Yes. Rectifier's pinout and the boards hole pattern only allows the it to mount one way.
Always refer to the circuit and trace things throughThe accessory sockets are 'before' the mains switch so the switch is the next likely suspect. With the amp unplugged try and measure continuity directly across the primary. You should have continuity between Brown and Green and between Orange and White.
Finding and exposing places for the DMM probe tips to connect with so as to make those continuity checks was easier said than done.
It was a bit awkward but I think I accomplished it. Had to undo some factory wire bundling to see where those wires went. The orange one required driving the probe tip through the insulation.
I noticed while doing this that both fuses had blown again. Not sure when that happened but I replaced them.
The result of the continuity check was:
Brown wire to green wire - no continuity.
Orange wire to white wire - yes continuity,.
I concluded that the wires you identifed by colour were those labeled on the schematic drawing of the transformer / power switch section?
ORG / WHT / GRN / BRN
The white wire and brown wire go to the fuses.
This excercise might have been pointless. The reason I say so is that I wondered why the power on lamp did not light.
After just now replacing the fuses again, I once more used the DBT and variac to see if the power on lamp lit up. But at about 50/60 volts, the 100W DBT bulb glowed brightly once again.
This is really frustrating.
That lamp on your diagram (PL1) is a neon and needs a high voltage to begin to work (70 to maybe 90 volts depending on type) so it may or may not work with a DBT or variac. It is also wired before the fuses and so on full mains would be expected to light whether or not the fuses were intact.
So again back to the beginning. Its a case of measuring voltage (such as across those emitter resistors) to see where the current is going.
A few days ago it all seemed OK and we know you had a mishap with the meter leads and shorted something. So we don't know if there is a 'real' problem or whether its something else
The 100W DBT was glowing brightly with the variac at full power when I re-measured voltage across the emitter resistors. Measurements are as follows:
R319.....0.0
R317.....0,0
R318.....0.0
R316.....0,0
R418.....7.0 volts
R416.....4.6mv
R419.....0.0
R417.....113.0mv
Twice during the time it took to make these measurement, something unexpected happened. The DBT dropped from full on bright to about 1/2 intensity. It remained so for 5 seconds or so then went full on bright again.
As mentioned this occurred twice over the space of about 5 minutes.
Next I measured voltages at Q515.....E: 1.4V C: 2.0V B: 2.0V.
And again, while taking these measurements, the DBT dropped in intneisty then brightened after a few seconds.
R319.....0.0
R317.....0,0
R318.....0.0
R316.....0,0
R418.....7.0 volts
R416.....4.6mv
R419.....0.0
R417.....113.0mv
Twice during the time it took to make these measurement, something unexpected happened. The DBT dropped from full on bright to about 1/2 intensity. It remained so for 5 seconds or so then went full on bright again.
As mentioned this occurred twice over the space of about 5 minutes.
Next I measured voltages at Q515.....E: 1.4V C: 2.0V B: 2.0V.
And again, while taking these measurements, the DBT dropped in intneisty then brightened after a few seconds.
It looks like something is going on with that second channel. Intermittently jumping between bright and dim... first though has to be an intermittent short possibly related to the mounting of the transistor that connects to R417.
Low voltages could fit with the negative rail being loaded heavily by a short in the mounting of a transistor.
Why not remove that output transistor (the R417 one) and see if it behaves normally. It will all work with one device removed.
Q551!
Low voltages could fit with the negative rail being loaded heavily by a short in the mounting of a transistor.
Why not remove that output transistor (the R417 one) and see if it behaves normally. It will all work with one device removed.
Yesterday afternoon was spent on a rail trail with a friend and our dogs. After a 12km hike then an hour drive back home, I was not inclined to do much of anything.
Today I removed the transistor associated with R417 which is Q407. When Q407 was out of circuit, the DBT still behaved abnormally. Getting bright at 50 to 60 volts. I gave Q407 a quick check since it was off the board and it appears to be OK.
When I measured voltages across those emitter resistors, R417 read 113mv.
R418 however read 7.0 volts. R418 is associated with Q410 so I pulled Q410.
With no Q410 in place, the DBT behaved much differently. It did not begin the glow until the variac was near full power. And when at full power, the glow was not intense, just orange.
Q410 did not test OK when I gave it the same quick check as I did to Q407.
