I have got few analog and digital signal that needs to be galvanically isolated. I am thinking to use H11F1M for analog signals. I am not sure how to rig-up the circuit using this device. Specifically bias resistors to be connected on LED and FET side.

And can the same device be used for digital signals also. And what precautions shall I take while using it for analog signals or digital signal.

EDIT1: I want to isolate RSSI generated from AD8367. RSSI vs Voltage graph on Page 10.

Apart from HCNR200 from Avago as a solution.

I got a application note from Silabs which claims to use Si8660 for isolation of analog signal as it has internal RF Modulator and Demodulator based on OOK. But I am not sure whether analog signal can directly be fed to this chip of I need to use some extrnal opamps as shown in the Evaluation Board of Si866x.

  • \$\begingroup\$ Is this a one-off or for production (i.e., are you interested in price differences of a few dollars)? Also, what is the voltage range of the analog signal? \$\endgroup\$ – Scott Seidman Oct 2 '13 at 13:18
  • 1
    \$\begingroup\$ What is the bandwidth of the analog signal, and with what accuracy do you need to get it accross the isolation barrier? Converting to digital then sending that accross isolation has many advantages. Why can't you do that? \$\endgroup\$ – Olin Lathrop Oct 2 '13 at 13:22
  • \$\begingroup\$ @ScottSeidman: It is for the production only. \$\endgroup\$ – AKR Oct 2 '13 at 13:26
  • \$\begingroup\$ @OlinLathrop:Analog signals are RSSI of some RF signal. \$\endgroup\$ – AKR Oct 2 '13 at 13:30
  • \$\begingroup\$ Do the analog/digital signals have to be galvanically isolated only from the control signal or also from each other? If the first situation is the case it is enough to use standard anlalog switch/MUX (4015, 4051, 4067 or similar) and only isolate the (digital) control signal(-s). \$\endgroup\$ – Curd Oct 2 '13 at 13:30

I think the analog part might be achievable by using your analog signal as a variable gain amp that has a a constant voltage source as an input, as suggested in the datasheet. I'm not real familiar w/ this IC choice, but it looks doable with the choice of IC (I just don't know how well!). If money were no object, I would have suggested something like an HCNR200 (at least for non-neg signals), but its for production.

  • \$\begingroup\$ Any views on Si8660 \$\endgroup\$ – AKR Oct 4 '13 at 17:09
  • 3
    \$\begingroup\$ I avoid silabs products because I didn't feel they were responsive to my needs during my only interaction with them. They refused to sample me because I'm at a university, and they wouldn't tell me about life cycle of a product I was interested in. Not that I hold a grudge! \$\endgroup\$ – Scott Seidman Oct 4 '13 at 17:13
  • 1
    \$\begingroup\$ @ScottSeidman I share your view on SiLabs: Manufacturers who are shy about sampling are not keen on my business, or that of my clients. So I make sure they don't get any of the business. A generous sampling policy such as TI's leads to a natural increase in incorporation of that manufacturer's parts into every design. It's business sense, not a grudge. \$\endgroup\$ – Anindo Ghosh Jan 3 '14 at 3:49

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.