| Products > Test Equipment |
| Test Equipment Anonymous (TEA) group therapy thread |
| << < (25325/27435) > >> |
| Zenith:
--- Quote from: bd139 on July 17, 2022, 07:54:19 pm --- Probably is. I try and break the law at least once a day. --- End quote --- But it's as well to know when you are breaking the law, so you can calculate the risks and minimise them. In this case it's a good idea not to be seen doing it if you can help it, and I'd say that phoning the police and telling them about an abandoned vehicle, needlessly draws their attention to the matter and puts you in the picture. |
| Neper:
--- Quote from: Zenith on July 17, 2022, 08:27:39 pm ---But it's as well to know when you are breaking the law, so you can calculate the risks and minimise them. --- End quote --- Aka the 11th commandment. |
| grizewald:
Datron 1071 - A very strange problem After fixing the PSU and putting everything back together again, I started testing the meter in earnest to see how accurate and stable it was. Sadly, I rapidly discovered that it was neither accurate or stable. :( The first thing I noticed was that when turned on from cold, there was a huge offset on the voltage ranges. For example, on the 10V range, there was a -700µV zero offset with the input terminals shorted. As the meter warmed up, this would decrease and sometimes go all the way to a few tens of µV positive, but not always. Once the meter had warmed up, I could use the input zero key to tune out the offset, but if I connected up a voltage reference for a few minutes and then switched back to a short again, the zero offset would again be way off. Or sometimes not! I noticed some white powdery deposits on the analogue board which looked like the result of poor cleaning after the repairs that had previously been performed, so I took the analogue board out, taped over all the pots to stop anything getting into them and did the job properly this time. This made absolutely no difference whatsoever. The meter still couldn't establish and hold a stable zero. In desperation, I powered up the meter, threw a winter coat over it and gave it a thorough roasting in case the instability was caused by moisture which had infiltrated it. Again, this proved to be a pointless endeavour. |O Utterly frustrated, I put the meter away in a corner and forgot about it for a year. ... The pressure from work had started to abate and I could feel the Datron mocking me from the corner of room. I started again by studying the service manual's functional description and circuit diagrams in detail. At some point, a lightbulb came on while I was looking at this part of the analogue board's circuit diagram: This jogged my memory about an excellent article by a gentleman named Karlheinz on amplifier.cd where he discussed his experiences repairing a Datron 1061A. (https://amplifier.cd/Test_Equipment/other/Datron1061/Datron_1061.html) The circuit in the diagram deals with input bias current compensation. Specifically, when calibrating the meter, one step of the process consists of connecting a shielded box containing a 10MΩ resistor in parallel with a 1nF capacitor to the input terminals and pressing the "Ib" calibration button. When you do this, the meter measures the input bias current, calculates a correction and stores that in the calibration memory. When a voltage or resistance range is subsequently selected, the digital board fetches the Ib correction value and transfers it to the analogue board over the opto isolators which implement the isolation barrier between the boards. The value itself is an 8 bit number which the analogue board then latches into buffers which function as the input to a D-A converter which is implemented in discrete components on the analogue board. The voltage generated by the converter enters the circuit pictured above on the Ib Comp line and is applied to the inverting input of op-amp M3. This causes the op-amp to drive current through the diode and resistor to the M23 opto isolator's LEDs and the right-hand transistor to sink current so that the voltage over R198 matches the input voltage on Ib Comp. The other transistor acts as a current mirror referenced to the negative bootstrap voltage (-BS) and effectively floats the Ib correction onto the bootstrap supply. This mechanism ensures that no cross current can flow and ensures that the DC isolator remains isolated from the correction voltage. If there's anything wrong with the M23 opto isolator, this circuit won't work properly and could well cause the kind of problems I was seeing. It wouldn't be the first time a faulty opto isolator has been identified as the cause of a problem in these meters and when you consider how this opto isolator is being used, both LED/transistor pairs need to have matched characteristics for the correction to be faithfully reproduced - this particular opto isolator is a specially selected part and is marked with an orange dot. Here is the relevant part of the analogue board: The opto isolator M23 is the 8 pin part with the orange dot at the lower left of the photo and the op-amp M3 is in the TO-99 package at the top of the picture to the right of the flatflex cable. Armed with the location of the components involved, I fired up my oscilloscope and prepared to start probing! The diode and resistor that pin 6 of the op-amp is connected to are just behind that red film capacitor next to the op-amp. You won't believe what I saw when I looked behind that capacitor... :-DD Whoever built the PCB in Datron's factory had obviously twisted the component leads of the diode and resistor together in preparation for soldering them but had gone to lunch or been otherwise distracted before they completed the final step! And so the board remained for some 35 years before the connection between the two components deteriorated to the point that it started to cause a problem! Due to them being hidden behind the film capacitor, I never noticed and neither did anyone else. I'm sure the connection was perfectly OK during QA testing and calibration, so it just sat there like a little time bomb waiting to trigger decades later. I gave the leads a gentle scraping with a scalpel and used plenty of flux when I finally completed the forgotten assembly step so I could be sure that the connection would be good after soldering. I then constructed the suggested "Ib source" required and performed the calibration step. The meter's zero offset immediately dropped to zero. After turning the meter off and allowing it to cool down before powering up again, the initial zero offset was within a few tens of microvolts and once the meter had fully warmed up, it was very close to zero without needing the use the "Input Zero" button. Since correcting this problem, the meter has proved to be exceptionally quiet, stable and accurate. Noise on the 10V range is around +/- 2.1µV which beats my Solartron 7061 (at +/- 12µV) by miles. Both meters agree within a few PPM on any reading which is excellent considering that I paid to have the Solartron properly calibrated couple of years ago. And finally, it's just a great looking meter! |
| med6753:
;D |
| tautech:
@grizewald Did you follow any of Defpom's vids on Datron repairs ? https://www.youtube.com/playlist?list=PL2z3V9RkHQE3JrgdtFlmEC6LDbg-5Ah00 |
| Navigation |
| Message Index |
| Next page |
| Previous page |