Hi.
I have a problem with ampeg svt 3 pro Power Amp.
There is anoying distortion and its coming from the power amp, as its not there if I take the signal from "PreAmp Out" socket.
Its like my speakers are broken and it occures on higher notes (on bass) low strings are sounding good.
I switched to another cabinet, but this problem is still there and its not the speakers.
Can you suggest, what could be the problem with it?
I changed the two tubes there, but no use.
Could it be the FETs?
What could be measured there without taking it all apart?
Here is the schematic for the power amp
ampeg svt 3 proPowerAmp.pdf download - 2shared
I have a problem with ampeg svt 3 pro Power Amp.
There is anoying distortion and its coming from the power amp, as its not there if I take the signal from "PreAmp Out" socket.
Its like my speakers are broken and it occures on higher notes (on bass) low strings are sounding good.
I switched to another cabinet, but this problem is still there and its not the speakers.
Can you suggest, what could be the problem with it?
I changed the two tubes there, but no use.
Could it be the FETs?
What could be measured there without taking it all apart?
Here is the schematic for the power amp
ampeg svt 3 proPowerAmp.pdf download - 2shared
I suspect you have power supply sag if is only the output stage. You need to clip lead a DVM to the power supply rails, and play a loud note, and see if the voltage sags. Warning, the note will have to be rather sustained, DVM's average the signal over about 4 seconds.
You could also have a coupler cap going bad, particularly if there are any electrolytic caps over 20 years old. These are aluminum cans wrapped in cardboard or plastic, with a plus on one lead or a minus in balls pointing at the other. Or, it can look like a peanut M&M, those are tantalum electrolytic caps.
Once my amps start sounding funny, and the caps are over 20 years old, I replace all the e-caps before I do anything else. Oddly, Salas and Enzo said something about that process happening at the incoming inspection desk of their amp shops. I date the new ones I install with a sharpie, the Nipponese don't date their caps with more than a one digit year and a week. I guess no consumer electronic device is supposed to last more than 10 years. Oh, BTW, Sony and Panasonic bonds went to Junk status this week. Funny thing, if you're going to buy trash may as well buy directly from the source of trash 100 miles to the west. You bought an Ampeg, worth buying and worth fixing IMHO.
Okay, looking at the schematic, C3 and and C5 are old electrolytic caps. Not a lot of electrolytic coupler caps, but the plate couplers c6 and c7 are in the position that a lot of caps fail, even plastic ones. This amp also has 1/4 phone plug to the speakers, which leads to a lot of blown output transistors if a phono plug gets pulled a little bit out. No protection circuit. If the Power Supply rails aren't collapsing, you might have to remove these FETs and test them one at a time by shorting gate to source and seeing if they will hold off 18v through 47kohms or something. (I use a car battery charger with a cap on it to produce the 18v, with a 47k resistor on a wire coming off to the amps scale of the DVM and a clip lead going back. Any amps, the transistor is leaking). Lots of shops bid all new output transistors on any guitar amp job.
You could also have a coupler cap going bad, particularly if there are any electrolytic caps over 20 years old. These are aluminum cans wrapped in cardboard or plastic, with a plus on one lead or a minus in balls pointing at the other. Or, it can look like a peanut M&M, those are tantalum electrolytic caps.
Once my amps start sounding funny, and the caps are over 20 years old, I replace all the e-caps before I do anything else. Oddly, Salas and Enzo said something about that process happening at the incoming inspection desk of their amp shops. I date the new ones I install with a sharpie, the Nipponese don't date their caps with more than a one digit year and a week. I guess no consumer electronic device is supposed to last more than 10 years. Oh, BTW, Sony and Panasonic bonds went to Junk status this week. Funny thing, if you're going to buy trash may as well buy directly from the source of trash 100 miles to the west. You bought an Ampeg, worth buying and worth fixing IMHO.
Okay, looking at the schematic, C3 and and C5 are old electrolytic caps. Not a lot of electrolytic coupler caps, but the plate couplers c6 and c7 are in the position that a lot of caps fail, even plastic ones. This amp also has 1/4 phone plug to the speakers, which leads to a lot of blown output transistors if a phono plug gets pulled a little bit out. No protection circuit. If the Power Supply rails aren't collapsing, you might have to remove these FETs and test them one at a time by shorting gate to source and seeing if they will hold off 18v through 47kohms or something. (I use a car battery charger with a cap on it to produce the 18v, with a 47k resistor on a wire coming off to the amps scale of the DVM and a clip lead going back. Any amps, the transistor is leaking). Lots of shops bid all new output transistors on any guitar amp job.
