Modding a cheap solid state amp for more clean headroom.

[Koreth]

So, I have this cheap little solid-state amp that was my first guitar amp ever. I've never really liked it. Take the gain knob or volume knob up at all, and the amp quickly goes into a very ratty, nasty distortion. Clean, I don't have much complaints, but getting the amp to stay clean, or at least transitioning more smoothly and gracefully into a dirt that isn't disgusting is the hard part.

Yeah, yeah, I know, get a tube amp. I have a tube amp I'm very happy with, but a 100W Mesa half-stack can't escape being loud. It would be nice to be able to use the cheapie solid state amp to practice while still respecting the neighbor's right to sleep.

So, I opened it up and started tracing out the circuit, hoping to find some things I could change to improve the headroom of the amp. I actually haven't finished tracing the circuit beyond the first gain stage, but I think I've found part of the reason why the amp sounds so nasty when turned up past 2.

I've attached a schematic. LTSpice didn't have a symbol for a for a jack, so I used an AC voltage source instead, since that's what a guitar signal is. I didn't put a ground symbol on the schematic, because as far as I can tell nothing in the circuit is grounded, it's all just tied to the negative terminal of the DC adaptor input.

I found a calculator for opamp gain, and if the math is right, with the gain pot cranked, the gain in this circuit is 370. That's way too much gain. Even a vintage single coil will send the op amp into hard clipping against it's meager 14V power rails.

The solution seems simple enough, change the value of R2 to a high enough one to reduce the gain. Problem is, how do I pick a value short of plenty of trial & error? I'd think I'd want the gain low enough that the opamp never clips, regardless of how hot a signal it's fed, and let the clipping diodes make the dirt. But what's a reasonable upper limit voltage wise to design for? Some pickups are pretty damn hot. I'd suck to have to have things that the amp was usless with a neck single coil but just right for a high output bridge humbucker, or vice versa. The other concern is screwing up the rest of the circuit (since I haven't traced it out, and don't know what the rest of it does). For all I know the rest of the amp circuit may need the first opamp stage running at max voltage all the time to get any output at all.

Am I over thinking things again?

[Steve Conner]

Yes, you are overthinking things. If there's too much gain with the gain pot cranked, then don't turn it up so far!

Turning the gain pot down has the same effect as changing the resistors.

If you're finding that the gain pot has no useful range at all, try increasing R2 by a factor of 10. (2.7k.)

Also try changing the 10uF cap for something smaller (1uf if you didn't change R2, 0.1 if you did.) This will focus the distortion more into the treble and upper mids.

It can be fun to replace the clipping diodes with various kinds of LEDs. You can also put another pair of diodes across the gain pot, then the op-amp can never hit its rails at all.

[Robert M. Martinelli]

Hi,
the gain of the circuit would be 370 only if there were no frequency compensating capacitors around. You have to take into account the different capacitive reactance the two capacitors have at the various frequencies to properly apply the "non-inverting stage" gain formula.

Just to give you an idea, and neglecting the other RC networks:

The 330 pf capacitor in parallel with the feedback resistor limits the gain at the highest frequencies, while the 10 uF capacitor does the same at the lower end of the frequency range.

1st case - Fin = 100 Hz.

Capacitive reactance, Xc, is 1/(2pi*F*C), so at this frequency the 330 pF cap behaves like a 4.8 MOhm resistor, and the equivalent (parallel) feedback resistor can be approximated to 100KOhm.

At the same frequency, the 10 uF capacitor behaves like a 159 Ohm resistor, and the equivalent (series) resistor is : 159+270=429 Ohm.

The gain at 100 Hz therefore is : 1 + 100E3/429=234

2nd case - Fin = 1000 Hz.

The 330 cap in this case acts like a 480 KOhm resistor, so the equivalent feedback resistor is 83 KOhm.

The 10 uF cap, OTOH, behaves like a 16 Ohm resistor, so the equivalent resistor is : 16+270=286 Ohm.

The gain at 1000 Hz therefore is : 1 + 83E3/286 = 291

3rd case - Fin = 10000 Hz.

The 330 pF cap now acts like a 48 KOhm resistor, so the feedback resistor now is 32 KOhm.

