MXR Micro Amp Linear Gain

[Elijah-Baley]

Hello guys.
After I discussed here Let's find a good Booster for the Triple Wreck about a part of a bigger project of mine I need a little technical help.
I have to talk about my intentions. I'm building a 5w solid state amp, the distortion preamp is a Triple Wreck. I found the ideal booster to replace the original booster of the TW: the MXR Micro Amp. The booster will work pre/post of the distortion using a switch.
The Micro Amp works really great with the TW, but I want to fix a little issue. The gain pot works especially in the end of the range and I would a more linear action.

This is the schematic:
.

An interesting analysis http://www.electrosmash.com/mxr-microamp

And a great post by Mark Hammer:

The Micro-Amp uses a single chassis-mounted control: a gain pot.  The gain pot serves to set the volume as well as the amount of gain applied.  The circuit has two resistances on the output, a 470R in series, and a 10k to ground.  With a 56k feedback resistor, and a 500k pot in series with a 2k7 resistor on the ground leg of the op-amp, the gain will range from 1.11 at minimum to 21.74 at maximum.

Because the ground leg resistance is used to set gain, there is a requirement to  step through resistance very quickly, going from max to min resistance, in order to produce appreciable changes in gain.  For example, cutting the pot resistance by half, from 500k down to 250k, only change the gain from 1.11 to 1.22.  That's why the need for a reverse log pot: to get through the first 90% of resistance quickly.

If we were to vary gain via the feedback resistance, things proceed rather differently.

Let's say we make the 2k7 resistor 10k, and get rid of the pot.  Now let's make the 56k feedback resistor a 4k7 fixed resistor in series with a 100k pot, so that the resistance varies from 4k7 to 104k7.  At 4k7 (pot at min), the gain is (4k7+10k)/10k or 1.47.  Now increase the pot to 10k resistance.  The gain becomes (10k+4k7+10k)/10k or 2.47.  Nudge the pot up to 20k resistance, and the gain becomes 3.47.

The lesson?  When the gain of a non-inverting op-amp is set by the feedback resistance, instead of the ground-leg resistance, it doesn't take quite as much change in resistance to have an impact on gain.

So why NOT use feedback resistance?  I'm sure others will point out stuff I never considered, but one of the frequent reasons is that the feedback cap cuts top end as a function of the feedback resistance.  Holding cap value constant, the rolloff moves downward as feedback resistance increases.  It works the other way when using the ground-leg resistance, trimming off the bass end as the gain is increased by reducing the ground-leg resistance.  Ideally, one would like to keep bandwidth constant as gain is changed, but people make their design compromises, based on practical considerations, and go with that.  In the case of the Micro-Amp, bass rolloff is so low at min and max gain ( 0.07hz and 12hz, respectively) that there is little cause for concern.  Dropping the 4u7 cap to 1uf, and leaving everything else in the circuit stock, moves that to 0.3hz and 59hz, respectively, which still leaves all usable low-end for guitars intact.

The stock 56k/50pf pair puts the high-end rolloff just under 57khz.  Increasing the 50pf cap to 100pf would move that to 28.4khz, and 220pf would roll off starting around 12.0khz, which doesn't sacrifice all that much when you consider the  frequency range of most guitar speakers (whose response drops off sharply above 6khz).

What about that 470R/10k pair on the output?  The 10k fixed resistor could actually be turned into a 10k pot, or perhaps a 5k pot in series with 5k1 fixed resistor....after all, what's the point of making a booster silent?  I gather the 470R series resistor is there for a cogent reason, and increasing it beyond 470R by using a pot and sticking more resistance in series, would likely alter the output impedance.  But then, if you're boosting a guitar and using the Micro-Amp to push an amp into overdrive, having more top end is not necessarily in your best interests.

So, putting this all together, it would not at all be unreasonable to replace the 500k reverse-log pot with a 3-position toggle, that made the ground-leg resistance either 41.7k (2k7 + 39k series resistor) for a gain of 2.3x, 9k5 (putting an 8k2 in parallel with the 39k) for a gain of   6.9x, and shunting the 39k resistor for a ground-leg resistance of 2k7 and gain of 21.7x.  Essentially, what that gives you is 3 ranges, each higher one roughly 3x the lower range.  Use the 5k pot+5k1 resistor arrangement noted above, and that will allow you to achieve all of the gain settings in the stock unit, from lowest to highest, but with more diabability and no need for an obscure taper pot to do it with.

