Oscilloscope Repair

[pierreretief]

OK, I get 210K and 162K for VR6, and I made a mark and changed it, at first it took the wonkyness away, but then I realised it just moved the trace up. Looks like its some kind of fine adjustment for the position knob, or it sets the position itself

Should I adjust it so that its in the middle?

 

[pierreretief]

As for the rest of the tests..I need to put that aside this week, since I got my test reslults back and it seems I have failed 2 tests, so I need to study really hard for the second one, but this weekend I will get all the measurements you need.

Thanks again, gonna try and get that second scope this week also

 

[shimniok]

You could try it in the middle, but sounds like no huge, obvious problem with VR6...

Will hang out until this weekend, no worries... best of luck on 2nd test!

Michael

 

[unclejed613]

the first graphs were somewhat confusing, the third graph says "Q10" isn't following a normal trend in Vbe. all of the Vbe lines should be somewhat linear but Q10 seems, out of the bunch, to have a serious foldback in Vbe. it doesn't necessarily mean that Q10 is the problem, but is close to it. since the trace doesn't go past one edge of the screen (i'll have to look back at pg 1 to remember whether it was top or bottom), the DC balance of the channel is already skewed, and that's what VR6 does IIRC. but the balance problem isn't the main problem, but does seem to contribute to it. one thing i would like to see is how much AC shows up on the power supply rails at the various settings of the position control (or maybe that's been done already?).

 

[shimniok]

The first graphs show Vbe and Vce plotted together per transistor for up, mid, and wonky trace positions. Q10 is part of the symptom, I think. I didn't plot Q13-16 which are a mess in terms of Vbe and Vce trends. Not to mention Q11. (We don't have Q12 info yet)

My thought is that oscillation in the amp (an old saying I heard: "amplifiers oscillate and oscillators amplify") occurs when it's driven to a certain point due to an imbalance in dc offset due to some bad component. Seems like this could conceivably explain all the oddball symptoms in all the trace positions... ?

I *think* we checked for AC on various sources with the trace wonky but I will have to go back and check...

The other possibility is that, given Q5/Q6's behavior -- Vbe and Vce reach their max/min values at the 'wonky' position. They don't go lower/higher in the 'down' position. So maybe the input signal dc offset is off earlier in the circuit and the amp is just attempting to amplify but oscillating when it does. Just upstream is Q5/6 is Q21 and a pair of resistors R83/85, 2.2kΩ each. Maybe something is wrong there.

I would really like to see the actual voltage ranges w/r/t ground on Q5 and Q6 emitters for up/mid/wonky/down. I'll add that to the list.

Michael

 

[unclejed613]

you should plot them anyway.... some plots will look odd. i still would like a verification on Q11's 2.6V Vbe, and it's polarity compared to the normal readings of Vbe on Q11.

 

[shimniok]

you should plot them anyway.... some plots will look odd.

My point for plotting the above was to show that those trannies seem relatively normal and well behaved. The other data would've made the plots even less readable.

i still would like a verification on Q11's 2.6V Vbe, and it's polarity compared to the normal readings of Vbe on Q11.

I think that's on the list. I suspect it is negative too. That would kind of line up with the dc imbalance theory, perhaps.

Michael

 

[unclejed613]

this schematic is MESSED UP!!!!!

i just noticed in the schematic a direct short from the +12V feeding the collector resistor, straight to the emitter...... one of the connection dots shown shouldn't be there. the dot to the right of R74 isn't supposed to be there....

 

[shimniok]

this schematic is MESSED UP!!!!!

i just noticed in the schematic a direct short from the +12V feeding the collector resistor, straight to the emitter...... one of the connection dots shown shouldn't be there. the dot to the right of R74 isn't supposed to be there....

I know... plus several components aren't labeled etc... crazy.

