Relay driver question....

[G. Hoffman]

So, on this relay driver from GeoFex





The first two inverter sections are working as a flip flop/latch, right?  So this assumes that you are using a momentary footswitch.  I'm working on a schematic to rebuild my amp, and I need a few relays, but I'm driving them from a my MIDI switcher, which by its nature will need to deal with all the latching stuff. 

So, my question, if I want to use this with a latching footswitch (or my MIDI switcher), I just need to drop the flip flop part of this, right?  (i.e., the first two inverters, and the related 1M and 100K resistors, and the cap.)

Actually, for that matter, can I just drop the first three inverters?  The signal from the switcher is already quiet well debounced (the actual switch element for this is 1/4 of a 4066, driven from a 74HC573 latch, driven from the uC - lots of places for debouncing, and the latch isn't too fussy about switch bounce anyway).  I've got three relays in the amp (two effects loops and a boost), so cutting down the number of Gates required would be awfully nice!


Gabriel

[head_spaz]

You can drop the left half, it's just the logic part of the circuit.
But you might want to include a couple of inverters, or a flipflop with complimentary
outputs, to guarantee that the logical "off " and "on" signals to the drivers cannot float,
as can happen with long cables, or iffy connectors.

[Processaurus]

"Latching relay driver" meant a driver for latching type relays (even though a momentary footswitch is made into a latching one with logic, funny mixup there).

Absolutely, skip the first 3 inverters if your incoming logic signal shares the same rails as the rest of the inverters will, as well as source enough current to light the LED.

I was just researching this circuit, to similarly use it with an existing logic signal representing the bypass. It's a twisty one with all the inverters, interesting how the logic state gets turned into a positive pulse by the inverter before it going low, and pulling the voltage through the 100K pullup resistor down, through the .1 cap.  Also a nice feature is the .1 across the driver transistor collector and base, this is aimed at slowing down the transient into the coil, to make less chance of a pop from the suddenly energized coil getting into the audio.

If the driving voltage waveform of the relay coil rises and
falls gradually, the inherent performance of the relay may not
be fully realized. Make sure that the voltage waveform
instantaneously rises and falls as a pulse

I wonder if the .1 cap gives the right ramp, NEC's datasheet below specifies 1ms max on the rise and fall time of the switching pulse.


Question of my own: can a 5v relay be safely kludged to run on 9v?

R.G. Keen had recommended the NEC EA2 series, on that recommendation and the price ($2.50) I was looking at the 5v part Mouser carries, # EA2-5TNJ.  

On the datasheet, coil resistance of the 5v version is 178 ohms, so, I'm guessing a 142 ohm (1/2 watt) resistor between +9 and the + on the coils would drop the voltage seen by the energized coil to 5v?  The AC resistance of the coil would be more at the transient as the coil got switched on, so the coil would regularly be seeing a higher voltage than 5v momentarily, but then again, the driver has a ramp up...

Or, would it be better to not fully turn the driver transistor on, but in an in between state, that would give 4v (9v minus the coil's desired 5v) on the collector when the inverter pulses it?

[Processaurus]

Ah, found this, other people like the resistor kludge it seems.

If you have 9V available already and also have 5V relays already, consider either putting a resistor in series with each relay coil so that the steady state current from +9V is still 28ma per coil, or putting in a 7805 three terminal regulator to drop 9V down to 5V just for the relays. Either way works. If you go with resistors, the added resistor in series with each coil is R = V/I = (9V-5V)/0.028 = 143 ohms. A 140 ohm resistor is a standard value and will work fine. The power is P = I^2*R = (0.028*0.028)*140 = 0.109W, which is pretty hot to run a 1/4W resistor over the long term. I'd use two 270 1/4W in parallel or two 75 1/4W in series, or go to 140 1/2W.

[R.G.]

"Latching relay driver" meant a driver for latching type relays (even though a momentary footswitch is made into a latching one with logic, funny mixup there).

Absolutely, skip the first 3 inverters if your incoming logic signal shares the same rails as the rest of the inverters will, as well as source enough current to light the LED.

I was just researching this circuit, to similarly use it with an existing logic signal representing the bypass. It's a twisty one with all the inverters, interesting how the logic state gets turned into a positive pulse by the inverter before it going low, and pulling the voltage through the 100K pullup resistor down, through the .1 cap.  Also a nice feature is the .1 across the driver transistor collector and base, this is aimed at slowing down the transient into the coil, to make less chance of a pop from the suddenly energized coil getting into the audio.

