Ampeg B-100 problems

[The Dude]

......Also Q5 gave me a reading using the diode test with the + on the emitter and - on the collector of 1.52V. I don't know if that is a problem but all the others did not give me any other readings then what I wrote above.

Was this out of circuit? If so, Q5 is bad.

[Jetbat]

Q5 reading was taken out of circuit. I rechecked the new transistors in the amp and they check fine.

I figured out the previous test I did was flawed.

I had been reading that unity gain stages can oscillate at very high frequencies. Intermodulation can cause it show up in the audio band as noise. That is one of the problems I am having. I don't know if this is what I am experiencing but the hiss and hum both were not there before I started working on the amp.

A new test I did is put a huge decoupling capacitor on the rail to see how that affected the ripple. I used a 10,000 uf cap parallel with C18. It lowered the B+ ripple from 1.2V to 120 mV and the output ripple from 260 mV to 34 mV. While the output ripple was reduced, the noise was unchanged.

Attached are two pictures of the scope at the same resolution, 50 mV per division. The B+ ripple is a nice thin line. The output ripple is all noise. That is heard from the speaker as a hiss.

[guitardad]

I would start over. Put new transistors in. You seem to changing parts in an effort to modify the orig design.

I can see changing caps that esr levels are bad but not just to try to limit noise. Good luck with that.

Did the amp work when you started working on it? You suggested a bad line cord (grounded orig according to schem). Did they cut the ground pin off?

NTE transistors can also be hit or miss in my experience. maybe find a better equivalent. It's a pretty simple circuit. Sometimes we can get too techie when we don't need to. I've done that. Then I just step back. You used replacement transistor, so make sure the pinout is exactly the same. On some subs or equiv's, you have to observe the pinout or ppooooooooof!

[Jetbat]

I am not trying to modify the design. Resistors and capacitors were changed with the same values. Capacitors were not switched to lower noise. Resistors were switched out in the first gain stage where current flows and the carbon composite resistors were removed for metal film in the output section to lower noise. Electrolytic capacitors were changed out because they are almost 40 years old.

The amp did work before I started. No hum or hiss. And it does work now but has a loud hum and hiss.

I changed some resistors, capacitors, the power cord and added a fuse compartment. The power cord was a two prong cord. No ground. The schematic show a ground but that is for the export model only.

[g1]

So everything is normal (voltages, bias, power output, etc.) except the noise level? Is the amp brighter sounding than it used to be?
Maybe you improved something about the frequency response that was killing the noise frequencies before?
Otherwise, I would suggest you buy a can of freeze spray and see if you can locate any components that impact the noise level.

[audiopete]

Did I see a couple cracked Mullard tropical fish caps there by Q4 and beside the 1000/35 cap? They are pretty bad for doing that.

[loudthud]

You might try a bias adjustment. Just tack one or two 47 Ohm resistors across the 12 Ohm resistor that is in series with the diode.

[Jetbat]

I'm fairly certain that C4 is leaking since more current is flowing through the first stage than is supposed to be. The voltage at the collector of Q1 is 17V instead of 19.9V. Did not get a chance to check that before leaving for the holidays. The coupling caps were replaced to make sure that was not part of the problem.

I've been studying the output section of this amp to understand how it works. One question I have is since the output of the amp is not at 1/2 the positine rail, (+1.1V from half), does that degrade the power supply rejection ratio?

[loudthud]

I've been studying the output section of this amp to understand how it works. One question I have is since the output of the amp is not at 1/2 the positine rail, (+1.1V from half), does that degrade the power supply rejection ratio?

No, that shouldn't affect the hum. One part to suspect is C10. Did you replace that? Any hum it lets through gets amplified by the output stage. Try just clipping another cap in parallel with C10 to see if that reduces the hum.

I think the bias current in the output stage is too high. Please try the following: Connect the black lead of your DVM to the plus side of C17. Now measure the Voltage at the Base and Emitter terminals of Q5 and Q6. Also measure the point between D5 and R24. Post the results.

[Jetbat]

I'm am still on holiday and away from the amp so I can not take readings at this time. I did take readings previously on R28 and R29. They were R28-112mA and R29-116mA.

The schematic for my amp is also for the G-100. I found a schematic for the G-110 which uses op amps for the preamp but has essentially the same output section. That schematic has the current through R28 and R-29 as 80mA for both. (Attached schematic)

In studying the circuit, I found the voltage divider at the base of Q2 sets the Q2 base voltage. When I calculate the voltage for the divider using the schematics power supply voltages the answer is around 1V higher for both schematics. When I used my readings I get .4V higher. The base current flowing through Q2 should only pull these voltages down by around 10 mV.

Confused by this I decided to try this circuit in a Spice simulation. Since I can't even find a data sheet for the transistors I had to use a generic model and adjust the 'IS' parameter to get .55 VBE at the current needed. I ran a simulation and the Q2 base voltage was higher than on the schematic but matched the voltage divider calculator.

I decided to lower R19 to see what happens. When I lowered it from 270k to 244k all the voltages matched up with what Ampeg labeled on the schematic! Yay!

I raised R19 back to 270k and raised the power supply to my reading of 48.7V. Most readings matched except for the bases of Q3 and Q4. They should have remained at 1.1V and .55V but the readings from the amp were 1.08V and .47v. Seems like the base of Q4 is pulling more current then it is supposed to.

[Jetbat]

Also I did replace C10.

[Jetbat]

I did measurements and they are from C17 to Q5 base .745V, emitter .112V.
C17 to Q6 base .714V, emitter .111V.
C17 to D1/R24 junction .013V.

I discovered that if I grounded the junction of R16 and C11 the noise dropped dramatically. A schematic for the G-18 (attached) which is the exact same schematic but with different component values had R16 labeled as 47k instead of 470K for the B100. I realized I did change R16 out when I originally started work on the amp. Checking through the original resistors I found out I must have swept up and thrown out one of the two 470k resistors when cleaning so I could not determine if that was the correct value. I decided to try 47k and see what happened. The hiss from the amp with the volume turned down went away. The ripple on the output dropped to 150mV.

I placed a variable resistor paralleled with R24 and adjusted it so Q5, Q6 just started to work. The ripple dropped from 150mV to under 100mV!

What would you suggest the bias should be set to?

[loudthud]

The voltages around the output stage look good and nothing strange is going on.

I discovered that if I grounded the junction of R16 and C11 the noise dropped dramatically.

This points to C2 as the suspect part. The hum is getting into the preamp stage and is just being amplified by the power amp. Reducing the bias just lowered the overall ripple on C18 by 1/3rd, but too much is still getting into the preamp. Make sure R9 is the correct value. I'd forgotten that you said that the volume control would silence the hum if it was turned all the way down.

[Jetbat]

The volume control would not silence the hum turned all the way down. C2 is a new cap and is 1000uf. R9 is correct. The ripple past it is not measurable it is so low. Not sure why R16 would make such a difference.