-
[SOLVED] EDC 520A working, but hot resistors and stability problems?
Posted by essele on 17 Mar, 2019 18:39 -
I grabbed myself an EDC 520A a couple of weeks ago and found a dry joint (well, entirely separated from the pcb joint), and a magic smoke precision resistor on the current circuit (caused by high voltage cause by the former joint) and once replaced and calibrated (thanks to the video from feedback.loop) everything appeared to be fine.
However....
1. The dry joint was on a quite bizarre resistor arch setup (photo attached) which I can't believe is original.
2. The stability was perfect once it had warmed up for an hour or so, then after two or three hours all went a bit wild (+/- 400uV on 10v range)
I initially suspected the zener, but it's actually rock solid ... I've checked with a heat gun and it's well within spec even with quite a wide range of temperatures.
So I then got out my dodgy Chinese thermal camera (which although it's not a Flir, is actually quite handy) and I discovered two different places where resistors were getting hot, and I mean 120 degrees C ... which to me seems wrong (interested to see if people think this is ok.)
Firstly there is a 10R in the 12v supply (but not on the schematic I have) prior to the 7812 that's dropping about 3v, so that's best part of 1W, it's looks fairly chunky (photo attached, it's between the two caps) but still seems like it's not coping.
Secondly there is the bizarre resistor archway, these are three x 4.3k, so just over 12k total, but they are dropping nearly 200V, so 3.5W or so ... but they are also well over 100 degrees.
I can't find a proper 520A schematic, but the 521 manual (which has quite a few differences) shows a 10K 5W resistor (although actually located in the power supply section) but schematically the same. I am really struggling to understand the workings of this section of the schematic (R307, Q301, Q302), but in all ranges I'm seeing nearly 200V dropped across that resistor arch with very little variation. I've ordered some new resistors and some MJE350's just in case.
I've watched feedback.loop's video, but his board is slightly different and doesn't show this 10K, so I presume it's in the power supply section, there are a few differences in the board so it's probably a different revision.
Then there's also a very bizarre set of mods on the back of the board, this looks like a strange attempt to deal with a different capacitor setup (photo attached) ... it's obviously working reasonably well, but I'm really struggling to understand what was going on.
And the other thing is that the schematic says +/-150V, the PCB says +/-130V, but I'm actually measuring +/-200V ... I did think it might be to do with the 110/220v switching, so I ran it back on a 110v supply (obviously with input switch set appropriately) and it was exactly the same.
So ... some questions I'd love to see if anyone has answers to...
- surely running components at 120 degrees C+ isn't right?
- does anyone actually have a 520A schematic?
- any ideas why I'm seeing 200V rather than 150 or 130 .. or is that right? That's definitely not going to help my hot resistor problem!
- any other help would be great!
I'm going to replace the resistor arch with a single 10K 5W component as per the schematic, 200V (measured) across 10K is 4W, so hopefully this should be ok.
I'm also going to beef up the 10R in the 12V supply ... I'm seeing 21V at the regulator, so I'm presuming the 10R is there to drop the voltage prior to the regulator, so I've ordered some 20R higher power versions as well.
This is a really nice bit of kit and very accurate, but for a limited time, so it's slightly frustrating. Any help would be greatly appreciated.
-
Hi,
there are differencies between models, even between different 520A devices. Here are pictures of one 520A.
Wictor
-
Hi,
I think this is best schematic for 520A model available online http://exodus.poly.edu/~kurt/manuals/manuals/Other/EDC%20521%20Operator.pdf
I have actually modified my unit so that I moved that big resistor away from that place to power supply side. It's heat in that original location was causing some drift issues if I recall correctly. If you see linked operator manual and part layout, you can see that R307 is located near big power supply capacitors. Attached image is from modified unit. I also added some diodes that are in later designs.
Wictor
-
For stability issues, check all socketed ICs. Their contacts are causing drifting issues.
Wictor -
Thanks Wiktor, that’s really useful.
I notice yours is 13K which would also reduce the power to 3W, so that seems like a better replacement, and obviously similar resistance to what I’m removing.
I’ll also have a look at how easy it is to place it back in the power supply area.
So it sounds like it does run bloody hot then?
