ESP Project 03 Bias/Oscillation Problems

[jackand08]

Hi everyone,
As part of a computer speaker system I am making, I have designed and built a PCB for ESP's project 03 amplifier (to use for the subwoofer). I used the standard circuit (the one with the bias control) and all of the parts Rod recommends in his article (BD139/140, 2N3055/MJ2955 etc.) and didn't substitute any of the parts.

The amplifier powers up fine, with the DC offset at the output rising as expected then dropping back down to 0V. The amplifier also passes signal (a 1kHz sine wave) fine too, however, there are a couple of problems:

1. The bias control has no effect whatsoever. I powered the amp up with it fully anticlockwise for lowest bias (to avoid destroying the output devices) and still, even with it at the minimum setting, measure around 190mV over one of the 0.47 ohm emitter resistors. I calculated that for 20mA through the output devices (as Rod recommends for +/-35V operation), there would need to be around 9mV over one emitter resistor however, the bias control will not let me adjust it down to this value (or any other value even). I have tried putting the bias transistor (a BC549) in the other way round and also a new transistor but to no effect.

2. The output stage is oscillating. I'm not quite sure at what frequency but it is too high to be heard. It is only low-level (i.e. not heating up the output devices too much) but is still there, regardless of whether a speaker is connected or not. I am waiting for the 10 ohm resistor for the zobel network from my supplier (it is on back order) so at the moment the zobel network is not in place. I was wondering whether this or the fact the output devices are mounted on heatsinks separate from the amplifier board (on around 10cm flying leads) could contribute to this oscillation.

Here is a shot of the oscillation on my oscilloscope:


Here is a photo of the PCB:


The layout is attached at the bottom of the post.

[amptech]

Jackand08...the zoble net work is a must in amp builds and can be from 4.7ohm@3watt type with 0.047 cap will work just fine.. here's what you can use 2.2 ohm to 10 ohms @2watts to say 10 watts and cap's will be 0.047 to .220nf @100v volts or more...

did you use limiter resistors as out lined in the set up? 190mv over 0.47 ohm dropper resistor is high..it should be 50mv here..or 75ma...you can also place a bulb limiter in the ac mains in and see if the bulb stays bright..if all is well it should light then dull out..

also your wiring for the o/p transistors should be very short or built into the same track lay out as they could and will pick up stray noise...
have you followed rod's advice for setting and trouble shooting..poor pcb lay out wrong solder links parts fitted wrong etc.. I did built this amp many many years ago just to see how well it sounded...
I would say recheck your work again go from each resistor/transistor etc.. and place the quick of checking current draw will be to place your test meter in one of the dc rails to check it's idle draw..

best of luck for now..

[jackand08]

Thanks for your reply,

I should have included in the original post that when powering up, I used my bench top power supply and increased the voltage steadily to +/-30V with a current limit set at 100mA. Because of the problem with the bias, this current limit was too low so the power supply would not let me go any higher than around +/-20V. I removed the current limit and everything was fine, no sparks or anything so continued testing. I had only set the limiter in case anything was badly wrong, so once it prevented me from increasing the voltage (having checked everything was OK), I removed the limiter. The amp draws around 350mA at idle from the power supply, even with the bias control set at (supposed) minimum.

At the moment, the main problem is the bias control. I am pretty certain that the oscillation is caused by the lack of a zobel network and the lead length from the output transistors. I can work around this whilst I am fixing the bias control.

[Tekko]

and the Vbe multiplier transistor is supposed to be mounted on the heatsink with the power devices.

[amptech]

what is voltage reading in the vas stage before it see's the driver's to the o/p transistors?

I'm relooking over the circuit diagram just now,and the only thing in the vas stage is bd140 and bc456 plus a few resistors and a bootstap cap in the line up.. I would have to dig my one out and check it as it's current draw was not as high as 350ma! with out the zobel r in there leave off any load at this stage just in case it fries your speaker.

[Tekko]

I also see things that doesent match rod's P3A schematic, no current source for the input stage.

Did you put this together from memory rather than rod's schematic ?

60-80W Power Amplifier

[Bone]

Originally Posted by Tekko

and the Vbe multiplier transistor is supposed to be mounted on the heatsink with the power devices.

No the vbe multiplier should be in contact with one of the DRIVER transistors in this circuit!

[amptech]

Tekko have a look at rods 03 amp he's not mounted this transistor to the main heat sink to track o/p's... replace the bc546 with bd139 and mount it along side the power transistors..

it's a basic amplifier circuit and works well...but this one has a little fault to correct it..

[Tekko]

Well rod does mention not to build this circuit and instead refers to the P3A, which i built on a piece of plywood with no oscillations what so ever.

[amptech]

what is being asked here is ...a little advice has to why it's not setting... possible's are incorrect pcb lay..transistors with that part of the circuit in the wrong...