A latching relay question

or

at

uld

el

maybe there can be unwanted coupling between the set and reset coils such that when you pulse say the set coil of RL1 it may feed some energy though the interconnected reset coils and create an unwanted reset to RL2. In your specific case since BOTH set and reset are connected in parallel, it probably won't be a problem. In other cases where they are only partially interconnected, there might be unwanted state changes.

From the construction of the relay, does it look like there could be significant coupling between the set and reset coils?

Mark

Don't the diodes act a balancing ballasts too?

Can't see why you can't just parallel them. The extra diodes make no sense.

We use this to drive small latching relays:

Just in case you didn't know, latching relays (at least the ones I've used) have a minimum physical separation - if they're too close, the magnetics can interfere.

Cheers

I like your scheme; much simpler that what I had in mind.

What is the purpose of Q1 and Q2 in your schematic?

Many thanks.

maybe there can be unwanted coupling between the set and reset coils such that when you pulse say the set coil of RL1 it may feed some energy though the interconnected reset coils and create an unwanted reset to RL2. In your specific case since BOTH set and reset are connected in parallel, it probably won't be a problem. In other cases where they are only partially interconnected, there might be unwanted state changes.

From the construction of the relay, does it look like there could be significant coupling between the set and reset coils?

Mark

The relays are sealed so there no way to determine how they are constructed.

Those are Supertex depletion-mode mosfets. They are current limiters. This is the front-end of a differential-input ADC that can measure microvolts to hundreds of volts, and the depletion fets keep the customer from blowing out the opamps by applying too much voltage.

We used latching relays to keep the power down and to eliminate the microvolts of thermal EMFs that you'd get from the relay coil power dissipation.

In my case they will be about one inch apart. Is that enough?

no. the relays resistance and inductance do a better job of that.

Well, it'll be in the data sheet, but yes, one inch will be plenty.

Cheers

I started changing my schematic to control the relays using your method and realized that I could do the same thing with two inverter gates. I must be missing something; what is the reason for using two 2-input nor gates?

There are a lot of relays on the board, and they are matrixed addressed in groups of 8, which saves uP port pins. That takes one TinyLogic chip per relay.

In the uP code, in ram, there's a table of desired relay states and a table of actuals. Any process can set or clear any of the "desired" bits and just go away. An ISR compares the tables, bytewise, and if any byte is different it outputs 8 state bits on the F lines and asserts a select line until the next interrupt, 10 milliseconds or something like that. That automagically updates the relay states, only when it needs to be done.

Diodes in series with each coil will prevent a sneak circuit from the COM t hough the coil of the first relay then back into the coil of the second rel ay through it to its COM ( if that is the configuration you're dealing with ). I can see where there could possibly be this interaction via magnetic co upling within the relay so that the relays end up interfering with each oth er if the turn-off dynamics are not exactly identical. So, it is not a bad idea, diodes are cheap, use them.

That's a different kind of drive where SET/RESET is encoded by polarity of the drive. The OP is working with the two-coil type.

drive. The OP is working with the two-coil type.

I guess so. The circuit is easily fudged for the 2-coil case. This gate thing is for addressing a large matrix of 1-coil latching relays. You could do an X-Y drive matrix, with series diodes, to mux 2-coil types. Otherwise, I can't imagine what the series diodes accomplish.

The diodes block current in the mesh of the two coils in series through the relay coil COM node. Obviously Panasonic ran into the problem before. If he doesn't have the headroom for diodes then he has two go with four separate drivers.

I switched to a one-coil relay.

Thanks again to all who replied to my query.