Pearl PE-10 dual parametric EQ, traced schematic + technical questions

[aion]

I picked up a Pearl PE-10 dual parametric EQ and decided to trace it since it's a very interesting circuit that has never been dissected before. Here's the PDF of the schematic with the CMOS switching removed.

https://dl.dropboxusercontent.com/u/10003621/pe10_sch.pdf
(asterisks by some of the resistor values are ones I need to pull out to measure... hard to tell the difference between blue, green and purple bands)

Couple of questions.

1. Can the "Level" pots be replaced with 20kW? (e.g. TS tone control) 50k's aren't available to DIYers but it's nice to have the "cut on one side, boost on the other, flat in the center" operation.

2. Is that first stage a BJT differential amplifier (pseudo-discrete op amp) just set to unity gain? It looks like what you would see in a Boss BD-2 but a lot simpler. It was weird to see two transistors with all three pins in parallel.

3. What's the story on the polarity of C9 and C13? This is definitely how they are oriented, but I would expect the positive side to face the emitter. I'd use a non-polar cap here anyway if I was DIY-ing one, but I was curious.

4. Same question with C3 - again, I would have expected it to face the other way.

5. Is there any way I could draw the schematic in a more straightforward manner? I traced it myself and tried to make it as neat as possible, but sometimes if you rearrange things here and there you start to see similarities with other circuits.

[PRR]

> Is that first stage a BJT differential

Not differential. Q1 Q2 are pure parallel, not diff. Not sure why they did it that way. In low-low-Z work, parallel devices give low hiss voltage, but not in guitar circuits. All-n-all I wonder if this stage started as something else and just kinda morphed.

> Can the "Level" pots be replaced with 20kW?

Without thinking in this heat, most likely you would want to change other resistors in the same ratio.

C9 and C13: the pot side sits at VA. The transistor side sits at VA minus Vbe and a tenth or so. The polarity drawn is correct.

C3 I am less sure of. Right side is again near VA. Left side is the idle-point of Q1 Q2 Q3 stage, which has its own bias network, and quite high impedances.

If the thing can be powered-up, and if you want to avoid thinking-- measure the DCV on both sides of the cap.

I wonder if the "bypass" bleed resistor really returns to VA?

> a more straightforward manner?

RANE has a paper with most of the graphic/parametric EQ topologies drawn and described. This is one of the "standards" tho I am not seeing which one. Perhaps band-pass into boost/cut network, but there seem to be other loops.

[snap]

looks state variable to me, just strangely drawn.

[Rob Strand]

To add to what PRR said,

- The eq will work fine with 20k pots;  it changes the behaviour of the ckt by a small amount.

- Yes, you could adjust the components to compensate for the pot change.
   You would have to increase all resistors around the two opamps IC1A and IC1A
    by a factor (50/20):  R5, R6, R7, R20, R31, R32 (and R8 and R30).
    and decrease the capacitors C4, C14, C15 by a factor (20/50).

The circuit is just another State-Variable filter form.  The main advantage of this form is it uses a single-ganged Q pot.    It's a fairly text-book design except for the fact R20 and R32 are not equal; which is OK - they are using both pre-emphasis + increase signal by 2 then de-emphasis + decreasing the signal by x1/2 to improve the SNR.

The Rane example structure is:

http://www.rane.com/n122fig23.gif

(with the exception of the centre-tapped pots)

[aion]

Very interesting, thanks so much for helping me make sense of this. I redrew the schematic (same link as in the first post - it's been updated) to make it match a little better with RANE's state-variable parametric EQ example (http://www.rane.com/n122fig24.gif) as it's pretty near identical once things are rearranged a bit.

Thanks for pointing out the boost/cut part as well. I can definitely see the two distinct circuit topologies that were combined. What's the difference between RANE's circuit where the wiper also goes to ground (or VREF) and this one where it doesn't?

Not differential. Q1 Q2 are pure parallel, not diff. Not sure why they did it that way. In low-low-Z work, parallel devices give low hiss voltage, but not in guitar circuits. All-n-all I wonder if this stage started as something else and just kinda morphed.

Having dissected a few other Pearl pedals from this series, I can say this wouldn't be the first time. The pre-gain stage in the OD-05 is strange as well and unnecessarily complex. I think the engineers had a knack for copying things that weren't intended for guitar and making really interesting guitar effects out of them (they seemed to be fans of Moog in particular), but some of the particulars are a little head-scratching at times. I get the feeling they sometimes just stopped at "hey, it works!" and didn't go through the exercise of "how many parts could I remove to cut down on production costs?" like BOSS or Ibanez might.

One more question: What's Q3 bringing to the table? Since this isn't a differential, and Q1/Q2 are redundant, is it really helping anything along?

I wonder if the "bypass" bleed resistor really returns to VA?

It definitely connects to VA, but I didn't really note very well that the "bypass" part then goes to a CD4007 switching circuit that's off the main board, rather than straight to the output jack by way of a hardware switch. I just took this bit out of my revised drawing since it's vestigial in a true-bypass arrangement.

[aion]

Here's the finalized version of the PE-10 derivative:

https://dl.dropboxusercontent.com/u/10003621/parametric_eq_sch.pdf

For layout purposes, I dropped one band from the circuit. It can easily be added back in, but I was going to have a hard time fitting six 16mm knobs in a 1590B/125B enclosure, so I decided to take a bit more 'modular' approach so a person could include two or three of them in series in a larger enclosure, and could even do some really interesting routing if they were so inclined (e.g. an FX loop box with pre and post EQ).

I also incorporated the Level control from the GE-7 (see this thread for more info) to counteract any perceived volume gain or loss from messing with the levels. This way it can also act as a booster, especially in front of a drive pedal. I elected to add it to the end of the circuit rather than the beginning (as in the GE-7) since the signal is doubled and later halved for noise control purposes, and I didn't want to risk overloading the EQ section of the circuit.

Haven't built it yet, but I wanted to share where I ended up. If anyone has any comments I'm all ears!