The 10 uF cap at 10KHz has a negligible Xc, 1,6 Ohm, so the equivalent resistor is 270+1,6=271.6 Ohm.

The gain at 10 Khz therefore is: 1 + 32E3/271.6 = 119.

4th case, - Fin = 20000 Hz.( Just for the record, my ears don't go this high, neither does a guitar )

The 330 pF cap now acts like a 24 KOhm resistor, so the equivalent feedback resistor is 19 KOhm.

The 10 uF capacitor is practically a short circuit at this frequency, so the equation divider will be 270 Ohm.

The gain will therefore be 1 + 19E3/270 = 71

You can clearly see that at the lower end the 10 uF cap acts as a gain reducer, while the same task at the higher frequencies is accomplished by the feedback cap.

You can play with these caps to "tonally shape" the stage, or try to alter (reduce) the gain, by altering the two resistors ( and the capacitors accordingly, to get the same frequency response ) but the waveform you'll get out of it will always resemble a square wave, due either to the op-amp clipping ( close to the rail voltages ), or to the two diodes on the output, that act as clippers, limiting the p-p voltage and clipping it at their forward voltage (0.6V Si - 0.3 Ge ). The harder the clipping, the higher the odd-order harmonic content, and the harsher the sound you'll get. You might then want to try with different clipping arrangements ( asymmetric clipping, LEDs etc ). Asymmetric clipping is less aggressive ( one half of the waveform being "less clipped" than the other ) and it softens the sound somehow as it tends to introduce even order harmonics.

At this point, I'd ask myself "is this amp really worth the effort?" - If not, I'd take the other ( simpler ) way around - get yourself ( or better yet build it yourself ) a small tube amp, you'll be glad you did it!

Hope this helps

Best regards

Bob

[R.G.]

...
Yeah, yeah, I know, get a tube amp.

Maybe. The issue there is that even a small tube amp is going to be plenty loud. When used in the clean range, SS amps do a good job of reproducing the signal fed to them by distortion pedals, etc. Think about it - the PA at gigs is feeding good sound to the crowd and it's almost certainly solid state.

So, I opened it up and started tracing out the circuit, hoping to find some things I could change to improve the headroom of the amp. I actually haven't finished tracing the circuit beyond the first gain stage, but I think I've found part of the reason why the amp sounds so nasty when turned up past 2.
...
I found a calculator for opamp gain, and if the math is right, with the gain pot cranked, the gain in this circuit is 370. That's way too much gain. Even a vintage single coil will send the op amp into hard clipping against it's meager 14V power rails.

The solution seems simple enough, change the value of R2 to a high enough one to reduce the gain. Problem is, how do I pick a value short of plenty of trial & error? I'd think I'd want the gain low enough that the opamp never clips, regardless of how hot a signal it's fed, and let the clipping diodes make the dirt. But what's a reasonable upper limit voltage wise to design for? Some pickups are pretty damn hot. I'd suck to have to have things that the amp was usless with a neck single coil but just right for a high output bridge humbucker, or vice versa. The other concern is screwing up the rest of the circuit (since I haven't traced it out, and don't know what the rest of it does). For all I know the rest of the amp circuit may need the first opamp stage running at max voltage all the time to get any output at all.

Am I over thinking things again?

Maybe. That stage you've drawn is a not-too-unreasonable imitation of the old MXR Distortion Plus. It's intended to do what it's doing. The designer may have had uninformed intents, though.

Steve and Robert have it right. I would do a somewhat simpler thing, though. I would pull up one end of R2 and solder in a pot of about 10K to 100K in series with it, connected like the gain pot so it's a two-terminal variable resistor. 20K would be the most useful value, but they're harder to find. Then play through it, and diddle the pot until you like it. When you get there, take a meter, measure the pot+R2 value and solder in the nearest 5% resistor value instead of R2. Once you've done that, the overall gain will be within bounds to your ear, and you can start tinkering with gain structure and clipping.

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.

The R2 machinations will get your gain down, the tinkering with cap values and such will tame the rattyness of the distortion to some degree.

On a similar vein, I picked up a "Rogue" brand 100W 2-12 combo from craigslist for $40. The complaint was "it's too loud and too ratty/SS sounding". It's got a very workable 100W SS power amp and OK if not great speakers in it. I pulled out the preamps and subbed in the preamp from an old UK Vox SS amp. Worked just fine. I'm still messing with the SS preamp, the the "soul transplant" part of the operation was a success. You can tame the grack from the amp.

