New mini-breadboard

[uranium_jones]

If I'm not mistaken, that would be a mechanical servo.

Is there an "electrical" servo? Because google is apparently not my friend today.

 

[Chuck D. Bones]

Sorry, I was unclear (at best). I had read the in the datasheet that one of the suggested applications for this opamp was as a voltage regulator and I somehow that translated to built-in VRs...dunno..anyway...unless I'm hallucinating it's being used as such in this case.

Any opamp can be used to build a regulator. One of the components in a regulator is a voltage reference. The CA3130 does not contain one, you have to supply one externally. Intersil's CA3130 datasheet shows some application circuits, two of which are voltage regulators. In the example below, IC2 serves as the voltage reference.

 

[fig]

IC what you mean. Thanks!

Follow-up if I may, if the correct TO package voltage regulator existed, could that be used, and if not why?

 

[Chuck D. Bones]

Are you asking about 78xx and 79xx series regulators? LM723? I used to design power supplies, so I could talk all day about that stuff.

 

[Chuck D. Bones]

google is apparently not my friend today.

You know who else says that? Rick Santorum.


There are all kinds of servos. In the most general sense, servo control means using a negative feedback loop to control or regulate some physical quantity, be it velocity, position, temperature, pressure, voltage, etc. Seems that the common parlance is to use the term "servo" to refer to closed-loop position or velocity control of a motor. You could call the cruise control in your car a servo.

In this case, we're regulating the average current in Q2. The control loop is very slow, so it doesn't react to wave shape or dynamics. It does react to slow drift caused by changes in temperature.

 

[cooder]

The CA3130 is a simple CMOS opamp. I'm not sure what you mean by "built-in output voltage regulators."

In this circuit, the opamp's inputs and outputs are operated near ground. Not all opamps have that capability. The CA3130 does. The TL072 does not. The requirements for the servo opamp in this circuit is that it must have very low input bias current (FET input) and the inputs and output must be able to operate at the negative rail (ground). The CA3130 is not the only opamp that meets these requirements, but it is readily available and relatively inexpensive ($2 ea at Tayda). The CA3130 shows up in other pedal designs; some by me, some by others, like Bjorn Juhl. This is the simplest implementation of the servo control. There are ways to make the TL071 work, but it takes more parts. There are other rail-to-rail opamps out there.

What he said....

Thanks for the insights and this looks amazing. Seems that it makes the idea of the Benson Germanium Fuzz with the 'inbuilt heaters' of the carbon comp resistors redundant in a very elegant and better way.
https://www.bensonamps.com/guitarpedals/germaniumfuzz

 

[fig]

Like this;

https://www.mouser.com/datasheet/2/389/cd00000444-1795274.pdf

 

[Chuck D. Bones]

I like regulators, but I like the temp controlled fuzz moar. A clever, if not extravagant solution. Could double as a foot-warmer.

I have seen precision crystal oscillators that have their own build-in heater to regulate the temperature of the crystal.

Engineering guy: "We could design a temperature compensated circuit."
Marketing guy: "Temperature controlled would sell better."

 

[cooder]

The idea of the temperature controlled' as in Benson Germanium Fuzz seems a bit possibly laden with inconsistency wouldn't it? Fancy and somewhat cool, but the servo option seems a lot better engineered me would think?

Over to Chuck....

 

[benny_profane]

The idea of the temperature controlled' as in Benson Germanium Fuzz seems a bit possibly laden with inconsistency wouldn't it? Fancy and somewhat cool, but the servo option seems a lot better engineered me would think?

Over to Chuck....

I would think that it would only work until the ambient temperature exceeds the temperature they’ve set the control for.

 

[jubal81]

Found an AC176 today, threw it in there and sounded/worked great. It's worked with every transistor I've thrown at it. I've got one that noticeably lower gain, but it still works and sounds pretty good.

Unfortunately, one of my AC127s crapped out. Was playing nice for about 10 seconds, then dead. My component tester thinks it's a resistor sometimes and two diodes other times.

 

[Chuck D. Bones]

They just don't make transistors like they used to...

Spent a little more time playing thru it. Changed VR2 to C1K because the bottom-end of rotation was useless. Retuned the TONE stack. Reduced R4 to get a little more volume and expand the BIAS control range. Added C10 to reduce the high-freq "trash" above 5KHz. Still need to install and test the CONTOUR control.

