Increasing gain in this circuit

[seten]

http://byocelectronics.com/skilletschematic.pdf

Theres a pretty noticeable volume drop when i kick the panning on for this circuit. I was gonna add a clean boost at the panning input but figured there was a good chance i could increase the gain just by changing a resistor value or something - unfortunately im electronically illiterate so thats where yall come in. A rough explanation of how the panning circuit works would be great too!

Thanks,
Seth

[seten]

Also, i would love to make the panning deeper - it may be my imagination but it feels like low frequencies are affected by the tremolo less than the high ones.

[antonis]

Make R68 & R72 22k and R48 & R57 330K..
(just because...)

[seten]

Make R68 & R72 22k and R48 & R57 330K..
(just because...)

Thanks! Are those changes all for the sake of  boost or something else too?

[antonis]

Boost mainly..
(but you never know..)

P.S.
Can't suggest anything for tremolo 'cause it's a phaser circuit..

[iainpunk]

the JFETs are used as variable resistors here, its just changing volume pots basically.
put a big-a$$ capacitor parallel to R50 and R58 to boost the gain and use a 22k trimmer in series with the new caps to lower its gain to taste

cheers, Iain

[antonis]

Are we talking about boost or distortion, Iain..??

Q8 stage gain is (R49/R50) / [1 + (R46/R48) * (R49/R50)]..
By bypassing R50, you effectivelly make gain almost set by R48/R46..
(a rough calulation yields for distortion..)

[seten]

Boost mainly..
(but you never know..)

P.S.
Can't suggest anything for tremolo 'cause it's a phaser circuit..

Isn't it a two in one phaser and a panning (double trem) circuit? Am I missing something?

[iainpunk]

Are we talking about boost or distortion, Iain..??

Q8 stage gain is (R49/R50) / [1 + (R46/R48) * (R49/R50)]..
By bypassing R50, you effectivelly make gain almost set by R48/R46..
(a rough calulation yields for distortion..)

that's where the trimmer comes in, to lower the gain until clipping is no more (unless you want that, its up to the user)

cheers, Iain

[Rob Strand]

Isn't it a two in one phaser and a panning (double trem) circuit? Am I missing something?

Yes.   The panner is the tremollo.


That panning circuit has a whole heap of small problems.   Adding gain boils down to how many part you want to change and how much gain you need to fix the gain loss issue.

The caps C16 and C19 and cut-off of a lot of lows they really should be 1uF
The caps C17 and C20 are a little marginal.

The 100k loads on the output are loading down the amp, which does lose some gain and swing.

So one angle would be to decrease R49/R59 to 10k, and R50, R58 to 1.2k.   However if you need more gain than that you will need to increase R48,R57 (which increases the gain) and also R47, R56 (which fixes the DC bias voltage on the transistor collector).

[seten]

That panning circuit has a whole heap of small problems.   Adding gain boils down to how many part you want to change and how much gain you need to fix the gain loss issue.

The caps C16 and C19 and cut-off of a lot of lows they really should be 1uF
The caps C17 and C20 are a little marginal.

The 100k loads on the output are loading down the amp, which does lose some gain and swing.

So one angle would be to decrease R49/R59 to 10k, and R50, R58 to 1.2k.   However if you need more gain than that you will need to increase R48,R57 (which increases the gain) and also R47, R56 (which fixes the DC bias voltage on the transistor collector).

1) Did you mean changing R51/R60 to 10k? Just wanna make sure since you said the 100k output load was an issue but then didnt recommend changing it.

2) Since im replacing C19 and C16 with 1uf should I replace C20 and C17 with a short or another 1uf?

3) If I make R48/R57 330k like antonis suggested then should R47/R56 be changed to 150k to maintain the ratio of R48/47?

4) I would really like as much swing as possible, near silence to unity gain if thats doable - any way i could easily accomplish that?

[seten]

Make R68 & R72 22k and R48 & R57 330K..
(just because...)

Could you explain why changing R68 and R72 would boost it? Isn't that the LFO section?

[Rob Strand]

1) Did you mean changing R51/R60 to 10k? Just wanna make sure since you said the 100k output load was an issue but then didnt recommend changing it.

No, changing the 56k's to 10k's drops the output impedance so there's not as much level drop driving the 100k's (R51/R60)
It also increases the swing.

2) Since im replacing C19 and C16 with 1uf should I replace C20 and C17 with a short or another 1uf?

Another 1uF is fine,  a short isn't a good idea as causes DC to appear at the output.

