The 32 OD Pure Germanium Overdrive pedal - design and build in words and pictures

[laurie]

I'm probably quite wrong in my thinking, but I sometimes can't help but wonder what the EQ stage would have been like in front of the clipping instead of behind it - mostly with an eye to being able to tighten up the bass just a bit when desired, while preserving the overall gain structure. Sometimes it's just a bit too loose on the bottom end, although most of the time that is also it's appeal.

Edit: my rationale is the the BSD is unique in having an active EQ - so with EQ in front of the clippers, you could decide how much clipping occurs for a given frequency range by driving more or less bass or treble signal into the clippers.

Makes me wonder if the dry blend is actually what I've been looking for

This one is a bit tricky... you can, of course, put the EQ in front of the overdrive stage, but then you would probably also need one after the overdrive because a lot of what you are hearing is generated by the overdrive stage and will be present even if you have EQ in front. An interesting thought, but two lots of EQ makes the pedal a bit unwieldy and expensive.

Perhaps the "add dry" will do what you are wanting... you will be able to try in a couple of months

[laurie]

Now that we are starting to build up the blocks it is informative to look at all the actual blocks:
.

Blocks1.jpg (56.3 KiB) Viewed 4646 times

[laurie]

Deconstructing the voltage amplifier stage...

The bias point of the voltage amplifier must be carefully set. In this stage, the bias point is critical... the voltage at point "C" (pin 3 of the transistors) should be close to 4.5V. This allows for the maximum voltage swing before clipping.
.

Bias1.jpg (42.96 KiB) Viewed 4646 times

.

Bias2.jpg (34.36 KiB) Viewed 4646 times

.

Bias3.jpg (31.36 KiB) Viewed 4646 times

[laurie]

So why is the bias point in the input stage not as important? Because the signal is much smaller (check out the signals on the scope pic earlier).

As long as the bias point for the input stage isn't too close to either the 9V supply or ground, everything will be OK.
.

Bias4.jpg (52.97 KiB) Viewed 4646 times

[laurie]

One way to get a stable-ish bias point is to use two resistors as a voltage divider connected to the base of the transistor. In the voltage amplifier stage, these are R21 and R22. There are other ways that guarantee the bias point, but they are more complex and aren't really needed in a pedal that will be distorting the signal anyway...

R21 and R22 are selected to get the right voltage at point "B" so that the voltage at point "C" is close to 4.5V. The ratio of R21 to R22 is what is important.

The exact values of R21 and R22 are calculated using the gain of the transistors. There are tutorials on the WWW that explain exactly how. See here for an example:
https://www.electronics-tutorials.ws/am ... amp_2.html

This configuration for setting the bias is relatively immune to variations in the characteristics of the transistors, so there is no need to "select" transistors for Q21 and Q22, other than them coming from the 60-to-75 gain bucket and having relatively low leakage. Coming from a bucket of 300 or so transistors rather than being individually hand selected makes life much easier...

So why not use bias resistors like this on the input stage? Because this configuration is much lower impedance than using the single resistor R12.

The input impedance of the voltage amplifier is (very roughly) 75k, mainly dictated by the 86k resistor. Much lower than the 300k of the input stage, so much more load on the guitar if the input bias was done this way.

[laurie]

C21 is needed because the DC bias point of Q11 is different to the DC voltage at point "B" on Q21. Without C21 the DC voltages in the two stages would mess with each other. C21 is an "AC coupling capacitor".

Capacitors have the characteristic of increasing impedance with decreasing frequency. This means that they block bass.

So C21 has to be selected to allow the lowest frequency you want in the pedal to pass through. 1uF (micro-Farad) has an impedance of about 3k at 50Hz. When coupling stages with low or moderate impedance I usually use a 1uF capacitor. It seems to work well and sounds good. Kemet or Panasonic metallized polyester seem to have the best tone.

C22 is there to limit radio frequencies. Common emitter amplifiers (like the one in the voltage amplifier stage) are susceptible to radio frequency interference. Because it is an inverting amplifier, C22 provides a negative feed-back path for high frequencies and this greatly reduces the problem.

C22 has an impedance of over 100k at 15,000Hz so it doesn't mess with your tone too much.

The value of C22 is a balance between stopping radio frequency interference (C22 should be large) and keeping C22 out of the audio frequency range (C22 should be small). 100pF (pico-Farads) is about the right balance. You could use 47pF or even 22pF for C22 and it would probably still be OK.

[laurie]

R23 and the combination of RV24 and C23 set the gain of the amplifier.

The gain of this amplifier is -R23/RV24

The "-" shows that this is an inverting amplifier.

So, ignoring C23 for a moment, the gain of the amplifier is about "5" (4700/1000). 200mV in gives one volt out.

C23 provides a "path to ground" for non-DC signals. If the wiper of the potentiometer RV24 (pin 2) is moved to connect to pin 3 of RV24, then RV24 effectively becomes "0" for non-DC signals (the guitar signal) and the gain of the amplifier becomes, theoretically infinite (4700/0).

It can't be actually infinite of course, but it becomes, essentially, the gain of the transistors. In this case, around 3600 for the Darlington pair. So 200mV in would give about 700V out. But because we have 9V supply, what actually happens is the output clips when it hits 9V.

The mathematics for this is "exponential"... as you creep the wiper of RV24 closer and closer to pin 3, the gain ramps up very quickly. Nearly all the gain happens in the last 10% of movement of the knob.

For this reason a type "C" potentiometer is used for RV24. Type "C" is called reverse-logarithmic. When you apply the reverse logarithmic potentiometer to the exponential mathematics, it becomes more linear.

So with a type "C" potentiometer the change that used to happen in the last 10% of the turn of the knob is now spread out over (perhaps) the last 40% of the turn of the knob. Making things much more controllable.

[laurie]

And now there is a design decision to make.

To ensure the voltage amplifier never clips, you can put a resistor between RV24 and pin 1 of Q22.

The math is easy... with a 9V supply and a 200mV input signal, the maximum gain you want is 9/0.2 or a gain of about 45.

So 45 = 4700/X therefore X is about 100 Ohms.

So putting a 100 Ohm resistor between RV24 and pin 1 of Q22 means the amplifier would never clip.

Or, you could leave the 100 Ohm resistor out so that the last bit of the turn of the Drive knob pushes the amplifier into clipping.

It is a germanium transistor amplifier so the clipping is "nice". So I left it out. I did consider being able to switch it in-and-out (hence that thought of a "boost" switch) but the boost switch would only have full effect if the Drive knob was turned all the way up. It was a bit silly to have a switch that only works when the drive was set to maximum, and a bit expensive to add a switch so I decided not to.

So in the schematic shown, the last 10% or so of the Drive knob adds germanium clipping distortion from the voltage amplifier. Gives a transition from overdrive (generated solely by the overdrive section) to overdrive with distortion (the overdrive generated by the overdrive section, and the distortion generated by the voltage amplifier) in the last 10%.

With the prototype I've found that pushing into overdrive with distortion then adding some dry signal sounds pretty good (to my ear... YMMV).

OK. That's the voltage gain stage explained away. I'll do the overdrive section next.

[laurie]

How is this going so far? A worthwhile exercise? Are things clear enough?

[The_Doc]

It’s a fascinating insight into something that I know very little about and I’m enjoying it a lot

Thanks for all the work you are putting into this Laurie. I can’t wait to have the pedal in my hands after learning how it was designed and built.