Traces on this board are not clearly visible because of the heavy masking that Sony used. But the traces for the output transistor jacks are bold and wide. I can see no break near any of the output transistor jacks.
There are still some fresh, new transistors at hand so Q410 can be replaced.
That DBT dimming / brightening I described, could that be caused by the failed Q410?
Today I removed the transistor associated with R417 which is Q407. When Q407 was out of circuit, the DBT still behaved abnormally. Getting bright at 50 to 60 volts. I gave Q407 a quick check since it was off the board and it appears to be OK.
When I measured voltages across those emitter resistors, R417 read 113mv.
R418 however read 7.0 volts. R418 is associated with Q410 so I pulled Q410.
With no Q410 in place, the DBT behaved much differently. It did not begin the glow until the variac was near full power. And when at full power, the glow was not intense, just orange.
Q410 did not test OK when I gave it the same quick check as I did to Q407.
Traces on this board are not clearly visible because of the heavy masking that Sony used. But the traces for the output transistor jacks are bold and wide. I can see no break near any of the output transistor jacks.
There are still some fresh, new transistors at hand so Q410 can be replaced.
That DBT dimming / brightening I described, could that be caused by the failed Q410?
Something very amiss there. Just think about that, 7 volts across a 0.47 ohm. The current would be 7/0.47 or 15 amps and power dissipation over 100 watts
Lets work with that. Leave Q410 out of circuit. Now disable the bias generator by linking it out (short C to E on Q404) and see if the amp can be powered up to full mains. If it can then check the DC offset at the output and confirm it is close to zero.
I did as suggested. Q410 is out of circuit and I bridged C and E on Q404.
When variac reached full power, DBT glowed a fairly faint dull orange.
Since the DBT was faint, and after a minute or so had passed, I checked for DC offset and measured 9mv. A few seconds after disengaging DMM probes from speaker terminals, DBT went full bright. It statrtled me so I turned off power.
So same story as before.
The only two things things I can think of at this point is that there may be an intermittent break (as @poundy suggested) and that would have to be here. The way to prove that would be to add the link between the base of Q405 and Q411.
The other thought would be whether the output transistors are genuine or whether they are breaking down under full voltage.
The only two things things I can think of at this point is that there may be an intermittent break (as @poundy suggested) and that would have to be here. The way to prove that would be to add the link between the base of Q405 and Q411.
The other thought would be whether the output transistors are genuine or whether they are breaking down under full voltage.
To me it seems that the main problem is that this can be caused by more than one reason and during these "two seconds" you are not able to do any measurements to define the defect area.
One way to get real time picture of what is going on is to hook up cheap DMMs across all resistors in the output area (emitter resistors of the outputs and drivers, base resistors of the drivers) and voltage rails and take photos of their readings in working state and fault state. That would show you what path is conducting and what not.
I have posted pohotos of my similar set in more than one thread here - one is at:
https://www.diyaudio.com/community/...n-circuit-biasing-problem.410497/post-7668515
You could also use hooked measuring cables instead of "crocodiles" - these are much more stable and take less space on the board.
Last edited:
A nice amplifier, but (like some other models) it has a serious design flaw in the Vbe multiplier (for BIAS adjustment). The flaw is that the variable resistor R is not located between the base and emitter of Q404, but is connected in such a way that only the center tap of RV401 is connected to the base of Q404. As long as there is no contact fault on this pot, everything works as intended. In the event of a contact fault, the maximum possible current that the power supply is capable of delivering flows through the output transistors. This leads to the immediate destruction of at least the output power transistors, which is obviously what happened with your model. Probably also other parts are damaged due this deficiency as describe above.
With this level of destruction after such an event, the first step is to modify the power amplifier in such a way so that it works without the output buffer (which then corresponds to a discretely constructed OP-amp or a line stage in a pre-amplifier). I have described how this is done in detail somewhere here - in any case, R420 from the NFB network must be re-soldered from the speaker output to the collector or emitter of Q404 (this is now the output) in order to check the control range of the bias voltage between C and E of Q404.
Before that, of course, R401 must be moved and connect between B and E of Q404. (R409 directly connected to base of Q404).
Only when all of this has been done, the DC conditions of input stage (LTP) and VAS (voltage amplifier stage) are correct and an undistorted audio signal is present at the new output, can the output buffer be connected with new transistors (preferably only after a longer time of observation phase in this state and setting the bias voltage between C and E of Q404 to 1VDC).
