Intro and first project: preamp and distortion based on velleman PCB kit

[ryanpg]

Hi all. So I'm familiar with electronics enough to wire an electric guitar or do mods to basic circuits (have done some mods to a DS-1 and a Valve Jr. amp).

I decided to get a bit more familiar with building/designing circuits. I found a Velleman mono preamp kit at Frys for $12 bucks (http://www.vellemanusa.com/products/view/?id=350480) and decided to get it and try to adapt it to guitar. I tried to do my homework and learn about basic HPF bandpass filters and basic op-amp operation. With some direction and tips from a buddy who knows circuit design, I started to design some basic changes to the circuit - stuff I can add/adapt without a breadboarding or etching anything (I figure I can cut PCB traces as needed).

I now need to run it by you knowledgeable folks before I heat up the soldering iron.

Here are my objectives:

1) increase input impedance to instrument levels (1meg)
2) adjust HPF to eliminate sub 80hz input signal
3) add hard diode clipping to the first stage of the lm358 (I have other op-amps on hand too tl072, tl062, and RC455HP - but I don't think they're pin compatible)
4) optional - add a cap in parallel with rv5 to create a bandpass?
5) optional - add a basic tone control (LPF) after the second stage of the op-amp
6) optional but probably should be required - add true bypass

The things I'm not sure about are:
1) the hard diode clipping circuit - is it right? It's adapted from http://www.generalguitargadgets.com/richardo/distortion/index.html
2) I'm thinking I need to bump up the input impedance, but I'm fairly lost on the HPF values

I would LOVE any feedback or advice, especially related to the above two questions. Thanks!

Here's the original schematic:



My proposed mods:


The reference schematic:

[ryanpg]

I got some more feedback from my friend:

• C2 to 22nf
• R2 left at 100k
• R7 10k
• Change the added 50nf to 47nf - more common value
• There may be issues due to the reference schematic being non-inverting and my kit being inverting making that 1k and 50/47nf circuit useless anyway
• C3 is essentially there only to keep the amp from oscillating

[ryanpg]

Oops. Caught an error - the circuit as described in my mod schematic adds "soft clipping" not "hard clipping." Don't know if that's something I should re-think. Will there be a volume drop with this soft clipping between the stages of the lm?

[ryanpg]

Another thought (Sorry for replying to my own thread, hopefully someone will join the "conversation" soon):

R1 and C2 create a HPF, and R2 does something to change that too! So to get a cutoff of 80hz I'd need R1 1meg and C2 2nf but what to do with R2?

Also, I think I can eliminate that 1k, 50nf circuit by going directly to ground (jumper it) as C3 and R5 in my circuit act as a HPF? Not sure though.

[PRR]

> increase input impedance to instrument levels (1meg)

The first stage is an inverter. The input impedance is R2 (and other stuff). 100K. (10K in your revision!) If you increase R2 to 1Meg (and R5 to balance), your noise level gets pretty high. Barely OK for a clean amp, hiss like my old air-hose when cranked to fuzz.

You really want the first stage *non*inverting.

All that "can" be changed. But it's all mis-wiring and cuts and air-splices.

In short: this PCB needs a lot of hacking to turn it around for fuzz. It may be better to get some Vero etc and copy some proven layout.

[ryanpg]

PRR, thanks for clarifiying the impedance issue. It makes sense to me now. A noninverting first stage would present a high impedance without the added noise.

I'm still going to go forward with building the preamp, but maybe I should do it in stages. First do what I can with the input impedance (since though not optimal, it's doable), and true bypass. Then go on to distortion and tone control if it seems worth it.

One thing you said that I don't get, "increase R2 to 1Meg (and R5 to balance)," what do you mean by R5 to balance should I replace R5 with a 1meg too?

Figured it out, the 1:1 ratio determines gain. So R5 has to be equal to or greater than R2. I think I can strike a compromise and change both to 500k - this way impedance is set and the same for all frequencies presented by the guitar. Right? Here's the thing, how does R1 affect impedance? It has to have an effect right? I know if R1 at 1M were parallel with R5 at 500k total resistance would be 333k, but it's not parallel, and not series. LOL, this should be simple! :-)

Also, any thoughts on the HPF and where to make changes to get to 80hz.

Thanks again!

Here's my revised (no diode clipping) schematic.

[PRR]

> if R1 at 1M were parallel with R5 at 500k total resistance would be 333k, but it's not parallel,....

Yes it is.

IC1A's "+" input is at (signal) ground. IC1A's "+" and "-" inputs are at the *same* voltage. Therefore IC1A's "-" input is at (signal) ground.

While the ends are different for DC, they are both at "signal ground" and therefore just parallel. 333K as you say.

[ryanpg]

Thanks again PRR, I thought about it for a while and came to that conclusion but I'm glad you confirmed it. I guess I could put a 2M resistor at R1 and end up with a impedance of 400k at all frequencies. I've read even a 10M pull down resistor is effective at stopping pops and clicks. (I'm guessing in this circuit I can probably just remove it altogether since there's no switch!)

