High Speed data collection (DAC and ADC)

Non-technical related questions should go here.
User avatar
lilltroll
XCore Expert
Posts: 955
Joined: Fri Dec 11, 2009 3:53 am
Location: Sweden, Eskilstuna

Re: High Speed data collection (DAC and ADC)

Postby lilltroll » Fri Apr 23, 2010 11:20 pm

This is the board from Sheldon that I used:

Image

But I wasn't happy about the noise level on the analog rail to the DAC and ADC on the card.
I built an extra card with antialias filter and connected it after the PGA, modified the input multiplexer ...

After all mods, I started to draw my own board, but instead I found XMOS.
So now, I would like to build an improved design of the AD/DA with better SNR + programmable antialias filters as a add-on module-card for the XC-2 with at least 16 bits of resolution.
The Sheldon-card has +- 10V maximum on the input and output, but it can be changed to higher sensitivity with help of the PGA.

You can also take a look at the (old) Master-thesis project.
He made an Ethernet oscilliscope.
Probably not the most confused programmer anymore on the XCORE forum.
User avatar
jrogers
Member
Posts: 13
Joined: Mon Apr 19, 2010 4:02 pm

Postby jrogers » Mon Apr 26, 2010 9:08 pm

lilltroll wrote:This is the board from Sheldon that I used:
...
After all mods, I started to draw my own board, but instead I found XMOS.
So now, I would like to build an improved design of the AD/DA with better SNR + programmable antialias filters as a
...
You can also take a look at the (old) Master-thesis project. He made an Ethernet oscilliscope.
First - I am very happy that there is already interest in high speed ADC/DAC!!

Next - I spent some time reading the things that you referenced (and a few other things), but I could not find the master's thesis... where can I find that?

I notice that you have a number of projects, are you still pursuing/interested in this sort of high speed ADC/DAC project? If so maybe we can team up? Are others interested?

Also I am very interested in the USB audio board ( I just purchased one) and I noticed that several of your projects involve audio. I won't take this thread off on a tangent; I will put my audio comments in those threads, but I wanted to just mention it here because they are somewhat related.

Thanks again for your interest and for your comments!!! - John
User avatar
shawn
XCore Addict
Posts: 238
Joined: Thu Dec 17, 2009 5:15 am

Postby shawn » Tue Apr 27, 2010 6:41 am

I like the idea of high-speed adc's and dac's for agile radio from 3ghz to 300 mhz, chopped to 100khz bands. I am also interested in professional sound so it go's with out saying.
User avatar
lilltroll
XCore Expert
Posts: 955
Joined: Fri Dec 11, 2009 3:53 am
Location: Sweden, Eskilstuna

Postby lilltroll » Tue Apr 27, 2010 10:40 am

jrogers wrote:
lilltroll wrote:This is the board from Sheldon that I used:
...
After all mods, I started to draw my own board, but instead I found XMOS.
So now, I would like to build an improved design of the AD/DA with better SNR + programmable antialias filters as a
...
You can also take a look at the (old) Master-thesis project. He made an Ethernet oscilliscope.
First - I am very happy that there is already interest in high speed ADC/DAC!!

Next - I spent some time reading the things that you referenced (and a few other things), but I could not find the master's thesis... where can I find that?

I notice that you have a number of projects, are you still pursuing/interested in this sort of high speed ADC/DAC project? If so maybe we can team up? Are others interested?

Also I am very interested in the USB audio board ( I just purchased one) and I noticed that several of your projects involve audio. I won't take this thread off on a tangent; I will put my audio comments in those threads, but I wanted to just mention it here because they are somewhat related.

Thanks again for your interest and for your comments!!! - John
Audio is 98% of what I am doing with XMOS and 95% of my professional work.
Since - I'm intrested in low-latency I'm very intrested in Hispeed DAC/ADC with good SNR. With the USB Audio card you can get down to ~70us total latency from in to out. That's good - but not good enought in many control applications.

I know almost nothing about the Ethernet things regarding programming - and Windows programming is a :twisted: in my World, but I would like to use it.

I would like to be able to just tap realtime signals with XC-channels - and be able to
show the realtime result in Windows, over both USB2.0 and Ethernet - Like taking the nice parts from LabView. Making control-GUI's is a powerful tool in LabView.