 

[unclejed613]

ok, to get 2.6V Vbe on Q11 (assuming it's the correct polarity, since the Vce is very low), there would have to be close to 15V applied at the base of Q11 through R65,66,67, which would mean that something's wrong connecting to the emitter of Q9, either that or something is sucking down the emitter of Q11, which should be at +12V. the more likely cause is something wrong between the +20V rail and Q9. maybe what needs to be done is to establish a baseline reading using ground referenced voltages at all of the transistor terminals with the position controls at center, and R6 at center. if everything were working properly, the readings on the top transistors should be close to those on the bottom transistors. because something isn't right, i suspect that this method might help find the problem. after establishing a baseline, then go in and get readings with the control up, down,etc... instead of rotating the pot for each measurement taken, put the pot at the top, and retake all the readings, put the pot down and retake the readings, etc.... this will give us a better clue what's going on here. maybe it should have been done sooner, but i thought the Vbe and Vce readings would give enough information. if it were a flaky transistor, it probably would have been plenty of data to find the problem. something tells me that just taking Vbe and Vce readings has created a situation where we can't see the forest through the trees.....

 

[shimniok]

I replotted Vbe and Vce DC readings. I chose the same color for each partner transistor (e.g., Q9/10 or Q13/14), Q9-19 use dashed lines, Q10-20 use solid lines. On Vce I used two scales -- Q13-16 due to higher Vce voltages, get their own larger scale. I also scaled Vbe in a reasonable range from about 0.5 to 0.8 -- so those readings that fly off the chart are obvious.

All the Vbe readings we have:
**broken link removed**

All the Vce readings we have:
**broken link removed**

 

[shimniok]

Ok, I went back through and re-read the ENTIRE THREAD to see if there was anything interesting that we kind of passed over.

Important points:

• In original post said that the scope trace started to go out of balance over a couple months and got progressively worse until the wonkiness appeared -- This sounds like a component going bad
• Unclejed early on suggested an oscillating amplifier or supply voltage problems.
• The wonky trace stabilizes to a line after about 2 minutes (mentioned at least twice in this thread) -- This seems more likely to be a damped oscillation to me, rather than power supply / voltage reference issue
• Some of the electrolytic capacitors have already been replaced (C17, 18, 20, 25) but this fixed nothing of course.
• We discovered AC on the Y+ pin and no AC on the Y- pin. An early measurement of R91 showed AC across it -- This is a 1k resistor between Q17 and Q18. This might be worth looking into further.
• We are seeing a little bit of Vce/Vbe AC on the Y- side.

Stepping back... in troubleshooting, usually there's only one root cause, rarely two, almost never three. I think the most logical explanation is that one component that has gone bad, resulting in the wonky trace, offset trace position, and oddball voltages on the Y+ side.

Originally we wanted to see if any transistors showed up as obviously bad. We found some obvious misbehavior in the circuit doing this but we now need to capture more measurements as unclejed said.

Michael

 

[shimniok]

Let's put the prior to do list on hold for a moment and do the following.

Per unclejeds request let's get a baseline of voltages for the transistors with the VR6 balance and the vertical position in the center. We need to do two things. One, we want to see where the DC imbalances are in the circuit. Two, we want to see where AC is and isn't in the circuit.

So, for One:

1. Set VR6 to the physical middle (check resistance from middle to each leg; resistances should be within say 5-10% of each other)

2. Set the vertical position knob to the physical center (as close as you can get it). The trace willl probably be up towards the top. We'll measure DC voltages so if the trace is a little wonky that is probably ok.

3. For each of the following transistors, measure DC voltages for Vc, Vb, and Ve -- that is, measure DC voltage with respect to ground. Red lead on the transistor pin, black lead clipped to chassis ground, ie, the metal frame.

a. start with the inputs and outputs: Q5, Q6, Q13, Q14, Q15, Q16
b. work your way out to Q9, Q10, Q11, Q12, the rest of the amp
c. then all the remaining transistors: Q7, Q8, Q17, Q18, Q19, Q20

What you're looking for -- the readings for transistor pairs should be fairly close. E.g., Vc for Q13 and Q14 should be fairly similar. If you find a pair of transistors for which the reading doesn't measure the same it is possible that an upstream (or downstream) resistor is more or less resistive than it is supposed to be -- thus it drops too little or too much voltage, causing an improper voltage on C, B, or E of one or more transistors. It may not be that simple. You may have to trace back a ways. Also, one resistor may affect many transistors, like R79 which is in the feedback path for the amp, so it perhaps affects Q11, Q17, etc., and it also may deliver less or more voltage to R66 which may appear mismatched with R92. Make sense?

Anyway, you could compare measurements at mirror points in the schematic/circuit like R76/R77, R67/R68, R65/R70, R72/R73, R79/R51, R55, R92, or R63/R64, and so on. See what I mean? Might as well also check L101/L102 voltages. If it is a single resistor (or one of those two inductors) that went south and which is just way off causing the DC balance problem as well as the amp oscillation, then we might be able to track it down this way. But start with the tranny measurements to see what we are dealing with first, then if you have time, keep searching.

Take meticulous notes.

Michael

 

[unclejed613]

actually the oscillation stabilizing to a line after 2 minutes could be a thermal effect. an unstable condition would be sensitive to any changes, and start up again any time a voltage change were input to the amplifier, or if a slight adjustment to the trace position were made. it would be another step to determine the cause of the thermal drift. put the amp into it's unstable condition and allow it to stabilize, then go through the circuit one component at a time with freeze spray and find out which component is thermally unstable. when you hit the unstable component with freeze spray, the amp will go back into oscillation. be aware that there is usually a 1 or 2 second delay from when you hit a component with freeze spray to when the component temperature actually drops below room temp. this is due to thermal lag of the component case, especially the plastic case of transistors. as you get closer to the source of the problem, you may get the trace jumping or drifting, but not completely "wonky" until you get really close to it. if you think you have found the offending component, verify it by allowing it to heat up again until the trace stabilizes again and hit other components around it to see if you get a sharper reaction from any of them. when you find the bad component, it's behavior should be the most radical with a temperature change, going back into oscillation almost immediately with the drop in temp, and taking the longest to stabilize again.

 

[shimniok]

actually the oscillation stabilizing to a line after 2 minutes could be a thermal effect. an unstable condition would be sensitive to any changes, and start up again any time a voltage change were input to the amplifier, or if a slight adjustment to the trace position were made.

Yes, the underlined bit above is precisely what happens. An earlier post describes this behavior -- after the trace stabilizes, any change to the vertical position and the wonkiness starts again.

Does that mean it is, or isn't, a thermal effect?

Michael

 

[unclejed613]

that might not be a thermal effect then. there may be a thermal factor to it, but if it were totally or mostly a thermal effect, it would only act up when cold, or after the unit warms up, or within a narrow temperature range. if it begins acting up again with a voltage change, then the problem is that it's marginally unstable. either there's a path for too much positive feedback (most likely) or some form of negative feedback (either true global negative feedback or local degenerative feedback) has been reduced to cause the instability. a reduction in local degenerative feedback is also likely. positive feedback in some form is the most likely cause, either through the power supply or through a degraded bypass or ground path. this type of feedback is usually the result of a dried out electrolytic cap in the power supply or at the bypass points for one of the amplifier stages (or even the noise bypasses for one of the current sources).

degraded local feedback would be happening within one of the amplifier stages, and would be evident in the circuit we're looking at as an imbalance of gain between the inverting side (Y-) and noninverting (Y+) side. the stage causing the problem would have too much gain, and this increase in gain is because the components that limit the gain of each stage have changed in value. this could be either a transistor, diode, resistor or a cap. there are more possibilities here, some of them unlikely (resistors rarely drop in value, and capacitors rarely increase in value, but transistors and diodes may exhibit decreases in Vf at their junctions, or increases in noise currents).

degraded global feedback is also a possibility, and if accompanied by a certain amount of phase delay, can cause oscillation at a frequency determined by the delay. such oscillations are usually at the upper frequency limits of the amplifier's response, which doesn't seem to be the case here... components that tailor the low frequency and DC response of the amplifier, however can cause low frequency oscillation or reduce the feedback at low frequencies enough to cause oscillation, so any components that limit the low frequency gain through the global feedback loop should be looked at (which in this amp is very simple and not likely, but worth checking anyway...)

so, likely places where positive feedback can occur are IC402, C441 (+12V source)
IC405, C406, D417, (+20V source), IC403, C413 (+5V source). C405,C404, C403, C401 (+120V source).... and the bypass points, C16, C17, C18, C20, C22, C25.

likely places where local degeneration could be degraded:
D10, D11(actually these are voltage references for the cascode sources), R67, R68, R74,R75, Q13, Q14, Q15, Q16 (these are the cascode current sources), and the resistor string between Q17E and Q18E).

the "global" feedback network isn't really very "global" since it doesn't go all the way back to the input pair in the preamp IC's, but goes from the deflection plate connections to the previous stage. R78,R81, R66, R65 form the feedback loop for the + side, R81, R82, R71, R70 form the feedback loop for the - side

 

[shimniok]

degraded local feedback would be happening within one of the amplifier stages, and would be evident in the circuit we're looking at as an imbalance of gain between the inverting side (Y-) and noninverting (Y+) side. the stage causing the problem would have too much gain, and this increase in gain is because the components that limit the gain of each stage have changed in value.

Based on the Vbe numbers so far... it looks like we are seeing Q13 fully off with the position "middle" to "up" and full on at about the "wonky" point and below (along with AC on Vce of each), along with the wonky trace. We're not sure wtf is going on with Q11 with the trace "up"

It also looks to me like some/many transistors of both sides sort of range only between the up and wonky positions with little change in Vce or Vbe below "wonky".

So to me it seems less like too much gain and more like the Y+ amplifier is being driven outside of its normal range of operation, or has a narrower than designed operating range, like it is oversensitive. While both sides are amplifying across a position range less than the full sweep of the position knob ... and/or the entire thing has an incorrect dc balance... ?

Does any of that make sense?

I don't know enough to know if that fits with any of the types of problems you described. Any thoughts on that?

Do the above symptoms, or the other symptoms we've seen so far, suggest anything?

For example, we know C25 voltage dips with the trace all the way up. But that's up, not down when it goes wonky. We *think* there was AC on R91 between Q17 & 18 -- but need to verify that.

Is it ok to proceed in measuring the voltages at the 'midpoint' to see if we have an imbalance?

We do have measurements for some of the above, like the bypass caps, no AC found, and also several of the 'lytics were replaced. Diode tests were done on most of the diodes and they checked out.

Michael

 

[unclejed613]

what i'm thinking is that whatever has caused an imbalance between the two sides (resulting in the DC balance being way off) is also causing a change in gain of one of the stages. the gain imbalance has shifted the operating points of some of the transistors into a nonlinear portion of their curves. part of the reason of using a cascode amp in the first place is to linearize the amp by reducing the changes in Vce across the transistors acting as amplifiers. as you can see from your chart, this works well for Q19 and Q20, but not as well for Q11 (and there's no plot for Q12). the majority of Vce change is supposed to be across Q13, 14, 15 16, which it is, but there seems to be problems with this action not being completely linear, and so we have a change in gain between mid and down positions, and this sudden change in gain seems to accompany the oscillation. whether it's caused by one of the transistors saturating too much drive) or going to an open condition (not enough drive) isn't entirely clear, but neither condition should exist inside the linear portion of an amplifier's normal operating limits. yet we have both conditions occurring here with Q13 saturating at the mid point, and Q14 going open circuit at the wonky point.

 

[pierreretief]

Hey guys, Thank you so much for all the effort. I am sorry I havent had time to do the rest of the tests. wanted to do it the weekend, but had a rewrite and I really needed to put my attention into it. I prob could not have even have done it, since I wrote it last night, engineering maths, and it really went bad. Anyway, so I got the cash and I am going to buy that second scope tomorrow, and this weekend I am gonna get all the measurements. Thanks again so much. I am sure we will get this thing fixed some time or another

 

[pierreretief]

I am going to do the measurement list earlier on this page. Is that correct?