please help with an oscillating amp

[ballai.tibi]

Hi diyaudio members!

I'm attempting to build David White's Mos100 amplifier.

http://www.wnaudio.com/amos100.pdf (details and schematic)

I have a mono stage ready, but it's oscillating.

I have connected a 40W bulb in series with the AC line, the VR is turned fully anti-clockwise (0 ohm).

If I power it up with a dummy load it draws about 20 mA in the positive supply line.
The output measures 0V AC/0V DC.
But, if I try to adjust the bias, as soon as the VR is turned to the point where the amp will start drawing more then 20mA it will start to oscillate even if I turn the VR back to 0 ohm. The output reads 0VDC/2.7V AC.

Using a speaker as the load the same thing happens as I power it up (the amp appears to be stable, no output and draws ~20mA), but as soon as I touch/tap the speaker, I get the same result. 2.7V AC in the output, drawing about 350mA in the positive supply line.
I suppose the oscillation is in the inaudible frequency range.

Other then that, the amp appears to be working, of course I get the inter-modulation distortion since I can't set the bias.

I also tried running it without the 40W bulb in series with the AC line.
I didn't have to tap the speaker, or to touch the VR in order to get it to oscillate. I didn't have a chance to check how much current it was drawing or the output, since the 10 ohm 3W resistor in the zobel network was giving off smoke so I just unplugged it.


I used the same parts as in the schematic except for the output mosfets (I used BUZ901/BUZ906) and instead of the 2N 3704 I used a BC 337.

Unfortunately I don't have a scope, and I don't know anyone who does...


Any help would be greatly appreciated!

Tibi.

[lineup]

I assume you have the R22, C18 filter attached.

The first I should try
is increase C14 to 47pf and 100 pF is needed

If this does not help I would add C15.
Maybe try something like 100-470 pF

[lineup]

Also the PCB can matter.

I would put R1 and R2 and R5 close to input pins

[gain]

im not familiar with this amp, but some quick things to check for

- zobel network (if used). check resistor and cap with meter and verify good (low R) ground connection to zobel.

- output coil and resistor (if used). doesn't seem suspect though since you are using a dummy load.

- check all components, circuit traces and grounds in the nfb circuit. same for the input network infront of the LTP.

- check decoupling caps to make sure RF or other high freq junk isn't being injected into the rails.

- miller cap. if defective or wrong value or installed improper it will push the P2 cutoff point higher.

[MJL21193]

Originally posted by lineup

The first I should try
is increase C14 to 47pf and 100 pF is needed

I would try removing C14. I had a similar oscillation problem and that was the cure.

[ballai.tibi]

Originally posted by lineup
I assume you have the R22, C18 filter attached.

The first I should try
is increase C14 to 47pf and 100 pF is needed

If this does not help I would add C15.
Maybe try something like 100-470 pF

Yes the R22 and C18 filter is attached, R22 gets pretty hot when it's oscillating.

Thanks for the tips, I'll try increasing C14, I don't have a 47pf capacitor handy, for testing I suppose I can use 2 or 3 22p caps in parallel.

Originally posted by lineup
Also the PCB can matter.

I would put R1 and R2 and R5 close to input pins

I'm afraid that the main cause here is the PCB itself, I couldn't find the original PCB, so I designed my own, but I don't have much experience with designing PCBs.

I've attached the PCB. Let me know what you thing please.

P.S. The OPamp I used is an OPA604 not TL... as in the PCB.

[MJL21193]

Originally posted by ballai.tibi

I'll try increasing C14, I don't have a 47pf capacitor handy, for testing I suppose I can use 2 or 3 22p caps in parallel.

Seriously, try it first without it.
You won't miss that filter anyway.

I wouldn't count on it being poor board design. I've had stuff running stable on pieces of plywood.

[ballai.tibi]

Originally posted by gain
im not familiar with this amp, but some quick things to check for

- zobel network (if used). check resistor and cap with meter and verify good (low R) ground connection to zobel.

- output coil and resistor (if used). doesn't seem suspect though since you are using a dummy load.

- check all components, circuit traces and grounds in the nfb circuit. same for the input network infront of the LTP.

- check decoupling caps to make sure RF or other high freq junk isn't being injected into the rails.

- miller cap. if defective or wrong value or installed improper it will push the P2 cutoff point higher.

thanks for the tips.

I just put the amp together yesterday so I don't think it's component failure, however I wouldn't rule out that some components although new could be defective.

- The zobel network reads a ~10ohm resistance to the ground so that seems to be in order.

- Not sure what I should check for in the output coil/resistor... I wound that myself, 17 turns of 0.9mm enameled copper wire on a 10 ohm 3W resistor.

- I re-checked the components and circuit traces, everything seems to be in order

- Sorry, but I'm not sure which capacitor is the miller cap in the schematic, would that be C4? between the input signal and the ground?


thanks.

[ballai.tibi]

Originally posted by MJL21193



Seriously, try it first without it.
You won't miss that filter anyway.

I wouldn't count on it being poor board design. I've had stuff running stable on pieces of plywood.

thanks for the suggestion MJL21193. however I've already tried that even before lineup suggested increasing the value, but it didn't solve the problem.

With 3x22pf as C14 though the amp appears to be stable with no input signal, I was able to remove the 40W bulb from the ac line and set the bias.

If I feed it an input signal it works, no distortion, no oscillation as long as the output voltage is under 1.5 volts, as I increase the input signal it starts oscillating again.

Even so, I feel that we've made some progress. just not quite "there" yet.


I'll have to get some caps in the 100-470pf range tomorrow to try and add C15 to see if that makes a difference.


Sorry if my reply's appear to be out of order, since I'm a new member I have to wait for each of them to be approved by a moderator.

[homemodder]

Just had a quick look at the schematics. This outputstage has got voltage gain, its going to be complicated getting this amp stable, messing that much around with c14 is not the way to go, maybe go 47pf. Agree with lineup, put some capacitance on c15 and mabe try decreasing R19 a bit to lower the gain, you dont need that much. What transistors are you using for TR4 and 5, theyll have to be pretty fast like the ones shown on schematic.