Tektronix AM503B, A6302 fault diagnosis (error code 588)

[matthuszagh]

I received an AM503B and A6302 from 2 different ebay sellers, both listed as functional. However, the unit doesn't work: probe degaussing leads to error code 588, I see significant amounts of noise, and I see no discernible signal when I attempt to use the probe with a DC current. Since they were both listed as functional I can return an item that's faulty, but my trouble is I don't know whether the probe is to blame or the amplifier is to blame. I contacted both sellers. The probe seller said the unit was taken from a working environment but they didn't perform further testing. I got the sense from the AM503B seller that they didn't perform any testing.

The manual doesn't contain much info on error code 588. The only thing I could find was:

Error occurred during internal automatic adjustment. Repeated errors in the range 550–594 indicate the Amplifier needs repair

which seems to implicate the amplifier.

I've attached 3 images of the oscilloscope screen. The noisy one shows the oscilloscope screen when the amplifier is displaying the error code. The output looks identical when I select AC coupling on the amplifier. The one with an 840 mV offset is the DC coupling, and the noise-free one is the ref coupling. If I remove the probe then all 3 coupling modes look like the ref coupling from this image. Any thoughts on which part may be at fault here? If I had to guess I'd think that the probe is at fault since the noise disappears when I unplug it, but I really don't know enough about how this unit works to say this with any confidence. All help appreciated, thanks!

[alm]

The AM503B/AM5030 service manual is not very helpful, telling you you need a probe to do any tests. But I wonder if the instructions in the AM503 manual work, since they share the same probe interface. I'd try the steps in chapter 2 (performance verification) regarding injecting a signal into the AM503B (see figure 3-1 for the adapter they use to inject a signal). Not all will work, because the AM503B has an automatic system that adjusts offsets during degaussing (which will fail without probe). I have a AM503B that worked last time I checked, so I could help you test, but I don't have time to run through all the steps right now to see which work.

Testing the probe without the amplifier is probably more difficult, since especially the DC part relies on the hall sensor providing the signal to null the magnetic field that's induced by the DC current by cancelling it through a bucking coil. You might be able to measure the transformer part using a resistive load (it won't be very flat). The schematic in this manual. might be helpful here. It could just be a loose wire inside the probe. I'd try looking at the schematic and checking resistance between pins. Again, if you have specific measurements you'd like to verify, I can give them from a known-good probe. But you're not likely to diagnose a drifty Hall sensor (sign it has been overstressed by a shock). I would also do a mechanical inspection of the ferrite faces: they should be mirror flat without any cracks or chips.

If you have someone in your neighborhood that have a working probe and/or amplifier (AM503 / 503A / 503B / 5030 will all work with this probe), that would be ideal.

[matthuszagh]

The AM503B/AM5030 service manual is not very helpful, telling you you need a probe to do any tests. But I wonder if the instructions in the AM503 manual work, since they share the same probe interface. I'd try the steps in chapter 2 (performance verification) regarding injecting a signal into the AM503B (see figure 3-1 for the adapter they use to inject a signal). Not all will work, because the AM503B has an automatic system that adjusts offsets during degaussing (which will fail without probe). I have a AM503B that worked last time I checked, so I could help you test, but I don't have time to run through all the steps right now to see which work.

Thanks so much for directing me to this. I attempted "2. Check AC Dynamic Range" from the 503 manual. I wasn't able to get any signal on the monitor oscilloscope or DVM. It's a bit difficult to feel confident that the special adapter is making an electrical connection, but I did try wiggling it around a bit and no change, so I think it's probably not a false negative. I also wasn't sure what amplitude to use. I started at zero and worked up in 1mV steps. When I saw nothing went up to 1V, but didn't want to go further. Both steps (i and q) involving the DC level control did work though. I read 59 mV on i and 57.5 mV on q.

Testing the probe without the amplifier is probably more difficult, since especially the DC part relies on the hall sensor providing the signal to null the magnetic field that's induced by the DC current by cancelling it through a bucking coil. You might be able to measure the transformer part using a resistive load (it won't be very flat). The schematic in this manual. might be helpful here. It could just be a loose wire inside the probe. I'd try looking at the schematic and checking resistance between pins. Again, if you have specific measurements you'd like to verify, I can give them from a known-good probe. But you're not likely to diagnose a drifty Hall sensor (sign it has been overstressed by a shock). I would also do a mechanical inspection of the ferrite faces: they should be mirror flat without any cracks or chips.

