Author Topic: advice on dual pt2399 design  (Read 1240 times)

PraiseIommi

advice on dual pt2399 design
« on: December 02, 2019, 05:05:21 PM »
I'm trying to make a delay/echo pedal by cascading two PT2399's in order to achieve delay times from short slap-back to at least a second long, with less noise than trying to use a single chip as I've heard this can be achieved by cascading. I hand drew a schematic, so excuse me if it looks a bit crude, but I would like to know if the way I am connecting two of these PT2399's is correct, and any other advice that might help me improve this design. Thank you for your time, hope everyone has a nice day. 


PRR

Re: advice on dual pt2399 design
« Reply #1 on: December 02, 2019, 11:58:43 PM »
https://www.diystompboxes.com/smfforum/index.php?topic=123487.0

That thread has notes on a proven dual-PT design from an experienced designer. Always study what has been done before.

deadastronaut

Re: advice on dual pt2399 design
« Reply #2 on: December 03, 2019, 05:07:05 AM »
take a look at this 'delay'  uses 2 x pt2399's, and modulation.  docs on my site.....

https://www.youtube.com/watch?v=3V4kzKXovT8

https://deadastronaut.wixsite.com/effects/dreamtimedelay
http://www.youtube.com/user/100roberthenry
https://deadastronaut.wixsite.com/effects

chasm reverb/tremshifter/faze filter/abductor II delay/timestream reverb/dreamtime delay/skinwalker hi gain dist/black triangle OD/ nano drums/space patrol fuzz//

anotherjim

Re: advice on dual pt2399 design
« Reply #3 on: December 03, 2019, 05:23:04 AM »
You have got much of it right.

On the output, a series control isn't a very effective way of doing it -  it can't turn down to silence - but you may want it that way, but it will depend on the load impedance of whatever the output plugs into. It will only go quiet when the 10k pot resistance is significantly larger than the load impedance.

Wiring both pin6 delay controls together is novel. I don't know if it works well that way. Wouldn't surprise me if it can work like that, but I've never seen it done before.

From IC1 pin14 to IC2 pin15 there is a DC path that might introduce a DC offset in IC2 input, but you will probably get away with it. Is the 180R/100n filter necessary?

The Equinox design that Paul points you to shows how the filtering can be very much simplified - it doesn't really need all those multiple-feedback filters on all of the PT2399 ins and outs, but that doesn't mean it won't work if you put them in if you want. You might think, that if the delay chips internal clock frequencies differ, it will cause heterodyne whine to be audible, but I've never actually heard that happen with PT2399 circuits, even with simplified filters.



Forever famous for Slug Tape.

PraiseIommi

Re: advice on dual pt2399 design
« Reply #4 on: December 03, 2019, 04:41:03 PM »
Thank you for your reply,

I apologize but I am not familiar with the term series control (I am familiar with what elements in series means), can you please explain. What might you recommend as a better alternative for a master volume control on the output, would a voltage divider with a pot be a better option, and what are some generally good values for resistance?

Would a simple 10uF capacitor (either electrolytic or non) suffice to solve the dc problem from pin 14 IC1 to pin 15 IC2? The 180ohm resistor and 100nF cap I was thinking would form a Low-pass filter for the output from IC1, and I based that on what I saw on the PT2399 datasheet in the surround/delay example, if it is unnecessary I will remove it. I chose those specific values because I calculated them to have a frequency cut-off of 8.8KHz and that is close to the approximately 8.5KHz Fc of the Multiple feedback filters in the circuit, I can adjust R to get closer to 8.5KHz (which I saw noted on Electrosmash as a good Fc for its "wide frequency range and reasonable low noise"), but is this beneficial or neccessary to try and make all of the low pass filters have the same cut-off frequency?

I hadn't looked at the other designs before I made this reply, but I will be sure to look at them, also I had never heard of Heterodyne before, so thank you for introducing this concept. Hope you have a nice day.


