hi there, i have built two identical amplifier boards based on MJL21193/94 ouput drivers, one works, the other is being impossible!
as it currently stands, i have checked everything i think i possibly can!!
the transistors all seem OK (i even replaced all 4 output drivers) and all parts have correct polarity.
the resistors are all in the right spots (checked with ohm-meter) and all the diodes are right way around. no soldering errors at all.
when i apply power, with VR1 set to full anti-clockwsie, there is no voltage at all across the 'safety resistors' and VR1 does not change this when adjusted.
voltage at the output = ~0.45V (as is documented it needs to be)
now the reason i am posting is becasue i have FINALLY found something different between the two boards:
the working board has a 300 ohm resistance across the input
the non-working board has over 20 mega ohms!!!
i guess this is not good?
can anyone help me troubleshoot this? i assume there is meant to be a low resistance across input like on the working board?
attached is a circut diagram and schematic if it helps.
i note that the diagram says there should be voltage of 0.62V across the 47ohm resistor next to Q6 and a voltage of 0.9V across the 68ohm resistor next to Q7
i measure 0.71V next to Q6 and only 0.1V next to Q7 (my mains voltage is a bit higher than 240V BTW)
thanks for any direction you might have to offer 🙂
as it currently stands, i have checked everything i think i possibly can!!
the transistors all seem OK (i even replaced all 4 output drivers) and all parts have correct polarity.
the resistors are all in the right spots (checked with ohm-meter) and all the diodes are right way around. no soldering errors at all.
when i apply power, with VR1 set to full anti-clockwsie, there is no voltage at all across the 'safety resistors' and VR1 does not change this when adjusted.
voltage at the output = ~0.45V (as is documented it needs to be)
now the reason i am posting is becasue i have FINALLY found something different between the two boards:
the working board has a 300 ohm resistance across the input
the non-working board has over 20 mega ohms!!!
i guess this is not good?
can anyone help me troubleshoot this? i assume there is meant to be a low resistance across input like on the working board?
attached is a circut diagram and schematic if it helps.
i note that the diagram says there should be voltage of 0.62V across the 47ohm resistor next to Q6 and a voltage of 0.9V across the 68ohm resistor next to Q7
i measure 0.71V next to Q6 and only 0.1V next to Q7 (my mains voltage is a bit higher than 240V BTW)
thanks for any direction you might have to offer 🙂
Attachments
When you measure impedance/resistance at the input where exactly are you measuring? If it is right at the start of the amp then because the input is AC coupled through the cap you will never get a stable reading here, so you can ignore both your readings mentioned above.
Sorry I can't help more, but if you have no voltage across your safety resistors that means a lot of current is trying to be consumed by the amp, so the output must be short. Try removing the very last output pair base connection from the board and see if they pass any current by measuring the voltage across their emitter resistors.
Which SC design is that? It looks nice.
Sorry I can't help more, but if you have no voltage across your safety resistors that means a lot of current is trying to be consumed by the amp, so the output must be short. Try removing the very last output pair base connection from the board and see if they pass any current by measuring the voltage across their emitter resistors.
Which SC design is that? It looks nice.
richie00boy said:
yeah, at the start next to the 2.2uF electro, i just thought they should measure the same (or similar),
oh well... i thought i was onto something 🙁
but there is no heat anywhere?
OK, will try tomorrow, it is getting late here..
it is the 175W into 4ohm design from April 1996
thanks for your advice...!
If there really is no heat ANYWHERE and really not even millivolts across the safety resistors then you cannot be getting any power onto your board.
When you can't set the bias, in my limited experienc, it's because for some reason the VAS transistors are not turning on. The reason is probably not ibn the VAS itself but in the input section.
If I had this problem I would recheck the value of each resistor, then do a systematic point to point continuity check touching, whenever possible, the component leads rather than the pcb traces or solder joints. It might be helpful to remove the caps temporarily to measure resistance.
With the caps in place but the amp not connected to a power supply, an AC test signal applied to the inputs should appear at the base input transistor in the LTP but no further along the signal path. It should also appear at various other locations in the input section before it reached the differential pair. See if you get the same AC reading for both good and bad amps.
By this point you will have eliminated some possabilities and may even be closing in. You might also try temporaritly removing the diff pair transistors when you do the above inorder to isolate the input components from the rest of the amp. If youget the same readings for both amps that way you at least know the problem is not located in the pre-LTP input components.
If you are still looking after this, start comparing quiesent voltages around the diff pair, the current mirror and the CCS. Don't for get to look for bad or backwards diodes, PNP vs NPN, etc.