Last edited:
The parallel value of a capacitor and a short is a short.Bulging top electrolytic caps are caps that have failed already, and are ready to leak slime water. You need to buy a DVM (not autoranging) and measure the voltage on them. Hot performance is more important than any measurement you get cold.
I take back what I said about no short protection, you have a speaker disconnect relay and something going on on page 2 and 3.
The print I've got doesn't have the power supply drawings and doesn't show C18. The download only has page 1. But, that is possibly where your problem is.
I take back what I said about no short protection, you have a speaker disconnect relay and something going on on page 2 and 3.
The print I've got doesn't have the power supply drawings and doesn't show C18. The download only has page 1. But, that is possibly where your problem is.
Last edited:
Hi.
So I measured some voltages.
Max drop on thouse caps was around 1V.
Between 0 (GND) and +Vcc
+69,0-68,0
Between GND and -Vcc
-69,4-68,7
I fed the input with a signal generator and this I got with the full sweep.
My ears are killing me now.
Full schematic for power amp
http://www.2shared.com/document/qDGM11XC/ampeg-svt3-pro-pwramp-42601h15.html
So I measured some voltages.
Max drop on thouse caps was around 1V.
Between 0 (GND) and +Vcc
+69,0-68,0
Between GND and -Vcc
-69,4-68,7
I fed the input with a signal generator and this I got with the full sweep.
My ears are killing me now.
Full schematic for power amp
http://www.2shared.com/document/qDGM11XC/ampeg-svt3-pro-pwramp-42601h15.html
Do you have DC coming out of that amp speaker terminals? That is a bad sign you have output transistor problems.
I use 5 ohm 200 watt resistors for a load when I am power checking my amp. I got them from a surplus house for $5 apiece. I have 4. I put two in series on each channel (on a stereo PA amp) and put a 4 ohm car radio speaker from the charity resale shop in series with two back to back >1000 uf caps across 5 of the 10 ohm ( on high powered amps). The speakers play music right through the DC problems if I have DC problems, or most recently, made a big click when the DC jumped out of control. You can get new 10 ohm resistors from farnell for about $20 each. The tapped ones can be set for 8 ohms, I have two of those also I bought 25 years ago.
If your power supply is not sagging at full power, then the quality check of the output transistors might be in order. Particularly if you have DC out the speaker terminals. If you don't have a 47k resistor and a wall transformer that will put out 12 VDC or 18 VDC (I have a race car set transformer at 18 VDC) to do the leakage tests listed above, then a first test is to short g to s with a clip lead and use a DVM ds. However, this is at 2 v and is not much of a test. Watch static on the gate, it can blow them up. I suggest a humidifier in the room, or a hot bath.
If you have a meter with a 2 vac scale, you can poke around looking for AC signals, but most DVM's have a 200 VAC scale which is useless.
I use 5 ohm 200 watt resistors for a load when I am power checking my amp. I got them from a surplus house for $5 apiece. I have 4. I put two in series on each channel (on a stereo PA amp) and put a 4 ohm car radio speaker from the charity resale shop in series with two back to back >1000 uf caps across 5 of the 10 ohm ( on high powered amps). The speakers play music right through the DC problems if I have DC problems, or most recently, made a big click when the DC jumped out of control. You can get new 10 ohm resistors from farnell for about $20 each. The tapped ones can be set for 8 ohms, I have two of those also I bought 25 years ago.
If your power supply is not sagging at full power, then the quality check of the output transistors might be in order. Particularly if you have DC out the speaker terminals. If you don't have a 47k resistor and a wall transformer that will put out 12 VDC or 18 VDC (I have a race car set transformer at 18 VDC) to do the leakage tests listed above, then a first test is to short g to s with a clip lead and use a DVM ds. However, this is at 2 v and is not much of a test. Watch static on the gate, it can blow them up. I suggest a humidifier in the room, or a hot bath.
