MEMS relays are back!

I looked at the data for the MM1200 – 6 Channel SPST Micro Relay There do seem to be a lot of restrictions relating to hot switching and floating nodes. The description "ideal switch" is perhaps going a bit far. John

One issue I would be concerned about is the fact that it actually consists of many relays connected in parallel. When they are opened, they can not be opened simultaneously, but rather open sequentially, even if very close together in time. This would tend to concentrate the current flowing until the last relay contact is opened, with all the current flowing through that one contact. I suppose this is limited by parasitic inductance and the consistency in timing of the contacts opening. If they all open at the same time to within a nanosecond, then I don't see much problem.

I suppose this issue is reflected in the life rating which is pretty good, at 3 billion operations, so not really a problem.

I would find a lot of applications for those operating in the sub-mA/3V range. Purchasable or "I had a dream" Silicon Valley startup vapourware?

Best regards, Piotr

A lot of MEMS relay companies have arrived with great expectations and then quietly died. I'll wait for Digikey to have parts, and try some.

MEMS seem OK for resonators and accelerometers, but contacts seem to have welding/sticking issues.

Here's another one:

Seems like they could make interesting filters, like the old Collins mechanical filters.

So will I. One remark about the 6 SPST BGA: they require +79V external supply. Without reading the datasheet, I guess it is an electrostatic relay, not magnetic. Quite interesting, but wait... they can make a MEMS relay and fail to make a built-in charge pump?

Best regards, Piotr

That website smells of dotcom.

The main issues with MEMS switches are materials (good materials for MEMS fabrication generally aren't great electrically and mechanically) and the tendency of the contacts to weld or vaporize under fairly mild abuse, because they don't have a lot of mass to take the heat.

If the Menlo folks have figured out how to fix those things, I'm all for it.

Cheers

Phil Hobbs

Did you miss the part about where you have to turn off the current through the contacts before you are allowed to switch the MEMS relay? In their power relay they use a parallel MOSFET to do this. Given that the MEMS relay has not that great on-resistance, you could probably leave that bit out, and just use the MOSFET. In the RF relay, there is no MOSFET so you have to turn off the RF power yourself, e.g. set modulation of the signal such that the carrier goes away, then you can switch it.

Amusingly, 79V, not 77V and not 81V. Perhaps they only test it at one voltage and so don't know whether the parts work at other voltages.

Or worse, maybe it really cares.

It likely does:

78V min, 80V max. This accuracy requirement calls for a built-in boost converter even stronger. It is not exactly trivial (in terms of part count and board space) to provide such a power rail.

Best regards, Piotr

That tolerance makes the gadget sound really flakey. As mems relays have historically been.

This one appears to exist:

And it has the charge pump. But just 7.2 years of continuous on lifetime?!

Best regards, Piotr

What a horrible part!