That’s the same schematic I have, but there are quite a few differences to the actual board, and good point on the socketed IC’s ... I’ll check them also. Thanks. -
Hi
I can measure the voltage across the resistor and temperature also, if you want. -
200 Vdc seems high for the high voltage rails. I do not remember them being that high. Are the filter capacitors working? What is the ripple voltage and what is the voltage (ac and dc) across the 50-Ohm filter resistors?
As previously stated, there were several revisions to the 520A. The picture above shows 2x 470uF in the positive 150 Vdc rail and 1x 220uF in the -150 Vdc rail (the middle one is for the -150 Vdc rail). This was later changed to 470uF in all three positions while the 520A was still being manufactured (looked similar to the 521A diagram). An update to the 521A power supply changed this to be symmetrical for both the positive and negative rails with 2x 470uF capacitors in both the positive and negative 150 Vdc rails.
Your twist tab capacitors seem to be a mod. I think someone has drilled out holes for them. There should be 2x 470uF in the positive 150 V rail with a 50-Ohm 5-Watt resistor between them. The negative rail just had one 470uF capacitor with a 50-Ohm series resistor between it and the rectifier. The 5-Watt resistors used in the power supply filters were later changed to 30-Ohm.
R307 was originally a 13 kOhm 2-Watt carbon resistor (which makes me believe the 200 V is way off) but it was revised to a 10 kOhm 5-Watt WW. It was also moved to the other side of the barrier. When placed in the power supply against the rail, a clear insulator was placed over one of the leads due to being bent across one of the runs.
If you still have stability problems, I'd first replace the OP07's with OP177's and I'd also replace the quad op-amp, then I'd begin troubleshooting. Also, those precision black resistors can go bad. -
Interesting ... I'm not at home now, so will measure again later, but from memory...
I was seeing 150VAC from the transformer, which matches pretty well with 200V rectified. I haven't looked at ripple yet.
I'm pretty sure there are 10-Ohm resistors on the 150V rails between the rectifier and caps but I'll need to check the exact location, and then a 100-Ohm after the cap on the positive rail only (but that looks like it's all tied up with the cap mod.)
I'll definitely switch out R307 (and hopefully move it), but it would be good to know what voltages you are seeing on those 150v rails, I don't think there's much I can do to lower them, but it might give me scope to put a 15K in for R307 instead to lower power a bit more. I'll also look again at the filters. -
Hi,
I did some measurements from my 520A. All measurement done with following settings:
- range selector 10V
- polarity switch 0
- resistor temperatures measured with infrared temperature meter, when 520A had been powered on about 10 mins
- fluke 287 used to measure voltages
Measurements marked to attached images.
-
Wiktor - that is super helpful, thank you.
I've just done an equivalent, don't have time to properly mark up a diagram, I'll try to do that tomorrow, but the summary is:
5v supply -- ok, I'm getting 8.3VAC vs your 7.1VAC
DAC Supply -- ok, I'm getting 33.9VAC, so very similar to your 33.5VAC
0ther 15v -- ok, I'm getting 39.4VAC vs your 38VAC
Now the interesting one... the 150v supply...
I get 298VAC vs your 233VAC, that's quite a difference. So I get +192V (0.02VAC) and -196V (0.3VAC)
I have 10-Ohm resistors in both the positive and negative (well 12 and 10), they don't get particularly hot (42 degrees) ... the 22K's however are over 100 degrees! (1.6W at my values vs 1W at yours) I assume these are just to discharge the caps, so I could put larger values in??
I also have a 100-Ohm in the positive line before another cap, a bit like the extra 50 you show .. but I'm confused by the cap, it seems to have two different values to the case (haven't looked too hard.)
Now, even more interestingly I have another supply, it's 2 blue wires, 17.5VAC, this has a 10-Ohm resistor after the rectifier that runs over 160 degrees ... now that just can't be right. It's supplying a 12V regulator that then seems to go to the MPU and front panel. I think I'll need to check to see if there are any problems on there, everything seems to work ok (although I haven't checked GPIB.)
I really don't understand why I would have a different winding for the 150V supply, but how else could this be so different? Is there any way any of the caps could cause this? Can a dodgy filter cap before the rectifier do anything like this? -
Early 522's came with 22 kOhm bleeders on the 150 Vdc rails but later came with 47 kOhm bleeders.
520A's that have a 6 digit BCD decoder for an MPU came with a 12 Vdc power supply. When the MPU was replaced with a 6502 the 12 V supply was removed. It's hard for me to believe that it needs to source 0.3 A. I would try to check the filters with an LCR meter.
