looking for DSP with fast AD/DA (not audio)

Hi,

i am looking for a DSP to control my experimental setup. I need to
measure several signals simultaneously at rates up to 200 kHz, do some
digital filtering (PID-like) and drive one or two analog outputs at
rates up to 30kHz. As a physicist and DSP newbie I prefer to spend my
time on my set-up and with  programming the DSP, so the electronics
should preferably be plug&play.

I found a nice system (Adwin-gold, www.adwin.de) that does most of
what I want. It is based on the Analog 21062 and comes in a sealed box
with lots of BNC plugs. Only problem is price: 5000 Euro without the
extras.

-does anybody know other products that are similar? I looked at the EZ
boards from Analog, but they are mostly oriented at sound (usually
only 1 channel and sample rate is just to slow for me).
-does anybody have experience with the Adwin?

Thanks,
Bas

You would need t be a little more specific. For example, what do you mean by
"simultaneously" - do you need parallel sample-and-hold amplifiers, and if
not (i.e. "burst" sampling is employed) are all inputs set to the same gain?
Also, does acquisition need to be continuous - gap-free, or can there be
some inter-burst period?


Bas wrote in message <8fe46d4f.0405030618.3085d3fd@posting.google.com>...
>Hi,
>
>i am looking for a DSP to control my experimental setup. I need to
>measure several signals simultaneously at rates up to 200 kHz, do some
>digital filtering (PID-like) and drive one or two analog outputs at
>rates up to 30kHz. As a physicist and DSP newbie I prefer to spend my
>time on my set-up and with  programming the DSP, so the electronics
>should preferably be plug&play.
>
>I found a nice system (Adwin-gold, www.adwin.de) that does most of
>what I want. It is based on the Analog 21062 and comes in a sealed box
>with lots of BNC plugs. Only problem is price: 5000 Euro without the
>extras.
>
>-does anybody know other products that are similar? I looked at the EZ
>boards from Analog, but they are mostly oriented at sound (usually
>only 1 channel and sample rate is just to slow for me).
>-does anybody have experience with the Adwin?
>
>Thanks,
>Bas

Bas wrote:
> Hi,
> 
> i am looking for a DSP to control my experimental setup. I need to
> measure several signals simultaneously at rates up to 200 kHz, do some
> digital filtering (PID-like) and drive one or two analog outputs at
> rates up to 30kHz. As a physicist and DSP newbie I prefer to spend my
> time on my set-up and with  programming the DSP, so the electronics
> should preferably be plug&play.
> 
> I found a nice system (Adwin-gold, www.adwin.de) that does most of
> what I want. It is based on the Analog 21062 and comes in a sealed box
> with lots of BNC plugs. Only problem is price: 5000 Euro without the
> extras.
> 
> -does anybody know other products that are similar? I looked at the EZ
> boards from Analog, but they are mostly oriented at sound (usually
> only 1 channel and sample rate is just to slow for me).
> -does anybody have experience with the Adwin?
> 
> Thanks,
> Bas


Hi,

I'm involved in a similar problem.  Here is what I have been looking at:


http://www.innovative-dsp.com/


-- 
____________________________________
Christopher R. Carlen
Principal Laser/Optical Technologist
Sandia National Laboratories CA USA
crcarle@sandia.gov

bassw@hotmail.com (Bas) wrote in 
news:8fe46d4f.0405030618.3085d3fd@posting.google.com:

> Hi,
> 
> i am looking for a DSP to control my experimental setup. I need to
> measure several signals simultaneously at rates up to 200 kHz, do some
> digital filtering (PID-like) and drive one or two analog outputs at
> rates up to 30kHz. As a physicist and DSP newbie I prefer to spend my
> time on my set-up and with  programming the DSP, so the electronics
> should preferably be plug&play.
> 
> I found a nice system (Adwin-gold, www.adwin.de) that does most of
> what I want. It is based on the Analog 21062 and comes in a sealed box
> with lots of BNC plugs. Only problem is price: 5000 Euro without the
> extras.
> 
> -does anybody know other products that are similar? I looked at the EZ
> boards from Analog, but they are mostly oriented at sound (usually
> only 1 channel and sample rate is just to slow for me).
> -does anybody have experience with the Adwin?
> 
> Thanks,
> Bas

We have a 2 channel board that can sample at 192kHz. Its called our 
dspstak c192k22. Our dspstak 21262sx DSP engine is based on a third 
generation SHARC (ADSP-21262). 

We also have a 4 channel I/O module that samples to 96k. You can use two 
of them for an 8 channel system. 

Details are on our web site: www.danvillesignal.com

-- 
Al Clark
Danville Signal Processing, Inc.
--------------------------------------------------------------------
comp.dsp conference July 28 - Aug 1, 2004

details at http://www.danvillesignal.com/index.php?id=compdsp
email: compdsp@danvillesignal.com

Who says you can't teach an old dog a new DSP trick?

