Categories
Analysis Boost Effect My Designs

Building a Simple Op Amp Booster – SPICE

I went through schematics for MXR Micro Amp and my Thunder, so now I’ll go through some SPICE analysis for them. I said the circuits are similar and should have pretty similar sound. Let’s see how they look when simulated and if we can learn anything.

Schematics

Here are the schematics in SPICE:

MXR Micro Amp schematic in SPICE
MXR Micro Amp in SPICE
Thunder booster schematic in SPICE
Thunder in SPICE

The schematics use TL072 model which does not come with LTSpice. I got the one I’m using the one for PSpice from TI’s web site. Also, I use a potentiometer library I described in one of my earlier posts.

And here’s a zip file with all the circuits and libraries bundled up:

Frequency Response

In the previous post I said the frequency response is pretty much the same between the pedals. The response is flat for the whole audible spectrum. Here’s how it looks like when I sweep the pot:

Voltage Sources and Commands
MXR Micro Amp
Thunder

If you look at it, the frequency response is pretty similar. From bottom to the top, each graph is a 10% extra pot rotation. While the response is the same, the gain law (how gain changes depending on pot rotation) is different. Let’s look into that next.

Gain Law

So the spacing between the graphs on pot sweep is different between the pedals, so let’s see how does it look like when plotted against “perfect” gain law:

Commands to plot gain law
Commands to plot gain law
Gain law plot for both pedals
Gain law for both pedals side by side (click for larger image)

You can see how much Micro Amp deviates from the ideal response, Thunder is much closer. R4 value determines the shape of the gain law plot and how close to the ideal it will be. To be honest, I had 680 ohms resistor handy and I used it, the book I used for inspiration suggests 620. Anything reasonable (470-1,000 ohms) will work:

Gain law plot from SPICE for some standard resistor values
Gain law with different loading resistor values

In the end of the day, it might or might not matter. I expect Micro Amp will be slowly increasing volume and then at some point, it will just go up very quickly. Thunder would have that transition smoother.

Transient Analysis

Let’s go through some transient analysis. I’ll show just some basic stuff. Both circuits work exactly the same as long as they don’t hit op-amp limits:

SPICE commands for running transient analysis
Keeping simulation within range
SPICE output of transient analysis
SPICE output of transient analysis

Pffff, those colours, I have to figure out a better theme colours. Anyway, sine wave in azure (lower graph) is input signal, green is output signal. Nothing out of ordinary there. The upper graph is just shifted by 4.5V as expected since that’s where Vref is (purple line). Light blue is input to the op-amp (to the non-inverting input) and red is output of the op-amp before going through coupling capacitor.

What happens if I set gain to full, what is my headroom, when does it start clipping?

Headroom

Maximum headroom space we can get out of a 9V supply without some wizardry is … 9Vpp. TL072 I’m using is not a rail-to-rail amp so that is not attainable. Actual headroom is about 6Vpp, roughly. It’s buried here in the data sheet:

Excerpt from TL072 datasheet showing Electrical Characteristics
Excerpt from TL072 datasheet

Maximum peak output is at worst about 3V less than power supply – look at ±15 supply (that’s a dual supply), minimum output is ±12, typical about ±13.5. So, worst case scenario, headroom is 3Vpp, typically 6Vpp (±3V at output). Who knows if this can be translated to single supply like that but let’s go with 6Vpp.

Let’s assume that we have 6Vpp to work with (that’s 3V peak). For MXR Micro Amp, maximum input before clipping is going to be about 3V/21 or about 140mV peak. For Thunder, that is: 3V/15 or about 200mV.

When playing a single note over a single coil there’s about 0.03V (30mV) peak output, but it could go to 200mV for chords when strumming hard. At least initially, after attack, signal decays relatively quickly from there.

All of this means that with single coil pickups (depending on pickups) there should be no clipping and distortion. Humbuckers are different and even single string picking will probably clip some at highest gain.

Clipping

Here are the commands and input sources for testing how output looks like when op-amp starts clipping:

Transient analysis SPICE commands for max gain
Transient analysis commands for max gain

I chose input sinusoid to have amplitude of 0.3V for MXR and 0.4V for Thunder to compensate the difference in gain so I can compare the output better. I don’t want one or the other to clip less and give me different result. Roughly, this would probably correspond to a hot humbucker output with strings being attacked hard (or just playing double stops and chords).

Transient response with clipping. MXR on the left, Thunder on the right

The response looks pretty much the same. The upper waveform is the output of the op-amp, and lower (green) is output (with input signal in azure). If I zoom in and add what would be expected output if there was no clipping we can see how much of the clipping actually occurs:

MXR on the left, Thunder on the right

Not much of a difference but it looks like a lot is being clipped (which is expected, 0.3V * 21 = 6.3V peak, that’s double the available headroom). If I zoom even more, there’s a slight difference in the output coming from the op-amp and after it goes through coupling capacitor for Thunder, for MXR they look exactly the same:

MXR on the left, Thunder on the right

The difference comes to the value of the output capacitor 1uF vs 15uF. Cut-off frequency for the high-pass filters formed by output capacitor and 10K resistor on the output of both pedals is 15Hz and 1Hz … that’s not going to make a difference in sound, but it will in the shape of the output.

Conclusion

The pedals should sound the same. Op-Amps are accused of having pretty harsh distortion when they clip, I’ll definitely try it out.

I would not be looking into waveforms and trying to guess how that sounds in reality. The input is just a simple sine wave, that’s not what guitar produces and sends to the pedal. Also, when clipping occurs, harmonics of different frequencies will be generated, and some of them won’t be in the audible spectrum but will show up in the diagram.

This exercise is useful to see how the circuit behaves, not how it sounds. There are ways to load guitar sound, then run it through LTSpice simulation and hear the output … I’d be cautious drawing any conclusions from that, but I’ll try it out at some stage.

If 3V peak (or 6V peak-to-peak) of headroom is too low for you for some reason, there are things to try that could improve this.

Using a different op-amp might increase the headroom – say OPA2134 would make this 7V peak-to-peak. It is a drop in replacement for TL072 with supposedly better characteristics, but it is 7 times more expensive.

Using 18V supply, if you have a power supply that supports it, would instantly improve the headroom to 15Vpp. The only thing to worry about is:

  • The power rating of resistors. They should be fine, even the one limiting LED current, that one would draw more than double the current but it should still be fine
  • Voltage rating of capacitors. Electrolytic ones, I’ll specify 25V in the bill of materials for their voltage rating so that should be fine. Ceramic ones should be just fine, they are normally specified for higher voltages
  • Op-Amp will be just fine
  • The only remaining thing is potentiometer. Keep in mind they usually have very low power rating, but current through pots should be in order of couple of milliamps max, so that should be all good too

On to the breadboarding part.

Extra Reading

Using LTSpice is easy once you get going, here’s getting started with LTSpice from Sparkfun.

Leave a Reply

Your email address will not be published.

This site uses Akismet to reduce spam. Learn how your comment data is processed.