I am trying to learn how to repair trickier things... previously I've just stuck to soldering physical connections.
However, a music school asked if I could fix their little powered mixer, a Yamaha EMX66m
On of the input channels distorts quite badly. I have narrowed down the issue to what I think is a pair of transistors that work as a buffer right before the first stage of opamp amplification.
I'm not totally sure how to diagnose problems like this (and while I have a vague idea about how transistors work, I am not 100% sure I understand), but I am doing my best. I figure this kind of goofing around is one way to learn. Anyhow.
The other channels work okay, though and there isn't anything else wrong with the amp, so I compared them. It seems like the signal is rectified after Q101, but I am not what would cause that. In the other channels, I get the full signal at that point. I have a little signal generator with a 1k sin wave to provide a source signal, and that is going into the line in on the channel. I haven't verified that the XLR side of this channel does the same thing, but that didn't seem likely to me.
Anyhow, this is the part of the circuit where that is happening:
If I look at TP5 with an oscilloscope, this is what I see:
On the other channels, I think that is not rectified.
And this is what is on pin 7 of the opamp IC101:
Anyhow, any pointers or things I should look at would be greatly appreciated, as I am not 100% sure where else I can look to see why this is working as it is.
Thanks!
However, a music school asked if I could fix their little powered mixer, a Yamaha EMX66m
On of the input channels distorts quite badly. I have narrowed down the issue to what I think is a pair of transistors that work as a buffer right before the first stage of opamp amplification.
I'm not totally sure how to diagnose problems like this (and while I have a vague idea about how transistors work, I am not 100% sure I understand), but I am doing my best. I figure this kind of goofing around is one way to learn. Anyhow.
The other channels work okay, though and there isn't anything else wrong with the amp, so I compared them. It seems like the signal is rectified after Q101, but I am not what would cause that. In the other channels, I get the full signal at that point. I have a little signal generator with a 1k sin wave to provide a source signal, and that is going into the line in on the channel. I haven't verified that the XLR side of this channel does the same thing, but that didn't seem likely to me.
Anyhow, this is the part of the circuit where that is happening:
If I look at TP5 with an oscilloscope, this is what I see:
On the other channels, I think that is not rectified.
And this is what is on pin 7 of the opamp IC101:
Anyhow, any pointers or things I should look at would be greatly appreciated, as I am not 100% sure where else I can look to see why this is working as it is.
Thanks!
Transistors in the common emitter configuration boost current of the signal. Q101 Q102. Usually transistors work or they don't. Other possible transistors problems are pops (loose weld) or hiss (dirt on the die). You don't have those.
I'd say the most likely problem is one of the time fuses, C120 C121. Electrolytic caps of limited life are put in to destroy the product at a defined time so the customer will have to buy another one. See if TP1 & TP2 have the rectified waveforms.
If not, there is possibly a bad solder joint between the input jack & TP5. Bad solder joints can act as a diode (pass signal only one way). Push on the various solder joints with a chopstick, to see if the problem goes away when you do it. Or if bangs occur, which means the joint you just pushed is the bad one.
It is remotely possible that R116 R117 R112 R113 is open or not soldered, or C108 is bad or not soldered.
When I buy electrolytic caps, I buy ones with service life >3000 hours, so I don't have to do the job again. In 10 uf cap you can usually buy 10000 hour service life ones. Digikey & Newark will show service life in the cap selector table. Mouser makes you download & read each datasheet. Don't buy e-caps from non-authorized distributors like ebay, alibaba, or local stores (RadioShack was horrible, sold garbage grade caps). Be sure to mark the board for + before removing the old caps. E-caps put in backwards can blow the vent open.
When testing products with the power on, use only one hand, with the other meter probe or scope ground connected by alligator clip to the analog ground. >25 v across your heart can stop it. Wear no jewelry on hands or wrists. 1 v at high current can burn your flesh to charcoal. Wear safety glasses, parts can explode. Desoldering, the liquid can splash into your eye.
I'd say the most likely problem is one of the time fuses, C120 C121. Electrolytic caps of limited life are put in to destroy the product at a defined time so the customer will have to buy another one. See if TP1 & TP2 have the rectified waveforms.
If not, there is possibly a bad solder joint between the input jack & TP5. Bad solder joints can act as a diode (pass signal only one way). Push on the various solder joints with a chopstick, to see if the problem goes away when you do it. Or if bangs occur, which means the joint you just pushed is the bad one.
It is remotely possible that R116 R117 R112 R113 is open or not soldered, or C108 is bad or not soldered.
When I buy electrolytic caps, I buy ones with service life >3000 hours, so I don't have to do the job again. In 10 uf cap you can usually buy 10000 hour service life ones. Digikey & Newark will show service life in the cap selector table. Mouser makes you download & read each datasheet. Don't buy e-caps from non-authorized distributors like ebay, alibaba, or local stores (RadioShack was horrible, sold garbage grade caps). Be sure to mark the board for + before removing the old caps. E-caps put in backwards can blow the vent open.
