Precision current control with Digital Potentiometers

[Taylor]

2 seconds = 16khz sample rate = 8khz max frequency without aliasing (see Nyquist).

1 second = 32k sample rate = 16k max frequency

Steep filters are easy in the FV-1, so you can filter aliasing without losing as much of the instrument signal's bandwidth. Not so with the PT2399, so longer delays will always mean more grunginess and/or more loss of high frequencies.

You're right that DIYers are scared of SMT.

BTW, if 14 bit is too low bit depth for you, run fast from the PT2399, which AFAICT is 10 bits.

[cloudscapes]

I thought the 2399 was 1bit, but at several magnitudes of the speed of conventional sampling?

[Taylor]

Well, I did see something about that while googling, but I also saw something from Princeton saying it was 10 bits. They don't say anything about it, even in the more detailed datasheet. In any case I think it's fair to say that the PT2399 sounds much lower fidelity than the FV1. That's neither bad nor good necessarily, of course.

[MoltenVoltage]

the FV1 doesn't leave a platform for a microcontroller product.

[Taylor]

_______________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________

[octfrank]

On the 14-bit issue: Keep in mind we are talking a floating point format so on real loud sounds a few bits will be truncated (but you aren't gonna hear it because, well, it's loud) and on soft sounds you get the full 24-bits of the MAC/converters.

As far as adding a pure delay as a default program, not going to happen. The programs are in a ROM in the FV-1 so changing them requires at least 2 new masks for production and if you have done custom chips you know what a mask set costs!

[Taylor]

_______________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________

[MoltenVoltage]

For those interested, the relationship between frequency and delay time on the PT2399 is inversely proportional +/- < 1% from my readings.

[puretube]

For those interested, the relationship between frequency and delay time on the PT2399 is inversely proportional +/- < 1% from my readings.

Q.E.D.  ...

[slacker]

At a slight tangent, has anyone ever tried running a suitable speed clock signal into pin 5? Depending on where this is tapped from and whether or not it's buffered you might be able to completely swamp the internal clock signal and control the speed that way. I believe circuit benders do a similar thing with crystal driven circuits.

[JKowalski]

At a slight tangent, has anyone ever tried running a suitable speed clock signal into pin 5? Depending on where this is tapped from and whether or not it's buffered you might be able to completely swamp the internal clock signal and control the speed that way. I believe circuit benders do a similar thing with crystal driven circuits.

I feel like this forum has come around to this question (PT2399 tap tempo) a hundred times or more...   

The answer was a definite no I believe, it is not like a conventional clock - the adjustment terminal is a completely different beast.

[ElectricDruid]

I feel like this forum has come around to this question (PT2399 tap tempo) a hundred times or more...   
The answer was a definite no I believe, it is not like a conventional clock - the adjustment terminal is a completely different beast.

It *does* alway seem to come back to this...at least, it will until someone comes up with the killer solution!

Part of the problem is the poor documentation. Even the better version of the PT2399 datasheet (the one with the block diagram) doesn't include everything.
http://www.synthdiy.com/files/2007/PT2399.pdf

I think Slacker might be onto something. The datasheet shows the clock output pin 5 as unconnected. But presumably it is connected to the VCO output, which is also the clock *input* for the main delay block (whatever it really is - they're not saying). So perhaps you could feed a clock *into* the clock output, whilst tying something to pin 6 to ensure the VCO is disabled. Or alternatively feed 240V into pin 6 - that'd disable the VCO...
The datasheet also gives no clue at all as to why the Little Angel Chorus trick of feeding modulation voltages into the pin 2 ref works. It's frustrating, really.

T.

[octfrank]

I always wonder why people don't link to the datasheet at the manufacturer's site which is usually more recent, like:

http://www.princeton.com.tw/downloadprocess/downloadfile.asp?mydownload=PT2399_1.pdf

Which does show more info.

[slotbot]

2 seconds = 16khz sample rate = 8khz max frequency without aliasing (see Nyquist).

1 second = 32k sample rate = 16k max frequency

Steep filters are easy in the FV-1, so you can filter aliasing without losing as much of the instrument signal's bandwidth. Not so with the PT2399, so longer delays will always mean more grunginess and/or more loss of high frequencies.

