PT2399 defeated by.. a few LEDs?

[puretube]

What is:

the bank of 5 warm white LEDs i'm using

?
Is it: 5 LEDs series-wired (or parallel?) directly to the powersupply? (without current-limiting resistor).
Or is it: 5 LEDs, each with its individual series-resistor wired to the PS (parallel to each other)?

[anotherjim]

I think you can increase the 10R in the 15v feed. 47R or 56R 1/2W resistor. This will take some of the heat off the 5v reg and reduce the power-on surge currents.

[HeavyFog]

What is:

the bank of 5 warm white LEDs i'm using

?
Is it: 5 LEDs series-wired (or parallel?) directly to the powersupply? (without current-limiting resistor).
Or is it: 5 LEDs, each with its individual series-resistor wired to the PS (parallel to each other)?

It's 5 LEDs, each with their own CLR (about 1k1), wired in parallel to each other. They're connected to the +15v line after the 10R/10uf filter, on the same rail that sends DC power to both delay boards and the VU driver circuit. I've tried connecting it before the filter, and also to the -15v line, and both resulted in the hum.

[HeavyFog]

I think you can increase the 10R in the 15v feed. 47R or 56R 1/2W resistor. This will take some of the heat off the 5v reg and reduce the power-on surge currents.

The way I've had it set up has the +15v feed going to the 5v regulator sourced from before the RC filter on that line. One quirk though is that the filter for the 5v line is actually post regulator. The filter is located on the board right there with the relays since I originally intended to source it from the 5v line from one of the delay boards, and i just didn't remove it

[anotherjim]

Argggh! Then add a 10R resistor in the feed to the relay 5v regulator. Use that to find the current taken by the relays and the voltage at the regulator input pin. Increase the resistor above 10R as necessary so the relay regulator gets around 9v at the input pin with maximum current drawn from the relays.
If they are fed with 5v, 1k series resistance is ok to run the LEDs off.

[HeavyFog]

Argggh! Then add a 10R resistor in the feed to the relay 5v regulator. Use that to find the current taken by the relays and the voltage at the regulator input pin. Increase the resistor above 10R as necessary so the relay regulator gets around 9v at the input pin with maximum current drawn from the relays.
If they are fed with 5v, 1k series resistance is ok to run the LEDs off.

I've measured the input voltage at about 14.8vDC, which is the normal voltage on the "+15v" rail on my psu. I can actually measure a drop of about 0.1v on this rail when the relays are energized. I measured the current draw from the relays as well and it comes down to 26mA, also pretty close to the expected 28mA from the datasheet.

The LEDs are running off the +15v rail post RC filter. Also tried it with the -15v rail but same result.

[pinkjimiphoton]

my first thought is disconnect the leds and see if it goes away.

some led's can put out ungodly amounts of hash into the power, for no apparent reason in my experience.

i've had a lot of fuzz circuits where white and uv leds in particular led to audible interference with some circuits.
if you can temp disconnect the power to the backlight just to confirm its NOT the led's, i'd be curious.

i believe sometimes leds can be damaged just enough to cause all kinds of weird noise issues.

if the noise goes away, decouple them more, i guess, if no change, my bad, but ya know its not the led's contributing to it.

it could be a shite 2399 too

[HeavyFog]

my first thought is disconnect the leds and see if it goes away.

some led's can put out ungodly amounts of hash into the power, for no apparent reason in my experience.

i've had a lot of fuzz circuits where white and uv leds in particular led to audible interference with some circuits.
if you can temp disconnect the power to the backlight just to confirm its NOT the led's, i'd be curious.

i believe sometimes leds can be damaged just enough to cause all kinds of weird noise issues.

if the noise goes away, decouple them more, i guess, if no change, my bad, but ya know its not the led's contributing to it.

it could be a shite 2399 too

When i disconnect the LEDs it does stop, but i don't know if the core issue is the LEDs or the relays. Gotta try and narrow it down further. I've got some more of the same ones i used in so it try breadboarding it to see if i can replicate the noise with other colours.

Definitely not the pt2399 though, if i remove it all together from the socket i can still hear the noise, albeit quieter. Definitely something to do with the power supply

[HeavyFog]

Haven't had much time to look at this lately because of work, but i have managed to make some progress with it. What i found is that aside from disconnecting the LEDs, i can also make the hum mostly stop by disconnecting one of the 2 relay coils from power. After that, increasing the current limiting resistors on each LED to 1.8k from 1.2k eliminated the hum. If i put the other relay coil back in after that change the hum comes back at full volume. Still hard to say what the source is, the relays or the LEDs, but im gonna keep investigating it, might try giving the second relay its own PSU

[bartimaeus]

could you switch to a Single coil latch type relay? that would totally eliminate the extra current draw when the pedal is active. i realize you'd have to redesign your relay circuitry to do that, but could be worth it.

[merlinb]

When i disconnect the LEDs it does stop, but i don't know if the core issue is the LEDs or the relays.

The core issue is drawing too much current from an inadequate supply. Doesn't matter if it's LEDs or relays.

[HeavyFog]

Well, i finally figured out what was going on here. The problem wasn't anything to do with the delay, relays, or LEDs, but rather the power supply. The synthesizers.com QPS6 I'm using seems to have a quirk where if you have enough of a load on the +/-15vDC lines, you need to also have some load on the +5vDC line as well, or else the +/-15v rails start oscillating, which is the source of the hum i was hearing. Usually this hasn't been an issue because the pilot light on the main modular cabinet runs off the 5v line, which is a sufficient load to keep thing stable, but the delay doesn't use the 5v line at all so when i run it without the main synth cabinet also powered i get the hum!

The solution was to simply put a 470ohm to 1K value resistor between the +5v pin on the power input to ground and suddenly it becomes stable! No idea why the power supply acts like this or why the LEDs in particular seem to be enough to destabilize it, but it's probably why dotcom discontinued the original QPS6.

[pinkjimiphoton]

it was an lfo!

[HeavyFog]

Ha, more or less. I really should upgrade to a better power supply. This one has given me too much trouble

[anotherjim]

Quite a few switchmode multi-voltage supplies really only regulate one output voltage, often that is the +5v. The other voltages are not actually regulated but go up or down depending on how hard the switcher works to keep the +5v happy.

With little load on the +5, the switcher starts and sees this voltage overshoot so stops until it drops then repeats and the other outputs go up and down with it. With enough 5v load it will settle to output 5v and the other outputs settle down too.

I don't think any of us were expecting a synth to have this type of power supply. A mixer I had to fix had one and although it regulated the +15v instead of the +5v which relied on ordinary voltage regulators elsewhere.

[PRR]

> power supply. This one has given me too much trouble

Amplifying on a untrustworthy power supply is like building your house in the swamp. If you can't trust your foundations, everything is dubious.

[HeavyFog]

Quite a few switchmode multi-voltage supplies really only regulate one output voltage, often that is the +5v. The other voltages are not actually regulated but go up or down depending on how hard the switcher works to keep the +5v happy.

With little load on the +5, the switcher starts and sees this voltage overshoot so stops until it drops then repeats and the other outputs go up and down with it. With enough 5v load it will settle to output 5v and the other outputs settle down too.

I don't think any of us were expecting a synth to have this type of power supply. A mixer I had to fix had one and although it regulated the +15v instead of the +5v which relied on ordinary voltage regulators elsewhere.

That would explain it. First red flag for me should have been learning it's actually a rebranded medical power supply. Didn't have any issue with it when I first got my synth built, but as I've built onto the system more and more issues have been popping up with power, so it may be time to bite the bullet and just order a better one. Would probably go a long way to improving the thing