Analog switch, single supply, bipolar input , low noise, simple

[Rooby]

Hello

I'm looking for a simple analog switch (e.g. mosfet) to switch a pipolar (around +-1V) analog signal on/off but only using a single supply voltage of max 8V.
Preferrable device with low noise as possible.
I have used a J175 p-channel JFET with a driving the gate with 0 (ON) or 8V (OFF).
But J175 is quite an old device, high Ron, not low noise.
Does anyone can suggest other devices, preferable SOT23 package. for an analog switch ?
Thanks

 

[JMFahey]

Search for an analog gate, similar to CD4016/4066/4053

 

[Rooby]

Unfortunately such gates are usually not for bipolar input. The input must be 0 to Vdd.

 

[SomekPoland]

I've seen people and companies using SSR-type circuits to switch analogue signals. They use two MOSFETs connected by sources and something to drive them with. you can but those double MOSFETs easily.

 

[soekris]

For JFET I can recommend 2SK3557, for opto mosfet switch I can recommend TLP3341.
I use both myself.

 

[SomekPoland]

For JFET I can recommend 2SK3557, for opto mosfet switch I can recommend TLP3341.
I use both myself.

Thank you for dropping some tips!

 

[Ken Newton]

Bipolar input-signal, unipolar supply-voltage.
So, you are planning to D.C. level-shift the input signal?

 

[PRR]

not for bipolar input. The input must be 0 to Vdd.

Isn't that uni-polar?

CD4051/52/53 is bipolar in the switch and unipolar in the control. THD is low if the load is >>>1k. However it may not be "modern" enough for you.

 

[JMFahey]

Unfortunately such gates are usually not for bipolar input. The input must be 0 to Vdd.

Of course, that´s why you bias gates to half Vdd and use input/output coupling caps.
Do you know how to use them?
Suggest you fully read datasheets and application notes.

I have used a J175 p-channel JFET with a driving the gate with 0 (ON) or 8V (OFF).

It has the same limitations as the gates I suggested.

 

[jan.didden]

We have this thing called 'Google'. You should try it out!

Gives stuff like|:

https://www.mouser.be/datasheet/2/256/MAX364-1513927.pdf

and

https://www.mouser.be/c/semiconduct... switch&mounting style=Through Hole&instock=y

There is a whole ecosystem out there with switches running on single supply yet still able to handle bipolar signals. Extremely clean and sub-1ohm on resistance.

Jan

 

[KSTR]

There is a whole ecosystem out there with switches running on single supply yet still able to handle bipolar signals. Extremely clean and sub-1ohm on resistance.

I'm not aware of many integrated (MOSFET) switches where you can go below the substrate voltage. There are some specialized parts what would fit the 1Vrms(?) requirement, however, from TI, OnSemi and others, for cell phones etc, in micro packages.

The OP gave little information about the requirements of the application, like source and output resistance, attenuation (is it about disconnecting electrically or is it about muting?), the control signal, if AC coupling would be tolerable etc.

 

[jan.didden]

You are right, they run from single supply for the logic but would need a negative supply for negative signal levels.
Thanks for bringing that up.

Jan

 

[KSTR]

Availability is a problem, now. Only found very few on Mouser etc, like:https://www2.mouser.com/c/semiconductors/switch-ics/analog-switch-ics/?q=TS5A2236&instock=y

Depending on the requirements, back-to-back (source-connected) NMOS FETs might do. I've use 2N7002 for similiar tasks. Non-audio, though, and a little of control voltage leakage didn't matter. Gates pulled to 12V via 1MegOhm and shunted to sources with an optocoupler.

 

[Rooby]

Thanks for your suggestions but I need bipolar (+/-1V) input. I do not want to couple by caps to half the supply voltage for other reasons.
With a p-channle JFET this is possible since it conducts when gate is at 0 V even with negative input voltages on source/drain as long it will not more negative as the threshold voltage. In my case 1V is still below thresh. For a positive voltage on gate (8V in my case) it will block even with 1V on source/drain, since 7V is still above thresh.
Unfortunately with n channel JFETs you need a negative gate voltage. Even if I would couple to half of the supply voltage to 4V . This will then end up in 3V min gate-source voltage which is to less for JFET threshold.
Also analog gates like 4051 no bipolar input. In the datasheet it says on recommendetd input voltage 0V to Vdd, so no negative input. I also expect they are noisy.
Also MAX364/MAX36 input is only positive for single supply.
TS5A22364 could be a alternative but I need to supply it with around 4V to get +/-1V range, since input range is Vcc-5.5V to Vcc. Not so good...
MAX20336 could be interesting but it is BGA, not usable for hand solder and also need 5V max supply. not 8V as in my case.
Unfortunately p channel JFETs are rare and difficult to get, so I guess not many alternatives as J175.
The following figure in the link is what I'm currently using
https://electronics.stackexchange.com/questions/177398/jfet-audio-switch-does-not-work-in-simulation

 

[PRR]

analog gates like 4051 no bipolar input. In the datasheet it says on recommendetd input voltage 0V to Vdd, so no negative input.

The analog side is Vdd to Vee.

 

[Rooby]

The analog side is Vdd to Vee.

Yes but Vee must then be negative but I have only positive supply (0 to Vdd=8V).

 

[jan.didden]

How did you make that supply? Is there a transformer and rectifier?

Jan

 

[Ken Newton]

Yes but Vee must then be negative but I have only positive supply (0 to Vdd=8V).

Would you elaborate on why you are avoiding an input-coupling cap.?

 

[PRR]

Relay.

 

[JMFahey]

^^^^ THAT, or its poor cousin, an LDR

Both are floating and couldn´t care less about signal polarity.