Debugging fuzz face squeal

Started by billings, June 26, 2006, 07:49:25 PM

Previous topic - Next topic

billings

I'm building a fuzz face clone.  This is my second one for a friend; the first one worked fine once I debugged it and got the sound coming out right, but no such luck with this one.  I've made a lot of changes to how I wired it up, though, and added some stuff.

This is pretty much the circuit I'm using:  http://www.geofex.com/Article_Folders/fuzzface/fftech2.gif  Here are the changes I've made:

1) Instead of PNP germanium, I'm using PNP silicon, MPS6517.  I measured the gains, and both are in the area of 140.
2) Instead of using a 2K linear pot on the fuzz knob I'm using a 1K rlog pot with a 1K resistor to ground.
3) The output 0.01uF capacitor is hooked up to a SPST switch that connects another 0.04uF capacitance in parallel.  In retrospect this is a little silly.
4) At the fuzz wiper there is a 22uF capacitor rather than a 20uF capacitor as is somewhat strangely marked in the circuit
5) I've also added the fuller mods at the input and by the 1Kohm resistor - so the 1K resistor is actually a 470ohm resistor in series with a 1K pot.

The problem is this:  the pedal makes exactly the sound it's supposed to make, except that there is a high pitched tone or squeal when you plug it into an amp.  This is mainly a problem with all knobs dimed;  it is mitigated by the fuzz, input resistance, and volume knobs, and cannot be heard when the guitar is playing.  It's worse on single coils than it is on humbuckers, and worse on humbuckers than it is with the guitar unplugged, but occurs in all cases.  On top of that, when you have just the cord plugged into the pedal, the tone/squeal cannot be heard at the lowest volume levels, but comes down into the audible range and get lower as the volume at the amp goes up.  This is without changing any knobs on the box, this is just by manipulating the volume on the amp.

I've tried adding a 100pF cap between the collectors of Q1 and Q2 and between the power supply and ground; the latter has no effect, and the former only seems to eliminate the higher frequencies.  I would increase the value of the cap even more, but even at 100pF it sounds like crap.

I've checked that everything going to ground goes to ground.  I've checked that everything not going to ground doesn't go to ground.  I've gone through many of the connections in a haphazard manner on a low ohm setting with my multimeter, to no avail.  The transistors appear to be fine, but here are the voltages:

Q1 - E 0.00 B -0.60 C -1.39
Q2 - E -0.79 B -1.39 C -2.36

Sorry if I'm giving too much information, I'm new at this and a little stymied...

Gus

Could be layout or even the switch.

petemoore

  What is the battery voltage?
  Q1 - E 0.00 B -0.60 C -1.39
Q2 - E -0.79 B -1.39 C -2.36
  2) Instead of using a 2K linear pot on the fuzz knob I'm using a 1K rlog pot with a 1K resistor to ground.

  This increases gain but..
  Q2's Collector looks low, should be about 1/2 the battery voltage.
  THe 18k, Q2's Collector resistor may bring the transistor to better bias if reduced to...less...around 8k2, or whatever achieves optimal bias, [1/2v] for Q2C.
Convention creates following, following creates convention.

Dragonfly

Quote from: petemoore on June 26, 2006, 09:48:40 PM
 
  Q2's Collector looks low, should be about 1/2 the battery voltage.
  THe 18k, Q2's Collector resistor may bring the transistor to better bias if reduced to...less...around 8k2, or whatever achieves optimal bias, [1/2v] for Q2C.

yep...could just use a 20k trimmer in place of the 18k...that way you can bias it to taste....

as for the squealing , i'd start with checking the layout...keep leads as short as possible...

AC

petemoore

  Squealing is what a FF does...if...
  The gain is too high, this can be adjusted a number of ways, starting with turning down the gain knob, or reducing the resistance from Q2E to ground [max gain set].
  The gain of the transistors is another factor.
  Guitar volume turned down or resistance before input...
  a cap across Q2's B/C junction...and/or...take a look at the Axis Face at Fuzzcentral, notice the .047uf cap, use one or adjust two smaller ones to taste...
  Not any one thing really, just a FF circuit.
Convention creates following, following creates convention.

Dragonfly

pete...youve built more ff's than any human should !

rock.

AC

Dragonfly

btw pete...sometime you should try putting a germ diode between q1's emitter and ground.... works great in silicon & hybrid faces...havent tried it in germs yet....