Al Clark wrote:

> bassw@hotmail.com (Bas) wrote in
> news:8fe46d4f.0405030618.3085d3fd@posting.google.com:
> 
>> Hi,
>> 
>> i am looking for a DSP to control my experimental setup. I need
>> to measure several signals simultaneously at rates up to 200 kHz,
>> do some digital filtering (PID-like) and drive one or two analog
>> outputs at rates up to 30kHz. As a physicist and DSP newbie I
>> prefer to spend my
>> time on my set-up and with  programming the DSP, so the
>> electronics should preferably be plug&play.
>> 
>> I found a nice system (Adwin-gold, www.adwin.de) that does most
>> of what I want. It is based on the Analog 21062 and comes in a
>> sealed box with lots of BNC plugs. Only problem is price: 5000
>> Euro without the extras.
>> 
>> -does anybody know other products that are similar? I looked at
>> the EZ boards from Analog, but they are mostly oriented at sound
>> (usually only 1 channel and sample rate is just to slow for me).
>> -does anybody have experience with the Adwin?
>> 
>> Thanks,
>> Bas
> 
> We have a 2 channel board that can sample at 192kHz. Its called
> our dspstak c192k22. Our dspstak 21262sx DSP engine is based on a
> third generation SHARC (ADSP-21262).
> 

I checked but I didn't find a hint in the web reference you gave, if 
your boards are using Sigma/Delta-ADCs. I guess so.

Such ADCs contain a delay line which delays the converted signal for 
some milliseconds (the incorporated FIR filters require this).

If Bas wants to build PID controllers, this delay is critical. 
Especially if very direct/fast controlling is required, this might 
lead to instablility.

Bernhard

Bas wrote:

> Hi,
> 
> i am looking for a DSP to control my experimental setup. I need to
> measure several signals simultaneously at rates up to 200 kHz, do
> some digital filtering (PID-like) and drive one or two analog
> outputs at rates up to 30kHz. As a physicist and DSP newbie I
> prefer to spend my
> time on my set-up and with  programming the DSP, so the
> electronics should preferably be plug&play.
> 
> I found a nice system (Adwin-gold, www.adwin.de) that does most of
> what I want. It is based on the Analog 21062 and comes in a sealed
> box with lots of BNC plugs. Only problem is price: 5000 Euro
> without the extras.
> 
> -does anybody know other products that are similar? I looked at
> the EZ boards from Analog, but they are mostly oriented at sound
> (usually only 1 channel and sample rate is just to slow for me).
> -does anybody have experience with the Adwin?
> 
> Thanks,
> Bas

Hi Bas,

The ADwinGold seems to have sufficient fast ADCs (the 12bit 
converters should  be able to deal with 200kS/s). 
If you are forced to work under Windows, and don't want to do low 
level programming, it's certainly an interesting approach.

As another approach, I'd propose to check PCI measurement cards 
(National etc..) which can be used in a standard PC, and then, 
instead of using DSP programs, do all the signal processing on the 
PC processor.

For most applications which deal with signals of 200kS/s (even in 
multichannel applications), nowadays PCs should be fast enough 
to do standard signal processing.

The only "difficult"  issue is to overcome the real time 
restrictions of the OS; however it is possible even on Windows.

If you are free to work with a different OS, try Linux (if 
necessary: extended with a real-time package as RTlinux) on a 
standard PC, select an appropriate PCI card (for which Linux 
drivers are available), and you're certainly done with most of the 
hardware issues. Programming on the PC platform might be much more 
convenient.

If you really want to work on a DSP basis, Sharc is certainly a good 
choice. Maybe you should combine any Sharc board with an ADC eval 
board (the SPORT IO channel provides a good means to couple them) 
and, if necessary, with an additional DAC board.
By daisy-chaining you should be able to link at least 8 or 16 boards 
without difficulties. Check the delays which the converters insert, 
if they are part of the controller loop.

Bernhard 

> You would need t be a little more specific. For example, what do you mean by
> "simultaneously" - do you need parallel sample-and-hold amplifiers, and if
> not (i.e. "burst" sampling is employed) are all inputs set to the same gain?
> Also, does acquisition need to be continuous - gap-free, or can there be
> some inter-burst period?

To be a little bit more specific with my wish list:
-I have 2 input signals that are in quadrature. They need to be
sampled exactly at the same time to get an accurate phase measurement.
Sample rate up to 150 kHz at regular intervals (sustained, so no burst
modes), 12-bits.
-At least 1 but preferably 2 or 3 extra input channels that are not so
time critical. Should be able to achieve sample rate of 30 kHz at some
times.
-at least 2 output channels, 16 bit would be nice. Rates up to 30 kHz
-a handful of digital IO lines
-PID-like digital filter running at 30kHz
-a data pipe to a host computer to download a measurement every now
and then. Not so time critical, but must be able to empty the memory
faster than the ADC can fill it. Matlab or LabView interface if
possible.

The end goal is to stabilize a piezo tunable laser diode to some
reference interferometers.