I disassembled the probe and took some pictures (attached). I didn't notice anything unusual, like disconnected cables, etc. but I also didn't perform any electrical tests on components inside for fear of damaging something.

If you have someone in your neighborhood that have a working probe and/or amplifier (AM503 / 503A / 503B / 5030 will all work with this probe), that would be ideal.

I'm in San Jose. If anyone in the Bay has these items and is willing to help out I'd be very appreciative to be able to drive over and test my items.

[matthuszagh]

Don't know if this is useful, but I also took pictures of the inside of the AM503B. The battery has some green stuff that is maybe corrosion that could be of note.

[MarkL]

I would first clean up the battery holder and replace the battery.  I bought a pair of AM503B amplifiers last year and they both had green crusty batteries and battery holders and displayed 5xx errors unrelated to "357" or "LO" (meaning: your battery is dead).  Sorry, I don't recall the exact error number they were displaying.

After cleaning and installing a new battery in each, they operated normally, albeit uncalibrated since the cal info is stored in the battery-backed NVRAM.  I haven't gotten around to re-calibrate them yet.

Broken wires in the probe cable are common as the insulation gets increasingly brittle with age.  I've seen breaks at the ends in both the probe body and Amphenol connector, and once I had a break inside the cable itself.  (I now have six of these probes in various states of working and disrepair trying to make a couple of working ones.)

I'd suggest unplugging the sensor from the probe head and doing resistance checks to make sure the hall and coil elements are ok. See this post about the A6302 (buried in a TCP202 thread):

  https://www.eevblog.com/forum/repair/tektronix-tcp202-current-probe-repair-schematic-and-suggestions-needed/msg1664789/#msg1664789

If ok, I would then verify the continuity of the all the connections from the probe head to the connector pins.  The pinout is in the P6302 manual posted by alm, above.

Also, look out for the ball bearing that sits in the probe latch.  It's really easy to lose.  You've probably already encountered it, but if yours is missing you'll need to replace it if you're going to keep the probe.

Do you get anything when probing an AC signal?

[jonpaul]

Bonjour à tous

Just seeing this now

Have many AM503, A, B, P6301, 6302 etc.

the amps and manuals are totally different 503/A/B and B manual omits all details.

Backup battery corrosive and mem cal lossis,a,big issue.

99% of ebay sellers are unwilling or unable to test these, thus any ebay purchase is risky and should be at a low price.

On the forum I have seen no indication of suceed to.fix,AM503B.
The probes are susceptible to cable termination break,at stressed ends and especially to dropping whick cracks,the ferrites cores

Testing Hall  effect circuits in the probe can cause damage.

At moment, I  am hours from SJ but no time to meet till the Summer.

Suggest that you return the amps and probes....happy to continue via PM.

Désolé pour votre problème

Jon

[alm]

Thanks so much for directing me to this. I attempted "2. Check AC Dynamic Range" from the 503 manual. I wasn't able to get any signal on the monitor oscilloscope or DVM. It's a bit difficult to feel confident that the special adapter is making an electrical connection, but I did try wiggling it around a bit and no change, so I think it's probably not a false negative. I also wasn't sure what amplitude to use. I started at zero and worked up in 1mV steps. When I saw nothing went up to 1V, but didn't want to go further. Both steps (i and q) involving the DC level control did work though. I read 59 mV on i and 57.5 mV on q.

Connecting a function gen via a BNC RG-58 cable with center to pin M (via a 22 Ohm resistor) and shield to pin J, and a 1 KHz sine wave, a 1 Vpp 1 kHz square wave (at the function gen with the output termination set to 50 Ohm, so probably 0.5 Vpp at pin M since the 22 Ohm resistor will act as a 2x attenuator) will produce a clean 2 mVpp square wave on my scope (input set to 50 Ohm) regardless of the attenuation setting on the AM503B. A 100 Hz 2 Vpp sine wave measured 3 mVpp (including some noise). If you don't get any AC signal out, then I think that's a clear indication your AM503B is faulty.

