PT2399 help needed

[dschwartz]

Hi all!
I'm working on a dual pt2399 delay in series.
It basically works OK, I'm controlling the time of both chips via a 100k voltage divider pot driving two 2n5089 npn (one for each chip ) between the vco and gnd, with a 47k resistor in parallel to limit the max time. I limit the lower time by limiting the voltage divider swing. This works OK but sometimes the chip gets latched up..i think i solved that by using a capacitor on the time control to let the oscillator start before  the transistors gets the voltage.

Anyway, my issue is that I'm getting wobbling repeats, specially when i pick hard..the pitch of the repeat wobbles down and up on attack..weird.. does anyone know why this could be?
I'm using 1 lm78l05 to power both chips..and the usual filtering and bypass caps. Each chip has its own 47uF power cap very close to them.
Any ideas?

[BluffChill]

I wonder if you could somehow introduce the latch-up protection as used by deadastronaut, distorque et al? It's a transistor based technique that stops the delay time getting too short (what usually causes latch-up)



http://distorqueaudio.com/hardware/projects/dimension-p.html

I'd have to see at least a fragment of schematic to know how to implement it though

[ElectricDruid]

From what you've described, the delay time depends on the voltage to the base of those two transistors, and that voltage is created by a voltage divider across the main power rails, right? So any variation on the main power rails will affect the delay time, and hence the pitch. It sounds to me like when you overdrive the pedal, the current draw goes up and the voltage goes down, making a pitch shift. I doubt this is the LM78L05 - it should be able to provide plenty of current for a couple of PT2399s. Is the "delay time" pot/voltage divider on the +5V supply or the main +9V supply? (I'm betting the +9V supply since the other one is regulated).

But that's just guess based on a few widely-spaced facts!

HTH,
Tom

[dschwartz]

I wonder if you could somehow introduce the latch-up protection as used by deadastronaut, distorque et al? It's a transistor based technique that stops the delay time getting too short (what usually causes latch-up)



http://distorqueaudio.com/hardware/projects/dimension-p.html

I'd have to see at least a fragment of schematic to know how to implement it though

Thanks, i used a capacitor in a similar fashion, but only 330nF..too short, I'll try with a 100uF maybe that will work.

From what you've described, the delay time depends on the voltage to the base of those two transistors, and that voltage is created by a voltage divider across the main power rails, right? So any variation on the main power rails will affect the delay time, and hence the pitch. It sounds to me like when you overdrive the pedal, the current draw goes up and the voltage goes down, making a pitch shift. I doubt this is the LM78L05 - it should be able to provide plenty of current for a couple of PT2399s. Is the "delay time" pot/voltage divider on the +5V supply or the main +9V supply? (I'm betting the +9V supply since the other one is regulated).

But that's just guess based on a few widely-spaced facts!

HTH,
Tom

I'm using the 5V regulated voltage for the time pot..i had the same hunch..but maybe it's something else.

Btw...i had some latching issues and powered the board repeatedly...when it came back it was very noisy! It was very quiet before that and now on longer delays it sounds awful, a lot of oscillating noises and whine...did i messed up the chips?

[anotherjim]

If the time control transistors have too much current gain, that could make it overly sensitive. try them in reverse beta?
Also, make sure pin2 of the PT2399 has a decoupling capacitor.
Another anti lock-up to remember is to tie pins 3 & 4 together at ground.
Pin6 VCO control is a current sink and normally reads close to the same voltage as vref on pin 2. Some fixed series resistance in the pin6 connection seems to be necessary  - you shouldn't short it to ground at any time.

[dschwartz]

If the time control transistors have too much current gain, that could make it overly sensitive. try them in reverse beta?
Also, make sure pin2 of the PT2399 has a decoupling capacitor.
Another anti lock-up to remember is to tie pins 3 & 4 together at ground.
Pin6 VCO control is a current sink and normally reads close to the same voltage as vref on pin 2. Some fixed series resistance in the pin6 connection seems to be necessary  - you shouldn't short it to ground at any time.

Well yeah, the 2n5089 has a high HFE ..
I have Agnd y Dgnd separated.. I'll check what happens if i just join both pins with a solder blob..
In theory, pin 6 should not go to ground, but ill check my sims if i went too far with the minimum time on the voltage divider.

[BluffChill]

If it's noisy when it comes back on, check your battery. If your battery is dying and the supply dips below 5V, PT2399 can get pretty crazy!

[ElectricDruid]

I'm using the 5V regulated voltage for the time pot..i had the same hunch..but maybe it's something else.

Ok, if you're using the +5V regulated voltage, I don't really see why there would be a problem.

Agnd and Dgnd usually need tone connected *somewhere*. It's just that it helps if that's one place and back at the power supply rather than close to the chips. That way any noise getting onto the ground has to go "the long way around" to get back.
But I'd definitely try connecting the grounds to see what happens. It might not be "ideal" from a noise p[oint of view, but it might give you a clue what's going on.

HTH,
Tom

[dschwartz]

Thanks so much for your replies..
Update
I added a 220uF to the voltage divider and it seems to work..i have to test some more.
I spotted a short on the output opamp, and that seems to be the cause of the background noises..its pretty quiet now..

And joined together pins 3 and 4 on both chips..no more wobbles!!  Btw that wobble could be an interesting effect..like a shitty tape delay option..you never stop discovering things with this chip.