A successful Superfuzz experiment

[Mark Hammer]

Energized by a recent build where I stuffed 6 different "classic" 2-knob distortion/drives into a single enclosure with a rotary switch to pick one, I decided to take on a trio of classic 3-knobbers in a single box.  One of them is a Superfuzz, using this schematic:

The perfboard I'm using doesn't have a lot of room, so I've had to resort to various space-saving tricks, like vertical installation of 1/8W resistors and such.  To save even more room, I thought I would try using a 2SC1538 matched NPN-pair for the "mirror" you see in the middle of the circuit.  Worked like a charm!  Sounds great, and provides decent octaving.  I also used Dino/digi2T's trick of inserting a small value resistor (220R in this instance) in series with the diode pair to soften the clipping, while still retaining the Q&D limiting they provide.  Also worked like a charm.

But but but the Superfuzz is NOT a "3-knobber", you say.  Correct.  I subbed a variable resistance to ground from the 100nf cap in the midscoop filter, like ZVex does with the Octane (Madbean Poindexter).  Works as well as the Tone switch, and means you can forego the Tone switch and the 47k/10k divider.

[Mark Hammer]

Actually, let me add to what I wrote.

Holy crap, the octaving is fantastic!  I can get very robust octave-up below the 5th fret, something that was well out of reach in previous Superfuzz builds.  This is at least as robust an octave-up unit as the Foxx, if not better.  I'm impressed.

As a sidenote, the remainder of the transistors are plain vanilla 2SC828,

[rankot]

Is HFE important in this case or not?

[Mark Hammer]

No.  It's not amplifying the signal in any discernible way, merely combining two opposite-phase copies of the signal.  The reason why I thought of using the 2SC1583 (and 1538 was a typo...sorry) is that the two  transistors in that 5-pin package are very precisely matched.  And when it comes to octave-up generation, the two complementary copies of the input need to be very closely matched to optimize that.  Normally, I would simply check the values of the emitter/collector resistors on the phase splitter, and resistors on the emitter and collector outputs.  But since I already used 1% components there, I decided to try something that would go a step past that level of matching.

Those who have followed posted info and drawings of the Superfuzz will likely be familiar with drawings like this one, that include a trimmer for adjusting the balance between the biasing of the two mirror transistors.

I've built that and still got nowhere near the robust and reliable octaving I get with use of a matched pair of transistors.  I should emphasize that it does not have to be a 2SC1583.  There are a number of available transistor-array packages that could probably work well, like the CA/LM3046 (regularly used in analog synths) or LM3086.  Both of those chips have 5 discrete transistors, including a pair with a common emitter.  With the addition of another discrete transistor, I'll bet they could make pretty decent octave-up units.

[digi2t]

 Hmmmm.... this calls for an investigative visit. Perhaps I'll distract you with a Bi Phase while I noodle with your SF.

[Rob Strand]

Very cool Mark.   I've been going through a lot of old era devices lately because I've been separating out the schematics for that era from the later ones.

I wonder if adding a 8.2k to 10k in series with the bottom 470R would further help matching.  I suspect it might because as there's a natural imbalance caused the output impedance of the phase-splitter working against the bias dividers of the "diffamp".

[Rob Strand]

I wonder if adding a 8.2k to 10k in series with the bottom 470R would further help matching. 

Not so simple.  With some signal levels it seems to make things worse!

[Mark Hammer]

Apparently the virtues of using a matched pair for the doubling has come up here before: https://www.diystompboxes.com/smfforum/index.php?topic=110775.0

...although the thread seemed to meander to other, less electronic, connotations of matched pair. 

[Rob Strand]

Apparently the virtues of using a matched pair for the doubling has come up here before

I actually went through exactly what is going with this pedal the other day but I didn't post anything because it was a little long winded and techy to explain.

When we match transistors it can mean different things because they have more than one parameter.  "Matched" transistors often matches Vbe for a given current (as well as gain to a degree).   As it turns out Vbe only has a small effect for the Super Fuzz.  What is very important is to match the hFE.

The reason can be explained as follows:

From the perspective of the phase-splitter we want the signals at the collector and emitter to be equal magnitudes.   To simplify the argument we will ignore the effect of the transistor's output impedance at the emitter (re) and the fact the collector current doesn't quite equal the emitter current.  Under these assumptions:  The gain for signal at the emitter of the phase-splitter is 1 So for matching we need the signal at the collector to have gain 1. And for that we need the load on the collector to equal the load at the emitter.

Now, the load on the emitter is:   the 10k emitter resistor, the 22k bias resistor, the 100k bias resistor, and (most importantly) the input impedance of the rectifier transistor all in parallel.  The load on the collector is
the 10k emitter resistor, the 22k bias resistor, the 100k bias resistor and the input impedance of the *other* rectifier transistor all in parallel.   We could throw in the effect of the 470R's we well.

The key point here is the input impedance of the rectifier transistors need to match.   That impedance is actually quite low compared to all the other resistances making up the load impedances.     The input impedance of a transistor stage is directly proportional to hFE (actually hfe+1).   So if the rectifier transistor gains don't match the input impedances don't match and that stuffs up the balancing of the signals on the collector and emitter.

With this in mind the, mod of adding the resistor in series with the emitter output of the phase splitter obviously doesn't help matching the collector and emitter *voltages*; which is exactly what I saw.

Another issue with unmatched hFE on the rectifier transistors is it makes each rectifier transistor operate at a different current and that makes gain for each arm of the *rectifier* unmatched.   This is much less of an influence than the input impedance issue.  Typical the side with the high gain transistor will have a higher rectifier gain and higher voltage gain so it make both mismatching mechanisms on that arm worse.

Some solutions were,

To combat the impedance imbalance issue: 
- Add buffers to each output of the phase splitter, or
- Add a buffer to the input of the rectifiers, or
- Make the rectifier darlingtons, or
- Add a trim pot to match the input impedance
- (less effective) reduce the resistance of the base bias networks for the rectifiers, and the lower phase splitter resistors.

To combat the bias point imbalance (which affects the rectifier signal gains).   These I didn't fully explore.
- Use a current mirror in the emitters but bypass the emitters with caps!
- Possibly use separate emitter networks for each rectifier to stop hogging of the bias current.
- Use lower base-bias resistors.
- Use a trim pot to balance the collector current.

Ideally I wanted to come up with a scheme so that when you adjust the collector currents to match it automatically tweaked the input impedance as well.

Feedback bias might work but it cannot have AC feedback, only DC feedback.  The AC feedback linearizes the signal and we lose the octave!!!

[zbt]

I wonder why VC = 6V and VE = 3V,  not VC = 7V and VE = 2V
according to https://www.electronics-tutorials.ws/amplifier/phase-splitter.html

is phase-splitter it self intend to be distort, compare to Acetone FM2 with 100K resistor?

[antonis]

It's a bit more complicated 'cause, for AC load line, Emitter/Collector equivalent resistors are 10k // [470R + (22k // 100k)].. 

[zbt]

try to draw it, correct me if I'm wrong



if I m not mistaken, if the 150K replace with 100K then

VB = 100/(220+100)*9 = 2.81
VE = VB-0.6 = 2.2 --> 2V
VC = 9V-VRC = 9V-(VE/RE*RC) = 9V-VE = 6.8 --> 7V


        FY6     Uni     Super   Ibanez  Aceton  Kay Fuzz   
Q3
RA      220     220     220     510     220     47      K
RB      150     150     150             100             K
RC      10      10      10      10      10      10      K
RE      10      10      10      10      10      10      K
Rin     0.47    0.47    0.47    1       1       0       K

Q4/5
RA      100     100     100     150     100     100     K
RB      22      22      27      22      15      10      K
RC      10      10      10      10      10      10      K
RE      2.2     1.8     1.8     1       1.8     1.8     K

Q3 VBE (0.6)                     
VB      3.65    3.65    3.65            2.81            V
VE      3.05    3.05    3.05    3.97    2.21    -0.6    V
IE=IC   0.3     0.3     0.3             0.22            mA
VRC     3.05    3.05    3.05            2.21            V
VC      5.95    5.95    5.95    5       6.79    9       V
Q3 ref  3 – 6   3 – 6   3 – 6   4 – 5   2 – 7      
VB                      3.27                    1.85   
VE                      2.72    3.97            1.84   
VC                      6.23    5               2.54   
                     
Q4/5                     
VB      1.62    1.62    1.91    1.15    1.17    0.82    V
VE      1.02    1.02    1.31    0.55    0.57    0.22    V
IC      0.46    0.57    0.73    0.55    0.32    0.12    mA
VRC     4.65    5.68    7.3     5.51    3.19    1.21    V
VC      4.35    3.32    1.7     3.49    5.81    7.79    V
        4.5     3                       6               V


assuming for 2 - 7V, we can get good octave at louder signal

[amptramp]

There might be a simple way to get matched transistors you can use for a push-pull frequency doubler.

Use an LM318 op amp and connect the bal/comp pins (1 and 5 on the 8-pin device) and use this junction as the collector.  It does limit the input voltage between the inverting and non-inverting inputs with diodes but if you keep the level below where the diodes conduct, they have no effect.  Of course, if the diodes conduct, they give a fuzz effect, so that may be what you want anyway.  You don't use the output of the op amp since it would follow the difference in collector voltages, but we have shorted them together.

Since the transistors in an op amp are formed at the same time from the same process, they tend to match very well for gain and Vbe.

[zbt]

chips too expensive

what if for Q4/5, 100K replace with 56K and 22K with 12K

[antonis]

It will lower Base bias, hence Emitter voltage by about 35mV..
So..?? Are we looking for a stiffer voltage divider configuration..??

[zbt]

In my current understanding impedance imbalance mean zout <> zin
zout Q3 = 10K, so I try to make it same by (12*56)/(12+56)=9.88K

then I realize not include rpi (I learn from electrosmash) which
"(most importantly) the input impedance of the rectifier transistor"

my mistake, so I calculate with good ref VE=1.11 VC=3.1 VCC=8.88
IC=(VCC-VC)/10000=0,000578
IE=VE/1800=0,000616667
to make it simple I use 0.6 mA
rpi = 25/0.6 = 41,66


hfe  zin
100   4166   
150   6249   
200   8332
250  10415      
300  12498


then I go to "- Add a trim pot to match the input impedance"
what ever hfe we can adjust to 10K, by triming IE

zin=hfe*rpi
zin/hfe=25/IE
hfe/zin=IE/25
IE=hfe*25/10000=hfe*0,0025


hfe  IE (mA) RE(VE=1.11)  VC(VCC=8.88)
100  0,25    4440         6,38
150  0,375   2960         5,13
200  0,5     2220         3,88 ***
250  0,625   1776         2,63 *** 3.1
300  0,75    1480         1,38


If 3.1V so good wouldn't be hfe between 200 - 250, good for impedance?
assuming Q4/Q5 also low I guess is ok if zQ3 <> zQ4/5, (10K <> 4K4)
but maybe is even better if match

Replace RE 1K8 with 5K trimpot, so whatever hfe we can adjust impedance
and what ever VC also, learn from master Paul when learn bigmuffpi
so I feel ok, beside the range is 1.7 (superfuzz with trimpot 10K) - 7.79V (Kay Fuzz)

for hfe what I learn is, in circuit can be 100 - 700 range,
but at higher hfe I fill it more sensitive, so I tend to choose lower.

for
"The key point here is the input impedance of the rectifier transistors need to match."

been overcome by Sir Digi2t and Sir Mark,

I choose 1% resistor 99% luck

[Ben N]

Would a AS394 work, or is the hfe too high?
https://cabintechglobal.com/pdf/ALFA_RPAR_AS194_AS394.pdf
(They look pretty cool in their nifty 8-pin metal can, even if they are a bit pricey.)

[zbt]

I could not say about match pair transistor

but on solution
"- Make the rectifier darlingtons, or"

[puretube]

" ... or: (J)FETs"

[amptramp]

You could always use a precision fullwave rectifier with the same effect as the push-pull amplifier used in the circuit.  It requires a dual op amp but this should not be a hardship since these items are cheap.



In this schematic, the 4.7K resistor should be 5K which you can create by using two 10K resistors in parallel.  There are several similar op amp fullwave rectifier circuits but this will give balance as good as resistor matching allows.  In this diagram, ignore the ± supplies and assume the grounds go to Vcc/2.  Most op amps should work in this circuit.