Page 1 of 1

FV-1 clock edge

Posted: Fri Jun 25, 2021 6:57 am
by gary00k
Hello, I've got a guestion.
Let's assume I'm running FV-1 with external clock. Is FV-1 sampling signal with rising or falling clock edge?

Re: FV-1 clock edge

Posted: Fri Jun 25, 2021 8:09 am
by frank
Basically neither or both depending how you look at it. The converters are delta-sigma so they don't really have a clock edge you can reference externally. And even if you could access a clock there is nothing you can really do with it as delta-sigma converters are over sampling converters.

Re: FV-1 clock edge

Posted: Fri Jun 25, 2021 11:03 am
by gary00k
Well, that complicates things a bit.
But maybe let me explain; I wanted to make pseudo-stereo circuit based on a multiplexer / demultiplexer. External generator produces main clock signal for FV-1 and the second signal, divided by 2 for de/multiplexers, so they're synchronized. And the multiplexer simply switches L/R inputs and outputs. I cannot do it another way bacause I need first DAC to get wet signal and second DAC to get dry signal, to mix it out of FV-1.

I assumed that if the program is executed with every clock tact, I can substitute input signals every other sample.

Re: FV-1 clock edge

Posted: Fri Jun 25, 2021 2:53 pm
by frank
No way to do this on FV-1. In fact I do not think it is really possible on any device using a delta-sigmal converter. You need a converter like a SAR type converter and I do not think anyone makes one in the audio range anymore, they do not have the performance of a delta-sigma.

If you must run 3 channels of data into the DSP core then you need a processor that can handle 3 or more channels.

Re: FV-1 clock edge

Posted: Fri Jun 25, 2021 11:34 pm
by gary00k
I not sure if you know what I exactly mean.
Anyway, please tell me when the program is execuded? I mean, with rising or falling clock edge?

Re: FV-1 clock edge

Posted: Sat Jun 26, 2021 6:19 am
by gary00k
UPDATE:
Ok, nevermind. I made some tests with channels multiplexing on prototype and it sucks.
Thank you frank for answers.

Re: FV-1 clock edge

Posted: Sat Jun 26, 2021 8:40 am
by frank
gary00k wrote: Fri Jun 25, 2021 11:34 pm I not sure if you know what I exactly mean.
I do but delta-sigma converters do not work the way you need a converter to for this. You are assuming there is a sample and hold that occurs each sample period and delta-sigmas do not do that. SAR converters do work that way.

Re: FV-1 clock edge

Posted: Sat Jun 26, 2021 9:34 am
by gary00k
Yes, but if the program is executed with clock edge, it reads ADC registers in that moment.
And, if clock signals are synchronized like this:
Image
(assuming it's rising edge in this case) ADC register should have value of left or right channel in reading moment.
Correct me if I'm wrong.

Re: FV-1 clock edge

Posted: Sat Jun 26, 2021 10:04 am
by frank
In a delta-sigma (DS) it is not that simple, while we read a specific register at a specific point in time for the value at that time for the processing core the conversion from analog to digital is far more complex. You are thinking how a SAR converter works by taking a sample at the Fs rate, DS converters sample at a multiple of Fs, 64x or 128x so there is NO WAY to do what you are trying to do by multiplexing an input with a DS based device.

So you are wrong on the conversion part, you are thinking of a SAR and not a DS. FV-1 will NOT work as you want in this case and as far as I know there are no audio range SAR converters, the industry moved to DS as they have better performance. Your only choices are to use a different processor with more inputs and DS converters or a 2 input processor with a SAR converter (if you can find one) but the S/N will suffer since SAR do not have the performance of DS.

Re: FV-1 clock edge

Posted: Sun Jun 27, 2021 12:33 am
by gary00k
Ok, I see now.
Thank you.