Tube Optical-Isolation Unit

[jackinnj]

Nice article in AudioXpress this month with an analogue opto-coupler for your tube stuff.

These devices are very linear, the HCNR201's spec is 0.05%, but I found it to be 0.01% to 0.03%. They should be much lower noise than the photo-resistor analogue isolators --

The author used the Infineon (Vishay) IL300 -- note that the Agilent HCNR-200 and 201 can also be used but were probably not available to the author. The IL388 is a faster version of the IL300. Vishay has some excellent application notes on their website (as does Agilent.) Look for AN50 on the Vishay site.

Agilent also has an analogue opto-isolator which uses a modulator-encoder on one side and a decoder on the other side of the barrier -- Agilent says that their linearity is even better than the photo-diode device.

Newark Electronics carries both the Infineon and Agilent Devices, Digikey carries only the Agilent Devices.

[analog_sa]

analogue opto-coupler for your tube stuff.

Sounds interesting but what's the point? I have used tube gear for many years and never felt a burning need for an optocoupler. If isolation is indeed desired, what can be better than a transformer?

[jackinnj]

Originally posted by analog_sa


Sounds interesting but what's the point? I have used tube gear for many years and never felt a burning need for an optocoupler. If isolation is indeed desired, what can be better than a transformer?

Elimination of noise and ground loops at less cost.

Better linearity.

Less real estate consumed.

[Geek]

Sounds like a fabulous way to couple a PC to a tubeamp

[fdegrove]

Hi,

Thanks, Jack.

Looks worthwhile to take a peek at the AN Vishay provides....

Have you actually implemented this gizmo in a tube amp yet?
If so, what do you think of it soundwise?

TIA,

[jackinnj]

when I wrote the article on the MCUTracer for AX I used the HCNR-200 between the LT1014 first opamp and the microprocessor. I found a company unloading some of the more expensive HCNR-201's and IL388's and purchased them.

On a subsequent revision of the article I took out the analog opto-isolators because it added a bit more complexity and cost than I wanted -- you need two sets of opamps -- one in front of, one after the optical barrier and the devices are about $3 if you don't buy a lot of them (so you would need one on each channel). If I have boards made for the MCUTracer (there seems to be some interest out there) I will probably just use 4N35's on the ADC output.

If I go with a manufactured ADC board though I would use the HCNR201's because even with their added cost they are less expensive insurance than blowing a USB card.

You do have to play with the compensation a bit since the surface area of a photodiode makes it look like a big capacitor (well, some picofarads). Around my lab everything oscillates once in a while.

Analog Devices makes some nice isolation devices -- one of their big markets is medical and you don't want some cardiac patient to inadvertently get zapped.