neg voltage relay driver (for an idiot)

Ok.

I decided to try a TL072 for the comparator, since they are in abundant supply at my studio, as the console is full of them.

First of all, let me thank you again, John, for all the help and direction in this endeavor.

It does sort of work. I had to use 15k resistors instead of 10's, maybe that is why. I tried 33k in place of 20k also. I did not use the pull up resistors on the opamp output.

It works in that output 1 goes high when +24VDC is present, and output

2 goes high when 0VDC == GND is present, but output 1 goes high in both cases.

This could be ok, since I could put relay 1's power leads through an extra set of contacts on relay 2, thus disabling relay 1 when relay 2 activates.

Maybe this is what was supposed to happen? If so, I can totally work with it, but I wanted to see if I did something wrong.

Reading with a meter, pin 1 on the opamp goes to ~24V on both states, pin 5 goes to 24V on 0VDC==GND.

Any more suggestions?

Chris

Thanks John.

Did you see my other post with my results using a TL072? There was a little strangeness I though you might have some insight into...

Chris

--
Dunno.

All I meant to show by my meter readings was that both go high for

0V==GND.

The relays were both turning on all the time at first. The idle output voltage was about 1.3V. I just used a much bigger resistor before the relay driver and that apparently dropped in current enough to not turn on. They are very low current relays.

It really is quite predictable, it's just that the side triggered by

+24V is goes high in both states, at the opamp outputs.

I can get the radio shack comparator tomorrow and ground the extra pins as you suggested.

Thanks for the offer to send the chip. I wouldn't want to trouble you more than I already have been with all my questions!

Chris

--
That\'s because the TL072 (as well as the TL082 I suggested in error)
aren\'t rail-to-rail output opamps, so when their outputs are
supposed to go to 0V, they can actually be a couple of volts above
ground.  If they can source enough current under those conditions,
they can keep the relay drivers turned on.  It\'s iffy, so I wouldn\'t
try the relay inhibit scheme you mention next, because you can\'t
really tell which relay is going to stay on and which one isn\'t.

I had originally calculated the resistor value as 40k or so, using a similar formula that derives from the same thing. It is a pretty sensitive relay, maybe 6mA required I think.

Unfortunately, with the original driver circuit, this was still enough base current to collapse the power supply. It seemed weak to me, but raising the value to 80k+ got it working predictably. I just raised the resistors a few times until the voltage would hold up.

Shouldn't I be able to put a meter on the comparator outputs and read the outputs going high when the corresponding voltage appears at Vin? The relay driver can be left out until later, correct?

Chris

--
But it\'s not _supposed_ to, and therein lies the problem!

Ok. So the TL071 could work with a voltage divider calculated to go near to zero from the TL071's offset voltage.

However, that doesn't explain why both cmprtr outputs are going high when Vin is 0V. Could it be because I am using 15k's instead of 10k? I've checked the wiring about 10 times, but I will check it again. Now I'm at home and I think I have some 10k's here, or I could pick some up tomorrow.

Could the 0V input actually have a small voltage against ground, resulting in a small positive potential to ground? would that even turn on the top relay, were it the case?

Chris

--- Yes. With 24V on Vin, the output of the top comparator should be high and the output of the bottom one should be low.

With 0V on Vin, the output of the top one should be low and the output of the bottom one high.

With nothing on Vin the comparator outputs should both be low.

Notice, though, that with a TL071 in there, the outputs won't go to zero volts, and that's where the problem lies. That is, the outputs go high enough above zero volts to turn on the drivers unless you "starve" their inputs by increasing the value of the base current limiting resistor to the point where the transistors aren't passing enough current through the relay to turn it on. The problem with that method is that you're running the transistor wide open and subjecting its output to the vagaries of temperature and just about everything else you can imagine which can change its gain. For instance, the first time you power up the circuit both relays might be off, but after a while you might find that you can't turn off one of the relays. Possibly becuse it warmed up a little and its gain increased, allowing its quiescent "OFF" base current keep it turned on. The right way to do it is to make sure you use a device with an output voltage less than the "ON" threshold of the driver transistor. You could also use a voltage divider on the output of the TL071, and make sure that with the worst case voltage out, when it's low, the voltage into the driver won't be high enough to turn it on.

-- John Fields Professional Circuit Designer

--- After looking at the data sheet for the TL071, I found that they don't have rail-to-rail inputs (mea culpa for making that asumption) but, rather, have an input common mode range of +/- 11V with a +/-

15V supply, which means that you have to stay 4V away from the rails in order to have the thing work right. With a single 24V supply, that means (I assume) that both sets of inputs should be between 4V and 20V. The reference divider satisfies that requirement since its high output is at +18V and its low output at +6V, but if Vin is being fed into the working divider from a low impedance source, then its output will go very close to the rails when Vin is either at 24V or 0V. The easy way to fix that is to connect a series resistor between Vin and the input to the working divider, but we need to know what the source impedance of Vin looks like in order to calculate the value of the resistance.

If you could set Vin to +24V and then then connect it to the ungrounded ends of, first, a grounded 100k ohm resistor, second, a grounded 10k ohm resistor, and third, a grounded 1k ohm resistor, then measure and post the voltages read across the resistors I can show you how to calculate the value of that series resistor if you don't already know how.

-- John Fields Professional Circuit Designer

John, I rebuilt th circuit with the spec'd comparator and it works perfectly. A Professional Circuit, indeed!

Thanks a lot for putting in all the time to help me with this. Especially all the time fooling around with alternative designs when the original one was perfect.

Best,

Chris

--
My pleasure! :-)

--
That was presented as an option several posts back.

what about an LM339 instead, it's nearly twice the size. but should be dirt cheap and fairly easy to find,,,

Bye. Jasen