If you have a meter with a 2 vac scale, you can poke around looking for AC signals, but most DVM's have a 200 VAC scale which is useless.
Last edited:
I have found that sometimes when these amps get pushed real hard the gates of the mosfet's get stressed and this creates a sort of "fuzzy" torn speaker sound. if you look at the output with a scope often times the amp is oscillating and or ringing!
Also, be very careful with Ampeg mosfet bias pots. one way too far causes high bias, the other way too far causes oscillation!
Zc
Also, be very careful with Ampeg mosfet bias pots. one way too far causes high bias, the other way too far causes oscillation!
Zc
Hi.
I messed around with the Bias trimpot, and if I max this out, the sound will clear out.
And I can repeat this again and again. Signal generator Frequenzy is 586Hz sine.
With guitar attached I also didn't hear any problems with it.
Is it possible that the bias has changed by time?
I only had to turn the trimpot a very tiny bit to fix it.
I am worried that this is not the real problem and it will come up again in the near future.
But I have to say that thouse Electrolytes have to be changed also, they are going any time.
I messed around with the Bias trimpot, and if I max this out, the sound will clear out.
And I can repeat this again and again. Signal generator Frequenzy is 586Hz sine.
With guitar attached I also didn't hear any problems with it.
Is it possible that the bias has changed by time?
I only had to turn the trimpot a very tiny bit to fix it.
I am worried that this is not the real problem and it will come up again in the near future.
But I have to say that thouse Electrolytes have to be changed also, they are going any time.
Last edited:
Potentiometers have been known to lose contact with the wiper after a long time not adjusted. Perhaps that is what happened to yours. Zero Cool sounds like he knows what he is talking about. One way on the pot- too much ringing. Other way on the pot- not enough crossover distortion eliminating bias current. Be sure to look on the schematic of how to measure the bias current and make sure it is right. usually voltage across an emitter resistor, but this has source resistors. A little contact cleaner on a pot can't hurt. Warning, contact cleaner is flammable, no smoking, open flame, electricity turned on or off within 10 m while using.
Hi.
So I solved that bias issue.
But the problem doesn't end here.
Sound cleaned up, but now the amp burned down.
It was probably because it was too cold for use. We have a winter time here and its really cold outside and after the amp was transported it was directly powered on. ( can't stop blameing myself for letting this to happen) But I can't do anything about it now.
So what I found was that R27 and Q6 was burned.
What is the function of them and what could be the result or cause?
Have a look at the schematic
So I solved that bias issue.
But the problem doesn't end here.
Sound cleaned up, but now the amp burned down.
It was probably because it was too cold for use. We have a winter time here and its really cold outside and after the amp was transported it was directly powered on. ( can't stop blameing myself for letting this to happen) But I can't do anything about it now.
So what I found was that R27 and Q6 was burned.
What is the function of them and what could be the result or cause?
Have a look at the schematic
An externally hosted image should be here but it was not working when we last tested it.
Cold weather could make old dried up electrolytic caps act up, but you said you already tested those on the rail voltages. Cold could also make a big transistor contract on the heat sink and touch the loose mount screw or find the bad part of the heat sink insulator. Cold could cause any bad solder joint to open up.
Looks to me like Q6 and R27 are part of the DC bias current network. AP1 and Q2 set the bias voltage, then Q5 and Q6 amplify it in current I think. Or maybe limit how big bias current gets. Are you sure you don't have DC voltage coming out on the speaker terminals? if an output transistor blows, It may put +-65 VDC on these low voltage rated interior parts through the blown gate and the little parts blow up. Doen't mean the big transistors aren't hosed. You check a FET in high humidity (or with a grounded wrist strap) by alligator clip lead connecting gate to source, then seeing if the disabled fet will hold of a significant voltage of the right polarity for it's position. Like Q7-10 want +voltage on the drain. For OT test I use 17 VDC out of a capacitor on a car battery charger, then through a 47k resistor so I don't blow up anything, then through the 200 ma scale of my DVM. You want the minus of the battery charger on the FET source on Q7-10. Opposite on the opposite polarity ones. . Not much current should flow on a good FET. The 2 VDC out of the ohms test of the meter was not enough to predict OT's that were going to explode when I put rail voltage on them.
