Bipolar power benefits with fuzz?

[John Lyons]

What are the benefits to running a circuit with a bipolar power supply?
For a fuzz circuit at 9v that is.
For example the Ultra Fuzz here:



Can this be converted to a single supply without any down side?
Is it just a matter of switching to a single battery and a dual op amp
pin 8 v+, Pin 4 Gnd setup?
I've never looked into doing this so...

Thanks
John

[Processaurus]

Yes, also you'd need to make a Vref and send it all the places that are hooked in a DC way to ground.  Caps for power filtering can still go to ground, as can the volume control, as it's AC coupled by the output 10uF cap.  The opamp + inputs, as well as the .7v and -.7v regulated supplies (made by the 10K's and diodes) feeding the outsides of the sensitivity knob (a comparator threshold adjustment, looks like) would need to be moved from ground to Vref.

I doubt, to a designer that knows their circuit, there's a sound benefit to running any distortion off bipolar or boosted voltage, since within the circuit headroom is artificially limited.

[Nasse]

Opamps and poor performance of 9 volt batteries may cause some things. Ordinary opamps need certain minimum working voltage, there are some types that can do with less voltage. I believe if you use two 9 volt batteries, they will last much much longer than one, or one twice. Someone should test this, how long would a D+ work with one or bipolar.

[John Lyons]

Ok, so I'm still a little unclear on a couple things.
The outer diodes (at V+/V-) are polarity protection diodes
and the caps to ground are filtering for the PS correct?
Since the PS will be a single now I will just use one cap
to ground off V+ and one diode in series from V+.
So I can drop V- diode and 10uf cap to ground correct?

So what I'm unclear on are the inner diodes and connection
of the 10K series and 10K sensitivity pot. Do I still need those diodes?

Thanks for your help!

John

[Mark Hammer]

It's bipolar powered, not because it HAS to be, or works better if it is.  It's bipolar powered because that was the design standard for the EPFM book it was part of.  Once you were going to have a +/-9v supply anyway, take advantage of it.  You can ask him yourself, but my guess is that Craig did not scootch off in a time machine from 1979 to the present to suss out whether players would be amenable to bipolar modular units, or preferred independent battery-operated things.

[John Lyons]

Ok, that makes sense and was what I figured.
Just need a little help with the conversion.

John

[R.G.]

So what I'm unclear on are the inner diodes and connection
of the 10K series and 10K sensitivity pot. Do I still need those diodes?

Yes. You'll need to move those to be symmetrical around the new Vref. Those diodes plus the middle 10K pot provide an adjustable reference voltage for the (-) input of the second opamp. That second opamp is working as a comparator, whether that's a good application for it or not, there being no negative feedback path for it.  It has its negative reference voltage forced to be no more than one diode drop above the "ground" and no lower than one diode drop below the "ground" by the diodes. The 10K strung across the diodes lets you adjust the actual voltage seen on the (-) input.

Calling that "sensitivity" is an unusual choice. What it does is pick the place in the incoming waveform where everthing's going to go full positive versus full negative at the output of the second opamp. I guess that lets you set it just above the noise level from the first opamp, forming a kind of crude noise gate. Probably a good idea with an open loop opamp there. But it means that with no input, the output of the second opamp will sit as close to V+ (or V-) as it can. The sensitivity pot lets you not only pick which one, but how little signal will prod it out of being quiet by being clamped to one power supply.

Needless to say, this is an app which can NOT use a DC coupled output very well.  

There's always more than one approach. An LT1054 will ovbiously make this thing into a single power supply version while leaving it dual power supply operation. Whether that changes the elusive tone remains to be listened to.

[John Lyons]

Is this correct?



I'm still unsure of the diodes/10k pot section.

Thanks

John

[R.G.]

Is this correct?

No. There is nothing pulling the + diode up or the - diode down. There is no DC on the diodes at all, other than one end being connected to ground and one end being connected to a very high impedance input. What happens in this condition is unknown and probably unpleasant.

