Understanding EHX Mole schematic

[il_mix]

Hi, everyone!

I'm planning to build a EHX Mole clone. Found some schematics. Quite simple and useful, too.
Looking at the schematic, it is similar to the EHX LPB-1, and to the DIYStompboxes beginner project, too. Basically, it is a linear booster that works on some frequencies.
I've read something about linear booster here
http://www.beavisaudio.com/techpages/SchematicToReality/
So, here is a schematic (I made it with Fritzing, based on schematics on the web).


C1 is the input capacitor, that cut input DC and some frequencies.
R1/R2 are the bias resistors for Q1 transistor.
FIRST QUESTION: no input resistor here? Like the usual 10+MOhm impedances?
R4/R3 define the signal gain factor.
Than there is the output capacitor C2 and the volume control pot R5.
SECOND QUESTION: where is defined which frequencies will be boosted? I suppose C3/C4 are doing some sort of filtering here, but I'm not able to figure out how these capacitors really works.

I will make some test playing with componets on breadboard, but I prefer to get a good understanding of the circuit, first.

Thanks to eveyrone!
MIX

[PRR]

> C1 is the input capacitor, that cut input DC and some frequencies.

WHAT frequencies?

Do you know how to calculate capacitor impedance? Or to calculate at what frequency a cap has a certain impedance?

What is C1 working against? To simplify, R1 and R2 in parallel. 43K||430K is 39K.

At what frequency does C1 3.3uFd equal 39K impedance? On my abacus it looks like 1.2Hz. So 1.2Hz tones will be 3dB down, lower tones even more down, and 2.4Hz is about 1dB down. This network is "flat" for ALL audio frequencies.

Jumping ahead: we find a same 3.3uFd at the output, and an even higher-value resistor (pot). At high pot settings the load will also parallel the pot, but in guitar-cord work we can "assume" load will be 100K or more. Say 50K total. If 3.3u+39K was below the bottom of the audio band, then 3.3u+50K will be even a bit lower. So for practical purpose, there is no loss of the musical bass.

> no input resistor here? Like the usual 10+MOhm impedances?

"Optional". Not needed for basic function. Usually very nice if you will switch the circuit in and out live.

> R4/R3 define the signal gain factor.

But what is C4 doing?

Remember that caps are open for DC, short for infinite frequency, and in-between for in-between freqs. So we guess that C4 does little for bass, shorts-out treble, and want to know where it just-starts to work. I'll let you do the math... the answer looks odd.

[il_mix]

Hi Paul,

thanks for the reply!

About C1 (input capacitor), it is mainly (only) to cut DC, as expected. When I said "some frequencies" I was talking about very low frequencies, but it is interesting how you got that numbers from math.

About C2 (output capacitor), I supposed it was doing something more than just cut DC, but it seems that again it is just used to cut DC.

Out of curiosity, how are that kind of capacitors called? (capacitors used to cut DC from signal) I thought about "decoupling capacitors", but I don't think it's the right term ("decoupling" is about noise reduction, right?).

About the input resistor
> "Optional". Not needed for basic function. Usually very nice if you will switch the circuit in and out live.
Uhm... ok for the "optional" part (I will search for some reading about it). What about the "switch the circuit in and out live"? Are you talking about switching the stompbox ON/OFF? Isn't it what the stompbox is made for?!?

About C3/C4, I'll do the math. I suppose I have to review small signals circuits analysis, am I right? I think I've deleted that bank of memory...

[nocentelli]

Input resistor to ground helps prevent switch pop from true bypass switching. C3 and C4 will bleed off high frequencies to ground and +9v respectively, which is why the mole was sold as a bass booster: The input signal has all high end removed, the muffled signal is boosted massively, then C4 bleeds off any remaining high end. What is peculiar is that the 100n value is very high, and (without doing the calculations) will roll off signal right down to to quite low frequencies.

[il_mix]

> The input signal has all high end removed, the muffled signal is boosted massively, then C4 bleeds off any remaining high end.

...and here comes the other question.
What I expect as output is the input signal (almost) as is with the bass frequencies boosted.
What you're telling me, and what I thought about the pedal looking at the schematics, is that the effect is cutting high frequencies and boosting the low one, so in output I get boosted basses and (highly) dumped down high frequencies. Not cool...
Am I missing something?

Maybe I have to think about it in a different way.
I attenuate the high frequencies by, e.g., -6dB with the fixed RC filters (I didn't made any math yet. It's just a random number), while the low frequencies remains untouched (0dB gain).
Than, playing with the pot, I set the transistor gain to 6dB. The result is that the output signal have the high frequency with 0dB gain (same volume of input signal), and the low frequencies are amplified by 6dB. So, I actually have a bass booster.
Did I get it?

The problem here is that, using again the +-6dB of the previous example, I can't get, e.g., a +3dB gain on the low frequency AND 0dB on the high frequencies. Right?

NOTE: I've never tried/heard this pedal, so I'm just making assumption given the pedal "surname" (bass booster) and info read on various posts, where I've read that the high frequencies are output "as is".

[nocentelli]

It may well work as you describe (i.e. treble frequencies are brought back up by the boost) but if you want a controllable bass boost with flat treble, a more nuanced design might be needed. You could use a dual opamp, use the first to buffer the input signal and the second as a baxandall bass/treble control, where the  two frequency bands can be adjusted without interaction.

