Modding a cheap solid state amp for more clean headroom.

[Koreth]

Close. Mine looks a bit different, but it has skilk screenings on the board for the jack and pots to be mounted to the board like they are in your pic. Mine ar mounted to the chassis with leads running back to the board. However I do see a lot of the same parts in the same locations (or close) with the same component values, R1 10K, C1, 223, etc. Using one of the schematics, above, I'd trace out the circuit first before making any mods since yours seems to have a different revision of the board. If you find R2, one leg of it will be going to C4, a 10µF film cap, which goes to pin 2 of the opamp chip if it a JRC48558D like in mine. The part numbers and values might have been changed between your and my versions of the amp, but if it has too much gain, the resistor from the feedback loop to ground is the one you want to change.

[Koreth]

Update:

I settled on the 1k resistor at R2. I get a much fuller sweep across the gain pot now, from fully clean, to rather distorted. Mission accomplished in that regard. Now

With the 2k7 ground resistor, the amp's highs seemed a bit muffled making for a bit of a muddy sound. Using the useful math that Robert posted, and a few other references graciously linked to by Google, I put together a spreadsheet to calculate gain dependent on frequency with ten steps per decade and the values of the capacitors in the feedback loop voltage divider. With R2 at 2k7, gain at 10khz was a bit more than half of what it was at 100hz. Gain peaks at or near 600hz, and remains fairly flat till about 1000hz, at which point it drops off shortly, to just over half the gain at 10khz than what is had at 100hz. With R2 at 1kOhm, the gain peak shifts to around 900hz, and doesn't fall off as fast as the frequency goes up, leaving only a little less gain at 10khz than is had at 100hz. My ears confirm the brighter sound, both clean and at max distortion.

However, clean, it still isn't enough. I want the amps clean tone needs to be brighter still. I have two thoughts on this. One: lower the vale of the bright cap in the feedback loop till I have the highs I want. Two: Since the Tone controls seems to be a simple high-cut filter, play with the values there till I have the highs I want. On one hand, if I make the opamp stage bright enough clean, it is probably going to be nasty when distorted, but I could always turn the tone knob down. On the other hand, tweaking the tone knob won't make the distortion unbearably bright, but might not get back enough highs, as you can't add back what isn't there with a passive filter. That Fender clean we all know and love has is rather mid-scooped, and since the current design is effectively mid-boosted, I'm not sure I'm going how much I can accomplish either way before I have to reconfigure the tone knob to something like the Fender/Marshall TMB tone stack, which isn't worth the effort at this stage, IMO.

Another thing I noticed when testing the resistor values was that even when the gain pot is all the way down, there's still a bit of breakup on attack, even with vintage output single coils pickups. Sweep the volume past 4, and the volume stops rising as much, and the breakup on attack becomes more prominent. This makes me think the amp is still being over driven somewhere, either in the second opamp stage which drives the output chip, or the output chip itself. I think I'll want to finish tracing the amp's circuit, (or at least the signal path) before starting to play with the clipper diodes. When the clipper diodes aren't clipping, the amp should be completely clean, IMO.

[Robert8192]

Lots of good ideas for this little circuit.

Just one thing I found when I rebuild some
preamps. The Ibanez TK999ht uses Zener
diodes on this first stage to prevent it from
overdriving the second stage.
I added the same thing to another preamp
and it limitted the weird sounding distortion
out. Maybe can be used in your design.
They are 5.1V limiters.
Robert.

[Koreth]

So that design uses zenners in the feedback loop to limit how hard the next stage is hit. Don't the hard clipper 1N4148s in my amp effectively do the same thing? If I'm understanding correctly, they clip by sending any signal above their forward threshold voltage to ground, in which case, the signal into the 2nd opamp stage would never be greater than .65-.7Vp, with the gain and volume controls all the way up. Is this correct?

Assuming that's the case, then I suppose I should look at the 2nd opamp stage. It has R11 at 22k in the feedback loop and R10 at 4.7K hanging off that. R10 doesn't go to ground, but it is is tied to the same line which feeds the opamps's positive power rails, which I suppose could act as a ground if the negative power rail of the previous opamp stage could. If that's the case, then the gain in the 2nd stage would peak at 5.47 at 2000hz. So then the 2nd opamp stage would never hit more than 3.8Vp, which less than the 7Vp limit of the power supply. Does my reasoning make sense, or did I get confused again somewhere?

I should trace out and examine at the output stage, shouldn't I?

[Robert8192]

Yes, the diodes do almost the same thing.

The setup for your first stage is so different from
the one I posted, Zeners wont work anyway.
Tube King uses a tube in the second stage anyway.
Maybe add a 2nd stage in that is a tube, Say the easy to build
sub-mini tube stage involving a 6021, and we're in
business. LOL. Anyway, hope you acheive what
you're trying to do.

Robert.

[Koreth]

Okay, finally finished tracing out the rest of the circuit. Turns out the voltage divider on the negative feedback loop of the 2nd op amp stage does go to ground, and not the positive supply as I initially thought. The output chip had "UTC2003" silk screened on it. Google didn't seem to find anything relevant about a UTC2003, but it did find a TDA2003 made by UTC. The ratings on the TDA2003's datasheet seems to fall in line with First Act's claimed power output for this amp, as does the resultant arrangement of supporting components when using the TDA2003's pinout. Now that the amp's circuit is fully traced, I think it will be easier to go at modding this thing.