Connect the anode of the left diode (the one labeled "10uF") to +9 with a 10K resistor. Connect the cathode of the right hand diode to ground with a 10K resistor. The whole point is to make the middle of the two diodes read the same DC voltage as Vb, while one end of the 10K pot is at Vb plus one diode drop and the other end of the 10K pot is at Vb minus one diode drop.

[head_spaz]

Here ya go.... Simple.

[PRR]

Man, that thing must be NASTY.

"Sensitivity" is an odd name. "Asymmetry" may be another name.

The diode-chain can also make your Vref; probably wants heavier bypassing to handle two chores.

[PRR]

head spaz's plan is correct for a "floating" battery, neither side connected to system ground.

I was assuming one end of the battery connects to ground. This allows simple jack power-switching, sharing power with other sub-circuits, etc. Just a frill.

[John Lyons]

Ok, thanks fellas.
So, Paul
If I add a vref circuit, 10K/10K/100uf to your schematic that will do it?
Or are you saying that the diodes make the vref point (bottom vref label)?
I would like to connect ground via the input jack so that is the way to go I think.

thanks for the help.

[Top Top]

It is a cool fuzz, by the way. I had one from an EPFM/PAIA kit a while ago.

[R.G.]

If I add a vref circuit, 10K/10K/100uf to your schematic that will do it?

Yes, that's one way.

Or are you saying that the diodes make the vref point (bottom vref label)?

Yes, that's another way.

I would like to connect ground via the input jack so that is the way to go I think.

??

[John Lyons]

I was talking about input jack ground switching via the ring/sleeve RG...

[kvb]

can someone tell me what the nodes with the little + and - are?

[R.G.]

I was talking about input jack ground switching via the ring/sleeve RG...

Yes, I knew that.
Did you mean that you are doing this change over (i.e. dual to single power supply) in order to get input jack switching to work?

If so, you do know that there are other ways, and that while input jack switching works most of the time, it's not always a good idea, right?

[John Lyons]

Kevin
Those are just pin 8 (+9v) pin 4 Ground as in the typical op amp power/ground connections.


RG
The input switching is just what I usually use so it works as a happy byproduct  or standarization for me.
How is it not a good idea (sometimes)?

[R.G.]

The input switching is just what I usually use so it works as a happy byproduct  or standarization for me.
How is it not a good idea (sometimes)?

It is a handy trick as long as it works - which is mostly, and why very few people think of this as an issue.

The problem becomes apparent when you study grounding the right way, worrying about where the current goes, not the voltage. In an effect switched by the stereo jack trick on the input jack, 100% of all current to power the effect is carried in the ground lead of the input jack. So any voltage created from it traversing that wire is added in series with the input signal.

This is the source of little enough problems most of the time, because the wire is only a few inches long and the current is small. However, if you happen to be using an effect which has high gain and little input common mode rejection, as well as a lot of current use and particularly a bad contact at the jack, you can get noise or oscillation. Once again, it's rare. But I have traced a few cases of switch popping down to the LED current suddenly raising the input "ground" by a few millivolts and this then being amplified through the circuit.

As a result, I use the PNP transistor/resistor switch on the positive side. This is at GEO. I connect the emitter of a PNP to the + lead of the battery on the PCB, then the collector of the PNP to the rest of the +9V on the PCB. PNP off, no effect. PNP on, effect works. I then connect a 4.7K to 10K from the base of the PNP to the ring lug of the stereo jack on the input. The battery minus lead is hard-soldered to the PCB.

This works the same as the input jack trick without a PNP - plug in jack equals battery connected. However, what actually travels down the wire at the input to compete with input signal is the much lower - and constant! - current from the PNP base. This can't cause oscillation from feeding back circuit events into the input ground; that current doesn't flow there.

As I say, it's an unusual condition. But PNPs cost me $0.05 or less and resistors are a penny. That six cents gets me immunity from even the rare cases.