This may have been of no help at all but it is where I would look.
One more technique: I have uncovered some major "gotchas" simply by stepping aaway for a day or two. The answer sometimes just pops out of thin air while at the gym, fishing or knocking back a beer at the local.
If I had this problem I would recheck the value of each resistor, then do a systematic point to point continuity check touching, whenever possible, the component leads rather than the pcb traces or solder joints. It might be helpful to remove the caps temporarily to measure resistance.
With the caps in place but the amp not connected to a power supply, an AC test signal applied to the inputs should appear at the base input transistor in the LTP but no further along the signal path. It should also appear at various other locations in the input section before it reached the differential pair. See if you get the same AC reading for both good and bad amps.
By this point you will have eliminated some possabilities and may even be closing in. You might also try temporaritly removing the diff pair transistors when you do the above inorder to isolate the input components from the rest of the amp. If youget the same readings for both amps that way you at least know the problem is not located in the pre-LTP input components.
If you are still looking after this, start comparing quiesent voltages around the diff pair, the current mirror and the CCS. Don't for get to look for bad or backwards diodes, PNP vs NPN, etc.
This may have been of no help at all but it is where I would look.
One more technique: I have uncovered some major "gotchas" simply by stepping aaway for a day or two. The answer sometimes just pops out of thin air while at the gym, fishing or knocking back a beer at the local.
richie00boy said:
must be a new definition of "cannot" that i am un-aware of?
there is full supply +/- rail voltage at the collectors of each output transistor. (i assume that this is correct?)
and, as i stated, there is voltage at the two points as specified by the circut diagram,
sam9 - thanks heaps for those steps, i think i understand most of them, and will give it a go...
OK, let me rephrase that. If there really is no heat ANYWHERE and really not even millivolts across the safety resistors then either you cannot be getting any power onto your board OR your amplifier cannot be drawing any current.
As you can measure voltage on the board then I suspect the latter. This would point to the input stage LTP constant current source not working. If this doesn't work then the VAS will be open circuit and no drive on the output stage.
As you can measure voltage on the board then I suspect the latter. This would point to the input stage LTP constant current source not working. If this doesn't work then the VAS will be open circuit and no drive on the output stage.
What are the voltages across the two 220 ohms next to Q4 and Q5? - They should be the same.
And as a check-up: What is the voltage across the two 150 ohms next to Q1 and Q2? - They also should be the same.
Compare all these measurements to the other (working) board. That should give some hints.
And maybe a picture of both sides of the fawlty board itself, please. Just to see, if we can spot anything, that you have missed - you know; fresh eyes 😉
And as a check-up: What is the voltage across the two 150 ohms next to Q1 and Q2? - They also should be the same.
Compare all these measurements to the other (working) board. That should give some hints.
And maybe a picture of both sides of the fawlty board itself, please. Just to see, if we can spot anything, that you have missed - you know; fresh eyes 😉
boholm said:
both are 0.47V +/- 0.01V
both are 0.26V +/- 0.01V
very embarased to show the underside, as my desoldering skills on the output transistors leave alot to be desire 😱 but they all connect to their relative points correctly with continuity tests.
there is lots of shiny flux, but i have held the board up to light and the tracks are all OK, and there are no obvious bridges...
BTW - i have just bought a new set of input stage transistors to install tonight...
Attachments
and the top side
thanks so much for a fresh set of eyes!
my fiance keeps telling me that "you will work it out, you always do" but having someone to bounce this off is great!
she also keeps reminding me it is "probably something simple and silly" that i have missed, but i am sure it is all identical to my working board!
ALSO - richie00boy - what is a LPT constant current source? the documentation refers to:
the signal from Q1 connects to the base of Q7, part of the cascode stage comprising Q7 & Q8, with Q6 providing a constant current source to Q8
then says:
the output signal from the cascode stage is coupled directly to the output stage
it also mentions that Q9 acts as a "Vbe multiplier"
thanks so much for a fresh set of eyes!
my fiance keeps telling me that "you will work it out, you always do" but having someone to bounce this off is great!
she also keeps reminding me it is "probably something simple and silly" that i have missed, but i am sure it is all identical to my working board!
ALSO - richie00boy - what is a LPT constant current source? the documentation refers to:
the signal from Q1 connects to the base of Q7, part of the cascode stage comprising Q7 & Q8, with Q6 providing a constant current source to Q8
then says:
the output signal from the cascode stage is coupled directly to the output stage
it also mentions that Q9 acts as a "Vbe multiplier"
Attachments
Please measure the voltages at B, C, E of Q9 with the bias pot first in one then the other extreme.