AFAIK  You cant filter aliasing once it has occurred. Aliasing = a higher frequency appearing as a lower one after sampling (ie getting a new Alias). For example if you sample at 16 KHz then play in a signal at 10 kHz, wihtout external filtering that will show up as 6 kHz when you play the captured sound back at the same rate. Internally you cant filter out the 10Khz because it never even gets captured as such.

I thought the 2399 was 1bit, but at several magnitudes of the speed of conventional sampling?

i am also pretty sure this is how the pt2399 works. This is why you get huge volume changes when you rapidly increase or decrease the sampling time etc. Sony was actually working on a standard based on this principle. http://en.wikipedia.org/wiki/Super_Audio_CD and many other devices work the same way.

[Taylor]

What I was talking about is software sample rate reduction for longer delays. So I have a program (most of the clever stuff was done by another guy named Don Stavely) where only every n sample actually goes to the output. By doing a 1-second delay but using this software SRR, we can have any delay length we want, with the proportional loss in sample rate (whereas in the FV1 you can't clock it below 16khz externally). So even a 2-minute long delay is possible with ridiculously unusable amounts of aliasing.

But the slick part is that since this SRR happens in the code, we can digitally filter the signal before the reduction, so it is actually possible to stop the aliasing from happening. I still like some filtering on the end as well for the longer delays (16 seconds...) and this ends up sounding "analogish" I guess.

[slotbot]

Ah ok i get what your saying. For this it makes sense. What is your effective sampling rate at 2 minits delay? about 260 hz?

[JKowalski]

I feel like this forum has come around to this question (PT2399 tap tempo) a hundred times or more...   
The answer was a definite no I believe, it is not like a conventional clock - the adjustment terminal is a completely different beast.

It *does* alway seem to come back to this...at least, it will until someone comes up with the killer solution!

Part of the problem is the poor documentation. Even the better version of the PT2399 datasheet (the one with the block diagram) doesn't include everything.
http://www.synthdiy.com/files/2007/PT2399.pdf

I think Slacker might be onto something. The datasheet shows the clock output pin 5 as unconnected. But presumably it is connected to the VCO output, which is also the clock *input* for the main delay block (whatever it really is - they're not saying). So perhaps you could feed a clock *into* the clock output, whilst tying something to pin 6 to ensure the VCO is disabled. Or alternatively feed 240V into pin 6 - that'd disable the VCO...

I definitely remember reading about someone trying this and failing. It's probably a buffered output.

The datasheet also gives no clue at all as to why the Little Angel Chorus trick of feeding modulation voltages into the pin 2 ref works. It's frustrating, really.

T.

I know! Totally wierd.

Maybe we could somehow contact princeton and clarify things, though I doubt they would care to indulge us. Unless we pose as a buyer thinking about a 2,000,000 unit order... 

[slotbot]

what about using a frequency counter as feedback device? measure the clock out pin frequency and have the mcu adjust the R accordingly.

[Taylor]

Ah ok i get what your saying. For this it makes sense. What is your effective sampling rate at 2 minits delay? about 260 hz?

Yep!  Which means max 130hz representable without aliasing. I only have a 4-pole lowpass at the beginning and end, but with something more aggressive at the front end this could actually be usable as a bass looper. The more obvious use is just insane weirdness, though, which I know from your products is something with which you're familiar.

[slotbot]

The more obvious use is just insane weirdness, though, which I know from your products is something with which you're familiar.

lol. do you have any clips of this?

i tried to start making some less weird thngs, i never got into teh FV-1 yet but i am still farting around with dsPiC. i started to make a sort of tremelo ring mod combo here http://www.youtube.com/watch?v=eHDSSkvuBDM.

but not to hijack the thread....

I know pt2399 is dirt dirt cheap but i often wonder, if you think about all the effort you put into making a pedal, is the difference between a 50 cent or 5 dollar IC really that meaningful? obviously from mass production standpoint yes but interms of elbow grease etc to do one or two it really seems a bit silly to me.


edit: sorry i didnt mean using the pt2399 in general seems silly, actually its super awesome. but i meant to go through all teh troubl of developing trap tempo, whenits not that great a delay to begin with.