High Bandwidth Voltage Input Isolation

[Mic_Scale]

Hello,

I'm using the NI 6025E DAQ board and connect it through the SC2050 conditioning board to an analog I/O board, 5B01. I would like to isolate between the board and an analog signal coming out of of a DUT unit, and I need the isolation unit to be able to handle signals of Mghzs bandwidths.

I realize that I cannot find a 5Bxx unit with such high BW, but what do you recommend I should do instead ?

Michael.

[Jesse_O]

Hello Michael,

I see that you are looking for an isolation module to connect to your 6025E device that has a bandwidth in the MHz range.  Unfortunately, I think this type of setup will not work for you since the 6025E has a maximum sampling rate of 200kHz.  If you need to be able to sample data in the MHz range you will need to have a device that can sample much faster (i.e. High Speed S-Series or  NI-Scope products). 

Typically you do not see much signal conditioning for these types of devices since they have such a high bandwidth.  For example if you look at our SXCI isolated analog input modules they have a bandwidth of just over 20kHz.

I hope this helps.  If you have any additional questions can you please give us a little more information about the signal coming from your DUT (i.e. voltage and frequency range and what type of isolation you are looking for)?

Regards,

Jesse O.
Applications Engineering
NIC

Message Edited by Jesse O on 10-17-2006 10:44 AM

[Mic_Scale]

Hi Jesse,

thank you for your answer.

The reason I would like to get a high BW voltage isolation amplifier is because I work with pulses of the micro-second order. Although they appear to be too fast for the 6025E to sample in the first place, some of them are indeed not caught when sampling but some percentage is nevertheless caught and seen. That's why 6025E fits for out application.

But, we do require isolation between the DUT unit and the 6025E. And because the rise time of these pulses and length is so short, we need a wide bandwidth isolation amplifier.

Now, the fastest amplifier I've seen is called AD215 and has a BW of 125Khz. But perhaps you could advise on other solution, not necessarily opto-isolators but perhaps there are products that work on an isolated transformer base, that could reach to Mghzs order BW.

[charlesy]

Hello Michael,

You might consider data acquisition systems that use digital isolation instead of analog -- where the analog-to-digital conversion takes place in the isolated domain and the digital results are transmitted across the barrier to the host. This avoids the problems of noise and attentuation in analog isolators.
If you need unconditioned DAQ, take a look at the isolated M-Series: USB-6215 and USB-6218: http://sine.ni.com/nips/cds/view/p/lang/en/nid/203225
or if you want a PCI form factor: http://sine.ni.com/nips/cds/view/p/lang/en/nid/202499

If you need signal conditioning, here are an array of USB modules: http://sine.ni.com/nips/cds/view/p/lang/en/nid/13871
For higher performance, you can use the CompactDAQ platform: http://www.ni.com/dataacquisition/compactdaq/
For highest performance and embedded systems, we have CompactRIO: http://www.ni.com/compactrio/
The vast majority of modules for these platforms are bank isolated.

I hope this helps,

Charles

[AnalogKid2DigitalMan]

Hi Michael, how has it been going?

>3000VRMS isolation, 17MHz bandwidth opto-coupler based isolation amp (HPCL-4562):

http://www.avagotech.com/products/product-detail.jsp?navId=H0,C1,C5231,C4945,C4980,P88939

-AK2DM

[AnalogKid2DigitalMan]

I forgot to mention that that chip is for AC coupled signals- don't know if this suits your application, but you did mention the use of transformer isolation methods so I presume this is OK since xformers are inherently AC devices

Message Edited by AnalogKid2DigitalMan on 10-18-2006 09:41 AM

[Mic_Scale]

Hi, very good thank you

how are you doing AnalogKid ?

Thank you, this sounds like a very good solution to our problem. Now actually the signal we use is  TTL that we connect to an analog input and sample it as a video signal. This signal is an "error map" which is later overlayed on a video sampled image.

[Henrik_Volkers]

Just an idea:

If just need the edge of of a TTL pulse: Try a ferrit bead as a transformer. Primary (TTL) side 1-5 turns and a resitor to match the load (2k-10k) . Secondary (to your DAQ) 1-5 (+X) turns with 1k in parallel and two diodes to clamp the output.

Values are my first guess. A scope/breadboard and/or SPICE will help. Properly build, an isolation > 2kV can be reached.

Or try an ethernet transformer. (No old ISA 10/100 Cards flying around?)