LTP = long tailed pair and consists of Q1 and Q2. The current source providing drive to this and the VAS (voltage amplification stage) is based around Q3.
Well, not quite. I have had an actual experience with something like this. For reference output topology was standard EF. Full rail voltage was present ar the collectors but zilch at the emiitters and the RE resistors remained dead cold. The output devices were simply not being turned on. Base-emitter voltage was <50mV. The problem was way back in the input section as even the VAS was not being turned on.
The poster's original description sounded similar to this. Since he could not even set the bias, the chances are that the VAS transistors are not turning on either. Since a functioning input section (LTP) is needed to turn on the VAS, this is a good place to start looking. The fact that therte was something odd about resistance measured across the inputs reinforces this.
All this is possible even with rail voltage present. I'm not speculating - I have had a similar experience. Unfortunately, I know my problem and solution had a different source simply because they related to a topological feature not present in this schematic.
Hmmm 😉 the picture quality, lighting and angle is not at its optimum - did you use your cellphone 😕 😉 😀 
Well, it´s a bit hard to really see - mostly because of the angle. The coolers covers some areas, where the error might be.
Your ZD1: Does this have 3.3 volts across it?
After looking a bit more, it looks to me that there could be something fawlty aound Q7 and Q8 to give that result given in your first post. I would solder them out again and measure them thru´with a multimeter set to measure diodes. That will show if they are dead. Also check the soldering with an magnifying glass and good light (sunshine or a very bright lamp).
By the way the cooler on the transistor Q7 isn't touching anything else, is it?
Well, it´s a bit hard to really see - mostly because of the angle. The coolers covers some areas, where the error might be.
Your ZD1: Does this have 3.3 volts across it?
After looking a bit more, it looks to me that there could be something fawlty aound Q7 and Q8 to give that result given in your first post. I would solder them out again and measure them thru´with a multimeter set to measure diodes. That will show if they are dead. Also check the soldering with an magnifying glass and good light (sunshine or a very bright lamp).
By the way the cooler on the transistor Q7 isn't touching anything else, is it?
sam9 said:
Hmm, did you read all my post? 🙂 That's exactly (slightly longer version of) what I said after what you quoted 🙂
Could always ask the guys at SC as they are helpful and might have some info with any history of probelms that builders have had. You won't get a response overnight, but they will reply.
You could always take the drastic approach, the one I normally use, and replace all the semiconductors (a new VR1 could be handy as well), but could be around the price of a new kit 🙁. I've built this amp previously and watch the heatsinks on Q6 and Q8 as the semis can bend over and heatsink can touch other parts.... ended up fixing them to the PCB with some heat glue.
You could always take the drastic approach, the one I normally use, and replace all the semiconductors (a new VR1 could be handy as well), but could be around the price of a new kit 🙁. I've built this amp previously and watch the heatsinks on Q6 and Q8 as the semis can bend over and heatsink can touch other parts.... ended up fixing them to the PCB with some heat glue.
rich00boy,
Oops. Comes from trying reply in hurry while a lot of unreal;ated stuff is gong on in the background here.
Oops. Comes from trying reply in hurry while a lot of unreal;ated stuff is gong on in the background here.
boholm said:
um, it has full supply rail voltage - 59V
why is this?!?!
it still reads as a diode...
removed it.
Q8 does get a little warm though...
thanks for explainations richie00boy and sam9!
rabbitz, i have replaced ALL the semi-conductors (including Q7 and Q8) 🙁
ALSO - when i measure the resistance across the input - it starts at 1M ohm and quickly rises to 50M ohm and slows down rising(but does keep increasing)
does this indicate anything?
tinkerbell said:
It indicates, that there is something terribly wrong around Q8. Either the transistor is short-circuiting or there is a solder bridge (B and C of Q8 is connected by solder . . . is it called solder that stuff you use when soldering?)
tinkerbell said:
Bad idea
You just have to make sure it can't touch another component.tinkerbell said:
Q8 too? And it still fails = 57 volts across ZD1? How about replacing ZD1 and the 5.6 kohms/1Watt? AND once again urging you to look at your solder side with a magnifying glass and check for short circuits.
tinkerbell said:
Yes: Your input circuitry is working perfectly. No need to spend more time there. At first you measure the resistance of your input, and after that the 2.2 µF will start to charge thus raising the resistance. As it charges, less and less current will run and when it is charged, no more current will run thus giving your 50 Mohms and more.
- Status
- Not open for further replies.