2 general gainclone questions, and an lm1875 question

[Adam M.]

I'm trying to find the answer to 3 questions before I start on my next gainclone project, hopefully the members here have some answers...

1) Anyone know a source for cheap stereo pots that could be used as a volume control? I'm looking less than $5/part.

2)Would a painted pentium 2 heatisnk provide adequate insulation from the -vin that is on the back tab of an LM1875? The coatning may actually be a black annodize, as it doesn't chip.

3) Can the input cap and the corresponding 22K resistor be removed from national's LM1875 schematic with no adverse effects?

Thanks for the help.
-Adam

[chipco3434]

1. Radio Shack has a nice stereo pot made by Alps for about 3 bucks.

2. No.

3. I'm no expert, but most likely not.

[widman]

I bought a mixed bag of stereo pots from Electronic Goldmine and ended up with a couple of decent potentiometers.

widman

[homer09]

Originally posted by Adam M.

3) Can the input cap and the corresponding 22K resistor be removed from national's LM1875 schematic with no adverse effects?

Yes the cap can be removed, people have been removing it from LM3875 based gainclones.

Adverse effects: no DC input block, less stable input impedance/DC offset, possible (but unlikely) damage to speakers.
Positive effects: no cap! (and the nasties those bring in signal path)


I would be more careful removing the 22K res though.

Adverse effects: possible very high DC offset depending on value of pot used (measure before plugging in speakers )
Positive effects: ?

[Sherman]

Originally posted by widman
I bought a mixed bag of stereo pots from Electronic Goldmine and ended up with a couple of decent potentiometers.

widman

I did the same a couple months back. They had a deal with stereo pots in 10K and 50K for something like 50 cents each. I bought a bunch of each value and put the meter to them. The worst matched pots were no worse than what one would expect from any new pot and the best were matched within 1%. Quite a deal.

[Adam M.]

Originally posted by homer09


Yes the cap can be removed, people have been removing it from LM3875 based gainclones.

Adverse effects: no DC input block, less stable input impedance/DC offset, possible (but unlikely) damage to speakers.
Positive effects: no cap! (and the nasties those bring in signal path)


I would be more careful removing the 22K res though.

Adverse effects: possible very high DC offset depending on value of pot used (measure before plugging in speakers )
Positive effects: ?

Should I just leave the 22K resistor and remove the 1Mohm resistor, since its adddition to the resistance will be insignificant?

[homer09]

Originally posted by Adam M.


Should I just leave the 22K resistor and remove the 1Mohm resistor, since its adddition to the resistance will be insignificant?

The way it is set up in the data sheet is the following. The cap isolates your source from your amp's input, which is a practical thing. But because they are isolated and to keep DC offset reasonable, your source needs a load to work into (1meg R1) and your amp's input needs to see a load to ground that is not infinite in resistance (22ohm R2).

If you remove the cap, then yes, you are right, keeping the two resistors is useless, one will do the job. I think (but could not confirm as i have not tried myself) if you remove the cap, you could and should remove the 1Meg R1. The 22k R2 must stay or else you will have problems.

With only the 22k R2, your amp should have very similar DC offset as the data sheet's approach, but your source will be working into much lower impedance load (22k vs 1meg). Nontheless, most sources should still work easily into 22k and you should see some RF/EMI rejection improvement over 1meg setup.

This is if you are building a true amp (no volume control) to be fed by a preamp. If you are planning to introduce a pot for volume control, everything changes.


EDIT: R1 is 1000K ohms, not 1 miliohm!! so its resistance is NOT insignificant as compared to 22K R2.

[Adam M.]

I was looking at R1 from the idea that adding 1 meg and 22k in parallel gives you about 22k, which is where I was getting the insignificance from.

I know by using a pot as a volume control, it introduces a varying resistance to ground, and at maximum attenuation, the resistance to ground is in theory 0 ohms. What is a good way around this, since maximum attenuation will be much less than 100%?

Thanks,
Adam

[homer09]

Originally posted by Adam M.
I was looking at R1 from the idea that adding 1 meg and 22k in parallel gives you about 22k, which is where I was getting the insignificance from.

I know by using a pot as a volume control, it introduces a varying resistance to ground, and at maximum attenuation, the resistance to ground is in theory 0 ohms. What is a good way around this, since maximum attenuation will be much less than 100%?

Thanks,
Adam

Sorry misread your thinking about the insignificance of the 1meg.

If you use a log pot as a passive pre amp, things dont get too complicated. If you place a pot at your input of less than 50K you can drop the 22k resistor. Then all you would have is a pot in the signal path (no cap, no R1/R2).

Your source will see an impedance equal to your pot value. This is because the input of an op-amp has very high impedance, so the only significant path to ground to consider is between the two ends of your pot. SO, your source will be working into constant impedance and nothing to worry about there.

Now for the problem you mentioned, when resistance to ground is 0 ohms from your op-amp input to ground. This only happens at max attenuation as you said. Not a big deal, shouldnt cause any problems. But just to be sure, measure your DC offset while varying attenuation to make sure you have no weird problems.

passive pres arent perfect, remember that.

BUT, i guess, if you really want a way around the problem, you can put a resistor between your pot's ground pin and ground. this will theoretically offer a certain resistance to ground for your input at all times, but then your attenuation will never reach max and i think that is more annoying than a bit of DC offset.

Really though, with a pot of less than 50K and nothing else, you should have no problems, people have been using this config for as long as passive pres exist and there are no problems. enjoy!

EDIT: things get complicated when you want to fake-law the pot, but thats a whole different story

[Adam M.]

Originally posted by homer09

EDIT: things get complicated when you want to fake-law the pot, but thats a whole different story

What is the story when you law fake the pot? It looks like this would create the same situation as using a pot as a passive pre and leaving the 22k resistor (although according to Rod Elliott's weblage, the ratio would be off unless the pot had a huge resistance)

I can look at the curves that an ideal law faked pot produces, but what happens in practice?