Here is the topic: http://www.diystompboxes.com/smfforum/index.php?topic=95120.0

I found this mod, still didn't try:
- To use 100K linear pot
R6 stock: 2.7k; NEW: 4,7k.
R4 stock: 56k; NEW: 100k.
- To encrease the volume
R9 stock 470R; NEW: 100/120R.

But it seems have some problem. It sets the minimum volume higher than bypass volume, maybe reduce the gain, and probably it don't solve the linear issue of the gain control.

I still didn't understand if I can get a more linear action gain pot.

Thanks for helping!

[slashandburn]

Hey Elijah, what taper of pot are you using currently? I'm sure the micro-amp prefers rev. log.

Sorry if I'm stating the obvious. Seems that could be the reason for everything being bunched up at the end of the pot rotation?

[Elijah-Baley]

I'm using a rev log pot as the original schematic. It's a well known "issue" of the Micro Amp. For my purpose I just would like to have a better taper.

[slashandburn]

Ah my apologies. If it's a well documented issue I'm probably well over my head. Maybe there's an arrangement of parallel resistors that will give you a "super-anti-log"? I dunno myself though. Just thinking out loud.

[PRR]

Considering R4 is 56K, pot range from 50K to 500K is only a teeny change of gain. I would try 100K or 50K. Still Reverse-Audio.

If you lack a stock of various rev-audio pots, use normal-audio taper, wire backward, turn backward. Works the same except your hand is confused. But if you get a decent spread of response, get that value in Rev-Audio taper.

[Elijah-Baley]

Thank you. I think we can't do too much without to change response of the pedal: frequency, amount of the dB max and and minimum. My point is to get just a better taper.
The minimum setting doesn't have to low the volume (I think it is impossibile) and doesn't have to raise up the volume, but the gain/volume have to be the same of the unity. And I don't wanna lose any dB at max settings.
Maybe I gonna socket R4, R6 and maybe R9. I should have some log pot to try.

[Mark Hammer]

Another possibility is to just get a 12-position rotary switch and devise your own taper by tapping a series of fixed 1% resistors at different points.  Audiophile preamp designers have been using rotary switches for years, to precision-match channels.  However, where they might need 41 steps to provide left/right channel matching for amplifier volume, a 12-position rotary would give as much adjustment of gain as you would realistically use for the Micro-Amp.  It would take more space than a 12mm pot, yes, but not in any sort of way that would preclude building into a 1590B.

[Elijah-Baley]

Thanks for the advice, Mark , but I think that solution is not good for me, for two resaons. First: I need the Micro Amp to boost a Triple Wreck, which I have to use as preamp in a 5w amp. I'm putting in the enclosure also a clean channel and other stuff. Indeed the front panel has already a lot of pots. I need a regular pot for saving space. Then I found, in my case of course, the rotary switch a bit inconvenient. I prefer just a normal potentiometer to set the gain.

I found an interesting table:

Code Select Expand

R5(kO)   GAIN
0      21.74074074
50      2.062618596
100     1.545277507
150     1.366732155
200     1.27627035
250     1.221606648
300     1.185001652
350     1.158775163
400     1.139061336
450     1.123702231
500     1.111398448

Tell me if those number say something for you, guys.
I hope to try soon, but I'm starting to think that a 50k, but even a 100k, reverse log, could be ok, at least a bit better.

[dschwartz]

Other option is to use a linear pot.
One side of the pot on the feedback and the other at the shunt. You get a quasi logarithmic response

[Mark Hammer]

Other sorts of tapers are also obtainable by making the gain-adjustment pot part of BOTH the feedback loop and ground leg.  Stock, the MA gain pot is simply a variable resistance on the ground leg, with a fixed resistance in the feedback loop.    However, if we connect the pot wiper to the inverting input, such that one leg is part of the feedback path, and the other pot leg is part of the ground leg, you simultaneously add to the ground leg as you subtract from the feedback loop, and vice versa.  That can be used to produce other sorts of tapers and gain ranges.