AC

mac

Quote from: Dragonfly on June 26, 2006, 10:51:07 PM
btw pete...sometime you should try putting a germ diode between q1's emitter and ground.... works great in silicon & hybrid faces...havent tried it in germs yet....

AC

This is a nice mod.
I've posted this on a FF thread here and at fuzzcentral forum. Also, I've placed a germ diode from q2 base to emitter. It acts somehow like a cap from collector to base reducing the fizziness. It works fine on Si and Ge. Or, instead of a germ diode, a resistor of about 1M or more in the same place. Or use two 47k in the feedback loop and put the diode or +1M to gnd in the junction of the two 47k.
Some suggest that setting q2c voltage to 4.3V gives a more germ sound.


mac



mac@mac-pc:~$ sudo apt install ECC83 EL84

Gus

People look at this circuit.

http://www.diystompboxes.com/pedals/gusFuzzFace.gif

Try a 1.2K or 820Ohm(from the Vox) fixed "under" the gain pot.  Adding a resistor under the pot does not realy change the gain much if at all because it is bypassed even at min gain setting via the cap from wiper to ground.

Note the bias pot.  note the 100 ohms in the emitter leg note the collector value Look at the Vox 2 transistor FF type circuit

Its Si people you should be able to "DESIGN" it to work and bias no problem.  The bias on in mine is an external for changing the sound.

Look at the rocket the 3 tran the Hot Si.

Layout,Layout,layout

Look at the first stage feedback R 100K/ 47K this with the source Z feeding the circuit in parallel with the OPENLOOP gain of the first transistor is were the gain/touch comes from.

If anyone takes the time to do the math on "my" take take on the circuit they might learn something work out the gain of the 1st transistor with a high Hfe.  It does not "bend" a triangle wave as much as a low gain grounded emitter but it controls the stages openloop gain.

For FF types note Ic, open loop, closed loop. Hfe vs Ic, The top resistor and the collector voltage wanted set the Ic at idle

Mark H have you build that version in the above link?


billings

Today I finally got frustrated and decided to plug in the first FF I made to see if it had the same problem, even though it was wired/laid out much differently...  and lo and behold, it did.

Turns out that the direct cause of the squeal was that I had been testing the thing hooked up on top of my marshall combo with a patch cable.  When I hooked it up on the floor like a good industrious person the problem disappeared.  That explains why the pitch was changing with the amp's volume knob - it being a pretty noisy circuit to begin with, it must've been picking up whatever EM junk the marshall was emitting.

Much thanks for the assistance;  I'll take note of all the above in my next build.


mac

#10
Quote from: Gus on June 27, 2006, 10:40:13 AM

If anyone takes the time to do the math on "my" take take on the circuit they might learn something work out the gain of the 1st transistor with a high Hfe.  It does not "bend" a triangle wave as much as a low gain grounded emitter but it controls the stages openloop gain.

For FF types note Ic, open loop, closed loop. Hfe vs Ic, The top resistor and the collector voltage wanted set the Ic at idle



I wrote a public program to do the math of FF-type circuits. Well, the bias at least.

MacOSX
http://geocities.com/guitarfxs/exes/BiasCalculatorX.sit

Win
http://geocities.com/guitarfxs/exes/BiasCalculator.exe.zip

Read the note about decimal point bug.


mac
mac@mac-pc:~$ sudo apt install ECC83 EL84

Dragonfly

Quote from: mac on June 27, 2006, 10:13:39 PM
Quote from: Gus on June 27, 2006, 10:40:13 AM

If anyone takes the time to do the math on "my" take take on the circuit they might learn something work out the gain of the 1st transistor with a high Hfe.  It does not "bend" a triangle wave as much as a low gain grounded emitter but it controls the stages openloop gain.

For FF types note Ic, open loop, closed loop. Hfe vs Ic, The top resistor and the collector voltage wanted set the Ic at idle



I wrote a public program to do the math of FF-type circuits. Well, the bias at least.

MacOSX
http://geocities.com/guitarfxs/exes/BiasCalculatorX.sit

Win
http://geocities.com/guitarfxs/exes/BiasCalculator.exe.zip

Read the note about decimal point bug.


mac

cool little program...thanks !

mac

Quote
cool little program...thanks !


gracias



mac
mac@mac-pc:~$ sudo apt install ECC83 EL84

Plexi

Niche thread!
A few great tricks here...  8)
To you, buffered bypass sucks tone.
To me, it sucks my balls.