There are no schematics for the AM503B/5030, as far as I know, but there are schematics for the AM503A on TekWiki. The AM503A is certainly not identical, but at least the analog part should be quite similar from what I've read.

[matthuszagh]

Connecting a function gen via a BNC RG-58 cable with center to pin M (via a 22 Ohm resistor) and shield to pin J, and a 1 KHz sine wave, a 1 Vpp 1 kHz square wave (at the function gen with the output termination set to 50 Ohm, so probably 0.5 Vpp at pin M since the 22 Ohm resistor will act as a 2x attenuator) will produce a clean 2 mVpp square wave on my scope (input set to 50 Ohm) regardless of the attenuation setting on the AM503B. A 100 Hz 2 Vpp sine wave measured 3 mVpp (including some noise). If you don't get any AC signal out, then I think that's a clear indication your AM503B is faulty.

Thanks so much for running this test. Just to clarify, did you also use the 10x attenuator shown in the test setup (image attached)? From your description, it sounds like you may have omitted this in which case I may have just been injecting too weak of a signal (I included it).

[matthuszagh]

Do you get anything when probing an AC signal?

I went to test this (will get to your other points later) and I was able to successfully perform a probe degauss. This is the only time it's worked and subsequent attempts have failed with the same 588 error. However, I do see an AC signal at roughly the right frequency, though it's noisy. The amplitude also seems to be about half what it should be. Also at times the signal becomes so noisy it's basically unrecognizable. Degaussing does seem to clear it up (despite the reported error).

The AC signal is also visible when the probe amplifier is set to DC coupling, though no DC information seems to make it out (confirmed by adding a DC offset to the function generator).

I also tried injecting higher frequency signals in 10 MHz steps from 10 MHz up to 50 MHz. All signals are visible on a scope, but severly attenuated. I also see a lot of phase noise on the scope signal. This is not the function generator's fault; when I hooked it up to the 50ohm scope input I only measured a small attenuation compared to lower frequency signals. I notice large amplitude jumps in the 1 MHz -> 10 MHz and 10 MHz -> 20 MHz transitions. I have the regular (not XL) probe.

My main question at this point is whether the presence of this AC signal indicates the probe is most likely fine and that the amplifier is to blame. I understand there are obvious concerns to address with the amplifier (such as the leaky battery), but I'm going to hold off on that unless I decide to keep this which I probably won't since I paid for a working unit. I'm concerned that even if I repair the battery I may still experience issues, at which point I'll have lost the right to return it because I made modifications.

[TERRA Operative]

I have found leaking batteries in all 3 or 4 of my AM503B's, definitely worth looking into repairing that in yours.

I have been working on a way to reprogram the calibration constants and serial number back into these units without the GPIB interface that is basically unobtanium unless you temporarily pull one from an AM5030, but then you need to find an AM5030...
It uses a Bus Pirate, IC clip and Putty terminal emulator. You can do it manually in Putty, but my batch script will hopefully automate most of the process with menus etc. once I finish it....
Anyway, if you want the info on that, I can provide what I have. I also added some manuals and details of the AM503B calibration adapters to the TekWiki page just now.
https://w140.com/tekwiki/wiki/AM503

Also, I have like 5x AM503's and one AM503B spare here if you want one cheap (plus shipping from Japan whatever that will cost.....) I can give it a test too, to make sure it is working (but I haven't got all the bits yet for a proper calibration though).

[MarkL]

...
I have been working on a way to reprogram the calibration constants and serial number back into these units without the GPIB interface that is basically unobtanium unless you temporarily pull one from an AM5030, but then you need to find an AM5030...
It uses a Bus Pirate, IC clip and Putty terminal emulator. You can do it manually in Putty, but my batch script will hopefully automate most of the process with menus etc. once I finish it....
...

Is this is the same manual procedure, or a parallel effort?

  https://groups.io/g/TekScopes/files/AM503B%20Calibration%20without%20GPIB/AM503B_AM5030%20Calibration_V01.pdf

[alm]

Thanks so much for running this test. Just to clarify, did you also use the 10x attenuator shown in the test setup (image attached)? From your description, it sounds like you may have omitted this in which case I may have just been injecting too weak of a signal (I included it).

