Impedance matching for 8 ohm speaker

[Kian]

Thank guys for the very active discussion. I would also want to explore the option of putting an audio amplifier on my original setup with the multiplexer. That would also work right? Any part numbers to recommend so that I can try out these different options.

 

[unclejed613]

if you have the multiplexer connected to the input of the amplifier, and switching between line level audio signals. it would basically act the same way as an AV receiver input switch. you would definitely be able to drive speakers without having to worry about overloading the output of the multiplexer.

 

[audioguru]

The datasheet of the opto-mosfets switch does not mention distortion or polarity of the signal. It also does not mention any audio application.

 

[crutschow]

The datasheet of the opto-mosfets switch does not mention distortion or polarity of the signal. It also does not mention any audio application.

True, but an ON MOSFET is just a low value resistance so I see no reason is should add any significant distortion.

What do you mean "polarity of the signal"?
The MOSFET SSR I posted has two MOSFETs that can be connected in inverted parallel to pass an AC signal.

 

[audioguru]

Doesn't the polarity of the drain and source of the N-channel Mosfets affect their on-resistance? Don't they cause crossover distortion as one Mosfet turns on then the second Mosfet turns on as the polarity across them switches?

In the AC application of the IC, the Mosfets are in series, source to source.

 

[crutschow]

Doesn't the polarity of the drain and source of the N-channel Mosfets affect their on-resistance? Don't they cause crossover distortion as one Mosfet turns on then the second Mosfet turns on as the polarity across them switches?

MOSFETs conduct equally well in the forward and reverse direction when biased on (they look like an Ohmic, non-polar resistance), so I wouldn't expect that to be a problem.
In an SSR the MOSFETs are continuously biased on by the LED, which rather acts like an isolated gate-source voltage.

n the AC application of the IC, the Mosfets are in series, source to source.

You are correct. I had it backwards.
They are in series back-to-back so they can block both polarities of the AC voltage.

 

[audioguru]

Then why doesn't the manufacturer mention that the opto Mosfets make a perfect audio switch?

 

[crutschow]

Then why doesn't the manufacturer mention that the opto Mosfets make a perfect audio switch?

Beats me, but it's probably because not many people need to switch speaker connections with relays, so it's not a big market.

And I didn't say it would make a perfect audio switch (but it should make a reasonably good one).

 

[Nigel Goodwin]

Then why doesn't the manufacturer mention that the opto Mosfets make a perfect audio switch?

Probably because they don't?, adding extra semiconductors in the audio path (and outside the feedback loop) can only decrease quality.

Be interesting to see what effect it actually has?.

 

[audioguru]

It might make a modern FUZZ pedal.

 

[Nigel Goodwin]

It might make a modern FUZZ pedal.

That was my thought

If the MOSFET's are in series, then they are relying on the internal diode across the opposite MOSFET to conduct the signal, so it's like putting two parallel diodes in series with your speaker.

 

[crutschow]

Be interesting to see what effect it actually has?

Yes, I would be interested in seeing that.

If the MOSFET's are in series, then they are relying on the internal diode across the opposite MOSFET to conduct the signal, so it's like putting two parallel diodes in series with your speaker.

No. Are you not familiar with the fact the MOSFETS conduct equally well in both directions when ON?
So when the MOSFETs are ON in series back-to-back, the back-to-back substrate diodes are bypassed by the low on-resistance of the MOSFETs.

Below is the simulation of that for my understanding of how the opto works:
V2 and V3 are voltages generated by the opto MOSFET LED to each MOSFET gate-source junction.
The output closely follows the input AC when ON.

 

[Nigel Goodwin]

I've no interest in the 'guesses' that simulators provide, however there seems no suggestion of distortion levels in that simulation.

 

[audioguru]

The datasheet for the opto Mosfets shows a perfectly linear low voltage drop at the output when the load current increases. Then the distortion might be "only" 0.5% like an old vacuum tubes geetar amplifier at low output power.

 

[crutschow]

I've no interest in the 'guesses' that simulators provide

That seems to be a rather archaic and condescending view.
Simulators don't provide "guesses", they provide a simulation of the circuit operation based upon the models used.
I've use SPICE simulators for over 30 years to help me design circuits, and they've often spotted errors in my design before I built them.
The built circuits almost always operated very close to what the simulator predicted.
They've also help me trouble-shoot circuits already built which where not operating as intended.
I would never build a circuit without simulating it first.

I believe my opto model is a reasonable simulation of the actual circuit, but only testing with an actual device will answer that question, of course.

 

[crutschow]

The datasheet for the opto Mosfets shows a perfectly linear low voltage drop at the output when the load current increases. Then the distortion might be "only" 0.5% like an old vacuum tubes geetar amplifier at low output power.

So like a linear resistor?
Such a resistor in series with the speaker should cause no significant distortion, other than a small reduction in the damping factor.
The opto I referenced has only 0.15 ohm maximum on-resistance.

 

[audioguru]

I can see that your simulation sinewave has distortion at about +3.3V and -3.7V, something like crossover distortion. I cannot see distortion when it is less than about 1% so yours is pretty bad, maybe 1 or 2%. I can hear when distortion is 0.2% or more.

 

[crutschow]

I can see that your simulation sinewave has distortion at about +3.3V and -3.7V, something like crossover distortion. I cannot see distortion when it is less than about 1% so yours is pretty bad, maybe 1 or 2%. I can hear when distortion is 0.2% or more.

Not sure what you are seeing.
It may be just an artifact of the plot resolution.
The FFT plots below shows little difference between the input and output.

 

[audioguru]

Kian said that the maximum amplifier output is only +/- 6V peak so the power in an 8 ohm speaker is only 2.25W. At such low power then distortion would not be heard even if it is 2%.

 

[unclejed613]

That seems to be a rather archaic and condescending view.
Simulators don't provide "guesses", they provide a simulation of the circuit operation based upon the models used.
I've use SPICE simulators for over 30 years to help me design circuits, and they've often spotted errors in my design before I built them.
The built circuits almost always operated very close to what the simulator predicted.
They've also help me trouble-shoot circuits already built which where not operating as intended.
I would never build a circuit without simulating it first.

i've had much the same experience. i wish there had been such tools available in the 1960s through the 1980s, might have saved me a lot of burned components (and burned fingertips)...
it's been my experience with simulating circuits that the sim matches the actual build within about 10% (usually better). that's good enough to see if something will work at least. if i entered the actual component values as measured, the sim gets closer to 2% or better. the best thing about doing a sim, is i can check the collector current in a transistor and see if it will be running too much current. that saves a lot of "magic smoke" all by itself. i've even done some sims using vacuum tube models that i knew weren't the right one (substituting a 12AU7 model for a 6BF7 pencil tube) and the actual circuit worked fairly close to what happened in the sim. it does help to see before building a circuit what the effects of various tweaks and changes will do, or whether a circuit works at all....