When testing products with the power on, use only one hand, with the other meter probe or scope ground connected by alligator clip to the analog ground. >25 v across your heart can stop it. Wear no jewelry on hands or wrists. 1 v at high current can burn your flesh to charcoal. Wear safety glasses, parts can explode. Desoldering, the liquid can splash into your eye.
Last edited:
Another way you can check for connection problems would be to check the DC voltages of similar points on the good channel with the bad channel. Particularly the collector base & emitters of the Q101 Q102. Base & emitters should be 0.7 v apart, collectors should be a couple of volts higher than base. pins 6 & 5 of the op amp should be approximately the same dc voltage.
The circuit is a balanced input to single ended output amplifier with gain setting controlled by sw101
Since It’s a balanced input, what are you using for a source?
It also has a phantom supply +48VDC coming in as drawn (the red wire) that is required for a condenser microphone, not sure if the mixer has a switch to turn it off or not.
2sc3324 are toshiba sc-59 smt and of course they use yamahas favourite opamp the njm2068 I assume without looking it’s the “md” so-8 version as well, good luck
Since It’s a balanced input, what are you using for a source?
It also has a phantom supply +48VDC coming in as drawn (the red wire) that is required for a condenser microphone, not sure if the mixer has a switch to turn it off or not.
2sc3324 are toshiba sc-59 smt and of course they use yamahas favourite opamp the njm2068 I assume without looking it’s the “md” so-8 version as well, good luck
Last edited:
Thanks for the pointers.
I did change out the electrolytic caps in the signal path (since that was easy), but that did not fix it.
"Since It’s a balanced input, what are you using for a source?" I have a signal generator hooked to the 1/4" TS input. I did verify that it looks correct on the other channels.
The transistors seems to test okay- using the diode mode on my meter, they seem similar to the other transistors. However, I can't understand what else could be broken here, so I suspect it is Q101. That just feels like a guess, though.
I think I have found a similarly-spec'd transistor so I ordered one and will try replace q101 and q102.
Thanks for your time!
I did change out the electrolytic caps in the signal path (since that was easy), but that did not fix it.
"Since It’s a balanced input, what are you using for a source?" I have a signal generator hooked to the 1/4" TS input. I did verify that it looks correct on the other channels.
The transistors seems to test okay- using the diode mode on my meter, they seem similar to the other transistors. However, I can't understand what else could be broken here, so I suspect it is Q101. That just feels like a guess, though.
I think I have found a similarly-spec'd transistor so I ordered one and will try replace q101 and q102.
Thanks for your time!
So, what is the voltage at idle on base of Q101? Emitter of Q101? Collector of Q101? Is Vb-Ve>0.6?
Is Vc-Ve>1?
Diode test of DVM are at 1 or 2 volts. Transistor probably operates between +15&-15 v. Lots of difference between operating condition and the situation the diode scale of the meter tests at.
What are DC voltages at 5 & 6 of IC101? With no input.
Amateurs make enough soldering errors it is prudent to find out what is wrong before waving the iron around. A baseline of voltages will give you something to compare to if you make the situation worse (which I have done).
Is Vc-Ve>1?
Diode test of DVM are at 1 or 2 volts. Transistor probably operates between +15&-15 v. Lots of difference between operating condition and the situation the diode scale of the meter tests at.
What are DC voltages at 5 & 6 of IC101? With no input.
Amateurs make enough soldering errors it is prudent to find out what is wrong before waving the iron around. A baseline of voltages will give you something to compare to if you make the situation worse (which I have done).
I am also completely capable of making the situation far worse and would rather diagnose things before removing components as well 
So I took some measurements on Q101, Q102 and their counterparts on other channels and found several differences.
On Q101 and Q102 Vb - Gnd = -320mV and the other transistors' Vb are -175mV
On Q101 and Q102 Vc - Gnd = 7.2V and the other transistors' Vc are 9.3V
On Q101 and Q102 Ve - Gnd = -0.9V and the other transistors' Vc are -0.75V
On Q101 and Q102 Vb-Ve (as well as 201 and 202 on the other channels) are all ~ -0.6V
On Q101 and q102 Vc-Ve ~ -8.1V and on Q201, Q202 (and the other channels) is -10.1V
Also I looked at the voltage between IC101 pins 5 and 6 and ground, both are 7.16V. However on the other channels (including pins 1 and 2 on IC101 for channel 2)the pins are are all at 9.3V.
The power for IC101 seems okay? +15 and -14.6V on all the ICs including IC101 and the other channels seem to work okay.