1. More highs. When I run the amp clean, I dime the tone knob. When I play dirty, the tone knob goes down to tame the nastier high freq components of the distortion. That's all well and good, but when playing clean, diming the tone control isn't enough. Looking at the tone stack, it seems to me that C8 is responsible for rolling off the highs, and if I want more, I should change it for a lower value. Thoughts? Yes? No? Right know the amount of highs is short of "pleasantly warm." I have a .01µ, .0047µ, and .0022µ to hand. Which should I try first?
   
   
2. More clean headroom.There's still a bit of dirt when the gain knob gets turned all the way down that appears as the volume knob gets turned up. Since the Volume knob is before the 2nd opamp stage, I've a few thoughts here:
   1. Reduce the gain of the 2nd stage. Not sure how useful that'd be as the 2nd stage's gain is only a bit over 5. With such a low gain, I'm not even sure what the 2nd stage's purpose is, except perhaps to provide the output stage with enough current maybe? I'm guessing. I may want to reduce the 2nd stage gain when I start playing with clipper diodes with higher forward threshhold voltage.
      
   2. Tweak the voltage divider to ground on the 2nd stage's output. Obviously, this will attenuate the signal going into the output stage, but as is, it is shaving so little off that, again I'm not sure why it's even there. If I don't understand a portion of the circuit, I'm loathe to change it.
      
   3. Lower the gain of the output stage. This seems simple enough. Though the resistor values currently in the negative feedback loop are so low, I'm not sure if I can go lower with what I currently have.

Other thoughts: The C1815 transistor going from the positive supply to the op amp's positive rails is an emitter follower, right? What the heck is it doing there? Why is C16 so huge a value and why is it in series with the output stage's inverting input? Is it there to keep the output from frying the input or something? Since it is on the other side of the voltage divider, what effect does C16 have on calculating the gain of the output stage?

[Jazz P Bass]

Mod Mod Mod

Q1. Is it not a constant current source?
TDA2003. Did you read the datasheet? Page 8.
R15 & C16 are there for ripple rejection on the supply voltage.

[Koreth]

If Q1 is constant current source, why is it there? Why would you feed an opamp's positive supply from a constant current source?

Okay, rejection of ripple on your supply voltage is good, but if it is a concern on the negative feedback loop, then why wouldn't all opamps have a huge series capcaitor on the opamp input?

[Koreth]

Hello again. I'm still occasionally making some tweaks to this little SS toy amp, and have a new question.

One thing that sounds very nice in setups like this is to change R5 to about 2K to 10K to taste, then to place a resistor in series with the two diodes, taking the output from the end of R5 as it is now, and with the added resistance in series with the diodes only.

I'm trying to visualize what RG is describing here and made a couple circuit diagrams on what I think he means. Which is the correct implementation of what RG is describing? Either way I look at it, it seems to me both circuits would have the same effect: the input of the clipper diodes is taken from the middle of a voltage divider. Are both circuits going to have the same effect on the amp's sound or is the former going to sound different from the latter?

[Alex R]

I've heard the resistor-in-series-with-diodes version touted as a way of smoothing out diode clipping. I have tried it and in practice I can tell you it works, it sounds less harsh and more 'natural'. I look forward to struggling with the theory...

[J M Fahey]

Hi koreth.
They are *not* the same.
The left one is the right one (resistor in series with diodes)
The right one is resistors in parallel with diodes, a very different beast.

[Koreth]

Do they both still act as voltage dividers into the clipping diodes though?

[R.G.]

I'm trying to visualize what RG is describing here and made a couple circuit diagrams on what I think he means. Which is the correct implementation of what RG is describing?

Neither of them is what I meant. Try this:
Take two resistors. Connect them in series. Call the one you'll connect to the signal source R1. Call the one you'll connect to the diodes R2. Now connect a back-to-back pair of clipping diodes from the unused end of R2 to ground. Hook the unused end of R1 to your signal source.

Now imagine that Both R1 and R2 are 10K variable resistances. We can turn the knobs and set them to any value between 0 and 10K.

Re-capping, from the signal source that will be clipped, the signal goes through the series connection of R1, then R2, then the pair of diodes to ground. We take the signal out at the junction of the two resistors R1 and R2, NOT at the diodes.

First test: Dial R1 to 10K and R2 to zero. How does this sound? It sounds *exactly* like R2 is not there. And it's not. R2 is pretending to be a wire. The signal is simply the original voltage for voltages less than a diode drop.

When either diode starts conducting, current flows in the diode and that's how the diode limits the voltage that appears on the output. The current through the diode is **at all times** either zero for signal voltages too low to turn on a diode, or signal voltage minus a diode voltage divided by the total of all the resistance in series with the diode. And the diode has complete control of the output signal.

Test 2: dial R1 to zero, and R2 to 10K. The signal out is **completely** unaffected by the diodes. No clipping at all (this is a simplification, but a valid one). Effectively, we're listening to the signal source, and the diodes can't do anything to it through that 10K R2.

Test 3. Change R1 to 10K again, but now make R2 be 1K. As the signal voltage increases, there is no clipping up to the voltage which would turn on a diode again. But now, R2 is causing a voltage of (1K/11K)times the signal voltage to be added on top of the diode voltages when they're clipping. With some of the orignal signal there, the diode clipping can't be as harsh.

Test 4: R.G., R.G.! I know! Replace R1 and R2 with a 10K pot!! Yes, that works. But the signal level varies wildly, because the diode signal is only +/-0.7V (about) and the drive signal at the top of the pot may be many volts. So you not only dial out distortion, you dial IN a lot more signal. The two-pots setup helps you tame that. Try it that way if you like.

[R.G.]

Do they both still act as voltage dividers into the clipping diodes though?

No.

[Koreth]

Drew up a new diagram. Here's hoping I got it right this time.