Hello,
I am posting to ask a simple question about the common analog switch 4066.
Can the power supply (and the control signals) to this IC be completely isolated from the analog path that will be injected into the switch I/O? No ground connection between them. In other words, can it be used as a "relay" where the "coil" circuit (IC power supply) has nothing to do with the switching path.
I am posting to ask a simple question about the common analog switch 4066.
Can the power supply (and the control signals) to this IC be completely isolated from the analog path that will be injected into the switch I/O? No ground connection between them. In other words, can it be used as a "relay" where the "coil" circuit (IC power supply) has nothing to do with the switching path.
No. If you want isolation between the control side and the analog signal path then you could use something like this : TLP592A(F) Toshiba | Mouser
Its more expensive and slower to switch, but has a much lower on-resistance.
Its more expensive and slower to switch, but has a much lower on-resistance.
These analog switch chips use CMOS transmission gates, which are not isolated in any way.
CMOS chips have anti-static protection diodes on the pins, so the voltages on any pin must stay without the supply rails at all times.
There are many modern analog switch chips with far superior performance to the original CD40xx series versions, but they all use transmission gates.
CMOS chips have anti-static protection diodes on the pins, so the voltages on any pin must stay without the supply rails at all times.
There are many modern analog switch chips with far superior performance to the original CD40xx series versions, but they all use transmission gates.
Care needs to be taken with voltages in and out of IC's.
I did a USB project and accidentally drove an output 5 volts into a another IC powered off 3v3. The PC went nuts and said it was closing down to stop damage to it !
But compared to a 4066, it has a much larger off-state capacitance, so depending on the circuit, some treble may leak through when it is off. In any case, there are other optocouplers with MOSFET output, the trick is to find the best compromise between on resistance and off capacitance for the application.
Again depending on the application, the main risk is probably latch-up. Most ICs (including the 4066) are full of parasitic thyristors between supply and ground. An input or output voltage above the positive supply or below the negative supply can trigger these thyristors. Nowadays ICs are usually designed not to trigger as long as the input or output current stays below 100 mA, but I don't know what the limit used to be when the 4066 was designed.
If a parasitic thyristor should get triggered, the IC will burn out unless the supply has a low current limit and no big decoupling capacitors. You often see 4000-series electronic switches being supplied by resistors and Zener diodes, that's probably a good idea if there is any risk of the input going above the positive or below the negative supply.
There is a reason that we still use signal switches and relays in the 21st Century
Short answer, no.
Longer answer, the signal must stay within the voltages at the power pins of the 4066. At the 4066's maximum operating voltage of 15V, the signal must have a peak value of within + and - 7.5V (actually, probably a volt or so lower than the power voltages). This is usually how the 4066 is connected for audio, the power pins are at -7.5 and +7.5, the control voltage switches between these two voltages, and the signal (with a nominal DC level of 0V) can be DC connected to the switch.
This all presumes a line-level (from a CD/DVD player, TV sound "line output," FM tuner and such) signal. It's possible to switch a signal directly from a phono cartridge, but you'd get much better performance running it through a standard RIAA preamp first and switching the preamp's output. It's basically not possible to switch a "speaker level" signal with a 4066.
We can give you better help (with probably shorter responses, not mentioning things we think you might want to do but cant) if you're more specific about what you want to do.
Many thanks for all your answers.
Actually, I want to switch / route analog signals with no clicks or pops. The signal levels will be guitar levels - in the era of a couple of millivolts up to 1-2V or so.
Since I haven't found any means of doing it flawlessly (=every analog switch IC I read about and can buy does not seem to be good enough, and I don't have much time to try them), I decided to incorporate a "master" mute circuit at the output to noiselessly perform any switches when they happen.
Since I wanted to isolate the analog power supply rails from the switches, I found it preferable to switch using relays and a single isolated supply for their coils. The relays I am planning to use are Fujitsu RY-12W-K, instead of using 4066 ICs.
So the question could be rearranged: could the 4066 act as a reliably quiet analog switch (which hardly seems to be the case)? Or would the Fujitsu relays or other ICs perform better in that regard?
