RF relays

I take that back, they would probably span too long a line or arc to be convenient. Then again, there is no avoiding the fact that you need separation in order to attain a low capacitance.

There seems to be a new part, AQY221N3T, which I haven't seen before. It would make the chain even more compact.

Regards, Mikko

I'm building a spectrometer for another customer that uses an RC servo to move a grating, but it's too noisy and too large for this application, unfortunately.

Cheers

Phil Hobbs

ge

well,

so far

lay

Hey neat idea... can I steal it? Is the servo stable enough?

George H.

t -

So, you're saying it grates on him? ;-)

Dunno yet. It's for a pretty low-resolution application. You can also get programmable servos if the analogue control is too jittery.

Servo City sells nice hub parts that can easily be hacked^H^H^H^H^H customized to hold gratings

Cheers

Phil Hobbs

Trying to use "standard" relays for switching in / out would seem to be a lost cause. As mentioned, the capacitance between the reed and driving coil is (relatively speaking) huge. Furthermore, not mentioned, is the fact of inductive coupling from the coil to the reed. At Fairchild, in the daze of the uA709 and uA741, we were designing a test fixture to measure input current of (at the time) new FET-input op-amp. The idea was to use a standard sample and hold scheme; close a relay to connect DUT input pin to capacitor and measure capacitor voltage at a later specified time; voltage follower at capacitor top. We used the same op-amp for the follower, but it was hand-selected from a fair-sized pile. Capacitive coupling created a step voltage increase in the sampling capacitor (AC voltage divider), and worse, the inductive coupling compounded the problems. We wound up using a discrete non-protected gate MOSFET for the switch, as the leakage was comparable and the capacitance was less.

In this case, a well-cleaned reed relay firmly supported by air (NO standoffs) will have adequately low leakage and can be arranged with a wide-gap magnetic loop to minimize capacitance to the driver coil. A shield directly around the coil will also help reduce dV/dT coupling; another trick is to use bipolar (symmetrical) drive = = one end of coil driven negative and other end driven positive.

Hope this is of some help..

No, I think it'll work fine. The coil can be bootstrapped pretty effectively, at which point the coil-to-contact capacitance becomes a

With a latching relay, this is only a worry when the coil is energized, and of course one wouldn't expect a stable output at that point--the low frequency gain of the stage is changing by 50x.

Thanks. It isn't ideal, but I think it ought to work fine.

Cheers

Phil Hobbs

"Latching relay" machts nicht..i was referring to the turn-on and turn-off dV/dT spikes combined with the magnetic flux changes which can create problems in sensitive work like this.

Balanced / differential drive helps, combined with a biased drive, ie maximum current that guarantees an OFF condition, allows a much-reduced drive needed for an ON condition. Combined,one has the least induced signals.

Naturally, one can capacitively couple an appropriately shaped pulse of the opposite polarity to more-or less cancel the seen induced spikes.

A remote magnetic actuator is good only for reducing the capacitive component as the field needed for reed activation cannot be reduced. The sneaky drive scheme mentioned can "dramatically" reduce the magnetic spike components.

I'm not expecting the amp to be well behaved during switching, so as long as the switching spikes don't actually damage it, I'm not particularly fussed about them.

It's the stray capacitance during operation, i.e. between switching events, that's the killer here.

Cheers

Phil Hobbs

Then firmly weld the reed relay to the air and put the magnetic driver (includes any iron to control your flux capacitor*) as far as space permits.

• Could not resist..