Yes, after I saw how low the response was, I didn't bother with the attenuator. I'm not sure why the response doesn't change as I adjust the attenuator setting. Maybe it depends on sensing the probe?

My main question at this point is whether the presence of this AC signal indicates the probe is most likely fine and that the amplifier is to blame.

I'd say no. DC (and LF AC) and HF AC are two entirely separate circuits in the probe. The AC signal is just a current transformer. The main thing that might go wrong here is the ferrites not making tight enough contact, degrading the lower frequency response. You'd observe this as a valley in the frequency response between DC and higher frequencies.

The DC works through a thin-film Hall effect sensor that measures the magnetic field through the ferrite core. The AM503B then sends a bucking current through a secondary coil around the ferrite core such that it nulls the magnetic field as detected by the Hall effect sensor. And this Hall effect sensor is by far the most fragile part. If the probe is dropped, even if the ferrite doesn't fracture, the Hall effect sensor (which is also an excellent stress sensor) may be stressed, and gain a permanent offset that the amplifier can't correct. Or it may turn unstable, so the DC offset is drifting all over the place.

This would be the main thing I'd want to test before declaring the probe "healthy". It may well be that the amplifier is at fault here, but it would take a lot of effort to say if this part of the probe without a working amplifier.

[TERRA Operative]

...
I have been working on a way to reprogram the calibration constants and serial number back into these units without the GPIB interface that is basically unobtanium unless you temporarily pull one from an AM5030, but then you need to find an AM5030...
It uses a Bus Pirate, IC clip and Putty terminal emulator. You can do it manually in Putty, but my batch script will hopefully automate most of the process with menus etc. once I finish it....
...

Is this is the same manual procedure, or a parallel effort?

  https://groups.io/g/TekScopes/files/AM503B%20Calibration%20without%20GPIB/AM503B_AM5030%20Calibration_V01.pdf

Yeah, I'm involved with that one. It wemt onto the to-do list for a while as other stuff came up.
I'll have to get back onto finishing it.....

[matthuszagh]

Broken wires in the probe cable are common as the insulation gets increasingly brittle with age.  I've seen breaks at the ends in both the probe body and Amphenol connector, and once I had a break inside the cable itself.  (I now have six of these probes in various states of working and disrepair trying to make a couple of working ones.)

I'd suggest unplugging the sensor from the probe head and doing resistance checks to make sure the hall and coil elements are ok. See this post about the A6302 (buried in a TCP202 thread):

  https://www.eevblog.com/forum/repair/tektronix-tcp202-current-probe-repair-schematic-and-suggestions-needed/msg1664789/#msg1664789

If ok, I would then verify the continuity of the all the connections from the probe head to the connector pins.  The pinout is in the P6302 manual posted by alm, above.

I checked continuity of the wires through the cable and they seem to be fine: most measured 0.8ohm. The connections to the current transformer read a little higher: about 1ohm. I also measured a bunch of component values. Most everything measured ok except I couldn't get a good reading from LR15. I used my HP 4192A impedance analyzer to measure this. At around 1 MHz I was reading approximately 1uH in parallel with a large resistance (something like 5kohm). The schematic shows 0.395u in parallel with 56 ohms.

I would first clean up the battery holder and replace the battery.  I bought a pair of AM503B amplifiers last year and they both had green crusty batteries and battery holders and displayed 5xx errors unrelated to "357" or "LO" (meaning: your battery is dead).  Sorry, I don't recall the exact error number they were displaying.

After cleaning and installing a new battery in each, they operated normally, albeit uncalibrated since the cal info is stored in the battery-backed NVRAM.  I haven't gotten around to re-calibrate them yet.

I'm tempted to do this, as it seems like there's a good chance this is at least one of the issues I'm facing. However, I probably can't do something that creates a new error on the front panel.

I redid some DC measurements and it looks like I was wrong before. If I change the DC current I see a corresponding change in the scope waveform that looks more or less ok. However, the absolute value is still way off. The fact that at least some DC measurement capability is preserved seems to be consistent with the reasonable low-frequency AC measurements I was getting.