Cheers,
Bas

Bernhard Holzmayer <holzmayer.bernhard@deadspam.com> wrote in 
news:1932784.DBS32x3I3a@holzmayer.ifr.rt:

> Al Clark wrote:
> 
>> bassw@hotmail.com (Bas) wrote in
>> news:8fe46d4f.0405030618.3085d3fd@posting.google.com:
>> 
>>> Hi,
>>> 
>>> i am looking for a DSP to control my experimental setup. I need
>>> to measure several signals simultaneously at rates up to 200 kHz,
>>> do some digital filtering (PID-like) and drive one or two analog
>>> outputs at rates up to 30kHz. As a physicist and DSP newbie I
>>> prefer to spend my
>>> time on my set-up and with  programming the DSP, so the
>>> electronics should preferably be plug&play.
>>> 
>>> I found a nice system (Adwin-gold, www.adwin.de) that does most
>>> of what I want. It is based on the Analog 21062 and comes in a
>>> sealed box with lots of BNC plugs. Only problem is price: 5000
>>> Euro without the extras.
>>> 
>>> -does anybody know other products that are similar? I looked at
>>> the EZ boards from Analog, but they are mostly oriented at sound
>>> (usually only 1 channel and sample rate is just to slow for me).
>>> -does anybody have experience with the Adwin?
>>> 
>>> Thanks,
>>> Bas
>> 
>> We have a 2 channel board that can sample at 192kHz. Its called
>> our dspstak c192k22. Our dspstak 21262sx DSP engine is based on a
>> third generation SHARC (ADSP-21262).
>> 
> 
> I checked but I didn't find a hint in the web reference you gave, if 
> your boards are using Sigma/Delta-ADCs. I guess so.
> 
> Such ADCs contain a delay line which delays the converted signal for 
> some milliseconds (the incorporated FIR filters require this).
> 
> If Bas wants to build PID controllers, this delay is critical. 
> Especially if very direct/fast controlling is required, this might 
> lead to instablility.
> 
> Bernhard
> 
> 

Yes, I missed the feedback issue and the boards mentioned do use sigma 
delta converters which as you noted have significant group delay.

We have plans to manufacture SAR based dspstak I/O modules but nothing at 
present. We are always interested in suggestions.

-- 
Al Clark
Danville Signal Processing, Inc.
--------------------------------------------------------------------
comp.dsp conference July 28 - Aug 1, 2004

details at http://www.danvillesignal.com/index.php?id=compdsp
email: compdsp@danvillesignal.com

Who says you can't teach an old dog a new DSP trick?

Bas wrote:
>>You would need t be a little more specific. For example, what do you mean by
>>"simultaneously" - do you need parallel sample-and-hold amplifiers, and if
>>not (i.e. "burst" sampling is employed) are all inputs set to the same gain?
>>Also, does acquisition need to be continuous - gap-free, or can there be
>>some inter-burst period?
> 
> 
> To be a little bit more specific with my wish list:
> -I have 2 input signals that are in quadrature. They need to be
> sampled exactly at the same time to get an accurate phase measurement.
> Sample rate up to 150 kHz at regular intervals (sustained, so no burst
> modes), 12-bits.
> -At least 1 but preferably 2 or 3 extra input channels that are not so
> time critical. Should be able to achieve sample rate of 30 kHz at some
> times.
> -at least 2 output channels, 16 bit would be nice. Rates up to 30 kHz
> -a handful of digital IO lines
> -PID-like digital filter running at 30kHz
> -a data pipe to a host computer to download a measurement every now
> and then. Not so time critical, but must be able to empty the memory
> faster than the ADC can fill it. Matlab or LabView interface if
> possible.
> 
> The end goal is to stabilize a piezo tunable laser diode to some
> reference interferometers.
> 
> Cheers,
> Bas


Bas, your sample rate and control requirements are very similar to my 
engine valve servo application, which has led me to the SBC6711 or SBC6x 
boards from

http://www.innovative-dsp.com/

And for analog IO, the A4D4 4x200ksps A/D & D/A.

There are also PCI boards such as M6x, but I am inclined toward a stand 
alone system, with a USB or other connection to a PC, for transferring 
relatively low throughput config data.

What do you think of these hardware choices for your application?


Good day!


-- 
____________________________________
Christopher R. Carlen
Principal Laser/Optical Technologist
Sandia National Laboratories CA USA
crcarle@sandia.gov

> Hi,
> 
> I'm involved in a similar problem.  Here is what I have been looking at:
> 
> 
> http://www.innovative-dsp.com/

Thanks for the suggestion, i hadn't found that one. Looked at the
website and found that most of the DSPs are more than adequate for
what i need. I also like that fact that they seperate the DSP and the
acquisition boards, so you can pick your own speed.

Only problem i have is with price of their software. The hardware I
want comes down to about $4500, but the software they advice (some
developer studio by Texas Instr) costs the same!!! Does anybody know a
way to get around this ridiculous price? I only need it for a single
project, i am not a full time DSP developer that uses the software on
a daily basis.

Bas