If any OT is bad, all the parts on this board should be individually tested. At least that is what I did on my amp.Diodes were blown or shorted across, transistors were blown or shorted across, resistors were vaporized or opened up, a couple of PCB traces were vaporized. Capacitor checking was a bit of a trial. I had 50V rated ceramic capacitors on the driver board that were opened up by the 85 v rail voltage. If it had vaporized them they would have been easier to spot. My PV1.3k amp had a bad solder joint on the input to an op amp that would cause the amp to occasionally whang the B channel output to +170 VDC. The previous repair man handled it by pulling all the power supply connections to the B channel, leaving the output transistors bad, and putting a label on it "do not use B channel". My repair wasn't economic, took me a long time, but it is permanent.
Looks to me like Q6 and R27 are part of the DC bias current network. AP1 and Q2 set the bias voltage, then Q5 and Q6 amplify it in current I think. Or maybe limit how big bias current gets. Are you sure you don't have DC voltage coming out on the speaker terminals? if an output transistor blows, It may put +-65 VDC on these low voltage rated interior parts through the blown gate and the little parts blow up. Doen't mean the big transistors aren't hosed. You check a FET in high humidity (or with a grounded wrist strap) by alligator clip lead connecting gate to source, then seeing if the disabled fet will hold of a significant voltage of the right polarity for it's position. Like Q7-10 want +voltage on the drain. For OT test I use 17 VDC out of a capacitor on a car battery charger, then through a 47k resistor so I don't blow up anything, then through the 200 ma scale of my DVM. You want the minus of the battery charger on the FET source on Q7-10. Opposite on the opposite polarity ones. . Not much current should flow on a good FET. The 2 VDC out of the ohms test of the meter was not enough to predict OT's that were going to explode when I put rail voltage on them.
If any OT is bad, all the parts on this board should be individually tested. At least that is what I did on my amp.Diodes were blown or shorted across, transistors were blown or shorted across, resistors were vaporized or opened up, a couple of PCB traces were vaporized. Capacitor checking was a bit of a trial. I had 50V rated ceramic capacitors on the driver board that were opened up by the 85 v rail voltage. If it had vaporized them they would have been easier to spot. My PV1.3k amp had a bad solder joint on the input to an op amp that would cause the amp to occasionally whang the B channel output to +170 VDC. The previous repair man handled it by pulling all the power supply connections to the B channel, leaving the output transistors bad, and putting a label on it "do not use B channel". My repair wasn't economic, took me a long time, but it is permanent.
Last edited:
Q6 and R27 are part of the VI limiter circuit along with the ones on the positive side. They monitor the voltage across the source resistors and limit the drive to the output gates. You can remove them for troubleshooting purposes and the amp should run fine without them. Don't forget to replace them after you have the other problems figured out.
Craig
Craig
Fine with me, but now you'll have to troubleshoot it yourself or have somebody do it.
Start by building a lamp bulb current limiter, you'll need it.
Then you'll have to start measuring , visual clues do not help much.
Here's the schematic, converted to .gif to avoid downloading time and again from 2 shared or whatever.
PS: unfortunately the bias circuit design is not foolproof, Q2 *can* be shorted base to emitter (they should have added a resistor in series with the pot) biasing output transistors into death.
Start by building a lamp bulb current limiter, you'll need it.
Then you'll have to start measuring , visual clues do not help much.
Here's the schematic, converted to .gif to avoid downloading time and again from 2 shared or whatever.
PS: unfortunately the bias circuit design is not foolproof, Q2 *can* be shorted base to emitter (they should have added a resistor in series with the pot) biasing output transistors into death.
Attachments
Last edited:
Hi.
So I digged into this.
All the Mosfets exept one IRFP240 in the Power Amp section are gone.
I built a test circuit for them and run them with 12V PSU. All of them were shorted exept one. I guess this one just survived for no reason.
Also all the current limiting 5W 0.47 ohm power resistors exept R32 (for the MOsfet that survived) are gone.
So I will buy new Mosfets and at first run the amp only with one pair of mosfets for testing.
Is there a reason that makes sense why this could have happened? Can extreme cold startup cause that (-15 Celsius)?
The Bias I did for this amp was only AP1 Trim, this only evens the PP sholders and not the power. So I guess that if there is something seriosly bad with this amp or its just bad luck.
