Tube Echo - PT2399 & ValveCaster

[Franky]

I was looking in my components stock, and I found a PT2399, I wanted to make a little delay to put inside my guitar, but I came up to new projects for that guitar..

Anyway, I'm in the Valvecaster mood, and the "external" parts of the Rebote delay should be simple enough (opamp buffers and amplifiers) to be replaced by a valvy, using one triode as an input buffer, sending this output to the chip and to a mix pot, getting the delayed signal from the chip to the other side of the mix pot, and putting the last triode as an output buffer..

All we need to know is the input and output levels of the PT2399, to avoid saturation..

[frequencycentral]

A thought that had crossed my own mind. Obviously. Certainly the input triode could be set up so as not to saturate the PT2399 input. And it would be easy to attenuate the PT2399 output so as not to overdrive the output triode - but maybe a tiny bit of overdrive would be nice! I have PT2399's and plenty of dual triodes - if you knock off a schematic I will breadboard it too!

EDIT: Which Rebote are you looking at? 2 or 2.5? It strike me that 2 would be more suited to this idea. As a starting point 100K plate and cathode biasing resistors should achieve a clean buffer.

[Franky]

Euh, I don't know the version (I'm not currently having the schematic) but I think it's 2.5..

But any version can work, on 2.5 there are 4 buffers to use the feedback function without sending it back to the output.. So yeah, I came to a simplified schemo with 4 triodes, just replacing each opamp by a triode.

I've got to take a look at the lower version, maybe something is achievable with only one tube.

Do you think it's possible to cascade the PTs to get even more delay? Even if it gets really dirty, 580ms is really not enough for me..

[frequencycentral]

Yeah, version two just uses 2 opamps - so one dual triode would do it: http://www.tonepad.com/getFile.asp?id=50

Version 2.5 has improved filtering - I don't know how much difference that makes: http://www.tonepad.com/getFile.asp?id=98

Have a look at slacker's Echo Base PT2399 delay, I think Valoosj did a mod for cascading two PT2399's: http://www.diystompboxes.com/smfforum/index.php?topic=60662.0

[Franky]

yep, I had the v2.5.. The 2 opamps more are there to mix the feedback and the "direct first" delayed signal, which is done passively in v2.

so yeah, the aim is to mix the wet and dry signals at the input of the second triode, and to mix the straight and repeats signals at the input of the PT, passively (without the repeats taking the dry way)..

[frequencycentral]

Cool! I'll definately breadboard this one once you get the schematic together! Good luck!

[Franky]

Thank you, I don't really know what is the use of all thoses resistors and caps on the PT, but looking both schematics, there must be some way to make it..

Edit: my mistake, I didn't look well.. The 2 "opamps" at the bottom of v2.5 schemo are a part of the PT.. They were maybe used in the v2 (that would explain the cloud of res and caps around the right side), but not as clearly drawn as the v2.5..

So both version are the same, on the buffer point of view.. Anyway the schematic of the v2.5 is much clearer than v2, clear enough not to have to redraw all things up..

The first buffer is really easy to set up, clean output, maybe unity gain if it's possible, let's say 100k resistors on both cathode and plate. Then send it to the second tube stage via a mix pot (blending pot, maybe stereo: 2 volume pots inverted for Dry/Wet), and to the input stage of the PT (with all the filter thing)..

[frequencycentral]

Edit: my mistake, I didn't look well.. The 2 "opamps" at the bottom of v2.5 schemo are a part of the PT.. They were maybe used in the v2 (that would explain the cloud of res and caps around the right side), but not as clearly drawn as the v2.5..

Hah - me too! I missed that. So 2.5 is the way to go.

The first buffer is really easy to set up, clean output, maybe unity gain if it's possible, let's say 100k resistors on both cathode and plate. Then send it to the second tube stage via a mix pot (blending pot, maybe stereo: 2 volume pots inverted for Dry/Wet), and to the input stage of the PT (with all the filter thing)..

Yeah, from my experiments 100K cathode/plate should be close to unity using a 12AU7. As for the second stage, 12AU7 at 12 volts overdrive pretty easily - so maybe a voltage divider at the output of the PT2399 to attenuate it, and set up the second triode to recover any gain lost with a higher value plate resistor, say 220K.

[Franky]

Yeah, and we might even not need to put an extra voltage divider at the output of the PT, the blend pot works as two inverted volumes for dry/wet, but at it's output, you've got to mix the signals, with 2 resistors, and experiments some values for these to get the same (unity) gain at the input of the seconde tube stage.

About last tube stage, maybe a trimmer would be good on the plate to set up the gain as you want, clear or crunchy..


