# neg voltage relay driver (for an idiot)

Discussion in 'Electronic Basics' started by [email protected], Aug 27, 2005.

1. ### Guest

Hi. I have a need for a circuit wherein 2 relays respond to a voltage
state of either +24V or -24V. The off state for both should be when
the line is at 0V.

I tried hanging two relays off the line with diodes to block current
depending on the polarity at the time. This didn't work, probably due
to the source driver not providing enough juice, so I added an NPN
transistor driver to the +24V-responding relay and it works great. For
the positive 24V state.

Do I need to use a PNP transistor driver for the negative state
voltage? If so, can I get rid of the isolation diodes that I was using
when I hung the relays on the line directly? Should these be left in
for isolation from the driver circuit anyways? I am using a bypass
diode on each to snuff inductive jolts from the relay when powering
off.

The standard PNP relay driver circuit that I've found shows a circuit
for ON when the voltage = 0V, off for +24V or whatever. This makes
sense to me, but I am unsure about how to arrange the circuit so that
-24V appearing will cause the relay to switch on, and both 0V and +24V
will stay off (all this for the second relay that isn't yet working).

I know I'm 90% there with the circuits I have in front of me, I guess I
just hand getting my head around working with a negative voltage.

Thanks all,

Chris

2. ### John FieldsGuest

---
I'm confused. Do you have a single power supply that you want to
use for the relay coils and you want to turn one relay on if your
input signal line is at +24V, the other one on if the input signal
line is at -24V neither of them if your input signal line is at zero
volts, or do you have a single input line which can be at either
+24V, -24V or 0V and you want to use that as the supply to turn the
relays on with?

3. ### Andrew HolmeGuest

Here's one way of doing it (view in a fixed pitch font) :

V+
|
|
|<
.-------|
| |\
| |
| |
|
|/
.-----|
| |>
| |
=== |
GND |
| |
| |
___ | |/
Vin ---|___|-------o------|
|>
R |
|
===
GND

4. ### Guest

John, there is a single signal line that can be at -24VDC, 0V, or
+24VDC. But I have close access to +24V supply power if needed, which
is how I powered the NPN relay driver circuit for +24V state.

I guess I am looking for the inverse of:

http://www.uoguelph.ca/~antoon/circ/relay5.html

that operates when -24VDC appears at the input.

I'm a little unsure about Andrew's circuit. It apparently performs
amplification for both -24 and +24 signals, using 2 NPN's and a PNP?

Chris

5. ### Andrew HolmeGuest

Connect one relay (with backing diode) from the open NPN collector to +24V
Connect the other relay (with backing diode) from the open PNP collector to
GND

6. ### John FieldsGuest

---
Yes. With +24V on the input, the bottom NPN and its relay will be
turned on, but the ~ 0.7V drop across its base-to-emitter junction
will reverse-bias the middle NPN, turning it off. That will cut off
the flow of base current into the PNP, keeping the top relay turned
off. When the input goes negative, however, that will forward bias
the middle NPN, allowing current to flow out othe PNP's
base-to-emitter junction, turning it on, allowing the relay to be
energized. At the same time, the drop across the middle NPN's
base-to-emitter junction will reverse bias the bottom NPN, turning
it and its relay off. Clever circuit.

7. ### Guest

Excellent. Thanks a lot to the both of you.

The relays draw 6.3mA, 3.9K Ohms coil if I recall correctly. Should I
use 20-50k as Rin for the circuit instead of 10k? I was going to use
2N3904 NPN's, and the equivalnet PNP I guess, so something like Hfe of
30 gives requirement of .21mA base current to turn on the NPN. So I
doubled that and said around .5mA, so 48K Ohms? Or somehwere in
between...

10k does work in the circuit I am using now. I suppose it doesn't
matter that much, but I was thinking might as well keep the base
current lower if it is possible. Not that I have any idea why.

Chris

8. ### Andrew HolmeGuest

As you say, it's not that critical, but I suppose 10k is a bit low.

9. ### Guest

Thanks again Andrew. I'll throw a 20k in and see what happens. Now I
need to go buy a PNP, I think I have the other NPN's on hand.

Last question, the "backing diode" is just the diode strapped across
+/- on the relay, right? I know, that should be obvious.

Chris

10. ### Andrew HolmeGuest

Yes. Be sure you connect them the right way round or you'll blow the
transistors.

11. ### Guest

Arggh. So I am indeed an idiot. Andrew's circuit works fine...

It's too bad I was wrong about the circuit requirement. I did the
measurements a long time ago, and I didn't write my notes out
correctly.

So here's the difference.

Relay "1" should turn on when the input becomes +24VDC to ground.

Relay "2" should become active when input becomes 0V == GND. In this
state, the input also is -24VDC to the +24VDC source (positive meter on
input, neg on +24VDC line). This is where my notes were wrong.

Both should be *off* when input is floating, very high impedance to
ground. Basically unmeasurable.

Any more ideas?

Thanks again. Apologies for my bone-headedness.

