-6dB or -10dB pad for a high-input-impedance circuit?

[midwayfair]

I use the MOSFET booster as a single-transistor gain stage a lot:



I wanted to add a pad to the input of a circuit using this to prevent some distortion with bass pickups and line-level signals. I can't realistically use a voltage divider because it would require a 10M on the input just for -6dB, and that will cause an extreme loss of higher frequencies.

Am I basically stuck adding a buffer if I want the pad and don't want to suffer treble frequency loss? I'm not totally opposed to that, I'm just not sure if there's something else I'm missing.

One circuit I'm dealing with (the Bearhug - http://1776effects.com/wp-content/uploads/2015/03/Bear-Hug-V2-1.pdf) has the MOSFET hard wired to a lot of gain. But ... if I'm dealing with a signal big enough to distort the MOSFET even when the compression is way up and I'm already at ~1x gain, then the buffer would distort, too.

So my question becomes -- is there a good way to get a gain of less than 1 at the input of a high-input-impedance effect? Should I resort to higher voltage?

[R.G.]

R3 doesn't HAVE to be 10M does it?   

[Gus]

I would build a different circuit. 
A link from the past maybe you can get an idea from it
http://www.diystompboxes.com/smfforum/index.php?topic=97617.msg851183#msg851183

[midwayfair]

R3 doesn't HAVE to be 10M does it?   

No, but once it's below ~500K, the pickup loading starts to become an issue (if it's passive bass pups ... not a concern with line level instruments), and using a 1M divider on the input causes a really severely noticeable loss of treble even with an active circuit (I was using another boost for testing to get the signal up to "line" level).

Gus: As usual, I completely forget that boostrapping exists for exactly this reason. Does the 100K value of R1 there matter? Can I increase it to 1M? The R5/R13 divider is perfect for the pad if so.

[amz-fx]

C2 may have been causing your treble loss. This version is not quite -3db at minimum gain. If you need -6db then change R3 to 510k.

[merlinb]

So my question becomes -- is there a good way to get a gain of less than 1 at the input of a high-input-impedance effect? Should I resort to higher voltage?

You're pretty much stuck with using a high-resistance divider, followed by a low-capacitance buffer. You can add a compensation capacitor in parallel with the series input resistor to counteract the loss of HF, although this will load down the pickups at HF somewhat. You could probably tune this by ear to a happy compromise between bandwidth and loading, though.

[midwayfair]

followed by a low-capacitance buffer

Hmm. I'm pretty much stuck with adding an op amp or transistor for that, then, right?

If I'm adding active components, I think Gus's circuit is going to be the best solution for this, as long as it also works with a high input impedance so I don't lose the highs.

[merlinb]

If I'm adding active components, I think Gus's circuit is going to be the best solution for this, as long as it also works with a high input impedance so I don't lose the highs.

Well your existing circuit can be turned down to practically unity gain, no? So I guess you can use that. Just make the circuit like AMZ FX suggested, and add a compensation cap. You could increase the series resistor for more padding, no probs.

[midwayfair]

If I'm adding active components, I think Gus's circuit is going to be the best solution for this, as long as it also works with a high input impedance so I don't lose the highs.

Well your existing circuit can be turned down to practically unity gain, no? So I guess you can use that. Just make the circuit like AMZ FX suggested, and add a compensation cap. You could increase the series resistor for more padding, no probs.

The bearhug has a feedback cap, 22-100pF, between the gate and source of the MOSFET. I was under the impression that a big series resistor with the input would cut the highs with that cap. (I could remove it, but I like having it there in general.)

[merlinb]

The bearhug has a feedback cap, 22-100pF, between the gate and source of the MOSFET. I was under the impression that a big series resistor with the input would cut the highs with that cap. (I could remove it, but I like having it there in general.)

That's what the compensation cap is for -it cancels out the input capacitance of the amplifier, whatever that may be. The time constant of the shunt resistor and input capacitance should equal the time constant of the input series resistor and associated parallel cap. This gives you a perfectly flat frequency response, but a non-flat input impedance.

[PRR]

Seems to me, 470K+220K (3:1, -10dB) will give a fairly high Z and reasonable high-end with a BS170 or similar small MOSFET.

That Bear Hug's C4 47pFd may be superfluous with a hi-Z divider in front, see if you can live without it.

BTW, because of the variable-gain action (ask Miller), the effect of C4 (and MOSFET's Cgd) can't be compensated over a wide range of gain change. IMHO you should just minimize the C at the gate node, and set your divider Z to pass 5KHz-10KHz (all you need).

It also seems to me that the Bear Hug can go to unity-gain, so if you are overloading then your signal is just too darn hot. Arguably it would have more upside with 18V or 24V supply, but you probably should not NEED high supply voltage in guitar-cord work.

There are other compressors which may be better for you.