OK, let's go step by step.
Measuring is the way to go.
I said earlier that even without the servo correction (the 1/4 TL074) which is the icing on the cake, the output rail should be close to 0V all by itself.
Why? because if we have symmetrical rails, and symmetrical components, they should meet halfway, i.e. 0V or very close.
From top to bottom we have (as main voltage dropping elements):
R14 > Q3 > R10 > Q2 > Q4 > R19.
You'll notice that string is symmetrical, the "bottom side" is exactly the same as the "top side", only it's its mirror image (and so NPN instead of PNP).
We'll measure voltages across that string, expecting symmetry; if not, that points to the culprit.
1000 times better than "mesuring parts" at random which is slow, tedious, can introduce new errors and is not the shortest path.
And it may miss the real problem !!!
Please measure and post voltages at:
R14 top and bottom
R10 top and bottom
Q2 emitter
R19 top and bottom
When I say top and bottom, this refers to "as drawn", check what pin connects to other components , following the schematic.
In the actual PCB parts can be oriented any way, depending on PCB designer convenience.
I know this looks an endless task, but truth is: I do not know what the problem *origin* is, no Tech knows beforehand, and the only way to know is measuring.
If I had the amp on my bench, I'd "guess" what's wrong, and take some measurements to confirm or discard.
If not successful, guess another possibility (based on experience and what I find here) and test it ... and so on and on (until getting close enough).
I may have a lucky day and confirm the first or second guesses ; or it might take 10 or 15 tests in a real bad day.
Reality is that all those possibilities may be tested in, say, an hour.
But since we are communicating by Mail ... it may take many days.
So, please, be patient, I'm sure we'll repair this amp.
Good luck.
PS: your measurements tell me the main (officially +/-65V rails are symmetrical, (+/- 32V) , good, so now we proceed to the next step.
Measuring is the way to go.
I said earlier that even without the servo correction (the 1/4 TL074) which is the icing on the cake, the output rail should be close to 0V all by itself.
Why? because if we have symmetrical rails, and symmetrical components, they should meet halfway, i.e. 0V or very close.
From top to bottom we have (as main voltage dropping elements):
R14 > Q3 > R10 > Q2 > Q4 > R19.
You'll notice that string is symmetrical, the "bottom side" is exactly the same as the "top side", only it's its mirror image (and so NPN instead of PNP).
We'll measure voltages across that string, expecting symmetry; if not, that points to the culprit.
1000 times better than "mesuring parts" at random which is slow, tedious, can introduce new errors and is not the shortest path.
And it may miss the real problem !!!
Please measure and post voltages at:
R14 top and bottom
R10 top and bottom
Q2 emitter
R19 top and bottom
When I say top and bottom, this refers to "as drawn", check what pin connects to other components , following the schematic.
In the actual PCB parts can be oriented any way, depending on PCB designer convenience.
I know this looks an endless task, but truth is: I do not know what the problem *origin* is, no Tech knows beforehand, and the only way to know is measuring.
If I had the amp on my bench, I'd "guess" what's wrong, and take some measurements to confirm or discard.
If not successful, guess another possibility (based on experience and what I find here) and test it ... and so on and on (until getting close enough).
I may have a lucky day and confirm the first or second guesses ; or it might take 10 or 15 tests in a real bad day.
Reality is that all those possibilities may be tested in, say, an hour.
But since we are communicating by Mail ... it may take many days.
So, please, be patient, I'm sure we'll repair this amp.
Good luck.
PS: your measurements tell me the main (officially +/-65V rails are symmetrical, (+/- 32V) , good, so now we proceed to the next step.
Last edited:
So here is what I found.
I have to say that the voltage change during time, so they are not 0.1 accurate.
If I stoppend the measurement and checked some points again, then they were a bit off, but it should not make a huge difference I guess.