Chris

12. ### John FieldsGuest

---

+24
|
+--------+-----+----+------+-------+
| | | | |K |
| | | | [DIODE] [COIL]
| | | | | |
[100.0k] [10.0k] | [10k] +-------+
| | | | |
| +----|-\ | C
| | | >--+----B npn
Vin---+-------------|+/LM393 E
| | |
| | GND
| |
| |
| | +24
| [20.0k] |
| | +------+-------+
| | | |K |
| | | [DIODE] [COIL]
| | | | |
| | [10k] +-------+
| | | |
+-------------|-\ | C
| | | >--+----B npn
| +----|+/LM393 E
[100.0k] | | |
| [10.0k] | |
| | | |
+--------+-----+-----------+
|
GND

Depending on the relays, you may want to get comparators capable of
driving the relay drivers a little harder than LM393's, which are
spec'ed to have a worst case drop of 0.7V with 4mA of collector
current in their output transistors.

13. ### Guest

Looks great John. It is sort of what I was thinking, that a comparator
would be needed, and started looking them up last night. But alas, I
am worse than a novice at this stuff.

The relay driver from the first circuit, which worked for one relay,
required a rather large (80k) resistor on Vin. The smaller resistors
were apparently dragging the power supply down too much, even with the
relatively small base current.

With the 80k it worked like a charm. Are there changes to your circuit
that I could implement that would reflect the smaller base driving
ability of Vin?

Thanks yet again,

Chris

14. ### John FieldsGuest

---
Because of the 100k ohm resistors, all your Vin has to do is be able
to sink or source 24µA while staying away from the rails by less
than 6V.

If you have problems you could raise the value of the resistances to
1 megohm, then you'd only have to source or sink 2.4µA.

Where's your Vin coming from, anyway?

15. ### Guest

Because of the 100k ohm resistors, all your Vin has to do is be able
Ok. I can calculate the resistance required I think. 24uA seems quite
small, it should be ok.
Ahhh....the curiosity is killing you, eh? ;-)

It's the solo/PFL logic in a large scale recording console. I am
currently in the midst of removing and replacing the master section
with an aftermarket discrete summing amp and passive switching for
source/montor selection.

When an individual channel is SOLOed, the PFL ENABLE (Vin) line goes to
+24VDC. This is via a transistor pull circuit in each channel.
Anyways, if a channel is PFL'd (pre-fader listen, like solo with no
fader or panning), the Vin becomes 500 Ohms impedance to ground.
Otherwise, it is essentially floating.

The SOLO function requires that all other channels be muted. This is
achieved by sending -7.5V down another line in the master ribbon. For
PFL, there is a separate PFL mix buss that is swapped in place of the
L/R buss.

The reason for relays is because in the PFL situation, the audio itself
needs to be switched, and sealed relays are pretty much the cleanest
way to do this.

I suspect the reason the SOLO/PFL logic is so wacky is because there is
a SOLO/PFL switch on the stock master section that forces all SOLO's
into PFL functionality. PFL is pretty lame, really, so I am removing
that switch altogether. However, PFL ability is still needed in the
master section because grouping channels and effects-returns do not
have SOLO buttons, only PFL.

Some day I intend to change the PFL's in the effects returns and groups
to SOLO as well. For this, I'd need to bypass the PFL buss send in
each return and make a MUTE enable that engages the mute when it sees
-7.5V on the SOLO mute line, unless the current channel is the one
being SOLOed.

That should be enough detail. It's a Soundcraft TS12, in case anyone
else out there is searching for details about messing with a TS12's
SOLO/MUTE/PFL logic.

Thanks again for all the help,

Chris

16. ### John FieldsGuest

---
Actually, the deal is if your Vin is capable of supplying 24µA into
a 100k ohm resistor, and not dropping much below 24V, then you
shouldn't need a resistor at all. It's easy enough to try if you
want to, just ground one end of a 100k resistor, connect Vin to it
when Vin is at 24V, measure Vin, and see what you get. I'd be
willing to bet (almost that it won't move much...

17. ### Guest

Ok. I can calculate the resistance required I think. 24uA seems quite
Now I'm confused. With a low resistor value the other circuit we
discussed earlier (forgetting about the fact that the neg relay
wouldn't work), 4.7k, 10k, and 22k wouldn't click over the relay. When
I went up to 100k, the circuit required less current and was able to
prevent the power supply from collapsing. that is my understanding.

Are you talking about using no resistor somewhere in your new circuit,
or in the original one...?

Chris

18. ### Guest

The other problem is getting the LM393. I don't think I can get that
locally. The only place I could think of has 0 on hand. Does radio
shack stock something that will work? Mouser and Digikey both take
about a week to get to me, and paying for overnight shipping seems
ridiculous for an IC chip...

Thanks again,

Chris

19. ### John FieldsGuest

Radio Shack carries an LM339 which is a 4 comparators just like the
two in an LM393, so you could spare out the two you don't need
(short the inputs and the outputs to ground) and use the other two.
They also have a TL082 which you could use by getting rid of the
pullups and connecting the otputs to the bases of the relay drivers
through 3900 ohm resistors.