EEVblog Electronics Community Forum

Electronics => Metrology => Topic started by: vsmith on October 27, 2019, 07:40:10 am

Title: HP 3457a troubleshooting/documentation woes
Post by: vsmith on October 27, 2019, 07:40:10 am
Hi all

I picked up a couple HP3457a "parts" units on the auction site, and have started working on getting them to function properly again.

They both seem to have Controller board issues, the symptom being that they freeze at some point in the Power Up sequence.

In trying to troubleshoot the problems, I have run into some areas where the Service Manual (copies of which I found on the internet by Googling) leaves a lot to be desired.

For one thing, the Theory of Operation and Troubleshooting sections seem to me pretty vague about the power up sequence, and I am not seeing much about the way in which the analog board's data ends up being displayed on the LCD. And worse, I cannot find any schematic of the Front Panel assemble at all!

So my purpose in posting is two-fold:

1. Has anyone seen a copy of the service manual which does have a schematic for the Front Panel? I am assuming it is a "dumb" peripheral, and has no logic of it's own, but can't be sure without the schematic! Can someone check their hard copy of the manual?

2. Has anyone got any suggestions for narrowing down the cause of failure for such a symptom? (The unit never completes power on self test, and displays either random or logical but frozen text and does not respond to any keyboard input) Unfortunately, my diagnostic equipment for this is limited to a 'scope, no logic analyzer... I have checked the Power Supplies, and visually inspected the modules. One does have some minor corrosion in the GPIB connector and on the PWB where that connector is located, I have cleaned this and dried it, and believe it is "clean". Could a GPIB short/leakage completely stop the power on sequence?

Title: Re: HP 3457a troubleshooting/documentation woes
Post by: vsmith on October 27, 2019, 11:26:38 pm
An update to my post.

While trying different ways to isolate the problem, I did discover (through heating the board after cleaning some of the corrosion residue at the GPIB connector with distilled water) that the rest of the unit seems to be working. I got things to work for several minutes, checked various dc voltages, and even started checking the ohms ranges when it reverted back to its frozen state. So I am now fairly certain that the problem is in the Controller board. It seems that it is humidity related more than temperature, and I suspect that there are some leakage paths from liquid damage that I cannot see, perhaps underneath ICs. It appears also that the Front Panel is fine, and the search for a schematic for the front panel is now somewhat academic.

I am stuck, tho, in pinpointing just where on the controller board this intermittent problem is (It intermittently works, haha) The other symptom I neglected to mention in my first post is that when it does start up successfully there are several loud beeps, rather than the single mid level one normally heard at the completion of power up.

Any suggestions or similar experiences would be welcome!
Title: Re: HP 3457a troubleshooting/documentation woes
Post by: GigaJoe on October 31, 2019, 01:41:35 pm
could be a ram malfunctioning , not holding a correct data for cpu
Title: Re: HP 3457a troubleshooting/documentation woes
Post by: vsmith on November 01, 2019, 07:28:19 am
Thank you GigaJoe. That was a good suggestion, but I think it is actually the MPU itself. I ended up doing very focused tests using a small aperture heat gun and cardboard shields, and it turned out that the MPU was very sensitive too a slight rise in heat. Since that IC is really the only one that perceptibly self heats (even in a good unit they run warm) I ordered a replacement one on line. It should arrive here Friday, and I will swap it out. Will let you all know.
Title: Re: HP 3457a troubleshooting/documentation woes
Post by: Kleinstein on November 01, 2019, 09:57:05 am
Besides the chip itself, there could also be bad solder joints and ceramic caps (with a crack), that react to heat.
Title: Re: HP 3457a troubleshooting/documentation woes
Post by: SvanGool on November 01, 2019, 08:11:50 pm
Is the problem similar to this: ( ?
Title: Re: HP 3457a troubleshooting/documentation woes
Post by: vsmith on November 03, 2019, 09:20:48 pm
@Svangool, now that is an interesting thread! Thanks for linking that for me. And yes the symptom is kind of the same (its even the same GPIB address) except with mine, the communication between the analog and digital boards is still working. The failure mode on mine is that initially the unit works and then after a few minutes, it freezes and/or starts displaying a scrambled display and then freezes (meaning that it does not change any more and does not respond to keypress input). On restart it often gets to the HPIB adress and stops there. You have to let it sit "off" for 10-15 minutes and then it will start up again.

On that subject though, I did consider that (the optocoupler) could be the point of failure and was probing the output with a scope, but not really seeing the serial bus communication I expected (all I saw was low level noise). Yet the meter was working, at least for a little while. So that is a problem I am interested in also, why I couldn't read that traffic with a 'scope. (I was probing on the Digital side of the connector using the Digital ground, which is chassis ground as well. Is it basically a low impedance node at that point?

Meanwhile, I ordered a replacement uP and while I am waiting for it to arrive (actually I ordered two slightly different parts from two sellers) I decided to install a 40 pin socket, so that if things didnt work out I would not have to keep stressing the PWB with rework. And obviously, I put the processor back in to test the socket, and see if being elevated a bit off the PWB made a difference. And it did. The meter worked for about 10 minutes (at least 2x longer than before), tested voltage, current, resistance, range changes, etc. But then froze up like before. So I think it is the uP but these things are tricky and as Kleinstein pointed out (Thank you Kleinstein), the point of failure may be nearby the uP instead. Fingers crossed, it looks like Friday I should have a replacement. BTW these things (Motorola 68B09 type) are getting rather hard to find, which is one reason why I bought a couple. Also that one vendor was offering the "industrial" version (MC68B09CP) which has a wider operating range, and these were actually cheaper than the original MC68B09P part from other vendors. So I will try them all out. (The readily available E suffix is not compatible with this circuit)
Title: Re: HP 3457a troubleshooting/documentation woes
Post by: Johnny10 on November 04, 2019, 01:23:29 am
I am not going to tell you I have the answer to your particular problem. However, as I repaired my 3457 I had a couple of problems, one of which was a cracked trace along the edge of the board.

It was causing intermittent lock-up similar to what you are observing.

Check around the areas that can flex when the optional multiplex rear panel inputs are installed.

Lightly flex the board while it is operating.
After all my component level repair was finished this was the solution to my lock up problems.

Good Luck,

Title: Re: HP 3457a troubleshooting/documentation woes
Post by: vsmith on November 06, 2019, 12:28:33 pm
Reporting back on results of replacing uP IC: Success.

Still waiting for the MC68B09P and MC68B09CP on order. I did get a MC68A09P processor, sent in error by a vendor. And because I am impatient, I decided to try it out. (The difference between the MC68A09 and MC68B09 are the clock speed rating, along with some of the timing specs for various functions.) On the theory that the "A" chips were simply ones that didn't meet the test criteria for the "B" rating, I went ahead and stuck the A type in the socket (glad I installed that) and lo and behold, the meter powered up, and stays powered (20 minutes ands counting so far) and passed Self Test. So this one did have a flaky uP IC. And the 3457A design leaves room for some variation in uP chip performance!

Sadly, the calibration corrections stored in eprom have all been lost. Replacing the processor will do that in this design, it appears. It may be that the desoldering process was responsible for that data loss, even though the rework was done on the digital board after removing it from the chassis, and the backup battery was not disturbed. I'm not surprised, anyway. Once I get the "right" uP and install it, I will give the unit a fresh calibration. Interestingly, data such as the Cal Number (17 in my case) did not get lost.

I am still annoyed at the lack of documentation regarding the front panel.....