 

[Chuck D. Bones]

Attention Cooder:
I found a way to make a TL072 (or TL071, TL081, etc.) work.

1. Install a green LED between Q1-E & GND.
2. Increase R7 to 3.9K
3. Add another green LED in series with D2.
4. Reset the bias trim (TR1).

This moves the opamp's inputs and output far enough from GND for a TL081 to work. The cost of doing this is two more LEDs and some volume. There is also a risk that the opamp will latch-up during start-up because some JFET opamps do weird shit when their inputs get too close to GND. Me, I'll stick with a rail-to-rail opamp like the CA3130.

 

[jubal81]

Got mine updated to the latest version this morning.
I tried to set it up so the bias pot full on was loudest/smoothest and all the way off is 80's toy "8-bit" sound. Fun. As. Hell.
What I did is swap the bias pot the other way around, turn it all the way up, then adjust TR1 until the collector voltage on Q2 is about 4.5V. It takes a good bit of time for the voltage to stabilize, so patience is required.
You also have to be a bit patient when changing the bias control. It takes a little time to catch up. Not too long, but it's just not instantaneous.
1K for Gain control is great.
The other note I have is I still prefer the original tone control. It has less range, but for me the whole sweep was in the sweet spot.

 

[Chuck D. Bones]

I have mine set up so BIAS at 7:00 makes symmetric or close to symmetric clipping. That puts Q2-C at 5.66V on mine. Turning BIAS up increases volume, mid presence and even-order harmonics. Above 3:00, it gets sputtery. I could have just as easily oriented the BIAS control the other way. Since it's B-taper, it's easy to switch over. Changing the BIAS setting settles in under a second. The bias trim is slow as molasses. As it should be.

For the readers, the "original tone control" that jubal81 references predates the first schematic above. It had C4 = 10nF, R4 = 47K and VR3 = B50K. Like he said, a little less range. Check it out on TSC or breadboard it. They are very similar.

This is the v0.3 tone control. Green is tone at 0; blue is tone at 5, magenta is tone at 5.5 (flat), white is tone at 10. Note: the TSC and schematic ref des do not match. Can't be helped. I made R2 (below) very large so it behaves like it's not in the circuit. There's always a way to get a simulator to do what you want.


Whether C3 is 470nF or 1uF is mox nix.

 

[cooder]

Attention Cooder:
I found a way to make a TL072 (or TL071, TL081, etc.) work.

1. Install a green LED between Q1-E & GND.
2. Increase R7 to 3.9K
3. Add another green LED in series with D2.
4. Reset the bias trim (TR1).

This moves the opamp's inputs and output far enough from GND for a TL081 to work. The cost of doing this is two more LEDs and some volume. There is also a risk that the opamp will latch-up during start-up because some JFET opamps do weird shit when their inputs get too close to GND. Me, I'll stick with a rail-to-rail opamp like the CA3130.

Following this glued to the screen, and I'm not worried about getting a CA3130 for 2 dineros at Tayda since your reply made as expected total eyeopening sense. Not interested in possible weird shit that might happen with a cheaper opamp.
Thanks guys, riveting stuff.

 

[Chuck D. Bones]

As requested, here is a PNP version of the SFF that does not require a charge pump or isolated power supply. Same parts (except for the transistors) and same topology, I just reoriented everything that has DC flowing thru it. At the moment, I'm running a П28 for Q1 and a 1T30B for Q2. They're both low-leakage and HFEs are in the mid 60's. Nothing magic about those part numbers, they were in the parts tray so I plugged 'em in. I'll try some other PNPs and see how they sound. C11 was added in the previous revision; it turns the CONTOUR control into a TIGHTNESS control. The C50K could be replaced with an A50K, just swap pins 1 & 3.

 

[cooder]

Duuuuuuude! Celestial indeed! I wonder who the goofball was who requested that...

 

[Chuck D. Bones]

The reason this works is because from an AC point of view, ground and the power supply rail are the same thing. It's important that the power supply is clean (no hum or noise), that's always a requirement with a FuzzFace and most other discrete designs.

 

[jubal81]

The reason this works is because from an AC point of view, ground and the power supply rail are the same thing. It's important that the power supply is clean (no hum or noise), that's always a requirement with a FuzzFace and most other discrete designs.

Would it be worth it to add a small resistor in series from D1 to C9?