MPC50x analog mux for audio?

[paulb]

Has anybody looked at these as an alternative to relays or the Analog Devices muxes?
MPC507: 8-channel differential (or stereo)
MPC509: 4-channel differential

Pricing seems reasonable, about $8 CDN apiece.
http://www-s.ti.com/sc/ds/mpc507.pdf

[paulb]

OK then, how about any standard analog muxes for audio use?

[PRR]

> MPC507: 8-channel differential

These are second-source for a very old Harris Gates part.

The series has been available for about 20 years, so it is probably not going to vanish soon.

These used to cost a heck of a lot more than $8, though I do recall them falling past $25 quite a few years ago.

Voltage handling is very good.

ON-resistance is very high. You need to consider this in all phases of design. The last time I considered them, the ON-resistance promised more noise than I could accept in a microphone preamp. It is not a big issue in line-level work. However the resistance variation with voltage, while small, is not clearly specified. You typicially need a LARGE swamping impedance to mask Ron variations and get low THD numbers.

In fact Ron for this heavily-protected part looks higher than I remember on the Harris parts.

Note this TI tip:

"Keep loading impedance as high as possible. This minimizes the resistive loading effects of the source resistance and multiplexer ON resistance. As a guideline, load impedance of 10^8Ω or greater will keep resistive loading errors to 0.002% or less for 1000Ω source impedances. A 10^6Ω load impedance will increase source loading error to 0.2% or more."

They want you to use 100MegΩ load! The error they are discussing isn't what we call distortion: THD will be lower than static DC error. But clearly you want at least 1 Meg.

Ah, TI has good Ron curves. Not very precise: drawn flat/level from -10V to +4.6V, but you can eyeball significant error above +5V. This may not be too important at line-level, and you may not need a super-high load impedance.

[dhaen]

Yes, I used them in production about 10 years ago. They sounded OK to me and many others. Since then I've "regressed" to relays. They're faultless by comparison, but have a finite lifespan.

[paulb]

Thanks for the background. The datasheet does mention second-source, but a quick google didn't pin it on Harris.
I think NatSemi's audio handbook had a section on using analog muxes like the 40??, I'll check that out.

Paul

[PRR]

> quick google didn't pin it on Harris.

I don't know that Harris, or Gates, makes them any more.

I think Intersil got the government contract and the tooling.

The original part numbers start DG or ADG.

Google "509 multiplexer" got me these hits:

http://www.intersil.com/products/dev...o.asp?pn=DG409
http://www.intersil.com/products/dev....asp?pn=HI-509
http://www.intersil.com/data/fn/fn3142.pdf

http://www.xs4all.nl/~ganswijk/chipdir/f/mux.htm

[rho]

I got some samples of the MAX397 (Maxim/Dallas Semiconductor). How do they compare to the MPC507? Thay are pin-compatible.

[Johnix]

Originally posted by PRR
ON-resistance is very high. You need to consider this in all phases of design. The last time I considered them, the ON-resistance promised more noise than I could accept in a microphone preamp. It is not a big issue in line-level work. However the resistance variation with voltage, while small, is not clearly specified. You typicially need a LARGE swamping impedance to mask Ron variations and get low THD numbers.

Does this mean that you can't use them to source a power amp ?

Power amps have input impedance more or less around 50k

Having a 1k input source is still a problem ?

[Papageoff]

I got some sample of the mpc507 and was wondering if someone could answer the last question posted

Thanx

Originally posted by Johnix



Does this mean that you can't use them to source a power amp ?

Power amps have input impedance more or less around 50k

Having a 1k input source is still a problem ?

[officeboy]

So for multiple source/output switching is this a dead end then?