Gainclone sucking back Current

[rayban68]

Hey all hopefully you guys can help me here..
I made a PCB for a non-inverting gainclone, the circuit is exactly off the datasheet with zobel and 10ohm/inductor on output and 3.3uF dc blocking cap on the input. When I power it up off the bench supply, +/- 20v with no input or output connected, nothing happens. Once I connect the speaker (still no input) the current on the negative rail jumps to .75 Amps and continues slowing crawling up and up and up. (like around .1 amps per second) same thing happens if i have the input connected. Music does come outta the speaker if i put a signal in, and doesn't sound too bad, until the protection kicks in of course.

Any ideas?

thanks muchly guys / girls

 

[rayban68]

And the PCB

 

[boholm]

One of the inputs of the opamp must have a DC connection to ground. This could be achieved by moving R5 (value?) from the input side of cap C3 to +input of opamp.

 

[jan.didden]

Yes I agree. Your + input is floating. measure the outp[ut, you will see a large DC. Grounding the + input with 22k (ideally the same as the feedback resistor) will cure it.

Jan Didden

 

[rayban68]

Thanks muchly guys, I replaced the cap with a wire and the amp works properly now. got pretty hot tho, i'll have to make sure that it is bolted properly to the heatsink. I am using a 4x8" piece of aluminum with 1/2" high fins every 1/2", should be more than adequate....

 

[rayban68]

Ok now I don't know whats up but lets see if you do

I fixed one amp, it has a 10k linear pot in place of R5 (no log pots in stock at the shop in town) and i replaced the input cap with a wire link. This amp works ok at low volume, but when i turn the volume up it sounds pretty terrible. I'm gonna try replacing it with a 100k log pot that I have handy, just to see if it works better.

I also have another amp, same schematic with R5 as a 22k resistor and I was just gonna use the input volume control (computer, cd player....)
This one I replaced the input cap with a wire link, but it still draws 250mA on idle (both +/- rails), and when I connect the speaker I get the same sort of behavior as when the + input was floating.

Any ideas??

Thanks again.

 

[boholm]

First . . . you go out and buy a multimeter. Then you can check for errors . . . BEFORE connecting anything to equipment in question.

When you say "input cap" you mean the C3 one?

When nothing is connected to the input and you connect the speaker does it fail then?

 

[rayban68]

I have a multimeter, and I measure about 120mV on the output when i power it up, and this drops to about 60 mV after about 5 seconds. What other errors should I be checking for?

Yes, C3 has been replaced with a wire link as this was the easiest way to fix the problem with PCB.

This test was completed with nothing connected to the input.

no input/output, the bench supply shows 250mA at +/- 20V . still no input and I connect the speaker and the negative rail jumps to 1A.

 

[boholm]

I have a multimeter,

Oh, sorry. I apparently got the wrong impression reading your posts

and I measure about 120mV on the output when i power it up, and this drops to about 60 mV after about 5 seconds. What other errors should I be checking for?

All this sounds fine.

Yes, C3 has been replaced with a wire link as this was the easiest way to fix the problem with PCB.

OK, just making sure . . .

This test was completed with nothing connected to the input.

Fine.

no input/output, the bench supply shows 250mA at +/- 20V ..

Sounds right, . . .

still no input and I connect the speaker and the negative rail jumps to 1A.

Now it gets spooky . . . BTW, which IC are you using?

 

[rayban68]

Ya its pretty strange, I can't understand it.

I am using the LM3886TF . I removed the chip from the pcb, stripped off the caps and put new resistors and the caps on a new board, with the old chip. Same problem. So i ripped off the chip, destroying it in the process of course, and put a new chip in (my last one ) and same problem. Next I guess I could pull off the R5 and replace it with a 10k pot, but that doesn't really answer the question.


The 250mA that it draws with no connections is not very good, it seems to me that it should be like 30mA or something... The other amp that I have draws hardly anything at all when idle.

 

[Tensop]

isnt mute meant to be coupled via a resistor and capacitor to -VE, not gnd?

 

[jan.didden]

When you bridged C3, the input cap, you made it a DC amp. Any stray DC from the preamp or whatever you drive it with is amplified. The advice given was NOT to take out C3, but to put a resistor from the +input to ground (22k was suggested). That will solve the unbalance in bias while still maintaining the AC coupling.
Be really carefull by doing random things, unless you fully understand the implications. The chip amps may be cheap, the speakers probably not.

Jan Didden

 

[rayban68]

Yes, I was supposed to move R5, not short C3, but as it was easier to short C3 I did that instead. I figured that it wouldn't cause too many problems, since C7 ensures unity gain at DC. I also tried to connect a speaker with the input shorted and get the same results.

I will fix it properly (move R5, unshort C3) on my lunchbreak today and see if that helps at all.


I believe that R1 couples mute to -Ve and C1 to ground, as shown in the datasheet.

Thanks

 

[peranders]

Mr. rayban, in cases like yours and in many other also, I really recommend you to read the datasheet first and all application notes before you trust an unknown website and circuit design. Start with comparing and make up your own opinion. It can be hard to understand what the datasheet says but at least there usually is a typical application and connection.

Normally you have good info there to start with and if you want you can tweak" the design.

 

[rayban68]

I thought I had read all the materials

This circuit is identical to the datasheet (other than my mixup with R5 and C3 which is now corrected)

I don't really want to tweak the design, well not yet anyways , I just want to make it work properly.