LM3875 full voltage through speaker wires?

[ozerik]

I'm building a gainclone with the LM3875TF chips, following this design diyaudioprojects.com/Chip/Synergy-LM3875-Gainclone/.

In on a budget, so everything is sourced from an old receiver or PC power supplies. The power supply checks out, with + and - 32.5V coming out of the rectifying/smoothing circuit. My airboarding looks good to me, and matches this schematic , with the 50k pot omitted, and the smoothing caps are two 3,300 caps on each rail, plus bleeder resistors instead of the impressive array that guy uses.



So... when i checked for offset, i got close to full voltage through the speaker output! I feel like i must have made a giant noob mistake here.

I take that back -- a giant noob mistake would have been to hook it up to speakers and test it like that. But still, can anyone see my error? I only have one ground star, instead of separate power/signal...

[osscar]

if you do not use a volume pot - you must add resistor 22k between input and ground.

[Zigis]

If you have full voltage on output, probably there is no connection (bad soldering) between one power leg and PS.
If you have + voltage in output, check negative power connection, and if negative in output, check positive rail.

Yes, and don't forget about input R, like osscar say.

[ozerik]

A 100ohm resistor from input to ground fixed it! AWESOME -- thanks guys!

Also weird: there's two chips, and one of the outputs was at + rail voltage, the other was at -. I'm not an electrical engineer, but i can think, and this doesn't make sense to me.


But it's all fixed now! I can fine tune, and stick everything in place!

[ratza]

100 ohm is VERY small. Use 22k to match the NFB resistor.

[cliffforrest]

"Also weird: there's two chips, and one of the outputs was at + rail voltage, the other was at -. I'm not an electrical engineer, but i can think, and this doesn't make sense to me"

With all amplifiers there will be a (small to very small) current flowing into the input.

If there is no path for it to flow, ie a resistance to ground, the amplifier will do almost anything - usually go hard up against a power rail. Either one. This you have noticed.

The input resistor should always be connected, and if no pot is used connect it between the 220R input resistor and ground.

There is no point using a value below about 100K - it just loads the input stage unecessarily.

[kroto]

Originally Posted by ozerik

A 100ohm resistor from input to ground fixed it! AWESOME -- thanks guys!

Also weird: there's two chips, and one of the outputs was at + rail voltage, the other was at -. I'm not an electrical engineer, but i can think, and this doesn't make sense to me.


But it's all fixed now! I can fine tune, and stick everything in place!

100R is a bit heavy load I think for the stage before it, such as a pre-amp.
thus will increase the Distortion right?

[ozerik]

The schematic i saw calls for a 100 ohm pot. And it seems about perfect. It seems to drive my speakers just to the point of them distorting.

And i'm driving it exclusively from a headphones output. I built this because i'm on a budget, and need a power amp. The exceptional sound is just a bonus!

[ozerik]

Okay, whoops, i am wrong. It calls for a 50k pot. And maybe the resistor is 100k, not 100R. I mean, i did measure it...

[ratza]

How and what did you measure?