[il_mix]

That's what I was expecting from the schematic. Input signal split in two paths, one will be filtered (low pass) and that boosted, one goes clean (maybe, as you said, buffering it first) to the output stage. A sort of dry/wet control (simple pot?) will mix the two signals together.
This way we are adding to the clean signal, that still have the low frequencies, the boosted low frequencies. Is it possible that there is a (noticeable) out of phase effect by adding these two signals that go through different paths?

[nocentelli]

I'm not sure why you were expecting that. Ehx has always been about cool but very Cheap-to-produce fx. The Lpb1 gave a massive gain boost with a single transistor, and a handful of caps and resistors. Within a few years, they were knocking out many slight variations as treble boosts, bass boosts etc all using virtually the same parts in the same cases with different labels.

I've not seen many bass boosts or eqs with a separate dry path: with a bit of design, it's perfectly possible to have a control that only affects a particular frequency band and leaves the rest unaffected (e.g. Baxandall): This is effectively the same as having a dry path for the unaffected signal range, and doesn't require buffering the separate path, or components to mix the wet+dry back together again.

[il_mix]

Back on the topic!
Assuming that what came out from our "analysis" is right (high frequencies cut out, signal boosted (all frequencies), output volume adjust the overall gain), than we can cut the schematic in 3 stages:
- input stage (from input to transistor)
- boost stage (transistor, R3 and maybe R4)
- output stage (from transistor collector to output, maybe with R4)
Input and output stages are nothing more than band-pass filters. High-pass for removing DC just after input/before output, low-pass before and after the transistor (to have a low frequency boost).
I drew the schematic for small-signal analysis, and in some ways it represents what I've said.
Before posting everything, since I'm not sure about my schematic, I post just the input stage



We have the input RC high-pass filter, that cuts mainly DC, as seen before.
After this, before the transistor, we have (?) a low pass filter. We actually have the parallel capacitor, but where is the series resistor? I thought about the transistor input impedance, but I'm not sure it is right...

Output stage soon.

Thanks!
MIX

[duck_arse]

you can think of C1 and C3 as a voltage divider for ac signals. as the freq increases, the "resistance" of C3 goes down, taking more highs to ground. as the freq decreases, C3 resists more, so you get lows louder than highs.

[PRR]

> we have (?) a low pass filter. We actually have the parallel capacitor, but where is the series resistor? I thought about the transistor input impedance, but I'm not sure it is right...

Not the transistor. We are looking for a *series* resistance.

There is NO series resistance *shown* on that plan. (There is C1, but a C against a C does not discriminate by frequency.)

If we drive this with a Dead-Zero source impedance, there is no filter action. An infinitely strong source will flow whatever current C3 asks, even to infinite current at infinite frequency.

1) There are NO dead-zero impedance signal sources. Every electric source has some impedance.

2) We may not know what that source impedance is. It may turn out to be wise to pad-out the path to C3 with some known resistance.

But we may know some things generally. Guitar pickup is 5K of copper resistance. At higher frequencies it acts as an inductor, but the essential cable acts as a capacitor, so a guitar gets complicated. However it is usually moderated by the onboard volume pot, but that effect varies with setting. We can generalize that it is over 5K and under 500K, probably covering some large part of that range as frequency varies from 500Hz to 5KHz.

Everybody has boxes in a chain. The output of an effects box could be a poor 10-Ohm source (chip opamp with no damping resistance) or 100K-250K (many tonestacks or output pots).

On The Other Hand, we can look at the impedance of C3. I seem to recall that 0.01u (10n) is 1k at the top of the audio band (16KHz), so 10X more F is 1/10th the impedance, 100 Ohms. And if a cap is 100 Ohms at 16KHz, it will be 1K at 1.6KHz, 10K at 160Hz.

Right away we seem to have a problem. The 1K-10K impedance of 100nFd across the guitar band is very-low compared to the 5K-500K of guitar and most guitar-cord stuff.

Another way to look at it: Guitar cord cable has capacitance which cuts highs. Generally 30pFd per foot, or 300pFd in 10 feet. We usually like less than 10 feet of cord for best tone. (If more, we may turn to a buffer.)

100nFd is 100,000 pFd or like a Three Thousand Foot guitar cord!

That seem excessive, even absurd.

Is "100nFd" really right? Or a transcription error?

If this is a "Bass Boost", then yeah--- it is clobbering treble HARD, then boosting what is left (just bass). But how much treble is clobbered? Can not say until we know the source impedance. Which in real life could be low or high. So results will vary depending what is before it in the guitar-chain.

[duck_arse]

you can think of C1 and C3 as a voltage divider for ac signals. as the freq increases, the "resistance" of C3 goes down, taking more highs to ground. as the freq decreases, C3 resists more, so you get lows louder than highs.

see above.

[il_mix]

Thanks again Paul!
Really interesting post.

There is some thing that are still not clear...
If I got it right, you're telling me that the series resistor missing for the RC low pass filter is the source impedance (i.e. the guitar impedance). In this case, looking at the input stage schematic I've posted, this resistor will be placed between the source and C1. I was expecting something connected to the low pass section node (the one above C3) (that's why I thought about the transistor input impedance (impedance of the transistor base)). Am I missing something?

About the capacitor(s) value (100nF), searching for "ehx mole schematics" in Google Images gives several schematics/perfboards/PCBs, all using 0.1uF for the filtering capacitors.

While trying to well understand the circuit, I will make some "audio" test (hear what the circuit does) and some oscilloscope analysis (just heard about xoscope. Amazing!!!)