To cut the story short, I am currently favoring relays over 4066 since they require simpler power supply (single ended, and only one for 20 coils I will be using). I have read that relays usually don't provide noiseless switching too.
If you think / know the 4066 or a similar switch can be completely quiet when switching please share it!
Actually, I want to switch / route analog signals with no clicks or pops. The signal levels will be guitar levels - in the era of a couple of millivolts up to 1-2V or so.
Since I haven't found any means of doing it flawlessly (=every analog switch IC I read about and can buy does not seem to be good enough, and I don't have much time to try them), I decided to incorporate a "master" mute circuit at the output to noiselessly perform any switches when they happen.
Since I wanted to isolate the analog power supply rails from the switches, I found it preferable to switch using relays and a single isolated supply for their coils. The relays I am planning to use are Fujitsu RY-12W-K, instead of using 4066 ICs.
So the question could be rearranged: could the 4066 act as a reliably quiet analog switch (which hardly seems to be the case)? Or would the Fujitsu relays or other ICs perform better in that regard?
To cut the story short, I am currently favoring relays over 4066 since they require simpler power supply (single ended, and only one for 20 coils I will be using). I have read that relays usually don't provide noiseless switching too.
If you think / know the 4066 or a similar switch can be completely quiet when switching please share it!
Yes, keep them well below 15Vpp .
1) you CAN switch noiselessly, just be careful with grounding, switching waveform (hint: no *abrupt* edges which bcan feed through) and making certain DC at both sides of the switch is same value.
So it´s a ready made decision ... I wonder why you ask about 4066 then.
IF it´s quiet enough as a "main switch", it should be quiet enough for internal switching, wouldn´t it?
Yes, relays are conceptually simpler.
Relays can be noiseless, you´ll also have to be careful about switching signal, DC levels on both sies of switch contacts, etc.
All require somewhat careful planning.
Hey, even a fingertip operated switch requires care.
Yes, it can, see above.
More specific, I bet 4066 (and 4053, etc.) datasheets or application notes have some pages on noiseless switching, control signal debouncing, etc.
Go straight to Manufacturer´s data instead of endless reading Forum posts written by none too savvy and contradicting Members.
Go straight to the source.
Last edited:
If I was confident that 4066 will switch noiselessly, I would go for it! Pardon me for the misconception, I actually mean that since 4066 don't seem to be able to work better switch-wise, I would prefer the relay.
As far as switching the relay goes, if I needed to get them to switch noiselessly, I have also read that I need to eliminate DC difference between the terminals of the switch.
Isn't that possible using AC coupling (RC filters) plus a "bleed" resistor to ground in each terminal? The caps of the filters being very low leakage of course..
If you switch any signal at any non zero point you will get a click or pop.
For example if you switch a 1 volt peak sinewave when it is part way through the cycle then you will get a sudden output voltage corresponding to the point on the sine the switch is closed. That step change in level generates a sudden click.
For example if you switch a 1 volt peak sinewave when it is part way through the cycle then you will get a sudden output voltage corresponding to the point on the sine the switch is closed. That step change in level generates a sudden click.
I understand that, and for my application this is not a problem: my switcher should not make any pops when there is no signal present, if the user decides to switch while a signal is still "dancing" I would not consider any pops problematic, but something to be expected.
For balanced line level audio signals, I've made switching circuits with both relays and 4051 multiplexers that produce no audible switching sounds at normal volumes, except for the sound of the moving parts of the relays. I haven't a clue how far you get with millivolt signals and possibly higher source impedances, though.
I guess I'd go for something with very little or very slow charge injection, like a small photo-MOS optocoupler or a JFET with slowly switched gate voltage (that is, switching signal filtered with 1 Mohm-100 nF or so).
I guess I'd go for something with very little or very slow charge injection, like a small photo-MOS optocoupler or a JFET with slowly switched gate voltage (that is, switching signal filtered with 1 Mohm-100 nF or so).
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.