Generating a high-performance DAC?

Tip / Sign in to post questions, reply, level up, and achieve exciting badges. Know more

cross mob
lock attach
Attachments are accessible only for community members.
LRJ
Level 1
Level 1
First reply posted First question asked First like given

Hey everyone. 🙂

I'm currently working on a project that have very specific requirements, i.e making an Arbitrary Waveform generator, via a reconfigurable DAC.

 

They are as follows:
24-bit accuracy on DAC
16 differential channels for output (32 io-lines in total)
0 - 1MHz signal output - high quality. (less jitter, noise etc)
Remote initialization and reprogramming
Less than 1mS delay for user - from button-press to signal-generation

I have looked into a few designs and such, and from looking at various open source projects and Cypress AN64275 application note, I think some form of "Modulated Parallel IDAC" could be the answer.


I was thinking something like this:
https://media.discordapp.net/attachments/420639460175380481/894220085928001546/AWG-vague_sketch.png?...

The concept is inspired by one of CERN's recent projects, the Scopefun Oscilloscope and Arbitrary Waveform Generator, as well as the Zrna project, which also utilizes configurable logic connected to a microprocessor to generate Arbitrary Waveforms.

What say you, PSoC5 -gurus - Is it possible to do this?

 

 

0 Likes
1 Solution
Len_CONSULTRON
Level 9
Level 9
Beta tester 500 solutions authored 1000 replies posted

LRJ,

I personally think the PSoC5 is great.   However based on your requirements.  The PSoC5 will not support what you're looking to do.

Even if you were to use 32 PSoC5 (one PSoC5 for each Output), the PSoC5 will not effectively achieve 24-bit resolution.  Also the resident 8-bt DACs can update @ 100Ksps not the 1Msps you are looking to achieve.

Many of the other specs are within reach.  However based on my statements above, I'd say the PSoC5 is a deal-killer.

You might think about 32 24-bit DAC ICs connected to the PSoC5 but I assume this is also a deal-killer.

Len
"Engineering is an Art. The Art of Compromise."

View solution in original post

3 Replies
odissey1
Level 9
Level 9
First comment on KBA 1000 replies posted 750 replies posted

LRJ,

What is the intended application of such signal generator? Is this for sound/synth application or a general purpose signal generator? Is it strictly for sine/ramp/triangle generation or MIDI input required? You didn't mention need for synchronizing all 8 outputs.

General statement: the specs requested are very tight due to requested bit resolution and update rate. If you notice, the DACs are grouped into high-speed / low resolution (>1Msample/sec, <14bit), or low-speed / high-resolution (<370ksamples/sec, 24-32bit). The latter intended for audio applications.

    None of the projects you mentioned fits into the specs you interested. The AN64275 describes several ways to improve PSoC5 VDAC8 bit resolution to (9-11) bits in expense of the update rate. It won't work for you. 

    The ScopeFun is only dual 12-bit DAC, no help either. The Zrna project is based on Anadigm's FPAA with 8-bit / 300kHz output. Not even close.

    I would concentrate on narrowing down specs first. I believe that some professional equipment of such sort exist, priced in $40k range.

    If sine/ramp/tri output is all you need, i would consider using DDS chips from Analog Devices, which already have high-speed DAC built-in with 14-16-bit output. A number of such DDS can be controlled by a single PSoC or similar micro.

      As an upgrade, it is possible to utilize 8 PSoC chips, each connected to its own 32-bit DAC. The sync between the DACs might be problematic.

       A general approach to the project would be based on Xilinx or Altera SoC, using embedded Linux core for communication and wave synthesis, and FPGA for DACs interfacing. 

Len_CONSULTRON
Level 9
Level 9
Beta tester 500 solutions authored 1000 replies posted

LRJ,

I personally think the PSoC5 is great.   However based on your requirements.  The PSoC5 will not support what you're looking to do.

Even if you were to use 32 PSoC5 (one PSoC5 for each Output), the PSoC5 will not effectively achieve 24-bit resolution.  Also the resident 8-bt DACs can update @ 100Ksps not the 1Msps you are looking to achieve.

Many of the other specs are within reach.  However based on my statements above, I'd say the PSoC5 is a deal-killer.

You might think about 32 24-bit DAC ICs connected to the PSoC5 but I assume this is also a deal-killer.

Len
"Engineering is an Art. The Art of Compromise."

Cheers for taking the time to reply, Len. 🙂

Yeah, I'm starting to realize just how much one would need to compromise to achieve this, with a price-tag that makes sense for the contract. There really aren't many shortcuts when it comes to analogue design.

0 Likes