12AU7 Overdrive Pedal

[Scootsit]

Hello!
I'm not much of a guitar player, I've built a ton of different low voltage SS and valve amps though, and a friend of mine was looking for an overdrive pedal. I told him I'd give it a shot. I tried to mostly stick to decent components I have lying around. I'm not 100% on the output level or impedance of guitar pickups, nor the input impedance and sensitivity of amps. I was thinking a 12AU7 would be the perfect tube to run at ~24V, and use a pair of resistors with a regulator to get the 6V for the heaters. My plan was for it to plug in, and basically use an opamp to overdrive both sides of a 12AU7, in parallel. Doing some googling, I found that someone created the valvecaster which uses the gain on one side of a 12AU7 to overdrive the other side of the 12AU7 (since it's a dual triode; this can be seen here: http://damacleod.wordpress.com/2011/03/14/a-diy-valve-overdrive-pedal-goldie/). Using one side of the tube to drive the other would also limit the gain of the first side, which would limit how hard it clips.

I have a few questions. Is there any advantage to using a tube or an IC as I had planned? Second, I had planned to buffer the output, because again, I'm not sure what the input impedance of most amps is, but I figured it couldn't hurt. Is there anything I'm missing? I'll probably buy everything I need then experiment with different resistor values/gain settings/etc until I get it just right.

I have a schematic as a pdf, I'm not sure I can upload it. I should mention, I neglected to include output caps, I haven't forgotten them, I just didn't get to that point, and have a parts drawer filled with caps, so I wasn't worried about them making my BOM.

Thanks a ton!

Here's a pdf of my schematic: https://drive.google.com/file/d/0B7y3E9VTFdy4ZWNlNXlQV1BDYlE/edit?usp=sharing

[MaxPower]

I don't know much about tubes. Maybe this is useful:

http://www.valvewizard.co.uk/Triodes_at_low_voltages_Blencowe.pdf

In case the direct link doesn't work:
http://www.valvewizard.co.uk/links.html

Near the bottom in the Downloads section, there is a PDF: Triodes At Low Voltages. I have built something similar to the circuit shown in Fig. 5 ('The Tubester' as appearing in Poptronics of July 2002 if I remember correctly). Anyway, from my understanding, the output impedance is a bit high so an output buffer is probably a good idea.

[Jdansti]

I couldn't see your PDF, but here's 166 pages on the Valvecaster. http://www.diystompboxes.com/smfforum/index.php?topic=63479.0 

There are a lot of derivations within the thread and on other threads in the forum. Just do a forum search on "Valvecaster".

The Valvecaster was designed to work on 9V, but most people use 12V with better results, and some go higher. As you mentioned you'll want to keep the heaters at 6V in parallel (12V in series).

The tone control on the Valvecaster isn't that great and a lot of people either make it switchable or omit it.

The issue of impedance and whether to use a buffer is beyond my understanding, so I'll let someone else address that.

Good luck!

[Scootsit]

Maxpower: Thanks for the link, I read the PDF, very informative, and gives me some ideas.

It would seem that powering the tube to 24V should yield great results, and the opamp will make sure I can swing enough to overdrive it.

I'll take a look at the valvecaster thread. I guess I'm really leaning away from the valvecaster, since I'm not limited by current, I can drive two opamps and both triodes in the tube. I should also be able to get more gain that way, since what I've read about the valvecaster seems to suggest that the clip is pretty soft.

I just tested the link, it should work, if you PM me your email, I can email it to you. It basically is an opamp set at 5x gain, with a 25k pot on the gain loop, allowing it to go from 5x to 30x, this drives the grid of both triodes within the 12AU7, and the output is fed to a BUF634, so that the out is buffered. I want to use a single rail, so obviously there'd be caps between stages and on in and outs. That's about it.

Thanks!
Anyone else having problems with the link to the schematic I drew?

