DC offset like crazy... OPA541- HELP!

[trespasser_guy]

OK, a friend needed an amp, so I volunteered to throw one together for him. I did so- my first GC. Worked right off, sounded ok. Chips got really hot... (REALLY) so I heatsinked, and tried again. The sound is accurate, but I kept hearing a fuzzy sound in the background. DC offset. Like no other. close to 40mV, which I am sure cannot be good for a 4-8ohm speaker.

I am using OPA541, with +/-25V@4.7A rails. How am I getting this DC? How can I fix it? Add a cap on output?

Please help....

Thanks

 

[daatkins]

To be honest, I wouldn't worry too much about a 40mV offset: that's pumping a mighty 0.04^2/8 = 0.2mW into an 8 ohm driver! And if you're experiencing a noise then it must be AC, not DC. It could be oscillation so check with a scope. If you don't have a scope then try a 12v car indicator lamp (the flashing turn signal) with 0.1uF in series - the bulb should light if any oscillation above about 100KHz is present.

If you're running the chip in inverting inverting mode then a resistor equal to Rf from +ve input to ground should reduce or eliminate any DC offset. Some say this hurts the sound though....

Nice one,
David.

 

[trespasser_guy]

Ok *phew* i was over-reacting.

Believe it or not... nothing was added, nothing was changed, and now it sounds great. I guess the chips just needed to warm up a tad. (really warm... finger-burning hot) It actually sounds quite nice now... I am sure he will like it... cost me around $20. hehehe of course, I lose a transformer to him for this.... and I still need decent heatsinking (just a piece of 1"wide aluminum now) and an enclosure.

Well... I think I need to eat. I have a headache.

 

[Panelhead]

[B

If you're running the chip in inverting inverting mode then a resistor equal to Rf from +ve input to ground should reduce or eliminate any DC offset. Some say this hurts the sound though....

I am waiting on Plitron to send the transformers to try an inverting GC. But the resistor on the non-inverting input was in my beginning circuit. Do this really mess up the sonics.

George

 

[daatkins]

I'm not sure what you mean George: if you're building an inverting amp then the input resistor is on the -ve input, in which case a resistor from the +ve to GND will remove any DC offset. Personally, I don't think a resistor from +ve to GND will affect the sonics, but I haven't tried this so I can't comment. Any differences will certainly be small though.

If you're building a non-inverting amp then the signal will go straight into the +ve input. There maybe a resistor to ground in this configuration but this sets the input impedance. To remove DC offset from a non-inverting amp requires a capacitor in series with the resistor from -ve input to gnd. The value of ths cap will affect the low frequency response and the quality of this part is important- cheap electrolytics are to be avoided!!

Nice one,
David.

 

[trespasser_guy]

Well, now that I am not in such a panic...

I used a non-inverting cirtuit... 2.2uF input coupling cap, 22k input resistor, gain=23 (1k and 22k). 1000uF, 50V per rail, bypassed by .0022uF cap. After 2-3min. warmup (for some reason it HAS TO warm up) it sounds quite good. If you run it before it is warm, it sounds kind of fuzzy and distorted, but cleans up nicely. yellow LED, +/-25V supply@4.7A. I built it in a shoebox, to match my shoebox midbass horn speaker (which it powers to VERY loud levels, with tremendous bass-- now would you thinik that that bass came from a 3" woofer? no... but it does )

Anyway thanks for the help. If I had a digi-cam, I'd take a pic ot show my uhh, handywork, if you'd call it that.

 

[pedroskova]

If the chip is getting that hot, you may well have some oscillation going on. Does the OPA541 have built-in thermal protection?

 

[trespasser_guy]

I am pretty sure it does... It seems to work just fine when heatsinked. I don't know...

 

[Nuuk]

I am waiting on Plitron to send the transformers to try an inverting GC. But the resistor on the non-inverting input was in my beginning circuit. Do this really mess up the sonics.

Not in my experience Panelhead. I'm using that arrangement in my LM3875 inverting GCs and it sounds just as good as with the wire from non-inverting input to ground BUT with 0mV offset instead of around 28mV.

And let's be honest - it's no work at all to try both arrangements!

 

[Panelhead]

And let's be honest - it's no work at all to try both arrangements!

I am planning on only trying it with the resistor to ground from the noninverting input. My speakers have extremely fine wire in the voice coils and no crossovers.
A little dc could damage something if used long term. But I hate to think that the direct grounding may yeild better sonics and I am too chicken to try it.
But I read another post where the comment was the buffered input sounded better because it reduced the amount of ground hash feeding into the input. This is very possible. Some think it sounds worse, some the same, and others better. Sounds like you cannot go wrong no matter how it is hooked up.
I will be glad when the transformers arrive from Plitron. I also am going to try driving balanced which makes this a non-issue. The use of the DVR134 looks simple, cheap, and reportly sounds great.

George

 

[pinkmouse]

I have built GCs with LM1875, 3875, OPA548 and 549, and just about all the circuit variations, and I have never had more than 50mV offset on one that worked properly, so as long as your design is sound you should have no problems with vaporised VCs due to offset.

 

[Nuuk]

My speakers have extremely fine wire in the voice coils and no crossovers.

I empathise with you Panelhead but even with a straight wire to ground you shouldn't get more than 30mV.

One of my Arcam A60's puts out 28mV and has done for years. It hasn't harmed either my active speakers or my Goodmans 201's full-rangers, neither of which use crossovers.

Having said that, I will probably revert to speaker protection for a while when I connect these speakers to the GCs!

 

[miguel2]

I tried an inverted design using OPA541's and the chips were cold (room temperature), even after playing for a long time. From my experience, I could use them even without heatsink. Maybe you have something there...

Miguel