[PRR]

The 10uFd elco cap will be leaky, tend to pop.

You don't need anywhere *near* 10uFd to drive audio into a hundreds-K load. 0.01uFd into 1Meg is 17Hz, often ample. And 0.01uFd is likely a plastic-film cap, much less leakage than electrolytic.

[ryanpg]

My thought was to replace C2 (10u) with something close to 4nF (with R2 at 500k) to get a cutoff of around 80hz. I'm still not sure why/how the specs of the original circuit say there's a frequency response of 20hz-20,000hz.

But based on what you're saying, (0.01uFd into 1Meg is 17Hz) I think I'm getting turned around again. Is it C1 and R1, or R2 that sets the cutoff freq?

[Jdansti]

This might be getting ahead of the game, but if you get this amp to pass a guitar signal, you might want to experiment with various combinations of diode/LED clipping on the output. Easy to do without having to rewire anything. If you find a combo (or more than one combo) that you like, just make a small daughter board with the diodes on it and put it in series between the main board output and your output jack. Use a rotary switch if you find that you like several combinations.

You may or may not need the following for this project, but here are some good reference articles about diode clipping to have on hand.

R. G. Keene
A Musical Distortion Primer
http://www.geofex.com/effxfaq/distn101.htm

R. G. Keene
The Technology of the Tube Screamer
http://www.geofex.com/Article_Folders/TStech/tsxfram.htm

Mark Hammer
Diode Clipping FAQs
http://www.diystompboxes.com/pedals/diodes.html

Jack Orman
AMZ Saturation Control
Setting Distortion and Harmonic Levels
http://www.muzique.com/lab/sat.htm

Jack Orman
Tone Clippers
A New Method of Signal Clipping for Effects
http://www.muzique.com/lab/tclip.htm

Jack Orman
Mosfets and Zeners
More Diode Clipping for Effects
http://www.muzique.com/lab/zenmos.htm

Gary Burchett
New Clipper
http://www.home-wrecker.com/newclipper.html

[ryanpg]

Wow! Thanks John. That'll give me some more reading! I have a revised schematic with diode clipping on the output stage, but I think if I'm going to proceed I should deal with the input impedance and filter cut-off issue first and, like you said, get it to pass signal.

But yeah, thanks for the links again!

[ryanpg]

One thing I've missed (and no one else has commented on) is that the OUTput impedance of this amp is 100k. So in the original circuit, 50k input Z and 100k output Z. That doesn't make any sense to me.

All this to conclude that: by changing R1 to 2M and R2 to 500k and R6 to 10k, I'll have input Z around 400k and output around 10k.
Changing C2 to 0.004uF should give me a corner frequency of 79.6Hz (unless R1 plays into the HPF too, which would mean total resistance is around 400k so 99.5Hz).

R1 = 2M
R2 = 500k
R6 = 10k
C2 = 0.004uF

Input Z = 400k
Output Z = 10k
Corner Hz = 79.6Hz (or around 99.5Hz if the value of R1 operates as parallel to R2)

Whew! Thanks for walking me through these basics!

[Jdansti]

Glad you can speak Paul's language!   A lot of this is "Greek" to me!

[ryanpg]

Thanks John! But really, I can't "speak Paul's language," in fact I'm struggling to understand it!

Heh, to take your metaphor a bit further, I'm really just copying an pasting bits and pieces from other people's translations! I'll bet you understand more than me!

[ggedamed]

Ryan, R6 is in parallel with IC1B's output impedance. For op amps the output impedance is in the tens-hundreds of ohms range. So you can leave R6 alone, doesn't have affect much until you make it too low. Also, the usual op amps  will not be happy if you make the output load less than 10k.

[ryanpg]

ggedamed, same thing my knowledgeable friend said. . . he further suggests putting a 680 ohm resistor in series with the output to limit the amount of current drawn from the output of op-amp IC1B if accidentally shorted to a ground destination capable of sinking a great deal of current.

In his paraphrased words "This would prevent the op-amp from sourcing all that current, possibly to the point of extinction (aka "magic smoke")."

[PRR]

> the OUTput impedance of this amp is 100k.

No. What ggedamed says.

There's R6, and other stuff. Mostly C5 and IC1B in series.

Study negative feedback. IC1 output is a few hundred ohms (typically) *divided by* the excess gain thrown-away by negative feedback.

IC1 has audio gain over 1,000, but is fed-back for gain of (say) 10. So say 200 ohms, divided by 1000/10, is 2 ohms.

The exact output impedance varies, but will usually be "very small".

Pencil 10 ohms. +Plus+ the 10uFd cap. All that in parallel with the 100K.

Plus perhaps a few hundred ohms series protection, if you wish.