[Koreth]

Yes, you are overthinking things...

Good to know. I do that sometimes. (Okay, maybe a lot.)

If you're finding that the gain pot has no useful range at all, try increasing R2 by a factor of 10. (2.7k.)

I'll give this a shot. It also ocurred to me to check out schematics of other amps and dirt boxes that make similar use of a non-inverting opamp stage for resistor values to try in the loop to play with the gain structure.

Also try changing the 10uF cap for something smaller (1uf if you didn't change R2, 0.1 if you did.) This will focus the distortion more into the treble and upper mids.

the gain of the circuit would be 370 only if there were no frequency compensating capacitors around. You have to take into account the different capacitive reactance the two capacitors have at the various frequencies to properly apply the "non-inverting stage" gain formula.

(lots of useful math)

You can clearly see that at the lower end the 10 uF cap acts as a gain reducer, while the same task at the higher frequencies is accomplished by the feedback cap.

You can play with these caps to "tonally shape" the stage, or try to alter (reduce) the gain, by altering the two resistors ( and the capacitors accordingly, to get the same frequency response )...

Good to know, I'll give that a shot to tune the flavor of the dirt once I've got the the overall headroom vs dirt in this stage to my liking.

...the waveform you'll get out of it will always resemble a square wave, due either to the op-amp clipping ( close to the rail voltages ), or to the two diodes on the output, that act as clippers, limiting the p-p voltage and clipping it at their forward voltage (0.6V Si - 0.3 Ge ). The harder the clipping, the higher the odd-order harmonic content, and the harsher the sound you'll get. You might then want to try with different clipping arrangements ( asymmetric clipping, LEDs etc ). Asymmetric clipping is less aggressive ( one half of the waveform being "less clipped" than the other ) and it softens the sound somehow as it tends to introduce even order harmonics.

It can be fun to replace the clipping diodes with various kinds of LEDs. You can also put another pair of diodes across the gain pot, then the op-amp can never hit its rails at all.

I'd been thinking about playing with the clipping diodes, both the hard clippers at the output and putting soft clippers in the negative feedback loop. I did the asymmetrical clipping thing in my Boss DS-1 with an LED and some 1N00x rectifier diodes and I really liked the result. Said pedal now is great for adding some dirt to my tone, the level of which I can control with my guitar's volume knob. I'll probably look into playing with the clippers once I've got the gain and tone of the stage to my liking.

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.

Are you suggesting some kind of voltage divider to tweak with how hard the hard clipper diodes clip? That sounds like a cool idea.

At this point, I'd ask myself "is this amp really worth the effort?" - If not, I'd take the other ( simpler ) way around - get yourself ( or better yet build it yourself ) a small tube amp, you'll be glad you did it!

Asking myself honestly? No, this little First Act ME104 is not worth the effort. It is a cheap disposable amp that sells at Wal-Mart for $30 bucks. I'm doing this more sh!ts and giggles, and for the fun of experimenting. I never thought I'd become addicted to huffing flux fumes, but I guess I have. I simply figured I'd have more fun tweaking something useless into something more likely to be useful, and even if I don't I'll probably learn something useful in my future music electronics adventures.

I'd love to build myself a little mini tube amp, like a champ or princeton clone of some kind, but that's just not in the cards right now. My main source of income right now part time delivering Chinese food for almost enough money to keep gas in the tank and keep my creditors from taking me to court. Even if I pour as much as $30 dollars of parts into this little amp to try to make it sound better with negligible results, that's more results in the more immediate future than trying to build a tube amp. $30 bucks can get me plenty of parts to tweak this cheapie SS amp if I shop right. $30 bucks won't even cover the cost of tubes, never mind the iron or speakers. I look at it as having fun with what I have and can readily get right now.

[Robert M. Martinelli]

I'm doing this more sh!ts and giggles, and for the fun of experimenting. I never thought I'd become addicted to huffing flux fumes, but I guess I have. I simply figured I'd have more fun tweaking something useless into something more likely to be useful, and even if I don't I'll probably learn something useful in my future music electronics adventures.