3) If I make R48/R57 330k like antonis suggested then should R47/R56 be changed to 150k to maintain the ratio of R48/47?

Antonis's mod is fine too but doesn't give you the swing increase.  Both my suggestion and Antonis's mod boost the gain.  Both mods together will boost it a bit more.

The main issue is we don't know exactly how much boost you need.   There's a right value and you can over do it.

4) I would really like as much swing as possible, near silence to unity gain if thats doable - any way i could easily accomplish that?

Fine tweaking of all the mods.

Here's something close to best swing.   You can squeeze tiny amounts of swing here and there but this a good compromise  of many factors in the decision process.

[seten]

Thanks! Just finished scoping out the board, gonna make the changes tomorrow - last thing, don't I want the dc bias voltage at 4.5V? So (R48+R49)=R47? Wondering since the original schem has it a good deal lower than that - is that one of the aforementioned issues?

[Rob Strand]

Thanks! Just finished scoping out the board, gonna make the changes tomorrow - last thing, don't I want the dc bias voltage at 4.5V? So (R48+R49)=R47? Wondering since the original schem has it a good deal lower than that - is that one of the aforementioned issues?

You want the collector voltage to end-up *around* 4.5V.   With the emitter resistor present (1.2k in my schem) and the output loads present (100k's) the ideal collector voltage can shift from 4.5V.  You can see on my schem it's biased at 4.7V; I can't say tuned it that closely though.

How to achieve the correct bias point  for different circuits results in different formulas, sometimes more complicated than it's worth.

You can see how my values are reasonable.

Assume Vc = 4.5V
Collector current Ic = (9 - 4.5) / R49 = (9 - 4.5) / 10k = 450uA
Voltage drop across emitter resistor = 1.2k * 450uA = 0.54V
Base voltage = Vb = 0.65V + 0.54V = 1.19V
Treat R49 as a nominated value of 120k
Current down R49 = Vb / R49 = 1.19V / 120k = 9.92uA.
Required R48 value consistent with 4.5V collector voltage, ignoring base current,
R48 = (Vc - Vb) / current down R49 = (4.5 - 1.19) / 9.92uA = 333k

The base current will increase the drop across R48 and raise Vc above 4.5V, which is what we see with my 4.7V.   Probably more than 0.2V.  Since for hFE=200, base current Ib = Ic/200 = 450uA /200 = 2.25uA.  And 2.25uA through 333k will raise the collector voltage by 0.75V.  The spice model I used must have a gain higher than 200. 

The base voltage Vb = 1.19V is high enough to water down Vbe variations and the 0.54V across the emitter resistor isn't throwing too much swing away.


If you want you can do the same for the original, it's not bad.

[seten]

Ended up a bit above unity gain which I like. Gonna read up on common emitter amps and come back and see if I can understand that last chunk. Thanks so much yall!

[antonis]

Gonna read up on common emitter amps and come back and see if I can understand that last chunk.

It might help.. 
(or confuse you more..)



P.S.
We ignore voltage drop on R48/R47 Thevenin equivalent resistance for the sake of lazines  so consider R47/R48 junction voltage as Base voltage Vb..

[Rob Strand]

I guess the tweak algorithm to back-off the gain would be:
- reduce R48 (330k / 220k) to reduce the gain.
- tweak/reduce R47 (120k / 100k) to get around 4.5V to 4.7V;  do your best with standard resistors.

[antonis]

I guess the tweak algorithm to back-off the gain would be:
- reduce R48 (330k / 220k) to reduce the gain.

True but not so simple..
(in the mean of R48 tweaking percentage directly reflected to gain, e.g. 300k -> 220k, gain reduced to 66% ..)

"Open-loop" gain R49/(R50+re) is hardly 8, so we can't consider NFB loop gain (R48/R46) solely setting stage gain..
I'm not sure even if (R49/R50) / [1 + (R46/R48)*(R49/R50)] for gain calulation can give decent results for various R48 values..
It should work for grounded Emitter (about 180 gain) but here R50 already implies significant amount of NFB, drammatically reducing "open-loop" gain..

[Rob Strand]

True but not so simple..
(in the mean of R48 tweaking percentage directly reflected to gain, e.g. 300k -> 220k, gain reduced to 66% ..)

I was just quoting the values in the original and modded circuits not really recommending anything. 

Off hand the gain doesn't  quite go up or down in proportion to the feedback resistor (due to finite gain).   Nonetheless choices are limited.   It's probably going to end-up being 270k or 220k.