Here is the simplified schematic (I didn't write all the PT parts, way too long and useless):
http://frankyfuzzfire.free.fr/DIY/LAMPES/LowVoltage/TubeEchoSchematic.pdf

[slacker]

If you wanted to make it more tubey you could get rid of all the filtering components around pins 13,14,15 and 16 of the PT2399 and make those pins into buffers instead by putting a resistor between the in and out pins (15 -16 and 13-14) and then using resistors of the same value for the input resistor. You could then use tubes to do all the pre and post filtering, that way you'd have tubes in the feedback path which might give you some nice overdriven tones on the repeats.
That might be more interesting than just replacing the input and output buffers with tube buffers.

[frequencycentral]

Looks good so far!

R4 needs to be moved, (in the Valvecaster it's the grid leak resistor for the second triode) so it should be from 12AU7 B to ground, with another 47n cap (coupling cap) before it.

Wouldn't the blend be easier like this though:



Also - I don't think the 7812 sub circuit is needed if you have a good power supply - I've never needed a 7812 for any of my tube projects - just smoothing cap.

And don't forget the PT2399 needs a little 7805 circuit.

This might be useful to you: http://www.diystompboxes.com/pedals/tubedummy.html


EDIT: I was typing as slacker was posting. An interesting idea Ian, could you contribute a modified schematic? I'm not sure if you are suggesting additional filtering around the tubes to replace those removed from the buffers.

[Franky]

Thanks for the link, rick! Yeah I made this schematic as simple as possible, and very quickly so yep, there are a lot of mistakes..

What Ian was thinking about might be a second tube that would replace the integrated buffers, and then the filtering part would be stuck on those 2 more triodes. Good idea yeah, but really less compact than using the integrated buffers, one just have to wonder what's the best for him..

Your blending solution should work, but there is no ground reference, might be fixed with a 470k resistor from the grid of the second triode to the ground..

[frequencycentral]

I'm pretty sure Ian meant using just a single tube, He's bound to chime back in anytime soon though!

This project is getting me all fired up! Now if we could only work a tape loop in somewhere..............

[slacker]

My thinking was like this, replacing the input and output buffers (IC2a and b) with tubes might not add much tube magic the circuit. I guess you could overdrive the output buffer a bit, but basically a buffer whatever it's made out of doesn't really add anything to the sound.
My idea was leave the input and output buffers as they are using an opamp and instead replace the filters made out of the PT2399's inbuilt opamps with filters made out of the 2 halves of a single tube. You'd still need to use the first inbuilt opamp (pins 15 and 16) to get the signal into the chip, so just make that into a buffer, but looking at the schematic you could leave the second one out completely and just take the output from pin12 into the tube stage. Take the repeat pot from after the second tube stage and feed it back into the first.
That way the tubes are doing more and as they are in the feedback loop they might add some tube goodness to the delayed signal.
I don't really know much about using tubes for this sort of thing so I can't help with a schematic but there's some useful info here http://www.tubecad.com/articles_2001/Tube-Based_Crossovers/Tube-Based_Crossovers.pdf maybe do something based on the "single section 3rd order lowpass" example.

Hope that makes sense, if not let me know and I'll knock up a block diagram.

[Franky]

@ Rick: yeah, tape echo, we had some thoughts on a french forum about that, by hacking an old answering machine, but that's a heavy project.. 

@ Ian: you're definitely right, but besides the fact that buffers don't change the sound that much, putting opamps and tubes together might be a pain to power, (we need 12V for the tubes, 5V thru a regulator for the PT, and 6V for the opamps..). We can save some components and some place by replacing the input and output buffers by triodes, and of course we got to keep a tube for filtering issues.

About filtering now, it seems to be just some single pole lowpass filters cascaded, which can be plugged at the grid of a triode, but there are those 47k & 24k feedback resistors that are still a mystery to me.. 

[frequencycentral]

I use two 6111 submini tubes on my breadboard, running the two heaters in series from 12 volts (300ma draw). It's possible to tap off 6 volts from between the two heaters as a Vref for opamps.



If this tube echo idea ends up using two tubes, it would be possible to use 12AU7, running the centre taps of both tubes in series and derive 6 volts in the same way to the above (still 300ma draw). But I would build a 6111 submini version!

.................I do have a WEM Copycat Mark IV - I love it..........wish it was the tube version though!

[Franky]

wowow, what's that synth called Dalek's Handbag? Tell us more!! Looks like something from MFOS..

It's possible to get that center taped voltage, but I don't know if it's that good to draw some current from the tap to power the ops, cause it will overheat the upper heater..

[frequencycentral]

wowow, what's that synth called Dalek's Handbag? Tell us more!! Looks like something from MFOS.. 

Dalek's Handbag is my own design: http://www.diystompboxes.com/smfforum/index.php?topic=70222.0

[Franky]

Ah yeah, I already knew that sample by your Myspace page.. Awesome noise, I love it!

[frequencycentral]

Just out of interest, here's the Watkins Cpoicat schematic - tube version!

Fancy making a clone?