Yep.

If the driving voltage waveform of the relay coil rises and
falls gradually, the inherent performance of the relay may not
be fully realized. Make sure that the voltage waveform
instantaneously rises and falls as a pulse

I wonder if the .1 cap gives the right ramp, NEC's datasheet below specifies 1ms max on the rise and fall time of the switching pulse.

Probaby not, and intentionally so. The 0.1 cap was picked to give a guess at the least pop. The "inherent performance of the relay" includes switching times, and I'm guessing that any slowdown in the on/off pulses to the coils will NOT give the switching times NEC states. But this is a matter of objectives. As an effects user, I/we don't really care that the relay flips in 10mS instead of 2mS. We want the audio to be quiet. There's no "clik" line in the NEC datasheet, but that's perhaps the most important thing to us.

There is no magic in the 0.1uF value. Adjust for what (a) makes the relay operate (b) as quietly as possible.

[G. Hoffman]

Thanks all, I appreciate it!



Gabriel

[G. Hoffman]

OK, so another question on this.  I'm putting this all in an amp which has an opamp effects loop running at ±15V.  I've got 12VDC relays, so I need a load resistor on the relays.  Simple enough, but what I'm wondering is if I can put them on the emitters of Q1 and Q2?  And, can they share one resistor?  It makes the board design a lot easier!  (Well, putting it on the emitters does, sharing a resistor just saves a bit of space, but valuable space.)


Gabriel

[G. Hoffman]

So, looking at this schematic for this amp (which is almost done now), I'm quite liking it.  I particularly like the dichotomy of having both a tube rectifier AND a bi-polar solid state cascading regulator PS in the same amp.  One tube and one solid state effects loop,* a simple tube boost switch,  and if I turn off the bells and whistles, it is a 100% pure tube signal path that hits 1/2 of a preamp tube (no tone controls - I never use them), a PI tube, and two EL84s.  Basically, this is the amp I already have, but I don't have the effects loops yet, and since I can't seem to give up my delay pedals, but they interfere with the dynamics of my amp, I needed to do something about it.


Gabriel


*The solid state is for things that will run in series, where it really doesn't matter if the loop is solid state, because if it is on it will be running through other transistors, so who cares? and the tube is for my delays, which will be in parallel, so the dry signal is never has to touch a semiconductor, just vacuum!  Both loops are from Kevin O`Connor's The Ultimate Tone.)

[Processaurus]

OK, so another question on this.  I'm putting this all in an amp which has an opamp effects loop running at ±15V.  I've got 12VDC relays, so I need a load resistor on the relays.  Simple enough, but what I'm wondering is if I can put them on the emitters of Q1 and Q2?  And, can they share one resistor?  It makes the board design a lot easier!  (Well, putting it on the emitters does, sharing a resistor just saves a bit of space, but valuable space.)


Gabriel

That seems kind of wrong, having the emitters tied together, to a single resistor to ground.  Even a single resistor between V+ and the +'s of each of the coils seems weird, as there would be interaction of some kind between the coil being pulsed and the other coil minding its own business.  I would just splurge the $.02 and put individual resistors on each coil, to V+.  Or not worry about it, +15 is 125% of the rated voltage, the NEC relays at least specify they can take up to 150% of the nominal voltage, at room temperature.  It goes down, however, as it gets hotter.

[PRR]

> can they share one resistor?

You know Ohm's Law. Use it. One resistor for two loads? If it drops 3V for one load, it will drop 6V for two loads. (Actually not quite, but you see the issue.

Is your 15V regulated? Read the relay datasheet. Most such are rated 150% over-voltage. 14.9V may be within ratings. The coil will run warm. Whether this is still kosher when you have jammed things too tight inside a hot tube amp may be a different question.

{EDIT: Ben beat me.}

[G. Hoffman]

Makes sense.  The only reason I want the resistor on the emitters is the PCB is to avoid a jumper on the PCBs - or to use the already needed resistor as the jumper, at any rate.  I can find another way to skin that cat, though.  I have no problem using separate resistors, though - its just getting a bit crowded in there, and I wanted to save board space. 

For a dirt simple tube amp, this thing is getting awfully complex!  I'm also thinking I might want to add a 12V regulator for the relays - not because I'm so worried about getting them the right voltage, but more because I'm going to have an op amp effects loop on that power, and isolating it a bit more from the relays makes my love of over-engineering happy.

So, no on both counts, at least in the good idea sense.  Thanks a bunch!


Gabriel

[Processaurus]

For a dirt simple tube amp, this thing is getting awfully complex!  I'm also thinking I might want to add a 12V regulator for the relays - not because I'm so worried about getting them the right voltage, but more because I'm going to have an op amp effects loop on that power, and isolating it a bit more from the relays makes my love of over-engineering happy.

The slow down from the .1's on the transistors driving the coils should help smooth the spikey transient on the power rails, and opamps have good power supply noise rejection, especially with bipolar power.  I'd be more worried about the grounding, and star ground the driver circuit back at the PSU.

> can they share one resistor?

You know Ohm's Law. Use it. One resistor for two loads? If it drops 3V for one load, it will drop 6V for two loads. (Actually not quite, but you see the issue.

It's strange, in theory, you might actually be able to get away with the one resistor, because in this specific case the two loads are by design, never on at the same time.  But that's just an academic observation, I'm slowly learning, even in production, it's often not worth worrying about whether a resistor is extra, if you know it will work with it, and have a measure of doubt it will work without it.

[G. Hoffman]

I'd be more worried about the grounding, and star ground the driver circuit back at the PSU.

Well, I typically star ground everything, so that's no much of a problem!  Though I am thinking of making this amp on a PCB (though a PCB with turret lugs, so modding is easier, things are more reliable, and that whole over-engineering thing!), so I was thinking I might use a ground buss or two.  I'm waiting for Kevin O'Connor's TUT 3 right now, which has a big section on wire routing and such, so I want to read that before I really get into it.


Gabriel

[Processaurus]

Just was curious, why not use a non latching relay, since your amp isn't battery powered?  Then everything is dead simple.  Relay, transistor coil driver, diode across the coil.

[G. Hoffman]

Just was curious, why not use a non latching relay, since your amp isn't battery powered?  Then everything is dead simple.  Relay, transistor coil driver, diode across the coil.

Well, for one I like the idea of not having the coil radiating anything when it's not operating.  It probably doesn't matter much - if at all - but hey, why not, right?

But mostly, I wanted to try using these, because I'm pretty comfortable using non-latching relays, and I want to try something I'm not familiar with.  That's how we learn, after all.  I've used single coil latching relays, but never a dual coil.  Its not a big difference, but hey, every little thing you can learn, right?


Gabriel

[R.G.]

Well, for one I like the idea of not having the coil radiating anything when it's not operating.

N. B.
Coils don't "radiate" anything that can be picked up as audio. DC coils have a static magnetic field. Well, they do if you feed them DC, not pulsed hash.
You can't hear DC.

[jacobyjd]

Well, for one I like the idea of not having the coil radiating anything when it's not operating.

N. B.
Coils don't "radiate" anything that can be picked up as audio. DC coils have a static magnetic field. Well, they do if you feed them DC, not pulsed hash.
You can't hear DC.

but...but...there are a ton of people on TGP who say they can! 

[R.G.]

but...but...there are a ton of people on TGP who say they can! 

Yeah, I guess there are. 

It's perhaps a good idea that Mother Nature doesn't let us vote on modifications to physical laws.

Or rather, we can vote all we like. She just ignores the results, like the rest of our blathering. 

[jacobyjd]

Seriously though--the progression of this thread has been a good read--I've got some project ideas that have been sitting dormant until I can collect my thoughts on relay switching. This is why I love this place.

[G. Hoffman]

Well, for one I like the idea of not having the coil radiating anything when it's not operating.

N. B.
Coils don't "radiate" anything that can be picked up as audio. DC coils have a static magnetic field. Well, they do if you feed them DC, not pulsed hash.
You can't hear DC.

Makes sense.  Though, of course, making real DC from AC is never particularly perfect. 

I still stand by the "learn new things" point, which is by far the bigger issue anyway.  I mean, whats the point if you're not learning something new, right?  At least, that's half the point for me. 

If I just wanted the a bunch of amps and pedals, I'd save a lot of money by just buying them.

but...but...there are a ton of people on TGP who say they can! 

It's perhaps a good idea that Mother Nature doesn't let us vote on modifications to physical laws.

Try telling that to the general populace of the world; much less the internet! 


Gabriel