I think my skills would ends in a streaming channel end with ADC/DAC data from a custom made PCB board.
Could you make the rest with interfacing it to the Host?
I'm typically that EE boy, which I realized then I took the course in distributed OS, with all the computer science people 8-)
Probably not the most confused programmer anymore on the XCORE forum.
User avatar
jrogers
Member
Posts: 13
Joined: Mon Apr 19, 2010 4:02 pm

Postby jrogers » Tue Apr 27, 2010 10:19 pm

lilltroll wrote: Ethernet ... Windows programming...l"]
I think my skills would ends in a streaming channel end with ADC/DAC data from a custom made PCB board.
Could you make the rest with interfacing it to the Host? ...
Yes... I have done quite a bit of Windows programming, and quite a bit of network programming... Sounds like a good match!

I have done some more reading of the forum and I have joined a few groups including your (USB Audio Card).

How do we proceed?

I mean, I realize that there probably are no formal or strict rules, I guess I am more asking what your thoughts are on how I should get involved, etc...

John
User avatar
lilltroll
XCore Expert
Posts: 955
Joined: Fri Dec 11, 2009 3:53 am
Location: Sweden, Eskilstuna

Postby lilltroll » Wed Apr 28, 2010 1:54 pm

What about

http://www.xcore.com/projects/high-spee ... ac-and-adc

The nice thing is that I can digitally bypass my part to check if the A/D D/A is working, and you can place a digital signal generator in a thread and build you stuff, before the two parts is added toghether.

Here is an existing example with dual 12-bit 40Msps ADC.
http://archive.xmoslinkers.org/node/284
Probably not the most confused programmer anymore on the XCORE forum.
User avatar
lilltroll
XCore Expert
Posts: 955
Joined: Fri Dec 11, 2009 3:53 am
Location: Sweden, Eskilstuna

Postby lilltroll » Wed Apr 28, 2010 11:55 pm

I would like to use a DDS as a clock generator for the A/D D/A converters.
You will then be able to create any sampling frequency with a few mHz of resolution.
Makes things very easy to work with.

Image
Probably not the most confused programmer anymore on the XCORE forum.
User avatar
jrogers
Member
Posts: 13
Joined: Mon Apr 19, 2010 4:02 pm

Postby jrogers » Thu Apr 29, 2010 12:02 am

lilltroll wrote:What about

http://www.xcore.com/projects/high-spee ... ac-and-adc

The nice thing is that I can digitally bypass my part to check if the A/D D/A is working, and you can place a digital signal generator in a thread and build you stuff, before the two parts is added toghether.

Here is an existing example with dual 12-bit 40Msps ADC.
http://archive.xmoslinkers.org/node/284
This sounds great!!!

And I always like the idea of having tools/techniques/etc for simulating the "parts" of the system. Great for independent development and testing...

Great!!! I am very excited!
User avatar
jrogers
Member
Posts: 13
Joined: Mon Apr 19, 2010 4:02 pm

Postby jrogers » Thu Apr 29, 2010 12:25 am

lilltroll wrote:I would like to use a DDS as a clock generator for the A/D D/A converters.
You will then be able to create any sampling frequency with a few mHz of resolution.
Makes things very easy to work with.
This sounds great too!!!
(although I have to admit that I need to read up on DDS's... I found the data sheet and a "tutorial" http://www.analog.com/static/imported-f ... 2-2-99.pdf

But this is great! Please feel free to keep on selecting/suggesting hardware components. I am pretty sure that at this end of the spectrum (i.e. pure hardware end that is) that you will have most of the input, etc... but I will do my best to keep up so that I have a reasonable level of understanding of the components and the decisions that led to them.

As these ideas start to "firm up", it would be nice to have a "high level" description of the device... for example number of digital inputs and analog outputs, I am fairly wide open. One of each would be ok, but several would probably be better... I would also like to have circuitry on board (or easily interfaced) that would allow for a fairly wide range of input and output voltages. Also there is the question of maximum current of the analog out - in case it needs to drive something that (relatively) substantial current. Anyway... not trying to jump ahead, I just thought that this was a good time to briefly mention some of these things.

Great! Thanks!
User avatar
lilltroll
XCore Expert
Posts: 955
Joined: Fri Dec 11, 2009 3:53 am
Location: Sweden, Eskilstuna

Postby lilltroll » Thu Apr 29, 2010 10:55 am

My idea is to Copy&Paste a mono-channel design. So the card will have space for say 4 channels in, but if you choose to only fit components on one channel that channel will work alone.
That worked nicely for the Sheldon card.

Since the price is rather high for such types of converters, everything should be buffered. That will create possibility for large output current, and also - if you manage to fry the input or output you replace the (low price) buffer-stage.
Probably not the most confused programmer anymore on the XCORE forum.

Who is online

Users browsing this forum: No registered users and 0 guests