Analog switches

Status
Not open for further replies.
Jump to Latest
L

Luke123

Member
Joined 2004
#1
Hello

I've been using the SSM2402 audio analog switches for a long time with excellent results but I just discovered it's been discontinued since few years.

Any suggestion for an equivalent, Analog Devices or other?

I need to apply on ±18 V or less supply, below 100 ohms ON resistance, great isoltaion resistance, very low distortion, 2 x spdt function, pin DIP version, Rhos or not...

Thanks!

Luke
 
#3
Dug

Thanks for reply.
But the first is surface-mount.
The second one is spst and no-stock.
I already made that search from Digi-key online.
The only one I found was ADG436 but is way less than SSM2402 but not bad.
Could be the one but I'm really looking for a 2402 equivalent.

Luke
 
#5
Hello Dug

Hisshhh!...$10.+ each and no mention about distortion.
Off isolation of 72 dB...Mmmm, not great for muting function unle$$ you u$e two in ca$cade.
Next order from DGK, as they have 1000+ in stock, I'll try measuring one against 2402.

Thanks Dug!

Luc
 
#6
maxim makes a number of nice parts that i think are still available in dip package.
check out their site and look at max312 or max352 - check the part numbers for exact configuration of switches; i going from memory here, which is risky at best 🙂

mlloyd1
 
#7
Hello mlloyd1

Great. I've just checked for DG412 series and it looks interesting but it has poor isolation of only 68dB..$7.75 DGK.
And equivalent ADG412 from AD ...$5.08.
So all these cmos SW seem to have great Ron but poor Off Isolation (±65-70dB) and unknown distortion ratings.
And bipolar SW like SSM2404... ordinary Ron (60 ohms) but over-needed but superb isolation of 120dB and AD specify distortion ratings of them.
Possible to increase Off isolation on cmos but external TTL or something probably needed to achieve this spec , so more complex.
Great infos anyway... good possibilities...thanks!

Luc
 
Last edited:
#8
yes, i agree not much in the way of choices for a similar performance to the old SSM switches. i guess we just didn't buy enough of them for ADI to keep them around ...

you could also try discrete jfets. i think mcintosh was doing that for a while.

mlloyd1
 
#9
The Intersil ISL54405 is a fantastic switch for audio, but it's only surface mount and you need to use it to switch current, not voltage, by loading it into a virtual ground. I have a test circuit using it now - off isolation is over 140dB, unmeasurable distortion on an AP2322, low cost, reasonably low capacitance, and a nice slow switch function to prevent clicks.

I think you'll have to get used to SMT or stay in the 80s land of awful CMOS switches!
 
#11
switch current, not voltage, by loading it into a virtual ground.
Click to expand...

Wasted 10 minutes into stupid google, couldn't find a linkable image of this very common circuit, so this one will have to do with one important correction:

a) short R1 which is to the right of IC1

b) put an individual R1 at each of the 8 IC1 inputs , meaning each to the left of respectively a, b, ...h

that way the CMOS switches interrupt *current* getting into IC2 "-" input, which is the virtual ground..

2009618222931400.gif


Switches in general are more efficient at interrupting current (which need a conductive path to travel) than voltage, which can bypass such an open path by inductive or capacitive coupling which is inherent to any real world wiring.

By the way, the original example shown uses a virtual ground stage ... but connects the switcher in such a way that it switches voltage 🙄
The way I suggested corrects that 🙂
 
#12
monte:

that's an interesting part i hadn't heard of before and your AP measurements are impressive.
thanks for the tip.
too bad it's limited to 3v rails; kinda limits the application, although i'm sure it's great for PCs, cell phones, and tablets.
if i can't have +/-15v anymore, i'd at least like the +/-7v of the CDxxxx CMOS from my younger days.
ahh well, nothing like progress ...

mlloyd1

Monte McGuire said:
The Intersil ISL54405 is a fantastic switch for audio, but it's only surface mount and you need to use it to switch current, not voltage, by loading it into a virtual ground. I have a test circuit using it now - off isolation is over 140dB, unmeasurable distortion on an AP2322, low cost, reasonably low capacitance, and a nice slow switch function to prevent clicks.

I think you'll have to get used to SMT or stay in the 80s land of awful CMOS switches!
Click to expand...
 
#14
mlloyd1 said:
monte:

that's an interesting part i hadn't heard of before and your AP measurements are impressive.
thanks for the tip.
too bad it's limited to 3v rails; kinda limits the application, although i'm sure it's great for PCs, cell phones, and tablets.

mlloyd1
Click to expand...

Actually, this circuit can handle a ±50V input signal… with the 3.3V powered switch.

In my test circuit, I use a 10KΩ resistor in series with the switch input and place the switch output at the inverting input of the current to voltage converter stage. So, the switch has 1.9Ω resistance, and the feedback resistor is 3KΩ, so I can put ±50V into the input, limited essentially by the resistor's breakdown ratings and the voltage swing of the output amp. The 3.3V supply of the CMOS switch is not an issue - the switch sees 10mV or so of signal with a 50V input... Again… I'm switching _current_ and all of the voltage has been dropped by a factor of 5000 by the input resistor ahead of the switch. The inverting current to voltage amplifier that follows the switch converts the signal current back into a voltage and all is well again.

You have to compensate for the shunt capacitance of the switch, because it destabilizes the current to voltage converter stage, but it's not such a big deal. Each switch is around 27pF, so the product of the feedback resistor and the feedback cap should be greater than the product of the switches shunt capacitance and the input resistance. Not hard to compensate for at all. In this case, with a 27pF switch and a 10KΩ input, I'd use 100pF across a 3KΩ feedback resistor, and it's stable.

Current mode is a far better thing to do with a CMOS switch - you eliminate on-resistance variations with voltage, and also shunt capacitance variations with voltage, both of which generate distortion. Eliminate the voltage, and you eliminate most of the problems with a CMOS switch.

Off leakage also gets nice since you can shunt the input signal current to ground when the switch is off - it's a DPDT switch, so when it's on, you can wire it to feed signal current into the virtual ground, and when it's off, send the signal current to ground. So when the switch is off, there is very little signal voltage that could feed through to the output amplifier.

Again… current mode is the way to use a CMOS switch. Voltage and CMOS do not get along well, so twist up your circuits and use them the right way!! 🙂

Regards,
 
#20
Luke123 said:
Hello

I've been using the SSM2402 audio analog switches for a long time with excellent results but I just discovered it's been discontinued since few years.

Any suggestion for an equivalent, Analog Devices or other?

I need to apply on ±18 V or less supply, below 100 ohms ON resistance, great isoltaion resistance, very low distortion, 2 x spdt function, pin DIP version, Rhos or not...

Thanks!

Luke
Click to expand...

Still have those in stock (new/unused), also the SSM2404. If you want to trade for parts please send me a PM.
 
Status
Not open for further replies.
Top Bottom