The 521 schematic (on page 57 of the PDF I think you may be using) shows a 50-Ohm trimmer near the reference zener on the left side below a 2 kOhm and two trims (and labels them all R2). This was one of the first revisions in the EDC 520A. The original design was very unstable (a 10 kOhm trimmer IIRC in parallel with some value I don't remember). The TC of the 2 kOhm resistor will govern the TC of R2 while in the original design the TC of the 10 kOhm trimmer will govern the TC of R2. The two trims (which should also be low TC resistors) are selected so that the trimmer can be adjusted for 5.03 Vdc at the top of the ladder (junction of R117 and R107). -
Hi,
here are some photos of the transformer and wiring, so that you can compare to your unit.
Wictor
-
continues...
-
So I’ve now changed the 10-Ohm in the 12V supply for a 5W version and it’s now down to about 110 degrees from 160. I’ve also replaced the 7812 regulator and checked all the caps ... it really does look like the 12v rail draws 300mA, there are six or seven transistors and low value resistors on the MPU board (not on the schematic) and the resistors get fairly warm, as does the CPU.
The transformer looks to be wired ok, although the numbers are correct but the connections are in a different position to Wictors photos.
I’ve also swapped the bizarre resistor arch for a 15K 5W, but this still gets hot, up to about 120 degrees. I haven’t moved it yet.
I’ve reseated all of the socketed IC’s.
None of that seems to have had any impact. I will try replacing the OP07’s tomorrow.
The voltage across the zener is absolutely stable, so that’s not the issue, it’s something else, and my guess is that it’s impacted by the heat but I’m not 100% convinced.
I will try to find a way to record multiple hours of data from my meter as it can only do 1 hour on screen and the problems seem to take two to three hours to materialise. I may also try some manual heat screening once I can consistently reproduce it so I can narrow it down.
The other thing I find strange is that every time I desolder a component I realise that there wasn’t actually a track there, and they’ve used the component lead to make the connection to a nearby track. It all feels a bit Heath Robinson!
-
Well that's just typical ... I ran my meter in aquisition mode for 3 hours, and no problem ... it seemed to take just over an hour to stabilise, then it's been pretty good (obviously the calibration is now wrong, so I need to adjust.)
This was from cold (room temp) with the top and bottom covers on.
I've attached a very quick excel chart based on the data (x axis is sample number, and they are at 405ms intervals) ... since I stopped the data it's been pretty stable at 10.000672 to within 2 or 3 uV.
I'm wondering whether it's a bit cooler in here this evening and it's just not tipped it over the edge (room temp is 23), and I haven't changed any of the OP07's yet either.
I think I'll re-run the same thing tomorrow ... very confused as to why I was seeing the inconsistency yesterday, but also we weren't anywhere near 600uV out!
Am I brave enough to leave 130 degree resistors running overnight to get a full nights data? ... no ... not yet.
-
I finally got back to this and managed to sort the issue...
Firstly I swapped all the op-amps, but there was no change.
So then I removed all of the CD4051's so it meant that nothing would be interfering with the zener and resistor string so I could figure out where the stability issues were coming from. I was measuring the voltage at the second 1K resistor (just because it was the easiest one to get my test clip on) and I was seeing similar stability issues after a bit of time.
The zener itself seemed to be stable, generally to a few uV once it had warmed up, although I had my doubts about the label on the board, and did some iterative tests warming to zener with my finger to find the voltage where the drift was the lowest ... I managed to reduce this from 50uV to 10uV ... so a good result, although I don't think it was really causing much of the issue above.
My next test was to put a temporary shield over that nasty high power resistor to try to stop the heat from that spreading across the rest of the board, and to my surprise the voltage at the resistor string almost immediately smoothed out and became completely stable. Removing the shield it took a good 10-15 minutes to go unstable again, but it did ... and replacing the shield again corrected it.
So my theory is that the heat from that resistor eventually gets to a point where it causes somewhat random thermal eddies around the resistors which then induces the instability. (Could be complete rubbish of course, but it fits what I was seeing.)
I've now moved the resistor back behind the heatsink-wall, and everything is absolutely perfect.
Thanks for all the help with this ... sadly I'm probably going to have to sell it to fund my newly found Datron calibrator addiction, but it's been a very interesting learning experience.