[dschwartz]

Other sorts of tapers are also obtainable by making the gain-adjustment pot part of BOTH the feedback loop and ground leg.  Stock, the MA gain pot is simply a variable resistance on the ground leg, with a fixed resistance in the feedback loop.    However, if we connect the pot wiper to the inverting input, such that one leg is part of the feedback path, and the other pot leg is part of the ground leg, you simultaneously add to the ground leg as you subtract from the feedback loop, and vice versa.  That can be used to produce other sorts of tapers and gain ranges.

That's what I meant to say. You said it better

[shredgd]

I know this circuit very well because I built three of these, one for each of my pedalboards, and another two for a friend of mine.

You exposed two problems, one being the progression of boost and one being the gain at min. Let's start from the latter.

The problem of the pot not giving unity gain at min is because your 500k-labeled pot is often much less than 500k: I sometimes measure as low as 415k in 500k-labeled pots. This circuit gives you minimum gain when the resistance applied by the pot between the opamp and the 2.7k resistor is at its maximum.
In fact the schematic is best drawn this way:



Increasing the 2.7k resistor isn't a good way to fix this problem: if your pot real value is say 450k, you should increase the 2.7k to about 56k to get closer to unity gain when set at min, but that would give you much less maximum gain. In some of my builds I used a 1M pot to achieve unity gain at min (that's should be the value, if you do the math). However, this accentuates the problem of the progression of the boost even farther (you don't get much boost until past noon with the pot). A way to get a unity gain without buying a new pot, again sacrifiyng some maximum gain though, is to reduce the feedback resistor from 56 to 47k.

Regarding the problem of the progression of boost, you should ideally use a smaller pot. Again, however, using say a 250k rev log pot would mean reducing the feedback resistor to about 22k, sacrificing a lot of maximum gain.

My suggestion is the following: try to decide how much you really need an exact unity gain with the pot at min, and what maximum gain you need. Then experiment basing on the above to reach the best result (or compromise).

Giulio

[Elijah-Baley]

First fo all, thank you for your explanation. I know that the key components are all those we are mentioning. Just a thing: in your schematic, Giulio, the output cap is 1.5uF, it should be 15uF.

I don't want sacrifice the maximum gain, because having a boost less effective than the one I tried, my Micro Amp clone based on the schematic I posted, seems to me I'm losing a good setting of the boost, when it is in front od the distortion.
Actually, I don't have all the components to complete the new circuit in a short time, but now I learnt something more, and I want to try a 100k or 50k pot reverse log hoping I get just an insignificant and small boost at minimum setting. I don't mind of this when the boost is in front, but as I said, but when I put the booster after the distortion. Too much volume even when the boost is set to 0 it is a bit impractical.

I didn't understand what Mark and dschwartz exactly meant.

Other sorts of tapers are also obtainable by making the gain-adjustment pot part of BOTH the feedback loop and ground leg.  Stock, the MA gain pot is simply a variable resistance on the ground leg, with a fixed resistance in the feedback loop.    However, if we connect the pot wiper to the inverting input, such that one leg is part of the feedback path, and the other pot leg is part of the ground leg, you simultaneously add to the ground leg as you subtract from the feedback loop, and vice versa.  That can be used to produce other sorts of tapers and gain ranges.

That's what I meant to say. You said it better

Maybe is most simple than I think, but a schematic redra, in case, could be help me.
Thank you!

[samhay]

I have never been much of a fan of varying the gain of a non-inverting op-amp by varying the resistance of the 'ground leg', so let's take a step back:

As Mark and Daniel have already mentioned (of sorts), there are essentially two ways to vary the gain of an non-inverting op-amp: You can vary the resistance of the feedback resistor and/or the ground leg resistor.
These resistors are involved in low-pass and high-pass RC filters, respectively, so the choice of gain control will have consequences in terms of how the frequency changes with gain.

The MicroAmp varies the resistance of the ground leg between 2k7 and 502k7 (assuming the gain pot is actually 500k). This makes a high-pass filter with the 4u7 capacitor with frequency varying from 13 Hz to 0.07 Hz. Both extremes are essentially flat/passing all audio.
The 56k feedback resistor works with a 47p feedback capacitor to form a low-pass filter with a corner frequency of 61 kHz, which is considerably higher than any human can hear.