So What I noticed is that blue box number. It started to rise if I connected my DMM + probe to it. And it rised very fast (seconds) as I could not detect a start point. (my DMM is too slow)
Bias is set to low. Trim resistance max.
svtStep3 | Flickr - Photo Sharing!
I have to say that the voltage change during time, so they are not 0.1 accurate.
If I stoppend the measurement and checked some points again, then they were a bit off, but it should not make a huge difference I guess.
So What I noticed is that blue box number. It started to rise if I connected my DMM + probe to it. And it rised very fast (seconds) as I could not detect a start point. (my DMM is too slow)
Bias is set to low. Trim resistance max.
svtStep3 | Flickr - Photo Sharing!
Last edited:
The key is the -24.4 on the collector of Q3, why is that negative? Have you checked your source resistors on the positive side, R32,34,36, and 38? If those are open that might explain the imbalance.
Craig
Craig
If you had showed this voltage chart from the beginning, we would have saved a lot of time 😉
Pulling parts and measuring then "outside" is slow, dangerous, and it's easy to miss a mistake.
While making it work, shows what's not normal" at once.
Anomalies:
1) Q3 : shows 55.7V on emitter, 54.1V on base ; so Vbe -1.6V forward biased (the base on a PNP needs to be more negative than emitter) -> Impossible value.
The base in a forward conducting PNP will be around -0.7V from emitter (It's a diode after all) so we have 3 basic possibilities:
a) Q3 is open. Replace with a good one. BIG suspect.
b) Q3 is miswired. Some pins went into the wrong holes. 😱
c) because of messing with it, some pad or track cracked, 😱 same end effect.
d) to further confirm the above possibilities, Q3 stands 55.7+24.4V = 80V , yet passes no current, further indication it's open.
2) Q4 looks fine: 55-54.2= 0.8Vbe, forward biased, normal value.
3) Q5: Base and Emitter voltages look reasonable, but don't believe the 0V on its collector.
Please recheck.
4) Q6 C and E voltages look normal.
It has Vbe -25.1-28.2V=3.1V (positive) which in a PNP means reverse biased.
The value is possible and shows the circuit is trying (unsuccessfully) to correct those weird voltages.
Which it can't because Q3 is open, one way or another.
So, in a nutshell: check Q3, it may be open, miswired or tracks leading to it may be cracked.
Typically a hairline crack where it meets the pad, because too much soldering and unsoldering makes the pad contract and expand (relative to the cold track).
Check that and post results.
Worst case, we might have to momentarily pull protection transistors Q5 and Q6 (or lift one end of D28/29 , but first I'll wait for your results.
Pulling parts and measuring then "outside" is slow, dangerous, and it's easy to miss a mistake.
While making it work, shows what's not normal" at once.
Anomalies:
1) Q3 : shows 55.7V on emitter, 54.1V on base ; so Vbe -1.6V forward biased (the base on a PNP needs to be more negative than emitter) -> Impossible value.
The base in a forward conducting PNP will be around -0.7V from emitter (It's a diode after all) so we have 3 basic possibilities:
a) Q3 is open. Replace with a good one. BIG suspect.
b) Q3 is miswired. Some pins went into the wrong holes. 😱
c) because of messing with it, some pad or track cracked, 😱 same end effect.
d) to further confirm the above possibilities, Q3 stands 55.7+24.4V = 80V , yet passes no current, further indication it's open.
2) Q4 looks fine: 55-54.2= 0.8Vbe, forward biased, normal value.
3) Q5: Base and Emitter voltages look reasonable, but don't believe the 0V on its collector.
Please recheck.
4) Q6 C and E voltages look normal.
It has Vbe -25.1-28.2V=3.1V (positive) which in a PNP means reverse biased.
The value is possible and shows the circuit is trying (unsuccessfully) to correct those weird voltages.
Which it can't because Q3 is open, one way or another.
So, in a nutshell: check Q3, it may be open, miswired or tracks leading to it may be cracked.
Typically a hairline crack where it meets the pad, because too much soldering and unsoldering makes the pad contract and expand (relative to the cold track).