[GibsonGM]

I saw it, it looks interesting. But - gotta ask - if you're looking for high-gain clipping and are mixing in opamps, why not just do the whole thing with opamps or FETs?   These 12AU7 jobs are usually used more for driving the front of a tube amp (which can get VERY overdriven!), or warming up a signal.  They'll give you even-order harmonics when pushed, with are nice and warm...the opamp stuff, if overdriven, will tend towards odd-order harsh stuff. 

Seems to me that if you go beyond the PDF you posted above, you'll just be making square waves with little warmth...that's a lot of work to go to for that end product, IMHO.    You have clipping diodes after the tube - that would be the same as putting them after the opamp and increasing the gain.  Odd-order, harsh clipping.  Not that it's a bad thing >;o)  But not tubey.

To get really good (metal?) tube distortion generally requires more than four stages, and you'd ideally want them "all-tube".   And running at much higher voltages.   Most of the tone would come from your interstage shaping components...hard (impossible?) to get that in so few stages and have it sound deep, full, warm while thrashing....

I feel you can do more with a lot less work....see the BSIAB, using FETs.  Similar clipping characteristics to tubes.  9V, no transformer BS, all that. 

But your call, maybe with some work you can get something really cool going with the "mixed method" you're currently working on.  If it were me, I'd get rid of the silicon and go for 2 tubes, 4 stages, and think about the coupling/shaping....good luck!!!

[rutabaga bob]

Welcome to the forum!

Shoot, man...try the valvy - they're easy, and sound nice.  If you're using standard stompbox power (9 volts), you can use a 9AU7 and a voltage multiplier for the plates.  Let us know how it goes!

Larry

[Scootsit]

I'm definitely going with 24V. I guess that's not standard, but the friend that's using it doesn't care, and likes the idea I had for it.
I've made a few changes, and I think it will be very similar to the valvecaster, but with a bit more gain, allowing for some harder clips. I'm still working on it, and need to figure out the cap values, but I'll probably post the new one I'm working on tomorrow.

I'm considering using a dual rail supply, I'm not sure yet. It would allow me to get rid of the caps on the buffer and after any ICs. I'm not too worried about transfos, I've built a lot of stuff this way before. There are really efficient switching power supplies. And I've built stuff with other tubes for more hifi oriented stuff, and with enough PS filtering caps (in the schematic I uploaded, C1-3 should be sufficient for me to use a really high frequency switcher.

Thanks guys!

Am I correct in thinking that the output should be buffered? What is the input impedance of the average amp?
Oops, looks like MaxPower already responded. Thanks!

[Scootsit]

I updated the schematic, now it's got a 5x SS gain stage, followed by the gain stages of the tube. I'm still doing the math on the anode resistor and figuring out the regulator on the PS. I'm also thinking after the voltage divider, I'll probably have a regulator on the heater. Here's the new schematic. Any thoughts? https://drive.google.com/file/d/0B7y3E9VTFdy4U1hXQl9lU3BDSHM/edit?usp=sharing

[MaxPower]

What's the purpose of all of those capacitors in parallel (C5-C6, C11-C12, etc.)? I've never seen that (other than on the supply rails for op amps to prevent oscillation or whatever)?

[Scootsit]

Well, 5/6 are right after that opamp, which has a near 0 output impedance, so that cap has to be big, so I threw in a bypass. I bypassed all of the electrolytics with small film caps, it yields better top end frequency response.

[PRR]

> Any thoughts?

Input pins must go to a DC bias voltage, so they know where to sit. Your first opamp and both triodes lack this. Second opamp does return to ground.

If this is a single-supply plan, both opamps need to be biased to "half" the supply voltage. Plus blocking caps at inputs and NFB loop.

Voltage-divider heater supply is very dubious.

[Scootsit]

Excellent points, thanks. I need to address all of that. The heater supply is something of a placeholder. Right now, I'm thinking I'm going to get a 27v supply, and then put a voltage divider bringing it to 13.5v, with an LDO regulator set at 12.6v, I found one with incredible thermal characteristics. Then, I figured I'd use an LM317  on the main supply, bringing it to 24. I considered using a dual rail supply, but it seemed like more work than it was worth. I'm still figuring out the specifics of the power supply.