If I connect a 4-8 ohms loudspeaker under this condition - i.e. without a darlington output buffer - I only get an extremely small undistorted SPL due to the current control character, but for me this was loud enough to recognize that the front end (i.e. LTP+VAS) is working in the right manner.
concerning Vbe multiplier for BIAS adjust check out
https://www.diyaudio.com/community/threads/optimizing-the-vbe-multiplier.216385/page-3
and
https://www.diyaudio.com/community/...us-vbe-multipliers.117047/page-4#post-7000659
so as the mentioned URLs there.
Last edited:
That may be so but can not be the case here as we are testing with the multiplier shorted out. The shorted multiplier does not rule out broken print between Q404 and the drivers however and the next step has to be to place the shorting link between the driver bases.
Yes, it seems to be a sudden self destruct situation. The bulb should limit current to non destructive levels as far as the output transistor currents are concerned although they may still have to cope with dumping the charge in the reservoir caps... I just think something else is going on here whether that be a physical short rather than a real failure or the theory of secondary breakdown due to non genuine devices. I would whether running (with the variac) at say 70% of total supply would prove anything. If it was OK at that level and yet fails at higher voltage then we might have handle on what is happening.
I jump wired the bases of Q405 and Q411. DBT gets bright when variac is at 50/60 volts as it has done before.
I obtain transistors from Digikey.ca so I assume I am getting genuine parts.
I still say you have a bit of broken print. I know it's a laborious exercise, but I think it is worth a realy though check. Look at the print and gently move each transistor to see if there is any movement when you wiggle them. You have proberbly taken out and put back quite a few components by now, and that board can only take so much. It doesn't take much for the print to crack just in front of where the legs are soldered. Check none of the transistors are inserted incorrectly. It's easily done when you go round in circles👍
Can you measure where the current is going? Do the same as before and measure volt drop across the emitter resistors. I would expect it to be zero on all four with the driver base's shorted. I think we can rule out fake parts if they are Digikey supplied.
One outlier in all this would be an intermittent short on the output line. Its very unlikely but still worth looking over. The output could be isolated here (not anywhere earlier because you must retain the Zobel network for testing). You would then have to measure offset and so on from before the speaker sockets.
Variac was up full, 100W DBT glowing brightly when I measured those emitter resistor voltages again. I recorded tham as displayed in the DMM.
R418 0.0mv
R416 0.745V
R419 0.647V
R417 119mv
Something I realized just now that I didn't fully notice before.
The power amp board can be disconnected and removed with no unsoldering. All or the wires going to the board are connected with push on terminals. There are 9 or them. I would photgraph and make a diagram to show which wire goes where.
That would allow me to examing the trace side more closely and in better light to look for breaks. Or more likely I think, missing or damaged solder pads. Solder pads seem to be fairly delicate in this amp.
R418 0.0mv
R416 0.745V
R419 0.647V
R417 119mv
Something I realized just now that I didn't fully notice before.
The power amp board can be disconnected and removed with no unsoldering. All or the wires going to the board are connected with push on terminals. There are 9 or them. I would photgraph and make a diagram to show which wire goes where.
That would allow me to examing the trace side more closely and in better light to look for breaks. Or more likely I think, missing or damaged solder pads. Solder pads seem to be fairly delicate in this amp.
So those voltage readings show lots of current flow in the output stage.
When you have replaced suspect/failed output transistors how are they failing (in measurement). In other words what are you measuring on them to say they are failed? 99% or the time a failed output transistor (and this applies to most power devices) fail by going short circuit between collector and emitter.
Is that what you are measuring or do you see something else?
When you have replaced suspect/failed output transistors how are they failing (in measurement). In other words what are you measuring on them to say they are failed? 99% or the time a failed output transistor (and this applies to most power devices) fail by going short circuit between collector and emitter.
Is that what you are measuring or do you see something else?
A successfully troubleshooting procedure without the (in my opinion absolutely necessary) steps of independently examining the individual units for proper operation, it seems hardly possible to me - mainly because it is very likely that there is a cluster of independent errors, which also appear to only occur temporarily, such as contact errors and cold solder joints as well as interrupts in the PCB tracks.
Therefore (after ensuring that the power supply works in the right manner) what I mentioned in my previous post (#113) should be done.