So I digged into this.
All the Mosfets exept one IRFP240 in the Power Amp section are gone.
I built a test circuit for them and run them with 12V PSU. All of them were shorted exept one. I guess this one just survived for no reason.
Also all the current limiting 5W 0.47 ohm power resistors exept R32 (for the MOsfet that survived) are gone.
So I will buy new Mosfets and at first run the amp only with one pair of mosfets for testing.
Is there a reason that makes sense why this could have happened? Can extreme cold startup cause that (-15 Celsius)?
The Bias I did for this amp was only AP1 Trim, this only evens the PP sholders and not the power. So I guess that if there is something seriosly bad with this amp or its just bad luck.
Last edited:
Ouch !!! 
No, I wouldn't blame cold for it.
You simply biased it too hot, (literally), maybe didn't notice it because you did not have it on long enough (there is a lot of thermal mass inertia, after all) but under actual use it overheated to death.
I would replace *also* the surviving MosFet and ballast resistor, they have been certainly overheated and can't be trusted.
Consider all semiconductors to the right of C6/C7 suspect, including the bias transistor.
Not necessarily replace them "just in case", I try to mess as little as possible with PCBs, but have a couple spares handy.
They are cheap parts anyway.
The trimpot is also suspect, I'd replace it (and start by setting it to max value).
And ... good luck, the amp is repairable, just go slow and steady
No, I wouldn't blame cold for it.
You simply biased it too hot, (literally), maybe didn't notice it because you did not have it on long enough (there is a lot of thermal mass inertia, after all) but under actual use it overheated to death.
I would replace *also* the surviving MosFet and ballast resistor, they have been certainly overheated and can't be trusted.
Consider all semiconductors to the right of C6/C7 suspect, including the bias transistor.
Not necessarily replace them "just in case", I try to mess as little as possible with PCBs, but have a couple spares handy.
They are cheap parts anyway.
The trimpot is also suspect, I'd replace it (and start by setting it to max value).
And ... good luck, the amp is repairable, just go slow and steady
Hi.
I will change them all in this case.
Could you give me a little explanation of how this Bias trimpot actually works?
I think I have mis understood the function of this Trimpot.
I thought that what it does is regulates the input signal to keep it between the limits for Mosfets. But there is something else that I dont understand. For example. What impact it has on the current that all the mosfets are conducting?
I will try to simulate the schematic and read some stuff to educate myself.
EDIT: Something I already found out is that the distortion that I heard before was crossover distortion not Clipping distortion.
And what probably happened was that my biasing left all the MOSFETS on. Even if there was no signal.
So just to say it out lout. Biasing is needed to add Gate voltage just so that MOSFET is exactly on the point of turing on. That way there is no crossover distortion (read: crossover Dist is reduced).
So I ruined it, a lesson learned.
Now the main question.
How to get the bias set correctly? As I understand correctly then starting point should be TRIM set to 0 ohms.
Should I find out the point where all the Fets open up (without any signal) and then turn it back a little?
I understand that it would be more precise if I had a signal generator and a scope, but I don't have a scope so I can't see what is going on there. But I have a good DMM.
I will change them all in this case.
Could you give me a little explanation of how this Bias trimpot actually works?
I think I have mis understood the function of this Trimpot.
I thought that what it does is regulates the input signal to keep it between the limits for Mosfets. But there is something else that I dont understand. For example. What impact it has on the current that all the mosfets are conducting?
I will try to simulate the schematic and read some stuff to educate myself.
EDIT: Something I already found out is that the distortion that I heard before was crossover distortion not Clipping distortion.
And what probably happened was that my biasing left all the MOSFETS on. Even if there was no signal.
So just to say it out lout. Biasing is needed to add Gate voltage just so that MOSFET is exactly on the point of turing on. That way there is no crossover distortion (read: crossover Dist is reduced).
So I ruined it, a lesson learned.
Now the main question.
How to get the bias set correctly? As I understand correctly then starting point should be TRIM set to 0 ohms.
Should I find out the point where all the Fets open up (without any signal) and then turn it back a little?
I understand that it would be more precise if I had a signal generator and a scope, but I don't have a scope so I can't see what is going on there. But I have a good DMM.
Last edited:
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.