Quasi amp problems oscillation

[Tekko]

SO heres the schemo for the amp i've had problems with:

http://i.imgur.com/3cMIs.jpg

After that miniamp i decided to hook this up again but it still does the same thing.

When i give it a signal it oscillates, when i take the signal off the oscillation stop.

If i try to give it bias it first starts like interrupted oscillation, then as i get higher it goes into steady oscillation.

It's based on one of quasi's designs.

The drivers and bias transistor is BD139 and BD140, NOT 2N5551 and 2N5401.

[djoffe]

Here's a good approach to stabilize the amp.

1. start with just one pair of output mosfets. They vary so much in threshold voltage, that the multiple outputs will have problems unless closely matched.
2. drop the power supply down to +/-25, so if something bad happens, it won't destroy the world.
3. remove C9 and C8...they can be bad news...they may make Q11 and Q10 oscillate all by themselves.
4. It's way over-kill, but change C6 to 1000 pF...remember, our goal is to get it stable, then we can make it good. That limits GBW to about 1/(2*pi*100*1e-9)=1.59 MHz. With your feedback factor of 1.2/(1.2+39), your 3 dB freq will be around 47 kHz...ok for a first test.
5. Make sure power supply bypasses near the gain stages include 0.1 uF caps with short leads from both rails to ground.
6. Adjust bias Vbe multiplier for something reasonable...maybe 100 mA in the output devices.
7. Don't push the amp too hard...there's only one set of output devices there in this test.
8. If it looks ok, e.g. no oscillation, you can start backing C6 down until you see oscillation when driven.
9. If it still oscillates with C6=1000 pF, and everything else is done right, then you have a local (e.g. not a loop) oscillation. Most likely cause is Q11 M3 loop. If your 4 Ohm resistor is non-inductive, there should be no problem.
If your 4 ohm resistor is inductive, or if you remove the load, the Q11 M3 loop may oscillate. If so, you need zobel at output node...10 Ohms in series with 0.1 uF is a good place to start.

Good luck

[Tekko]

At 2x35V rails it runs just fine with no oscillations in sight. With 2x50V rails i never get above 12mA bias.

Without the caps on the drivers it oscillates even worse.

The oscillation is only on the positive side which i find weird, from 0v to around 5-8v.

[Workhorse]

Tekko,

82pF cap on both drivers B-E terminals is wrong concept. The Position of such caps should be at B-C terminals of Drivers.

Using BD139/140 at rails voltages more than +/-35VDC is again a blunder mistake because during full output voltage swing the VCE limits of drivers will get exceeded and they will blow in no time. Replace them with MJE340-350 Pairs.

Hope this helps,

K

[Tekko]

The drivers dont blow, if i reconfigure it to that miniamp it works just fine.

And if i put those caps between BC or BE doesent seem to make any difference what so ever.

[ontoaba]

Originally Posted by Tekko

At 2x35V rails it runs just fine with no oscillations in sight. With 2x50V rails i never get above 12mA bias.

Without the caps on the drivers it oscillates even worse.

The oscillation is only on the positive side which i find weird, from 0v to around 5-8v.

I ever told someone in this forum that hexfets are easily switched at high voltage. I'll suggest the same with that time, place a zobel network, connect the zobel ground close to c4 and c7, add 100nF parallel with them. Replace R4 and R3 with higher values, and make it with higher gain by replacing R6 with lower value.
Your quasi schematic has high drop voltages it is inefficient for high power (you have many parallel IRF, it must be high power). Also it need higher wattage transistor for its driver longevity not 2n5401.

[Tekko]

SUCCESS!!!

I replaced the BD's with IRF540 and IRF9540 and now it works just fine!!

Maybe those BD's are the problem. I'll order some MJE340's and MJE350's and see if those will work.

Even these IRF's are cutting it a bit short on voltage, beeing 100v devices.

[Tekko]

Heres a vid:
YouTube - High power Mosefet Amplifier