| Electronics > Repair |
| HP 8116A Error 42 (E42) Repair |
| << < (5/6) > >> |
| henmill:
Hello, and sorry to keep you all waiting, no doubt anxiously for my update haha. I have not gotten as much time as I'd like the past few days to dig into this old beasty, due to silly things like spending time with family :) But, I have managed to think a little bit more logically about how to step through testing and thus developed a crude plan. Said plan involved probing a handful of points I deemed informative: 1) U502.2 the inverting input of the offset control op amp driving the moving reference of the output (at least I think that's what's happening) p.s.) looking back, maybe I should have looked at the vernier control input at C501 2) The node at CR501 cathode / CR502 anode - output post R509 of offset opamp U502 3) Vbe of Q501 4) Vbe of Q502 5) The node where R528/529 meet which is the output reference potential So I checked these at a few different settings, first starting with no waveform, 0V offset, output disabled. Then with output enabled, everything was essentially the same, except output reference was closer to 0V So the meat of what I've gathered is just looking at a 3Vpp sine wave, 1kHz and stepping through different offset voltages. The first attachment shows those values as well as a screen cap showing exactly where on the board I was probing. My initial takeaway from these measurements is that U502 appears to be just saturating to the negative rail as soon as offset is set to 0V and below. Whereas CR501/502 node stays about a diode drop below 0V when offset is positive. I reckon this is due to something not working (duh) and it immediately driving as hard as it can negative in attempt to get something but alas. Now, I decided to also poke around with the scope, to see if I could find where in the chain I lose the bottom (top?) half of the signal. So I have some scope captures looking at the signal coming from the shaper IC at R511 (ch2 blue), then the other side of R511 (ch3 fuscia), and then the final output (ch1 yellow). You will see the signal come in and out on ch3 on R511, which really is the inverting input at U502.2. I bothered to do proper scope caps but didn't bother to setup measurements on all relevant signals but oh well. Maybe next time ;) Hopefully others can glean more from this exercise than me. But the big takeaway to me is that something is not working on the Q501 side probably, or something in the feedback network of the offset amp is buggered? Which isn't getting me much further but I think I am slowly absorbing the linear amp function and expect my eureka moment any day now... Until then, cheers! edit: note the "dots" on the drawing are the probe points. I also labeled the nodes base and emitter as well as relative polarity so it's a bit cluttered. |
| henmill:
Very quick update: I checked the Offset Vernier Control signal coming from the control board, and it lined up with what is shown in the manual. I don't have those numbers in front of me, but the important thing I observed is that the analog signal coming in is indeed dual polarity, ranging from (about) -2.8V to +2.8V for offsets +7V to -7V. So, I'm still betting it is some failed component on one side of the amp. Just having a little trouble figuring out how to determine that with everything in circuit. Thanks! |
| MarkL:
A couple of observations... U502.2 should always be 0V since the non-inverting (U502.3) input is ground. Note that when U502.2 is 0V, everything is ok. When U502.2 goes positive, it causes the U502's output to saturate negative, which explains the extreme negative readings you're getting on Q501's base (and causing Q501's collector to go positive). The collector of Q501 is clearly signalling for more negative output by going positive, but is being ineffective. Your yellow output trace is always at +250mV when things aren't working, and is probably what's causing the positive input to U502.2. The output should be 0V in static conditions, so something is dragging this high and not allowing it to get below +250mV, or the negative pull down is not working. I think what I would do at this point is start looking backwards from the output and check bias and compare voltage levels on the output drivers Q510 thru Q513, then Q508/Q509, and so on. |
| henmill:
Hello again and thanks as always for the feedback. (speaking of which, in a moment...) MarkL, you are absolutely right about your explanation of the offset control amp. I am slowly but surely starting to understand the components of the amplifier. Emphasis on slowly haha But to that point, I decided to take a crack at simulating the circuit for funzies. I started with just the input stages and was happy to see it more or less function as a buffer. Then just tonight I've built out pretty much the rest of the gain stage and final power output. I am not using opamps yet for the offset voltage and current control. And I'm happy to report it works! ....but I had to add in my own feedback network to set the gain (i.e., amplitude) instead of it kinda just maxing out. For some reason I am not understanding where the feedback network is and how it's controlled, so I'm asking for your help! Does it have something to do with Q590 and Q591 and their resistor networks? (this seems like the most logical). But feeling like a duntz I don't understand the schematic Number/Lettering scheme? Do