Ampeg SVT Solid State Amp-Need Help

[CJ]

I replaced all the power mosfets in this bass amp but it is clipping. There seems to be an imbalance in the bias voltages such that the source-gate voltages are incorrect,

Here are the voltages , maybe those two transistors with the red arrows are not biased right?

Also, what is the purpose of the opamp circled in read, some type of feedback maybe?

Thanks for any help!

PS: the spreader circuit in the middle of the two transistors has the abillty to make the offset voltage at the speaker terminal go to about plus 50 volts. Over...

 

[Khron]

The circled opamp looks like a DC servo to me.

 

[JohnRoberts]

The voltage drop across the source degeneration resistors looks light. 10mV across two of them is like 10mA class A operating current, light but should still work?

The DC servo is pegged negative, but MOSFET drive string is sitting positive. So figure out why.

I would be tempted to crank in a little more class A current. Bipolar amps generally like 25mA, mosfets generally more...

JR

 

[abbey road d enfer]

The voltage across the zeners seems incorrect to me. They should have 3.3V across.
Looks like Q4 & Q5 are not really conducting.

 

[TheJames]

SVT3?

 

[CJ]

Thanks for all the help! We have made progress!
I stuck in the original rail transistors back in there as i was worried about hfe on the new pair i put in, when i checked the old ones, they were matched to hfe = 122 and 123 which is pretty good for a NPN and PNP pair,

I also stuck in a new TL074 as i was worried that somehow it saw excess voltage on Pin 1 from somewhere. This seemed to bring in the output voltage on pin 1 down quite a bit.

yes, SVT 3 Pro
I bought 8 new mosfets from a guy in Israel who matched them well.
I also installed new 0.47 ohm di-generation resistors.

looks like when some mosets failed, they shorted gate to drain which then nuked one of the zener diodes , D7, in the clipping circuit. This caused an imbalance of the bias as the shorted zener (well it read about 160 ohms) now fed the good zener so we were limited to the forward drop of the diode plus the resistance of the bad one, thus the 0.93 volts between gate and source which caused the clipping.

Now the amp is pumping out power but i can not dial out all the crossover distortion. I ran out of 3.3 volt zeners so i used 3.6 volt zeners so maybe you are right about needing more conductance.

here are the new voltages and a shot of the crossover distortion.

if i get more current going through the spreader, i bet the distortion would go away.

the fan runs in series with the heater voltage for the tubes which i find to be a bit distressing.

How does that servo opamp function with just a cap in the feedback loop?
Would a lower voltage zener get those transistors conducting better?

 

[abbey road d enfer]

Still, the voltage across the zeners is even more incorrect.

 

[moamps]

That amplifier does not seem to have a global NFB applied, so it is necessary to bias the output transistors in class AB to make the crossover distortion acceptably small. Probably around 20-40mA per transistor, depending on the size of the cooler. To achieve this, the voltage on the Vbe multiplier (+ resistor's R10 voltage) should be around 7-8V. Those types of Mosfet have a knee in the Vgs characteristic at about 4V. In this sense, you should check the current sources around Q3 and Q4 because something is not quite right there, as someone has already noticed. Check resistors R20 and 22.
The IC is set to work as an integrator, or in simpler terms an "RC network" or lowpass filter here so that only the DC signal comes to resistors R12 and R18.

P.S.
Just as a curiosity, Mosfets IRFP240 and 9240 are not a good complementary pair at all, although many use them that way.

 

[JohnRoberts]

0mV across the source degeneration resistors is even less class A current than before.

As I posted before 25mA or higher class A current is probably expected.

Measuring around 2V across 3.3V zeners suggests an issue there.

Is that 3.85V measurement across the entire drive circuit?

JR

 

[abbey road d enfer]

Simulation shows that transistors Q3 & Q4 have only 0.7mA to drive the bias multiplier Q2. With trimmer at max this results in only about 4.5V across the 15k resistors, which is extremely low copmpared to the 6-8V that seem adequate for these MOSFET's.
I have a feeling the designers have opted for safety at the cost of x-over distortion.
Of course, comparison with a similar amp would confirm (or not).

 

[CJ]

with a bass plugged in it sounds horrible, so i need to get those fets turned on,

i will order the correct zeners and see if that helps,

maybe do a breadboard of the current sources and the spreader so i can see how this thing works,

what about an alternate current source , is there anything that will handle hi voltage and deliver current in a more modern fashion?

ah yes, the integrator, need to re read Mimms Radio Shack Opamp Handbook

thanks or all the help! lot of talent round here!

entire schemo>

 

[moamps]

Simulation shows that transistors Q3 & Q4 have only 0.7mA to drive the bias multiplier Q2. With trimmer at max this results in only about 4.5V across the 15k resistors, which is extremely low copmpared to the 6-8V that seem adequate for these MOSFET's.

Could you do a simulation with the trimmer set to 2k7? Thanks.

 

[JohnRoberts]

with a bass plugged in it sounds horrible, so i need to get those fets turned on,

couldn't hurt unless you turn them on too hard.

i will order the correct zeners and see if that helps,

looks like around 7mA of current feeding those zeners... you might try one on the bench

maybe do a breadboard of the current sources and the spreader so i can see how this thing works,

what about an alternate current source , is there anything that will handle hi voltage and deliver current in a more modern fashion?

I wouldn't start redesigning it until the you get the zeners up to their rated 3V

ah yes, the integrator, need to re read Mimms Radio Shack Opamp Handbook

pretty common DC servo topology.

thanks or all the help! lot of talent round here!

not fixed yet... good luck.

JR

entire schemo>

View attachment 96172

 

[moamps]

i will order the correct zeners and see if that helps,

Surely you have red (recommended), green or yellow LEDs somewhere. Two LEDs in series will replace a zener diode of about 4V.

 

[abbey road d enfer]

Could you do a simulation with the trimmer set to 2k7? Thanks.

Doh! I simmed it at max, which indeed results in the lowest voltage.
At 2.7k it sims at about 8V, as well as at zero. At mid-rotation (5k) it sims at 5.8V.
Anyway the problem seems to be that the zeners are not properly powered or duff.

 

[marucyun]

hello.
I often visit this site when I have a problem.

I am posting for the first time.

I was also annoyed by SVT3pro.

due to overcurrent
IRFP240 & 9240,
Some transistors, resistors and ZDs were dead so I replaced them all.

D2 and D3 (3.3VZD) after replacement are around 2V,
Both ends of R16 and 17 were maximum 5.3V.

After much thought, I realized one thing.

I used 1N4728 for the ZD that replaced D2 and D3,
It wasn't good.
It seems that the current was not enough for the 1W type.

When replaced with 0.5W 1N5226 (3.3V type, 3.6V when actually measured with 100Ω resistance to 9V battery),
Both D2 and D3 increased to 2.8V, and the R15R16 voltage also increased to 7V or more.

When I measured the D2 and D3 voltages of my friend's SVT3pro, it was 2.8V as well.

I hope that this experience will be useful to you.