ADC-Will this setup work ?

[brazingo]

I have been building a data acquisition system to be connected to a
pentium 4 system. But the problem is scanning rate of ADC (Analog to
Digital) is about 3MHz. But I need a higher rate. Will connecting 2
such ADCs in parallel and scanning them alternatively using my pc give
me a better rate like 6MHz (3+3). Also can this be extended ?
All opinions are welcome. And, thanks in advance.
BZ

 

[Bob Masta]

I have been building a data acquisition system to be connected to a
pentium 4 system. But the problem is scanning rate of ADC (Analog to
Digital) is about 3MHz. But I need a higher rate. Will connecting 2
such ADCs in parallel and scanning them alternatively using my pc give
me a better rate like 6MHz (3+3). Also can this be extended ?
All opinions are welcome. And, thanks in advance.
BZ

Ihave heard of this trick being used on high-speed digital scopes.
Note that each ADC needs a sample/hold that is at least as fast
as the overall desired sample rate would require. I don't think
you should try using the PC to provide the interleaved scans at
these rates; you'll need a dedicated clock with accurate phases.

Just out of curiosity, how are you going to move the data to the
PC at these rates? Or is this for some intermittent application
like a scope that can ignore data between trace updates?

Best regards,





Bob Masta
dqatechATdaqartaDOTcom

D A Q A R T A
Data AcQuisition And Real-Time Analysis
www.daqarta.com
Home of DaqGen, the FREEWARE signal generator

 

[brazingo]

Hi Bob,
I was going to use a USB 2.0 interface to connect this ADC system
to my computer - P4 2Ghz. I thought this should work fine given that
6MHz is just a fraction of 2000MHz. I even planned to use assembly
programing to get the clock cycles right with an optimum code.
I'm not too intent on 6GHz though. Anything around 600Khz to 1MHz
scan rate would do fine for me so that I can capture a sine wave of
atleast 100Khz with 6 to 10 samples per cycle respectively. I just
wanted to push this system to the limit and get 6GHz.
I was really surprised by your question. Am I missing something here
?
-BZ

 

[Bob Masta]

Hi Bob,
I was going to use a USB 2.0 interface to connect this ADC system
to my computer - P4 2Ghz. I thought this should work fine given that
6MHz is just a fraction of 2000MHz. I even planned to use assembly
programing to get the clock cycles right with an optimum code.
I'm not too intent on 6GHz though. Anything around 600Khz to 1MHz
scan rate would do fine for me so that I can capture a sine wave of
atleast 100Khz with 6 to 10 samples per cycle respectively. I just
wanted to push this system to the limit and get 6GHz.
I was really surprised by your question. Am I missing something here
?

I am not sure about your expertise in programming, so please
forgive me if this seems condescending. However, in a modern
protected mode OS (Windows XP, for example) you have essentially
no control over timing issues. The OS can jump in at any time and
run some other thread. When you start some operation, Windows
may wait several msecs before actually doing anything about it.
Some of these issues can be dealt with by Ring 0 drivers, but it
would be a big can of worms unless you are already a pro at it.

The normal way A/D systems work under protected mode is to
buffer everything. You would collect a block of data to your own
memory buffer, then signal Windows that it is full (while you
switch to filling another buffer), and Windows will transfer the
data whenever it feels like it. You have to insure that the rate
of buffer filling doesn't exceed the average transfer rate.

But it sounds like in your case you are making a scope, so
you could get by with a single buffer. When the buffer is
full (one scope-face sweep), you let Windows have it.
The few msec of lag would be no problem, since your
eyes don't need display updates all that fast anyway.
Note, however, that this supposes you do all the triggering
in hardware. If you want to do it in software, you have to
pass Windows the entire data stream, which means you
need to support sustained throughput. Or at least a big
enough block of data that you can be sure it included at
least one trigger event plus enough data for display.

Best regards,



Bob Masta
dqatechATdaqartaDOTcom

D A Q A R T A
Data AcQuisition And Real-Time Analysis
www.daqarta.com
Home of DaqGen, the FREEWARE signal generator

 

[Jasen Betts]

I have been building a data acquisition system to be connected to a
pentium 4 system. But the problem is scanning rate of ADC (Analog to
Digital) is about 3MHz. But I need a higher rate. Will connecting 2
such ADCs in parallel and scanning them alternatively using my pc give
me a better rate like 6MHz (3+3).

it'll give a higher rate (there shouldn't be bus speed problems at that
rate), but not neccessarily higher resolution.

maybe have a look into ADCs designed for video capture

 

[Jasen Betts]

Hi Bob,
I was going to use a USB 2.0 interface to connect this ADC system
to my computer - P4 2Ghz. I thought this should work fine given that
6MHz is just a fraction of 2000MHz. I even planned to use assembly
programing to get the clock cycles right with an optimum code.
I'm not too intent on 6GHz though. Anything around 600Khz to 1MHz
scan rate would do fine for me so that I can capture a sine wave of
atleast 100Khz with 6 to 10 samples per cycle respectively. I just
wanted to push this system to the limit and get 6GHz.
I was really surprised by your question. Am I missing something here

usb 2.0 is 480M bits per second, so it should work.


Bye.
Jasen

 

[brazingo]

Hi Bob,
I confess I'm not a pro at system programming. Just been
learning pentium programming limited to only real mode and also
restricted to DOS. I thought I could make myself a nice little scope.
Guess I overestimated certain things. Thanks for the inputs though.
I'll try working with the buffering concept.