adjacent letters connect across sheets? (i.e., A to B, E to F, etc.) So, first pic is those transistors in question, do the nodes labeled 1E both get tied to 1F? Surely not right? Next pic is the schematic of my simulation in LTSpice, followed by an example output waveform. One really big observation I've gotten already is that tweaking the resistor network for Q507 (labeled "Bias Adjust") has a huge effect on the offset of the output waveform. AND, I poked at those nodes the other night and found them all to be at the same potential, about 1.25V.... which I found curious. I measured across the resistors with the power off and it checks out about 1k across Vce and R536 measured about right too. So it's not a dead short through Q507. To me, this leads me to think maybe any one of Q503-Q506 could be screwy and not turning on to allow current through the Bias Adjust network. Is it worthwhile to measure resistance across the different nodes of each transistor with the power off to potentially assess a failure? Or check for diode Vf across base to emitter? Thanks in advance for any ideas! And let me know if you have any questions or suggestions on the simulation, or want to see different settings run! |
| MarkL:
--- Quote from: henmill on November 05, 2024, 05:25:45 am ---... But to that point, I decided to take a crack at simulating the circuit for funzies. I started with just the input stages and was happy to see it more or less function as a buffer. Then just tonight I've built out pretty much the rest of the gain stage and final power output. I am not using opamps yet for the offset voltage and current control. And I'm happy to report it works! --- End quote --- Cool! Can you share your spice file? --- Quote ---....but I had to add in my own feedback network to set the gain (i.e., amplitude) instead of it kinda just maxing out. For some reason I am not understanding where the feedback network is and how it's controlled, so I'm asking for your help! Does it have something to do with Q590 and Q591 and their resistor networks? (this seems like the most logical). --- End quote --- The feedback path from output to input starts at R541/R542 and continues through R528/R529. Q590 is associated with "ERROR FEEDBACK", and Q591 with "LEVEL CHECK" according to where they are connected. They are both TTL level signals, so the unit is probably using these signals for checking the output during self-test. --- Quote --- But feeling like a duntz I don't understand the schematic Number/Lettering scheme? Do adjacent letters connect across sheets? (i.e., A to B, E to F, etc.) So, first pic is those transistors in question, do the nodes labeled 1E both get tied to 1F? Surely not right? --- End quote --- No, not right. The "1E" refers to the schematic sheet number shown in the lower right corner of all the schematics. Sheet 1E is in the VCO and Width generator section (sec. 10.3). The number in the circle refers to the node number on the sheet. --- Quote ---Next pic is the schematic of my simulation in LTSpice, followed by an example output waveform. One really big observation I've gotten already is that tweaking the resistor network for Q507 (labeled "Bias Adjust") has a huge effect on the offset of the output waveform. AND, I poked at those nodes the other night and found them all to be at the same potential, about 1.25V.... which I found curious. I measured across the resistors with the power off and it checks out about 1k across Vce and R536 measured about right too. So it's not a dead short through Q507. --- End quote --- "Those nodes"? All of them? Can you be a little more specific? --- Quote ---To me, this leads me to think maybe any one of Q503-Q506 could be screwy and not turning on to allow current through the Bias Adjust network. Is it worthwhile to measure resistance across the different nodes of each transistor with the power off to potentially assess a failure? Or check for diode Vf across base to emitter? --- End quote --- I would be careful measuring resistance in-circuit. Depending on the voltage being used by your meter, it may be turning on junctions and you might not get an accurate reading, or there may be other passives interfering. Measuring Vf (Vbe) under power I think is a better test since you don't have to wonder about interfering components, but it still does not say if a transistor is 100% ok. --- Quote ---Thanks in advance for any ideas! And let me know if you have any questions or suggestions on the simulation, or want to see different settings run! --- End quote --- I would really encourage comparing quiescent voltages with the schematic sheet I posted of the working unit. The problem area will likely become very obvious. If it helps at all, below is a version of the schematic with blanks in all the measurement boxes if you want to write them down as you go. The source document is in LibreOffice Draw - if you want that let me know. Note that Q510/Q512 and Q511/Q513 are just drivers in parallel, and they are being driven as emitter followers by Q508/Q509, which are emitter followers too. So the bases of Q508 and Q509 should be about 2 diode drops above/below the output. From your readings, it sounds like there is an issue around Q507. Perhaps Q506 is not pulling down for some reason. Look at the voltage levels at Q506 and Q504 with respect to ground. |
| Navigation |
| Message Index |
| Next page |
| Previous page |