fast isolated signaling with photodiode matched pairs

Has anyone made an isolated linear signal circuit using photodiode matched pairs? As in AoE III, Figure 12.88.D, page 848, and Fig 12.90, page 849? We list in Fig 12.88 the HCNR201, IL300 and LOC110. They claim 0.01% linearity, much better than I need. I'm thinking if I run one of those blokes at a high photodiode current, it should be possible to make a very fast isolated signal transmitter.

This is for my RIS-764 project, which is a 1.2kV 10ns 20A pulse generator using a pair of new SiC MOSFETs (similar to AoE Figure 12.44.H, page 822).

This is to monitor the current though the flying output node. Note, that node includes a nifty +20V,-4V flying power supply, which can run the photodiode isolator's op-amp and LED.

I'll have to deal with the current generated by the flying node's more than 50kV/us slew rate into the isolator's coupling capacitance. That's 0.4pF for the HCNR201, which will create about 20mA for 5 to 10ns, swamping the photodetector's meager 0.5% output. I can add cancelling and a blanking circuit to the receiving end. I'll be happy with us, rather than ns, response speed.

--
 Thanks, 
    - Win
Reply to
Winfield Hill
Loading thread data ...

I think that Siemens/Infineon did it first. I looked at their dual receiver product at Cambridge Instruments in the 1980's. I don't recall that we ever used it - photo-diodes were a bit slow for what I had in mind.

Why not use a transmission line transformer wound with high-voltage coaxial cable?

AOE2 shows a 4:1 version, which might be handy for monitoring. 9:1 might be nicer, and 16:1 doesn't seem to be impossible.

--
Bill Sloman, Sydney 

For 10nsec pulse, the absence of a DC component wouldn't seem to be a problem.
Reply to
bill.sloman

JL, his colleague Jonathan Dufour, and I did a nanowatt photoreceiver that worked that way. It used a sort of built-up optocoupler for the current fee dback loop, to avoid using a feedback resistor, and formed it's output usin g a second PD illuminated by the same LED, with an ordinary op amp TIA. The CTRs were adjusted to be 0.01% for the feedback arm and 10% for the output arm, so effectively it had a photon/photon gain of 1000.

That approach has an automatic 3 dB noise penalty, because the feedback cur rent has full shot noise, so this one actually wired two photodiodes _in se ries_ to reduce the noise penalty to 10 log(1.5) ~ 1.6 dB and reduce the ca pacitance. (Auxiliary feedback kept the two from fighting.)

Bandwidth was about 3 MHz on the 100-uA FS range and 1 MHz on the 1 uA FS r ange.

Cheers

Phil Hobbs

Reply to
Phil Hobbs

My HV pulse circuit has several applications. In the case of ns-scale pulses, I'm not so concerned with measuring current, it'll be over shortly, so what if it's 5A, 10A or 20A? But for the slower applications, where real power dissipation is at play, I need DC and moderate speed measurements.

--
 Thanks, 
    - Win
Reply to
Winfield Hill

How accurate do you want it to be?

The most CMRR (pretty much infinite) would be a fiberoptic link. But it wouldn't be super accurate.

An LED link would be linear and cheap, something like toslink maybe. A laser link would be a lot faster, but lasers aren't great linear devices.

Laser+PWM or FM over fiber would be great, but probably overkill.

Free-space light over an inch or so might be fun. Run the light through a grounded metal tube. Somebody should make an optocoupler like that... LED, photodiode, two big grounds in the middle.

Do you need DC coupling? I've used current transformers at voltage/current levels like yours, testing my DSRD high-voltage pulser thing. Seemed to work fine, and had ns speed. Just a few turns of wire around a small permalloy powder core, single-turn primary.

--
John Larkin         Highland Technology, Inc 

lunatic fringe electronics
 Click to see the full signature
Reply to
John Larkin

I think I can adjust the matched-coupler circuit in Fig 12.90 to 1% accuracy pretty easily. I'm going to add a bias, so the LED coupler pair will still be operating at zero current. I'm hoping for well under 0.1% zero offset.

And yes, I need response down to DC.

--
 Thanks, 
    - Win
Reply to
Winfield Hill

I've added a few files to DropBox.

formatting link

--
 Thanks, 
    - Win
Reply to
Winfield Hill

I came up with this thing a while back, just playing around...not for any specific purpose. For some reason I decided I wanted a full wave rectified output.

Circuit and relevant waveforms:

formatting link
formatting link

Reply to
bitrex

Win, Many years ago (around 1980) I designed a DC-coupled multi-channel isolation amplifier for cochlear implant research.

It used a Burr-Brown hybrid opto-isolation amplifier for DC to a few hundred Hz and a broadcast quality toroidal audio transformer for higher frequencies up to around 100kHz.

The crossover between the opto-isolation amplifier and the transformer was first order, so the overall response was linear-phase. It happily transmitted aquare-waves at 20kHz with nothing worse than rounding of the corners.

I think the dynamic range was about 70dB (up to 20kHz).

The Burr-Brown device consisted of a driver op-amp with a pair of leds, one in series with each supply rail. The output was loaded with a resistor. (I dismantled one to find out how it worked as the data sheet was unclear.)

One photodiode optically coupled to each LED was used for driver amp feedback and the second photodiode coupled to each LED was connected to the receiver op-amp.

Obviously you are interested in much higher frequencies, but the same principle of a DC-coupled optical isolator combined with a high-frequency transformer with first-order crossover might work well.

John

Reply to
jrwalliker

Many years ago we used the IL300, it seemed to be not so linear but in our application it did not matter.

I had also used a chopper isolated amp, years later. It had better linearity but at the expense of the chopper hash.

Cheers

Reply to
Martin Riddle

Thanks! The idea to combine a current transformer with a low-frequency optically-isolated path is very interesting. Several people have pointed out that CTs are quite simple.

--
 Thanks, 
    - Win
Reply to
Winfield Hill

You might cut a slot in your transformer core, and use a Hall sensor to monitor the flux through the core. That would measure the droop on the transformer output - neatly filling in the low-frequency down to DC component that the transformer doesn't offer.

You could even couple the Hall sensor to a current driver on an auxiliary winding to keep the flux at zero, or at least constant and low ...

--
Bill Sloman, Sydney
Reply to
bill.sloman

I've designed signal isolators with both IL300 and HCNR201s, dc coupled. Offsets in this application are critical, so had to have user-accessible offset adjustments :( Have tried both dual-isolators (as in the appnote, one for positive-going signals, the other for negative-going) and a single isolator biased at midpoint. Performance not markedly better for any particular scheme, though maybe a bit better common mode rejection for dual-isolator approach. Usable -3dB BW (without any exotic effort) around 40-50kHz.

HTH!

Reply to
Frank Miles

Concept-wise...

...Jim Thompson

--
| James E.Thompson                                 |    mens     | 
| Analog Innovations                               |     et      | 
 Click to see the full signature
Reply to
Jim Thompson

Thanks, Frank, that's very helpful and encouraging. Offset pots, yes.

--
 Thanks, 
    - Win
Reply to
Winfield Hill

"Decades ago" probably doesn't pre-date me ;-) (I was consulting to Burr-Brown ~1970-71) ...Jim Thompson

--
| James E.Thompson                                 |    mens     | 
| Analog Innovations                               |     et      | 
 Click to see the full signature
Reply to
Jim Thompson

Burr-Brown (since bought by TI) did this decades ago. See the ISO100 datasheet for how. It's been obsoleted now but another concept to start from. Art

Reply to
Artemus

ElectronDepot website is not affiliated with any of the manufacturers or service providers discussed here. All logos and trade names are the property of their respective owners.