Single coil latching relay help needed

[Ethan]

I tried  the following driver circuit by Transmogrifox and it works.

http://www.geocities.com/transmogrifox/relay

However, it does not work when Vcc is 5v. I've been playing around with caps but no luck yet. 
I am using a NAIS 5v latching relay, single coil with a coil res. of 250 ohms and current req. of 20mA. ANy ideas what's up?
-ethan

[R.G.]

I'd bet that it doesn't supply enough current for long enough to flip the relay over. You need either or both of a bigger capacitor or a beefier driver.

That circuit needs some catch diodes, too, I think.

[Transmogrifox]

I designed it for a 460 (ish) ohm series resistance relay.  RG is right, it needs a bigger capacitor.  Put a 10 uF in there and I think it should trip.  I'm not completely convinced it needs catch diodes.  I thought about it a bit, and what I came to is that the outputs are only high impedance during switching. If it was only an inductive load, then there would be DC current in the coil at the time of switching, and it would be rapidly interrupted during switching, and would very likely generate a significant and potentially damaging kickback, however, the capacitor is completely charged and the system is static upon switching so there is not a rapid change in current through the coil.  I have not calculated it, but the only thing that could cause a damaging voltage potential for the inverter would be resonance between the coil and capacitor.  I think it is adequately damped, since I had been using 3 of them for about 3 months in my envelope filter before I applied 23 volts to the power supply and roasted the inverters (the absolute max on them is 18V, so they should not be operated above 15V). 

Anyway, try the larger cap.  If 10 uF doesn't work, then go up to 22 uF, then 47 uF....I have been able to operate it at lower voltages with larger caps so I'm pretty sure it will fire.

Take it easy.

EDIT:  It looks like you said you've been playing with caps...I should look up the datasheet for the relay you're using.  I can see one difference:  the NEC relay I designed for (EA2-5SNJ)  requires 10mA for switching, and it's actual typical switch voltage is about 3 Volts.  Another thing I looked into is the CD4049 datasheet rates the device to 5mA sink current, and about 2 mA source current at 5V.  At 10-15 volts, it's much higher.

I think RG is right again:  You need a larger capacitor AND a beefier driver--or you can wire all 6 inverters in parallel in the CD4049 and see if it works at 5V.  This will give you a rated 30 mA sink current (within spec) but only a 12 mA rated source current (might prove problematic).  I knew there could be some practical use for the other inverters  .  Give it a try since it's not that much more trouble just to jumper all the outputs and inputs together on the IC.

If you have an SPDT latching switch, then you can do away with the inverter and connect the center lug ("lug B") of the switch to the relay, and the lugs A and C  to V+ and V- respectively.

Otherwise, if you're using the SPST or a voltage to drive the input, then you need an active driver of some kind.  Maybe somebody else will chime in about a good driver IC for a higher current relay such as yours. 

Otherwise, get a 551-EA2-5SNJ from Mouser, and I know this works at 5V because I've tried it.

[Transmogrifox]

  Give it a try since it's not that much more trouble just to jumper all the outputs and inputs together on the IC.

Errr...ummm... Give it a try since it's not that much more trouble to jumper all the outputs together, AND all the inputs together...DON'T jumper all the outputs and inputs together...just all inputs to inputs, all outputs to outputs so you effectively have 6 parallel inverters.

I probably just made it confusing by catching myself in a little sentence structure error.  I hope I cleared it up before your CD4049 became a solder ball.  If you're soldering it, you can probably fold all the output pins down under the chip and solder a little wire between them all, then all the inputs to the top of the chip--same thing, then take your connections off the 2 sides...anyway

[Ethan]

Doh, I didn't read data sheet for 4049, it can't supply enough current! I will try parrallel for quick fix but I want to eventually hook trigger it with a pic.  That's why I am testing at 5v. Hmm your relays only need 10mA to actuate? Maybe I'll check those out. Thanks for your tips guys.
-ethan

[Ethan]

Works fine now. Excellent circuit.  I didn't scope any HV kickbacks at all (this was not using any catch diodes either).  I am using 470k pull down res. instead of 4.7k. So the current draw is like 10uA! I had to increased cap to 22uf tant also.  This slows down ramp voltage to about 20mS but still trips the relay fine.  I gotta try it with my guitar to see if this is noticeable.  This simplifies my original bypass circuit significantly since previously I was using a pic, back to back zener diodes, 4 i/o pins (switch,led,relay on, off) and some crazy code to put unused i/o in high impedance mode to protect pic from spikes. 

[Transmogrifox]

I'm glad to hear it was that simple.  I hope it ends up being a reliable circuit for you.  My intention was to reduce parts count and current draw in the switching department.