Check that and post results.
Worst case, we might have to momentarily pull protection transistors Q5 and Q6 (or lift one end of D28/29 , but first I'll wait for your results.
Changed Q3.
Nothing changed.
Q5 reading re checed and I still got 0VDC.
Diode is blocking - and cap is blocking DC.
Nothing changed.
Q5 reading re checed and I still got 0VDC.
Diode is blocking - and cap is blocking DC.
To be more precise, what does "gone" mean?
How did you find out?
EDIT:
But we *still* have a problem there.
It is impossible that you have 1.6Vbe there
Measure again, on the transistor legs themselves.
If still anything above 0.7V, pull that transistor (with amp off) and measure the BE diode in the multimeter diode scale.
Red to emitter and black to base.
Just for confirmation, read it the opposite way.
Are you using a real 2N5415 or some substitute?
Last edited:
Hi.
I found out by just measuring the resistance on thouse Resistors and found that there was no resistance.
Q3 is 2N5415.
I will change thouse resistors today.
BE Diode was fine.
I will read about oscillation today as I dont have a good understanding of it.
I found out by just measuring the resistance on thouse Resistors and found that there was no resistance.
Q3 is 2N5415.
I will change thouse resistors today.
BE Diode was fine.
I will read about oscillation today as I dont have a good understanding of it.
What does "no resistance" mean?
Open or short?
Did you measure them on the PCB or out of it?
Not an answer.
I asked:
I need a number, not an opinion.
Thanks.
Hi.
Desoldered D2.
Diode D2. Multimeter in a low resistance mode.
Anode-Red Cathode-Black: 707
Anode-Black Cathode-Red: 1787
Its a reading in resistance, sorry, don't have a diode mode.
It was difficult to understand if you ment D2 or you wanted me to measure Q3 BE in a diode mode. I measured D2.
Resistors were open. I first measured them on the board and then took them out, reading didn't change.
Desoldered D2.
Diode D2. Multimeter in a low resistance mode.
Anode-Red Cathode-Black: 707
Anode-Black Cathode-Red: 1787
Its a reading in resistance, sorry, don't have a diode mode.
It was difficult to understand if you ment D2 or you wanted me to measure Q3 BE in a diode mode. I measured D2.
Resistors were open. I first measured them on the board and then took them out, reading didn't change.
Last edited:
Ok, thanks, now I see.
Use "open" or "shorted" instead of "no resistance" which can have both meanings.
They must have blown open by a *strong* overload.
Since gates are insulated and do not pass current, those MosFets must have turned into blobs of molten metal, plus ballast resistors must have blown open too.
Talk about "overload".
That amp was probably "repaired" with a 20 or 30 Amp fuse pulled out of a car , a ball of crumpled aluminum paper or something like that.
Only way to achieve such a meltdown.
I wanted you to
, both ways.
But don't know whether I can trust your multimeter or not 🙁.
The Diode Scale *often* but not always, matches the 200 ohm scale, hope yours does.
To test it, get a clean unused diode (any one: 1N4002 , 1N5402, whatever) and measure it both ways.
I can trust that.
Does it have any label?
As said before, I trust numbers more than descriptions, if that makes sense.
Thanks.
EDIT: by the way, having R27 and Q6 burnt is consistent with having R40/41/42/43 burnt open, all point to a massive short on the negative side MosFets.
Last edited:
Hi.
So I truned the amp on. After changing resistors I got +49V on the output.
My DMM is looking like this http://www.ehighend.de/pics/peaktech/2010/2010-product.jpg
I measured it not in the 200ohm mode but this mod has a diode drawn on it and you can make it beep if leads are shorted. (sorry for this bad description)
Some observations (sorry, no number at the moment as I didn't want to risk it)
The Bulb was very dark now. Turning the Bias pot didn't affect the bulb.
Q3 BE.
