Help verifying new design (first one!)

[Haraka]

Hello everyone,

First of all this is a great forum  , well I need your help in verifying the design I tried for the first time...How it all started me and my friend who plays bass were of course fascinated with the Darkglass sounds and most of all with the Microtubes X7 one. Since we could not find a traced schematic with the aid of available info and literature we tried to recreate the X7 setup (without the Direct output section for now) what we have achieved (or we at least hope we achieved  ) is the following:

- the schematic is based on the B7K and Vintage microtubes schematic we found over the Internet
- we separated the signals
- low goes through a active LPF (set for 50 to 500Hz) and compressor, the high through active HPF (set for 100Hz to 1000Hz) in accordance with the descriptions of the X7
-the compressor schematic used is the "Dirt cheap compressor"
-EQ section was made using the parametric eq design similar to elliot sound products set for 500Hz, 1000Hz and 3000Hz something similar to the description of the X7
-Blend control was left out as each signal paths already has volume control implemented

If someone has the will and time to help us out if the design will work or we made a handful of mistakes in the whole mess 

P.s. seeing similar designs for darkglass on this forum some features have definitely been done the wrong way (we suspect) but still wanted to know if the design would work before we start prototyping it 

Thanks in advance!

Here is the segmented schematic:













 

[m4268588]

Doesn't the RV6 affect the EQ operation?

It looks like need to fix the U1B bias.
COMPRE_OUT and U5D output require series resistor.
3rd pin of the RV3 should be connected to VDD.
Is R19 unnecessary?

[Haraka]

Hi, first thanks for your comments! 

I hope you have some patience for my questions 

Regarding RV6 I think you are right, I can change the location it is not an issue (especially since yesterday we found a demo video explaining the X7 path a bit better including that the EQ section is after blend for both signals, link: https://www.youtube.com/watch?v=0o76kW8b_Ns the explanation is on about 2:23 so accordingly we would move it after blend and add additional controls for maybe 80Hz and 120Hz?)

The first question if we move the EQ after the "blend spot" and add 50k or 100k(COMPRE_OUT and U5D output - the value I don't know  ) series resistance and a 470nF cap to ground before the EQ for each branch is that ok?

Sorry I think you are right the RV3 should be connected with 3rd pin do VDD to keep with the other branch (mistake by just importing the schematic ) also consequently R19 was also imported but will be left out....

I don't understand how to fix U1B bias since he is used as an active high pass variable filter?

Additionally, after watching the video the guy said the compressor is a SSL-type compressor with VCA so I think the dirt cheap compressor part will be replaced by something like MBP's cupcake is that ok?

Are those the only issues?

Sorry for lots of the questions, we appreciate the help but we would like to understand the mistakes we have done 

Thanks once again!

[m4268588]

470nF cap to ground before the EQ

It's possible only when higher than frequency of the EQ.

• Freq. = 1/(2*pi*15k*Cap_Value)

Failure to do so will affect the operation of the EQ.
(470nF is too large)

EDIT:
Sorry. My mistake in the drawing.

[m4268588]

It seems to me that "blend spot" is correct for subtraction, not addition, but I can't judge.

[Haraka]

Hi, once again thanks for your comments!

Here is the modified schematic:















Changes:
- EQ section moved after the "blend spot" and expanded to frequencies: 80Hz, 120Hz, 500Hz, 1000Hz and 3000Hz
- U1B bias adjusted in accordance with your comment, should I also bias U2A in the LPF the same way?(c13 to VDD)
- Compressor output and main out changes made in accordance with your comments (added c33, two series resistances - r40 and r41, added r42 - 15k and c34 modified to 1000pF yielding f= ~10000Hz, modified c46 to 470pF and added the last U6D stage).
- R19 will be omitted (my mistake of leaving it there  )

Your comment "It seems to me that "blend spot" is correct for subtraction, not addition, but I can't judge.", I must admit I do not understand what you want to say?

Is the schematic now functional?

Thanks once again!

[m4268588]

- U1B bias adjusted in accordance with your comment, should I also bias U2A in the LPF the same way?(c13 to VDD)

There is no need. Gnd or VDD work the same, but Gnd is a slightly "better choice".

Your comment "It seems to me that "blend spot" is correct for subtraction, not addition, but I can't judge.", I must admit I do not understand what you want to say?

The signal output by the compressor is inverted in phase. "MAIN_OUT" not inverted, so I think subtraction is more appropriate than addition.

U6D is not required, but there's no harm in leaving it as it is.
Do the other forum members have any opinions?

[Haraka]

I really appreciate your help @m4268588!

Whilst looking the schematic I think you are right the compressor introduces a phase invert. It is also conformed in the following topic (the second op-amp stage is just for LED lighting) https://www.freestompboxes.org/viewtopic.php?t=8581&start=20

Should I just add a phase inverter after C18 like in the general schematic using the remaining TL074 op-amp (U2D) I grounded in my schematic (with R1 and R2 set to 10k for gain of -1? and should the non-inverting input be biased to VDD or GND?

Schematic


Thanks once again!

[antonis]

- U1B bias adjusted in accordance with your comment, should I also bias U2A in the LPF the same way?(c13 to VDD)

There is no need. Gnd or VDD work the same, but Gnd is a slightly "better choice".

Do the other forum members have any opinions?

U2A isn't biased via C13..
Its non-inverting input is directly biased from LPF_IN..

Should I just add a phase inverter after C18 like in the general schematic using the remaining TL074 op-amp (U2D) I grounded in my schematic (with R1 and R2 set to 10k for gain of -1? and should the non-inverting input be biased to VDD or GND?

To VDD for a signal symmetrical swing..

[Haraka]

Thanks for the corrections @antonis!

Sorry  , yes it's biased from the stage before (the rush to prototype a theoretically viable design got the better of me    ).

Thanks for the answer will add it with NI to VDD.

Can I ask do you think there are any other issues with the design?

Thanks in advance!

[antonis]

Can I ask do you think there are any other issues with the design?

Only minor ones..

e.g. absence of series resistor right after 9V_IN (to form LPF with C48), absence of 100nF ceramic(disk) decoupling caps from op-amps positive supply (as close as physically possible), use of AC ground instead of GND for some configurations (like R28/C22/RV5, RV6 & RV3..)
(the later is already pointed out by m4268588 for C13 routing..)

[m4268588]

No additional OP-Amp stage is required.

You can also be "addition" circuit like this.

[Haraka]

Thanks you guys for having the patience to help fix my mistakes!

Here is the revised schematic, is the design now in accordance with your comments?
The series resistor is 47R 1/2W, the LPF is about 34Hz (in accordance with the rule that the maximum frequency is 50Hz? And 1/2W just for dissipation reasons. If in the worst case scenario the pedal uses 100mA (I never understood how the current consumption is assumed  ) that yields power dissipation of 0,47W.

Thanks once again!

[antonis]

The series resistor is 47R 1/2W, the LPF is about 34Hz (in accordance with the rule that the maximum frequency is 50Hz? And 1/2W just for dissipation reasons. If in the worst case scenario the pedal uses 100mA (I never understood how the current consumption is assumed  ) that yields power dissipation of 0,47W.

Can't recall on what rule you refer on, but the "rule" for filtering full-wave rectified current is LPF's corner frequency has to be much lower than double the mains frequency (100 or 120 Hz..)

1/2W power rating for 470mW power dissipation flirts with thermal dissaster..
(go to 1W, at least..)

BUT..

100mA through 47R result into 4.7V drop..!!!
(only 4.3V left for op-amps supply - not enough for TL074s..)

P.S.
I'd suggest to measure your build current consumption (either directly or via a 1R resistor) and set R60 value for no more than 1V (better less) drop across it..
Then set C51 value such as RC time constant to be about 0.005..
(32Hz corner frequency)

[Haraka]

Once again @antonis thank you very much! (I cannot overstate this)

I read the rule here https://stompboxelectronics.com/2022/12/02/circuit-6-of-48-the-power-supply-filter/

Ok no issues I will empirically determine the values or the resistor and capacitor (after measuring the current consumption)

Is the rest of the design now viable? Can I start prototyping?   

[antonis]

I read the rule here https://stompboxelectronics.com/2022/12/02/circuit-6-of-48-the-power-supply-filter/

Confusing a bit for noobs..

What is refered as 60-cycle hum is actually a waveform of 120 Hz (or 100Hz for 50Hz mains frequency) with its amplitude margins (headroom) refered as "ripple voltage"..

The bigger the smoothing cap the narrower the amplitude margins (the lower the voltage ripple).
There are formulas for affordable ripple estimation but it's getting out of present thread scope..

Here, we suppose that our 9V_IN come from an already well regulated configuration and we just wish to "eliminate" any remaining ripple with the use of our LPF..

P.S.
As for the rest of the design, cross your finger and go for it..!!

[Haraka]

Thanks for all the help!

When I acquire the part, make the layout and test it I will report the progress!

[m4268588]

[Haraka]

I dont understand @m4268588

Should I modify the schematic in accordance with your pictures?

Thanks once again!

[Haraka]

I cannot see your pictures anymore?