Super low volume in muff overdrive clone. Unsure of cause.

[Vivek]

To me, the explanation makes sense only when there is a feedback resistor on the second Opamp.

Consider Diode incremental resistance as an almost short...
(that turns second amp into a buffer..) 

Yes

Diode was considered as an almost short when it is clipping

The main point is

When is the diode clipping ?

If there is no feedback resistor, the diode is clipping when that Opamp input signal is more than 0.65/200,000 volts peak

That is very extreme, severe clipping.

There is no difference between two very extreme clippings

Clipping a 100mV signal at 0.65/200,000 V has same frequency analysis as

0.65001/200,000 with diode resistance considered

2.5/200,000 with diode changed to LED (except the volume difference


So I feel, WHEN THERE IS NO FEEDBACK RESISTANCE on 2nd stage ,

Changing gain of first stage from 470 to 47 changes nothing

Changing diodes does not change tone, it only changes volume. If that volume is normalised, there is no change at all

[Vivek]

I'm trying to make Vivek alter his consideration from "no feedback resistance - no NFB" to "short feedback resistance - 100% NFB"..

In general, I love changing my opinion based on new knowledge

But the above was already considered

A) No feedback resistor and diodes not conducting = very very high gain (open loop gain)

B) No feedback resistor and diodes leaking a bit = very high gain


C) No feedback resistor and ideal diodes clipping = signal clipped at Vf diode / gain ie maybe 0.65/200,000 V

D) No feedback resistor and non-ideal diodes clipping = signal clipped maybe at 0.65001/200,000 V ie no real difference than ideal diode


What am I missing out on ?

[Fancy Lime]

@Vivek
You are missing the diode leakage. Real diodes act as their own parallel resistor. Well, actually more like a constant current device with a series resistor but that also reduces the gain, just not in a linear fashion. Using different diodes will give different gains. Don't expect much useful information in the datasheets on that, though.

Hope that helps,
Andy

[Vivek]

You are right !!!

At 0.5V forward bias, a IN4148 appears to be similar to a 5000 ohm resistor

at 2V negative bias, the 1N4148 appears to be similar to 400 Mega Ohm resistor


The theory is right !!!!


The devil is in the details.

So how does that affect our situation when there is no feedback resistor in the loop

I would say a 400 Mega Ohm resistor can safely be ignored in places with 2 Volt signals.

On the other hand, the internal series resistance will create a extremely small difference in the clipping characteristics due to the extremely high gain of that stage.

So both people are right

Those who state that when there is no feedback resistor, the parallel resistors of a diode come into play. Maybe the 400Mega ohm equivalent resistor changes the total gain from 200,000 to 180,000

maybe the series resistor inside the diode changes the output from 0.65V to 0.650001 Volt

and they are theoretically right !!!!!


and those who say that the series and parallel resistors of a diode does not matter when there is no feedback resistor are right too, since the changes between ideal diodes and real diodes in this situation could possible be of the order of 0.00001%


Sure a 400 Meg resistor will make a difference when its in parallel to a 5000 ohm resistor !!

Is it significant ?

[iainpunk]

At 0.5V forward bias, a IN4148 appears to be similar to a 5000 ohm resistor

at 2V negative bias, the 1N4148 appears to be similar to 400 Mega Ohm resistor

i had to check my old school notes and my numbers on the 4148 are a bit different (comments are loosely translated):
0,63V -> below 100 ohm
0,6V FWD -> 3,682 ohm
0,5V FWD -> 18,574 ohm
0,4V FWD -> 25,336 ohm
0V -> more than 1M, less than 2M (bad DMM keeps fluctuating)
-1V -> 9M      \
-2V -> 17M     } constant-ish current sauce
-3V -> 25M    /

they are 2 measurements i did combined in one table, the 0 and negative measurements were done at home as homework, the others are done at school with $800 DMM's, so there is a difference in accuracy

cheers, Iain

[Vivek]

I got my data from

https://2n3904blog.com/1n4148-reverse-bias-leakage-current/

But

17M ohms parallel to 5K is same as 400M parallel to 5K


Hence I still believe that when there is no feedback resistor, the effects of internal resistances of diodes can be ignored.

[amz-fx]

The corrected schematic for the Muff Fuzz is on my page at:

http://www.muzique.com/schem/index.html

Best regards, Jack

[Vivek]

Thank you, Sir Jack !

[iainpunk]

I got my data from

https://2n3904blog.com/1n4148-reverse-bias-leakage-current/

But

17M ohms parallel to 5K is same as 400M parallel to 5K

Hence I still believe that when there is no feedback resistor, the effects of internal resistances of diodes can be ignored.

no they can't be ignored, the pedal was designed with germanium transistors, which have anywhere from 10k to 500k leakage resistance, this sets the pre-clip gain of the 2nd stage lower and makes the specific Ge diodes more important for the sound. and using Shockley diodes doesn't do the trick, at least 1 diode should be Ge so the leakage can set the gain.

cheers, Iain

[antonis]

Hence I still believe that when there is no feedback resistor, the effects of internal resistances of diodes can be ignored.

no they can't be ignored,

Yes, they can be ignored..
(although I stand for the opposite..)

Who's coming next..??

[jdoughty]

Yes I see the earlier comments make sense if there was a feedback resistor on the 2nd IC



But without the feedback resistor,

Changing diodes = no perceptible difference except volume

Changing gain of the first IC from 470 to 47 = no perceptible difference at all

Hahaha, I can say, there is definitely no shortage of clipping in this pedal.

https://m.facebook.com/story.php?story_fbid=10164752297695221&id=903655220 did a post to my Facebook showing it off! It's a rough pedal, but it is fun

I'll inform my child he requires eye liner. I don't think he'll mind 🤣

[Vivek]

Really great idea to get the son involved !

It's wonderful bonding time.

[Vivek]

Are diode models in LTSPICE accurate enough to model leakage currents and effective internal resistances in Diodes ?