So the voltages around q101 and 102 are the same as a pair, but they are different than the other channels (which are consistent with each other).
One frustration I have is that I want to measure the R114 - R117 cause that seems relevent, but on this board a lot of the SMD resistors don't seem to want to be measured? Am I imagining that?
Anyhow, thanks for the responses, I am still not sure what the issue is but maybe yall have more observations?
So I took some measurements on Q101, Q102 and their counterparts on other channels and found several differences.
On Q101 and Q102 Vb - Gnd = -320mV and the other transistors' Vb are -175mV
On Q101 and Q102 Vc - Gnd = 7.2V and the other transistors' Vc are 9.3V
On Q101 and Q102 Ve - Gnd = -0.9V and the other transistors' Vc are -0.75V
On Q101 and Q102 Vb-Ve (as well as 201 and 202 on the other channels) are all ~ -0.6V
On Q101 and q102 Vc-Ve ~ -8.1V and on Q201, Q202 (and the other channels) is -10.1V
Also I looked at the voltage between IC101 pins 5 and 6 and ground, both are 7.16V. However on the other channels (including pins 1 and 2 on IC101 for channel 2)the pins are are all at 9.3V.
The power for IC101 seems okay? +15 and -14.6V on all the ICs including IC101 and the other channels seem to work okay.
So the voltages around q101 and 102 are the same as a pair, but they are different than the other channels (which are consistent with each other).
One frustration I have is that I want to measure the R114 - R117 cause that seems relevent, but on this board a lot of the SMD resistors don't seem to want to be measured? Am I imagining that?
Anyhow, thanks for the responses, I am still not sure what the issue is but maybe yall have more observations?
yes, but the voltages on Q101 Q102 are all low: Which means the pulldown side resistor is working and the pull up side is not. Or IC101 is doing something inappropriate.
Inability to measure resistors with DVM may mean surface mount resistors don't have conductive pads on top to be measured. Which means measurement should be taken on connecting parts with visible pads. Resistance measurement should be low due to semiconductors in circuit but not high. High indicates bad solder joint, wrong value part, or unusually value drift of a carbon film part.
BTW I find tiny solder joints are easier done with solder paste instead of flux core wire solder.
I was responding to above which to me is incorrect, either measure a working channel or do a sim to verify the theory
It will be interesting to find out what’s causing the bias issue, it could be bad component or connection issue.
Carbon film smt resistor, do they even exist?
Last edited:
Well, I replaced all the electrolytics as well as q 101 and q102.
I also noted that the opamp was sitting on its pads in a really funky manner. So I touched that up with a soldeing iron, as well as the various SMD parts on the board in the area.
I ended up sewing it up, as I think fixing it might be beyond me- after a lot of research I am still not sure why the signal looks like it does at tp5.
Anyhow, thanks for the responses- I wish I had a better follow up report.
Maybe I will return to it later, but for now I think that I'll just send it back to the user and tell them not to use that channel.
I also noted that the opamp was sitting on its pads in a really funky manner. So I touched that up with a soldeing iron, as well as the various SMD parts on the board in the area.
I ended up sewing it up, as I think fixing it might be beyond me- after a lot of research I am still not sure why the signal looks like it does at tp5.
Anyhow, thanks for the responses- I wish I had a better follow up report.
Maybe I will return to it later, but for now I think that I'll just send it back to the user and tell them not to use that channel.
It’s possible that the opamp is the faulty component, esp knowing that it is pulling the inputs lower than you measure on the good channel. The part is cheap enough to try out a new one, but it could be difficult to remove if you are not experienced or uncomfortable doing it. There are techniques to use to remove a dmp8 packaged part. Post a pic of the assembly for me to see what you are up against.
Last edited:
Well, I think that I could replace the opamp. Now I am curious about it just to try it LOL.
Replacing the transistors was a whole lot easier and faster than I had thought- I reently got some better liquid flux and solder braid and have been surprised at how well that works. If that fails I have a buddy with a couple other options including a hot air station...
Replacing the transistors was a whole lot easier and faster than I had thought- I reently got some better liquid flux and solder braid and have been surprised at how well that works. If that fails I have a buddy with a couple other options including a hot air station...
From an idea in elektor mag, take a piece of bare copper solid wire and form cut it so that the wire makes contact with the pins, on both sides. It forms a triangle, pointed up where you place your hot iron tip to heat wire to reflow joints. As the wire sits against the leads, flood with solder so that all the leads get attached to the wire, do on both sides. Now crank the iron temp, place on the top triangle to heat up, tin, and reflow happens once hot enough to remove the part. use some wick and clean excess solder off pads to place a new part and solder. Touch of solder on corner pad. Place part, align, Tack solder one corner pin to align press flat then solder the rest.
Good luck
Good luck
Last edited: