Relay oscillator?

[Svart]

I've cobbled together a circuit based on some articles I read that will reduce the current drawn by the relay's coil by pulsing it after an initial latching period of full current.

The question is..

The system, as designed, runs at 42khz during pulsing, do you think that noise from the coil could possibly couple to the contacts/conductors in the relay itself? I can move the freq up if needed but would like to utilize this in order to keep my PSU requirements low in case of large scale usage of relays in some boards I am thinking through.

It's looking like prototyping is the only way to tell for sure..

 

[bcarso]

Well---two kinds of coupling---electric and magnetic. Magnetic is the tougher one to combat but look at the way the loop area of the contacts intercepts the coil field to anticipate the situation. Depends a lot on what your switched circuit is like as to the importance.

Electric fields could be alleviated somewhat by having the power supply connected to the outside layer of the coil so that the driven part is inside. Also I would slow down the rise-fall times of the driving pulses in a tradeoff with efficiency.

I'd be tempted to look at the coil self-resonance and try to do something with that---making the pulses switch near current and/or voltage zero-crossings. This frequency may be too low though.

I'll be interested to know your results. I have a related application.

 

[Svart]

Yes I did mean magnetic coupling from the coil's field and the related RF noise from the collapsing field.

Interesting idea about driving the coil from the inside out. I doubt it would be easy to figure out which lead is the inner lead of the coil though..

I like the idea of the resonance switching. I'm trying to keep the freq high enough to not be totally within the audio bandwidth but low enough to not emit RF junk.

I suppose a prototype is in order.

I'll try to get one together soon.

:thumb:

 

[bcarso]

[quote author="Svart"]Interesting idea about driving the coil from the inside out. I doubt it would be easy to figure out which lead is the inner lead of the coil though..
:thumb:[/quote]
Drive a high frequency signal into each end, alternately, and hold a 10M scope probe near the outside of the coil, with the scope common referenced to the generator common. It should be very evident.

I do this a lot with axial-lead wound-film caps to see what the outside foil lead is.

 

[Svart]

awesome. I'll give it a try.

 

[bcarso]

Just for fun I pulled out a relay and measured the resonant frequency. Good news: it was around 50kHz. Not-as-good news: it was very low Q. But there still might be an advantage to quasi-resonant switching.

 

[mwkeene]

Ha, not to derail this but I have a somewhat closely related story:
When I started playing pedal steel guitar my father had me meet up with one of his colleagues from IBM who also played pedal steel. Turns out this guy had really big feet, too big to play some of the "inbetween" foot pedals individually, so what he did was fashioned solenoids to the rods that pull down the levers that you push with your feet. Then, he hooked up 5 or 6 foot switches that would each activate different combinations of these solenoids to pull down combinations of levers. Hit button one, you get pedal one and two pulled down, hit button two you get pedal 3 pulled down, etc... Long story short, he burned out a lot of solenoids in the process of building this, and he realized that after a few seconds at the current required to initally turn the solenoid on, the part would become too hot and blow, so he said he fashioned an "apparatus" to limit the current consumption after the initial turn on. Never said how he did it, but I bet this could be it.

Back to the topic at hand:
Could these relays recieve a limited DC current (maybe from a pass device) after the initial "ON" current or will DC not work? This seems like it would eliminate the possibility of frequencies being added to the signal...
Maybe I should get some relays and try for myself before asking stupid questions, huh? Guess I can't curb the temptation to prod you guys to talk about it.

-Mike

 

[bcarso]

[quote author="mwkeene"]Ha, not to derail this but I have a somewhat closely related story:
When I started playing pedal steel guitar my father had me meet up with one of his colleagues from IBM who also played pedal steel. Turns out this guy had really big feet, too big to play some of the "inbetween" foot pedals individually, so what he did was fashioned solenoids to the rods that pull down the levers that you push with your feet. Then, he hooked up 5 or 6 foot switches that would each activate different combinations of these solenoids to pull down combinations of levers. Hit button one, you get pedal one and two pulled down, hit button two you get pedal 3 pulled down, etc... Long story short, he burned out a lot of solenoids in the process of building this, and he realized that after a few seconds at the current required to initally turn the solenoid on, the part would become too hot and blow, so he said he fashioned an "apparatus" to limit the current consumption after the initial turn on. Never said how he did it, but I bet this could be it.

Back to the topic at hand:
Could these relays recieve a limited DC current (maybe from a pass device) after the initial "ON" current or will DC not work? This seems like it would eliminate the possibility of frequencies being added to the signal...
Maybe I should get some relays and try for myself before asking stupid questions, huh? Guess I can't curb the temptation to prod you guys to talk about it.

-Mike[/quote]

The current reduction-after-pull-in scheme is fairly old and certainly works. I believe the appeal of what Svart describes is even greater efficiency.

 

[Svart]

Yes greater efficiency.

It is essentially DC to the coil, just in pulses. the trick is to pulse fast enough but with shorter pulses to keep the coil stable but not with excessively long pulses that would allow the current to rise beyond what I intend it to use. I could form a filter for the coil averaging the DC pulses to straight DC but that would bring my average voltage down considerably and likely causing the relay to unlatch.

Answering your question directly:

Reducing current will normally reduce voltage but reducing voltage will normally cause current draw to rise, except in a case like this where the current flow is impeded and then they both will fall. (that pesky ohm's law deal) the coil uses the voltage/current to generate a small mag field and heat. That would easily consume a reduced current flow. we pulse it to keep the coil saturated (within reason) and magnetized but the pulses are short so the average current draw is cut down considerably.

 

[Svart]

Brad, Looks as if I'll be going with the SSM2402 and not the relays.

I'll send you the relay osc. schemo if you want to look at it.

:thumb: