level-related failure like I think you're describing, Scott. I don't think the issue is the "level-composing" logic, or even the monitoring/protection systems, but a simple hard stop at 60V out.