Q About PT2399 Supply Voltage

[Paul Marossy]

Dumb question time.  

I was looking at the PT2399 data sheet and it says that the absolute maximum supply voltage is 6.5 volts. So what would happen if you used an LM317 set to say 6V? Would there be any benefit to doing that? Seems like maybe you would get a little higher peak to peak voltage, which theoritically should lessen any distortion you might get.

Or is this wishful thinking and all that's going to happen is maybe more noise?

[peterg]

Paul

Take a look at some circuits using the PT2399 chip. A 7805 voltage regular is used for most stomp boxes running off 9V to keep the chip at 5V

[Mark Hammer]

It's a digital chip, Paul.  A bit set high is a bit set high.  The datasheet will declare something somewhat higher than +5V because there are a lot of ways to arrive at 5v, some a little more accurate or stable than others.  The 5V ceiling allows for reasonable error, while still keeping the magic blue smoke inside the chip.

Will a higher supply voltage help the analog part of the overall circuit?  Quite possibly.

[Pojo]

I can't say with 100% certainty but here's my take on it, I could be dead wrong. Since they're digital, the supply voltage is there solely to provide an 'on' reference, with 0V being 'off'...or 1 and 0. The noise you get with longer delay times is because of the limited memory size and I'm guessing there's an algorithm that lowers the resolution of the audio so it can store a longer sound clip within the available ram. So, I'm thinking as long as it's there's enough voltage for it to function and its low enough to burn it up, the exact amount doesn't make a difference.

Again, could be totally wrong here.

[Paul Marossy]

Will a higher supply voltage help the analog part of the overall circuit?  Quite possibly.

That was my initial thought, but I know that digital chips are pretty much designed to operate on 5V and going above that is a "do so at your own risk" kind of thing.

I can't say with 100% certainty but here's my take on it, I could be dead wrong. Since they're digital, the supply voltage is there solely to provide an 'on' reference, with 0V being 'off'...or 1 and 0. The noise you get with longer delay times is because of the limited memory size and I'm guessing there's an algorithm that lowers the resolution of the audio so it can store a longer sound clip within the available ram. So, I'm thinking as long as it's there's enough voltage for it to function and its low enough to burn it up, the exact amount doesn't make a difference.

From what I understand, it's the current that kills the PT2399s more than the supply voltage. The datasheet even says that they don't recommend less than 1K for the "R" resistor on start up. I understand that something less than that can possibly damage the chip.

[defaced]

I remember there being some discussions about supply voltage on PT2399s a while ago when there was alot of work with people building choruses with them.  Might be worth searching out.  I seem to remember that the higher voltage effected delay time, and over some level it just killed the chip. 

[psychedelicfish]

I don't think lock up is caused by over currenting the chip, I think it's to do with the clock not being able to start

[psychedelicfish]

Supplying the chip with 6V or so will increase the clock freq, which decreases delay time

[Paul Marossy]

I don't think lock up is caused by over currenting the chip, I think it's to do with the clock not being able to start

OK, makes sense. Someome suggested it could lead to "catastrophic failure" if you used smaller than 1K for "R" resistor. I dunno, I'm taking their word for it.  

Supplying the chip with 6V or so will increase the clock freq, which decreases delay time

Sounds like 5V is the best all around compromise then. The question is how much do we trust that data sheet? Some people have suggested that those noise figures listed are wishful thinking.

[defaced]

So going back to your original thought about headroom, do you have an issue with insufficient headroom in the circuit?  

4.5v to 5.5v is listed in the data sheet as the recommended operating voltage.  6.5v is the "guaranteed not to blow up" rating (at room temp no less), not the "run it here for the rest of its life" rating.  

Edit: thinking out loud.  The AD conversion is going to only have so many bits to represent a signal amplitude.  Assuming it uses all of its bits to do conversion, then "increasing headroom", is just going to force the chip to space out its bits and get poorer resolution (1v in a 5v system is 20%, so that might be noticeable).... yes?

[R O Tiree]

IIRC, it is applying power with less than 1k at "R" that causes the chip to lock up, sometimes catastrophically. After it has started successfully, you can lower "R", if that's your bag.

[Paul Marossy]

So going back to your original thought about headroom, do you have an issue with insufficient headroom in the circuit?  

4.5v to 5.5v is listed in the data sheet as the recommended operating voltage.  6.5v is the "guaranteed not to blow up" rating (at room temp no less), not the "run it here for the rest of its life" rating.  

Edit: thinking out loud.  The AD conversion is going to only have so many bits to represent a signal amplitude.  Assuming it uses all of its bits to do conversion, then "increasing headroom", is just going to force the chip to space out its bits and get poorer resolution (1v in a 5v system is 20%, so that might be noticeable).... yes?

Not a problem with headroom, was just thinking that if I could make something better, why not? 

IIRC, it is applying power with less than 1k at "R" that causes the chip to lock up, sometimes catastrophically. After it has started successfully, you can lower "R", if that's your bag.

Yes, this is true. I have experienced it firsthand. 

[merlinb]

Edit: thinking out loud.  The AD conversion is going to only have so many bits to represent a signal amplitude.  Assuming it uses all of its bits to do conversion, then "increasing headroom", is just going to force the chip to space out its bits and get poorer resolution (1v in a 5v system is 20%, so that might be noticeable).... yes?

The PT uses sigma modulation, not the "multiple levels" type of A-D conversion, so this doesn't apply. With sigma moduilation, the faster the clock, the better the resolution.

[defaced]

Cool, thanks for the explanation. 

[Paul Marossy]

Edit: thinking out loud.  The AD conversion is going to only have so many bits to represent a signal amplitude.  Assuming it uses all of its bits to do conversion, then "increasing headroom", is just going to force the chip to space out its bits and get poorer resolution (1v in a 5v system is 20%, so that might be noticeable).... yes?

The PT uses sigma modulation, not the "multiple levels" type of A-D conversion, so this doesn't apply. With sigma moduilation, the faster the clock, the better the resolution.

Looking at the data sheet, that must explain why the sound quality goes down as the delay time goes up - the longer the delay time, the slower the clock is operating and hence the lower the resolution. Interesting....