Optocoupling and Isolation question

I'm trying to understand isolation a little and I'm somewhat confused.

I've built vacuum tube amps with 600+v and the circuit is essentially isolated using input and output transformers which I see as a complete isolation.

I see many modern electronic circuits that use optocouplers to power large MOSFETs or TRIACs, but they seem to share the same power rail on both sides of the optocoupler. Does that not nullify the isolation? In my primitive thinking, it would be more isolated to use a small relay for the switching than an optocoupler and TRIAC. Or use some kind of 1:1 transformer or separate power rails for the two sides.

Is it a matter that one is trying to protect "minor" surges while "major" surges will still fry everything? Or is it all simply a cost factor?

Thanks

Peter

Optocouplers can be used for driving the high side of switching circuits. But there are many more ways to do this, transformers, solid state level shifters, capacitive coupling etc. Relays would be too slow though. It not a question of protecting from overvoltages but of shifting the level of a signal from one reference to another.

It all depends on the purpose of the isolation. If one of the "rails" is common, it is not fully isolated, and whatever "isolation" is there is probably for level shifting or, as you said, "minor surges". Optocouplers can be used for true isolation however. And, I submit that an isolation transformer is NOT "complete isolation".

profner:
I see many modern electronic circuits that use optocouplers to power large MOSFETs or TRIACs, but they seem to share the same power rail on both sides of the optocoupler. Does that not nullify the isolation?

Yes, that would nullify the isolation. Normally the power supplies either side of the optocoupler are isolated otherwise there would not be much point in using the optocoupler.

Russell.

I've built vacuum tube amps with 600+v and the circuit is essentially isolated using input and output transformers which I see as a complete isolation.

There should also be a power transformer to isolate the power line voltage from the circuitry (and to isolate the user).

Transformers do provide 100% electrical isolation. There is NO ELECTRICAL CONNECTION between the primary and secondary. Only a magnetic connection.

The power supply in just about every piece of electronic equipment has a transformer to isolate the power line voltage.

There is often a common ground. For example, in your (desktop/tower) computer the power supply output-ground (which is connected to the motherboard ground) is connected to the power-line & chassis ground.

I think the main purpose of the input/output transformers is impedance matching. I believe there are tube preamps without I/O transformers or with input transformers but no output transformer. The output transformer in a tube audio power amp converts a high voltage, low current signal into a lower voltage, higher current, signal to drive a (low impedance) speaker. From what I understand, some high fidelity tube amps use feedback from the transformer output to reduce transformer distortion and to improve frequency response. That means the audio output is NOT 100% electrically isolated from the tube-voltage.

On the input of a tube amp, a step-up transformer can give you "free voltage gain" since the input (grid) on a tube has very high impedance.

Transformers can transfer power. A transformer is like the transmission in a car or the gears & chain on a bicycle.

Relays & opto-isolators only control power, like the gas pedal in your car.

In a sense, any non-one-to-one transformer always performs impedance matching. The change the ratio of the voltage to the current. While the voltage/current transfer is n, the impedance ratio is n-squared.

OK, so I do understand more now. It seems designs I see are more level-shifting and calling it isolation which isn't exactly correct. I have an industrial controller at work that claims isolation between inputs and outputs for surge protection via opto-couplers, but uses a single shared power supply and I always doubted the effectiveness in the case of a true surge.

So as a followup question, at what point in the differing levels would one insert an optocoupler? For example, I have a design that takes the 5V digital output from the Arduino into a ULN2003 and activates small 12v coil relays. I used the relay since there would be no fast switching, and the switched signal could range from 12VDC up to 24VAC solenoids so I felt there should be some isolation from the switched device. Should there be an opto-isolator in there instead of the darlington array?

I also realize a transformer is not complete isolation, but I suppose I more meant physical isolation when compared to an optocoupler with a common-rail power supply.

Thanks again to all those who replied.

Peter

If a common signal ground is acceptable, and as long as ground loops are avoided, separate power supplies are often satisfactory when voltage and currents in one part of a circuit might interfere with other parts of the circuit. One way that ground loops are avoided, is by providing separate ground returns to the power supply for high current devices.

profner: Should there be an opto-isolator in there instead of the darlington array?

Optoisolators are mostly very low current. You can't use them like an Arduino's 20mA outputs. The optoisolator will always need a transistor to drive a relay (or whatever the relay was driving.)

profner: I used the relay since there would be no fast switching, and the switched signal could range from 12VDC up to 24VAC solenoids so I felt there should be some isolation from the switched device. Should there be an opto-isolator in there instead of the darlington array?

No, an opto-isolator isn't required. The relay should give all the isolation needed. Remember to put a diode across the relay coil to prevent any reverse voltage being generated when it is switched off.

Russell.