Will this switch pop?

[plumsolly]

I accidentally ordered the non-shorting version of a Grayhill for a use as dual 10k pot. The circuit is attached. Should I expect the switch to pop?

Thanks,

Ben

 

[dfuruta]

Not what you're asking, but the + input of IC1 needs a bias path to ground.

 

[plumsolly]

Hey there,

Yes - I added 470K to ground.

Thanks,

Ben

 

[PRR]

> I added 470K to ground.

That is way-large for 553x. Input bias can be a half microAmp. In half-Meg resistor that is a *quarter Volt* of offset.

This part-Volt is applied directly *to your headphone*. No, wait, first amplified by gain-of-eleven.... over TWO Volts (worst case) DC at your headphones.

Any why a Hundred uFd of cap to drive a 470K load? Bass response is 0.0034 Hertz! Even the quarter-dB point is 0.027Hz!

Make 553x input resistor like 50K.

Make C6 more like 1uFd.

The smaller cap will not leak so much DC. If you favor exotic-juice caps, you can use even more exotic cap and stay on budget. Even lose the 100nFd.

-3dB is now at 3Hz. -0.3dB is near 20Hz. There's no headphone alive with 0.3dB accuracy at 20Hz.

Actually, _I'd_ put some DC block before the THAT1200. Who knows what crap may come in from external sources? Most are benign, but a few leak bad. Then C6 is not essential.

If you must use that monster 100uFd cap, put it in series with R15. That kills the gain-of-11-at-DC factor. With ~~50K bias and DC-block at R15, the output DC should be 0.025V (probably less). This adds a bass-cut at 1.6Hz, -0.3dB at 13Hz.

You also need a low-ish bias resistor on IC1 because when your switch is between contacts and "open", the floating input will suck-up all the buzz in the room and half the radio stations in town. The lower the impedance the better. In this case the 2.5K impedance of your 10K string suggests staying well over 25K for <1dB "error" in your 3dB steps. Though if your ears are not calibrated in precise 3dB jumps, a little error may be unimportant.

 

[CJ]

you might try a 1meg or 10 meg resistor added to the switch so that it is across the contacts at all times,

is that right?  pultec does this?

 

[plumsolly]

This place is still awesome.

I went ahead and changed the input resistor to 47K.

I ordered a 2.2µF for C6.

As far as the switch pops, I went ahead and built it and they are tolerable.

I was thinking of switching the 5534 to an LME49710 because it is unity gain stable. I have a gain of 6 right now (R15 = 2k) and its too much. I would change R15 to 10k for a gain of 2. this will help with pops, too. Will still be pretty loud. The LME49710 also has input bias of only 7nA.

A question on putting a cap in series with R15: You still need one in series with the input, right? A cap in series with R15 with bring DC gain to unity, but you still have to block dc getting into the amp?

Thanks,

Ben

 

[plumsolly]

Though I'd share the final schematic and some pictures of the implementation. I am really happy with it. Thanks to everyone for all the help.

Ben