Red Emitter; Black Base : 638
The other way around OL. (Open loop, over-Limit)
DMM test 1N4007
Forward 573
Backward OL
DMM test 1N4148
Forward 600
Backward OL
So I truned the amp on. After changing resistors I got +49V on the output.
My DMM is looking like this http://www.ehighend.de/pics/peaktech/2010/2010-product.jpg
I measured it not in the 200ohm mode but this mod has a diode drawn on it and you can make it beep if leads are shorted. (sorry for this bad description)
Some observations (sorry, no number at the moment as I didn't want to risk it)
The Bulb was very dark now. Turning the Bias pot didn't affect the bulb.
Q3 BE.
Red Emitter; Black Base : 638
The other way around OL. (Open loop, over-Limit)
DMM test 1N4007
Forward 573
Backward OL
DMM test 1N4148
Forward 600
Backward OL
Last edited:
Reconnected Feedback and output has 0V.
I would like to leave this on my desk for couple of hours with the bulb connected, but I will measure everything and post here before I do that.
I would like to leave this on my desk for couple of hours with the bulb connected, but I will measure everything and post here before I do that.
No, the description is fine.
I also found your Multimeter Manual, but even without it, your diode (1N4007/4148) measurements confirmed it 😀
NOW I can trust your measurements 😉
Well, *this* is what I wanted to check (but I had to trust your Multimeter first): Q3 BE junction *now* measures loke a regular diode, around 600mV (similar to what you measured on those 2 others) .
BUT I was worried about what you had posted earlier, showing 1.6V (or 1600mV) BE at Q3 , which is impossible, unless the transistor is open. (Or "is not there" , as in cracked solder or tracks).
As you see, besides measuring parts, the faster and more accurate way is to "make the amp work" (or try to) and measure abnormalities.
Yes, they will probably point to a "bad part" ... or not ..... : bad connections, bad biasing, missing rail voltages, you name it.
You may spend a month pulling and measuring *everything* in a complex preamp .... while the real problem was a missing -15V rail or 1 broken pin or wire in a ribbon connector.😡
Well, a dark bulb is a dark bulb, in this case it's acceptable info, but later I'll suggest some voltage measurements.
Now my lunch break is over, will post later.
But it will be similar to your earlier post where you showed some key voltages .... which now will be different, of course. 🙂
EDIT:
GOOD!!!
Looks loke we are real close now.
Only stone in the path will be to check Bias to avoid both your original problem (crossover) and the last one (fireworks) but I feel we almost are done 😛
Last edited:
Yes the bias. I will not bias it before I know what I am doing, so I will spend some time reading.
After everything is successfully done...
I was thinking to write an article based on how you guided me through it.
Do you have any place in mind where it should be found in the future?
This has teached me so many new things that I could not even describe what kind of enthusiasm is bulbing inside me.
After everything is successfully done...
I was thinking to write an article based on how you guided me through it.
Do you have any place in mind where it should be found in the future?
This has teached me so many new things that I could not even describe what kind of enthusiasm is bulbing inside me.
Hi.
I think there is still something wrong.
If I turn the BIAS pot then voltage on the output rail starts to rise again. I will make some measurements.
I think there is still something wrong.
If I turn the BIAS pot then voltage on the output rail starts to rise again. I will make some measurements.
you need to be looking at the speaker output terminal with an oscilloscope. Input and master volume levels turned all the way down to zero. while watching the scope. if you turn the bias pot too far you will see oscillation very high in frequency. this is bad. turn the bias pot the other way. a scope is the only way you will know for sure!
To simplify what oscillation means. the amp has become unstable and has started to "squeal" for lack of a better word, very high in frequency, often way above the audio band. your meter may not be able to read that high in frequency so it just looks like DC. the amp may even produce some DC offset. The oscillation can be very high in level. can cause the output devices to overheat and die. which is exactly what I suspect happened to your amp (i have seen it many times). these amps are notorious for this!
- Status
- Not open for further replies.