[PRR]

> a 27v supply, and then put a voltage divider bringing it to 13.5v, with an LDO regulator set at 12.6v

Do you know how to figure Volts, Amps, Watts?

It's a useful skill. And tube heaters are HEAVY loads. Do the math, you'll see that a voltage-divider is either uselessly saggy or grossly wasteful.

Tube heater is 12.6V 0.150A or 84 Ohms. Let's adapt the popular 10K+10K voltage divider. Put an 84 Ohm load on a 10K||10K or 5,000 Ohm equivalent divider, it will sag to about 0.016 of the no-load voltage, or like 0.4 Volts. OK, go lower than 10K, say 10 Ohms + 10 Ohms. The output voltage is fairly stable, but the total divider current is over 27V/20r or 1.3 Amps. For a 0.15A load. Eating over 8 times the power actually used.

And _why_ would you want an LDO when going from 27V to 12V? If you persist in a voltage-divider approach, you can use a lesser division ratio, leave more for the regulator slop.

But unless the "27V" is wandering 23V to 31V, you *really* can just use a series resistor to power a heater. +/-10% won't bother the heater. Series-resistor is actually a wee bit kinder to the heater (the turn-on surge is limited).

The rest of the plan can work fine with R-C filtered raw voltage. The classic guitar tube amps don't regulate the plate supply. Opamps aren't fussy. The slight day-to-day variations in amp response are small compared to your day-to-day variations in mood, venue, band relationships, and moon-phase.

[Tony Forestiere]

The slight day-to-day variations in amp response are small compared to your day-to-day variations in mood, venue, band relationships, and moon-phase.

A true gem. Beautiful and succinct. Signature worthy. Thank you Paul. All of your posts are gold to me. 

[Scootsit]

> a 27v supply, and then put a voltage divider bringing it to 13.5v, with an LDO regulator set at 12.6v

Do you know how to figure Volts, Amps, Watts?

It's a useful skill. And tube heaters are HEAVY loads. Do the math, you'll see that a voltage-divider is either uselessly saggy or grossly wasteful.

Tube heater is 12.6V 0.150A or 84 Ohms. Let's adapt the popular 10K+10K voltage divider. Put an 84 Ohm load on a 10K||10K or 5,000 Ohm equivalent divider, it will sag to about 0.016 of the no-load voltage, or like 0.4 Volts. OK, go lower than 10K, say 10 Ohms + 10 Ohms. The output voltage is fairly stable, but the total divider current is over 27V/20r or 1.3 Amps. For a 0.15A load. Eating over 8 times the power actually used.

And _why_ would you want an LDO when going from 27V to 12V? If you persist in a voltage-divider approach, you can use a lesser division ratio, leave more for the regulator slop.

But unless the "27V" is wandering 23V to 31V, you *really* can just use a series resistor to power a heater. +/-10% won't bother the heater. Series-resistor is actually a wee bit kinder to the heater (the turn-on surge is limited).

The rest of the plan can work fine with R-C filtered raw voltage. The classic guitar tube amps don't regulate the plate supply. Opamps aren't fussy. The slight day-to-day variations in amp response are small compared to your day-to-day variations in mood, venue, band relationships, and moon-phase.

[/font][/color]


Yes, I do. I have built and designed quite a bit of stuff. In the past few years, I've gone through a number of moves, and had jobs that involved a lot of travel. In short, I've been away from building electronics and audio gear for about two years. I've finally got a bit more stability, so I decided to build this for a friend. I've never built any guitar stuff before, mostly headphone and preamps, and a few DACs. In short, yes, I do I apologize, it's been a few years, and the math isn't quite as fresh in my mind, and I can't just sort of see my errors like I used to.