Welcome to the "solder fume breathers" club! ( credits ----> Enzo ).

Be careful, as solder fume IS addictive!

All the best

Bob

[Enzo]

Watch out, that stuff will flux you up.

[Gunny]

Don't forget to try this amp connected to a different [bigger] speaker. You may be surprised at how much better the modded circuit sounds when you're not using the stock speaker.

[Koreth]

Don't forget to try this amp connected to a different [bigger] speaker. You may be surprised at how much better the modded circuit sounds when you're not using the stock speaker.

That was actually the very first mod I did to this amp. I had a couple 1x10s left over from a 20w SS amp that died on me, so I grafted some speaker wire and a jack onto the ME104. The 1x10 definately fills out the lows and low mids that ME104 was so severely lacking with that raspy little 4" speaker.

We'll see what the modified circuit sounds like.

[Koreth]

So, I decided it would be a good idea to trace out the rest of the amplifier circuit before making and modifications. My rationale was that I could mod more intelligently, and hopefully avoid any gotchas had from not understanding from how the different parts of the circuit interact with each other.

After spending about an hour tracing what lies beyond C6, and doubling back multiple times to retrace something that didn't make sense at first glance (like a 1M/1M voltage divider from the 2nd stage non inverting input to the 1st stage non-inverting input, WTF?), I have more respect now for the guys who took the time to trace out the now readily available circuit schematics of various FX pedals and tube amps. I'm taking a break for now. I'll have an updated schematic posted later.

An aside, some of the solder joints in this little thing are really sloppy. I'm surprised the amp works with the potential cold joints on the circuit board. Maybe that contributes to the less than ideal sound of the amp? Either way, when I start modding, I'll be reflowing all the bad joints.

[Koreth]

The more I trace this amp's circuit out, the more complex it becomes and the less it makes sense to me.

I've attached an updated schematic. If anyone's good with LTSpice and has some pointers on how to re-draw the schematic so its easier to follow or is willing to redraw it for me, lemme know.

[theWicked]

ME104 was so severely lacking with that raspy little 4" speaker.

I think it's funny that I am hooking a $200 processor up to this thing at it sounds good at living room volume. I struggle trying to suppress my 65w Peavey with a 12" in it.

What do you guys think a higher quality speaker would do? Same size, maybe a 2 or 3 way car stereo speaker. Seems like if it had a $30 speaker the thing would scream just fine. I like this thing because it was free, small, and cheaper than a microcube.

Here is what mine looks like. I sanded off all the silk screened junk by the knobs and scribbled my own (junk) "Orange" style symbols in sharpie.


I have faith in this thing that it will turn into SOMETHING someday. Some kind of Frankenstein amp with all sorts of diy boxes hanging off of it.

[Steve Conner]

The more I trace this amp's circuit out, the more complex it becomes and the less it makes sense to me.

As far as I can tell, R18 and R6 are a potential divider that creates a reference of half the supply voltage. This is a common trick to run op-amps off a single supply, you can find more details at RG Keen's geofex.com site.

So the node at the junction of R18 and R6 doesn't handle any signal, and they play no part in the amp's tone. Also you should find that the bottom ends of R6 and C6 are connected to ground, as are the left-hand ends of C16 and C20. Hence these are all bypass caps and not really part of the signal path either.

[Koreth]

Okay, finally got around to making the suggested mods. I changed R2 from 270 to 2.7K, and C4 from 10µF to 1µF. The result? Much better. The gain knob has a much more usable range now. It can go from totally clean, to a light crunchy breakup.

Though now I think I have the opposite problem from before. There's not *enough* gain. With the gain cranked to ten and using my axes high output bridge pickup, the amp only goes into a light crunch. That by itself is fine, but I prefer the sound of LED hard clippers to that of 1N4148s. LED's have a much higher forward voltage than 1n4148s do, so I'll probably have to up the gain on the opamp to get a get a decent clip out of LEDs if right now the 1n4148s are only lightly clipping. I was thinking red LEDs and changing R2 to either 1K or 1.5K. 1.5K is will be closer to 2x the current gain, and 1k, will be closer to 3x the current gain.

Maybe I'll try both and let you guys know.

[theWicked]

Just so we're on the same page, is this what the guts of yours looks like?