PRR

Re: advice on dual pt2399 design
« Reply #5 on: December 03, 2019, 11:07:40 PM »
> not familiar with the term series control

Your drawing shows a 10K variable resistance in series with the output. Obviously this can go up to 100%. But in a high impedance load (1Meg) it can only go _down_ to 99%. In a 10k load (rare in guitar work) it can go down to 50%, which is not a lot of drop.

If you instead wire it as a 'potentiometer' to ground (you have an example in the middle of the plan) you can dial 100% to zero %. 10K is a fine value I think.



anotherjim

Re: advice on dual pt2399 design
« Reply #6 on: December 04, 2019, 05:16:29 AM »
+1 ^ for the output volume control.

Yes, you can block that DC with a 10u cap. Missing out the cap is only a technical error in this case. Each end has a DC bias of about 2.5v (but not identical because different sources) and the circuit only has about x1 gain - so the small difference in DC bias won't change things in any significant way.

The output of a PT2399 is from pin 12, then you have a 2-pole active filter built on the opamp in pins 13 & 14. That then feeds another 2-pole to the input of the second delay in pins 15 & 16. So, I was wondering why it needed another pole.

I don't know how you intend to use your design. As it is, it will be fine in an amp or mixer FX send/return loop, but the input impedance will be too low to plug a passive guitar or bass into - it will need a buffer. But as I don't see a dry/wet mix control, I guess you do only plan on FX loop use where you only want the wet output.

Forever famous for Slug Tape.

tubegeek

Re: advice on dual pt2399 design
« Reply #7 on: December 04, 2019, 03:52:28 PM »
PRR's explanation is elaborated upon (but NOT improved!) here:
https://en.wikipedia.org/wiki/Impedance_bridging

The concept is "Impedance Bridging," which is an alternative to "Impedance Matching." Impedance Matching is a technique that is less frequently encountered in typical analog audio circuits than bridging. Matching provides optimal POWER transfer, Bridging optimizes VOLTAGE transfer, which is equivalent to saying it maintains the highest possible signal level when connecting an output to an input.

The main reason NOT to use the highest impedance input possible in "our" circuits is that the noise contribution is less with low impedance. So we trade-off signal level preservation with noise performance here.

The main reason to match impedances in "our" circuits is to maximize power transfer to a load, e.g., in a power amplifier connected to a speaker. But even in that case there may be some merit in bridging instead - that's an area of "designer's choice."

Gross oversimplifications, complicated topic.
"The first four times, we figured it was an isolated incident." - Angry Pete

"(Chassis is not a magic garbage dump.)" - PRR

PraiseIommi

Re: advice on dual pt2399 design
« Reply #8 on: December 12, 2019, 04:01:36 PM »
Thanks PRR regarding the volume control that makes perfect sense to me. Also, thank you anotherjim, I was just trying to follow the existing designs from the datasheet with the whole extra pole thing, I guess I was not aware that it was adding another pole to the opamp filter, I just assumed it was a like a final passive filter for the output of IC1, which may have been an incorrect assumption. This is my first attempt at designing a guitar pedal circuit, I'm a musician and engineering student and I wanted to do something related to both in a creative and productive way. My intent with this project is to create a delay pedal that has a variable delay control for quick slap-back type echoes to at least 1 second of delay time, as well as feedback(repeats) control, and volume control. I figured that since 1 of these chips is said to create some internal noise before it gets to that final threshold of delay time that two would be able to solve this dilemma. I am not quite sure how a mix control is accomplished or how it can be implemented here, but that is definitely a feature I would like to add if possible, I will definitely look more up about that and if you have any advice on the subject yourself I'd appreciate it. I was aware that a buffer would be required for the input, I did not include that in this schematic, but I was thinking about using an op-amp buffer set up for unity gain, since op-amps have high internal impedance values, what are your thoughts on this, is there another/better way to accomplish an input buffer? Thank you as well tubegeek, I found what you said to be very
informative. Hope everyone has a nice day, I'll be sure to get back to you after trying to make some improvemnets, thanks again.