You mentioned using the opamps at half the positive rail. The opamps I selected are pretty stable on a single rail supply. I considered using a dual rail supply, but it seemed to complicate things more than simplify them, and I'm not opposed to using DC blocking caps in the signal path. Is that what you were suggesting, or am I neglecting/missing something? I've made a lot of changes. I'm still trying to figure out the anode resistors, I'm not 100% sure how to calculate what they should be, most of the equations I've found have been pretty dubious. I'm sure you could provide some insight. Am I making any mistakes or missing anything?

I really appreciate your help! All of the errors you've pointed out have really helped me refresh my memory and get back into this stuff. Thanks!

Here's the latest: https://drive.google.com/file/d/0B7y3E9VTFdy4dFBZamMxc2p0akk/edit?usp=sharing

[ggedamed]

First things first: without a negative rail there is no positive rail. I mean when we're not talking dual supply (like +12V and -12V, everything referenced to the ground, which is equally far from any rail), we only have the power supply (i.e. +9V) and ground. In the single supply case, the ground takes the job as the negative rail. So we need a voltage that takes the voltage reference job.

The op amp will try to keep its inputs as close as posibble, voltage-wise. If there's any real amount of negative feedback (and almost all the time it is), the output will not be far from them. Even if your op amp wil be able to work that close to the rails (the ground, in your case) the output will have no headroom on the negative side of the signal. It is already at the lowest voltage, so it cannot go lower. Something like a half-wave rectifier, I imagine. That's why you want to keep the inputs biased at half of your single supply voltage. Check some schematics, you'll find most op amps inputs biased at half the total supply voltage, regardless if it's dual or single.

Good resources:
Single-ended and differential amplifiers
Op Amps For Everyone


[Scootsit]

First things first: without a negative rail there is no positive rail. I mean when we're not talking dual supply (like +12V and -12V, everything referenced to the ground, which is equally far from any rail), we only have the power supply (i.e. +9V) and ground. In the single supply case, the ground takes the job as the negative rail. So we need a voltage that takes the voltage reference job.
It
The op amp will try to keep its inputs as close as posibble, voltage-wise. If there's any real amount of negative feedback (and almost all the time it is), the output will not be far from them. Even if your op amp wil be able to work that close to the rails (the ground, in your case) the output will have no headroom on the negative side of the signal. It is already at the lowest voltage, so it cannot go lower. Something like a half-wave rectifier, I imagine. That's why you want to keep the inputs biased at half of your single supply voltage. Check some schematics, you'll find most op amps inputs biased at half the total supply voltage, regardless if it's dual or single.

Good resources:
Single-ended and differential amplifiers
Op Amps For Everyone


By adding a resistor on the input line, after the input caps, right?

[ggedamed]

Yep.

The standard is like here (search "What is V.R., V.B, VREF, 1/2V+ etc....?"). This creates the voltage reference which all op amps non-inverting inputs op are connecting to through usually large (100k - 1M) resistors.

Examples:
MXR MicroAmp Analysis - R7, R8, C4 are suppling the reference which is fed through R2 to the non-inverting input.
Tube Screamer Analysis - R33-R33-C16 for Vref, fed through R5 and R9 to op amps.

[Scootsit]

Sweet, thanks for all the great references, it's starting to come back to me! Next week, I'm going to be cooped up inside, so I'm hoping to order today so I can build then.

I've done a bit of googling, but I can't seem to find a good explanation of anode loading selection. Can you suggest a place to look for that (or explain it)?

Is there anything else you would change or improve upon? I think the regulator should work well, and leaving the heaters outside the regulator should diminish the draw significantly. I can't think of much else, can you?

Should I have a similar voltage divider between the tube and the buffer, it would make sense given what you said before, but the pot is a voltage divider, so I don't need it, right? But, the pot should probably be 100k, not 50k

[Scootsit]

It also occurred to me that the tube is the source for the buffer, the tube cannot swing lower than zero, so that wouldn't be necessary, right?


I've gotten a few requests to share the original schematic. I've removed the sharing from all but the most recent version of the schematic, in an effort to avoid any confusion. I have revised it again, I'll upload that soon.