The ground leg resistances works with the feedback resistor to set the gain, which varies between 1.1 and ~22.
To get a pleasant sweep in gain, we (well I at least) want the gain to increase exponentially ( 1, 2, 4, 8, etc. ) as was we turn the gain control up. This means we/I want the gain to be ~4.7** when the gain control is at noon/half way, which happens when the gain control resistance is only ~12k.
I very much doubt you will find a 500k potentiometer with such a taper. You could instead use a 100k rev. log pot (as PRR suggested), which will have a centre resistance much closer to this value. The only down-side is that the minimum gain rises to ~1.5.

What happens if we vary the feedback resistor instead?
If there is no additional (e.g. series) resistance, at min gain pot resistance, the gain will be 1.
This is the case regardless of the resistance in the ground leg and is one reason I much prefer this method of varying the gain.
If we want the maximum gain to be similar to that of the MicroAmp (~22), we want the maximum resistance to be about 50k, which gives a gain of 19.5. If you reduce the ground leg resistance from 2k7 yo 2k2, the max gain is now 23.7 - either are close enough that I doubt you can tell the difference.
How about the gain pot sweep? A cheap log pot often has ~20% max resistance when centered.
With a 10k feedback resistance (20% of 50k) and 2k7 ground leg resistor, you get a gain of 4.7 at noon - a much more comfortable sweep.
You don't have to change the feedback capacitor, as at max gain you will have a similar corner frequency. As you turn the gain down, this frequency will rise, but this is probably not going to cause any problems.

So, my 2c. use a 100k rev. log pot in the existing design, or vary the feedback resistance instead.

** gain of 4.7 is about half way along an exponential rise from 1 to 22.

Edit for typos.

[Elijah-Baley]

Very useful explanation, samhay.

I have to necessarily wait the next order (few months) to try something. Unless I decided to make some experiment with my Micro Amp clone pedal, which I have to open anyway one day, because I have to repair it, it is the first pedal I built and now seems have some problem, right with the pot. Turning the pot the boost feded away and come back (not the signal just the boosting). I never find out what is the problem.

Keeping the original schematic the only thing I have to see is about the minimum setting of the pot. If the 50k doesn't boost too much even to 0 it is ok, else I'll try the 100k, maybe won't boost at minimum setting, I hope, but also I hope it doesn't unbalance too much the taper, as it is the original 500k pot.

I read somewhere that the 500k rev log in the MXR Micro Amp it is enough weird that it can be a kind of mistake of the production, hypothetically. The right value was 50k not 500k. Or maybe the company had a lot of 500k rev log to use.
We'll never know.

[Mark Hammer]

The somewhat undesirable tonal consequences of achieving ALL of one's gain via either the feedback path OR ground leg is why I used a combination of both, and the multiplicative properties of gain distributed over two stages in the "Crank" circuit.

That's not a slag at either exclusive use of the ground leg or feedback path - there are times when either of those are the perfect solution - but if you don't want to invite the things that come from asking a lot of gain using ONLY one of those, using a bit of each works nicely.

[Elijah-Baley]

Forgive me Mark, I think I didn't get how I can use a control the gain in the two parts of the circuit you told. I'll try to reduce just the pot without change anything in the design, but I just want to know something more about your theory.

If it can help the discussion I have to say that in the amp - and even for the Micro Amp - the voltage will be at 12v, it should be give me a bit of headroom.

[samhay]

^like this



From here:
http://douglas-self.com/ampins/balanced/balanced.htm

[Elijah-Baley]

Thanks!

I want to talk about something else, now. Not so important, but what about the IC? I used in my pedal version of the Micro Amp the original TL061 without any problem of noise, indeed the Micro Amp is pretty quite and less noiser than my Brian May Treble Booster and the Fetzer Valve. Can I upgrade the Micro Amp replacing the low current TL061 with the low noise TL071? The power supply of the amp where I have to include the Micro Amp is has 2000 mA. It is the TL071 better for the Micro Amp how it could seem?

[Mark Hammer]

The 71 is quieter.  The 61 uses less current/  If you were running it off a battery all the time, a 61 might be a better choice.  BUt for quiet performance, use a 71.