My furnace has a design flaw that I am trying to work around and in order to do that I need to be able to tap into the thermostat connection to see when the heat is being called for. The thermostat closes a relay in the thermostat that then closes a 24 V AC circuit in the furnace. I'm looking to sense that switch closure by measuring the voltage across the heat wire (white) and the 24 v source (red wire). I do know I should make sure I've got isolation.

Here's a first pass at a schematic. I have a pc817 opto-isolator that I want to put into service and have calcuated that a 1.2 K resistor will probably limit the current to a reasonable limit. I don't need for this to be a pure DC output. I can always loop on reading the pin for a period of time.



Any suggestion on alternatives? Or corrections?

Thanks,
Jim.

 

I'd put the resistor on the LED anode so the voltage on the cap can't damage the LED. Its value should be chosen to limit the LED current to a reasonable value. Values for D1 and C1 aren't critical. You could use a signal diode and some small uF value for the cap.

 

In the old days the thermostat switched the gas valve, so current in the loop was high.
Modern furnace control boards are around 1k ohm input, so about 24mA at 24VAC. Some have a dummy load for higher input current 100-240mA to keep compatibility with mechanical thermostat's heat anticipator resistor.
If your monitoring circuit ends up across the thermostat, some current will always be flowing and can false trigger the control board input. It's important to ID the (typically) 24VAC common "C" white and 24VAC hot "R" red, and "W" is the demand for heat.

The circuit you have will work, it's not far off. I would use less current for the opto LED. 24VAC and 10k for ~2mA, I prefer a small bridge rectifier (so as to not present a 1/2 wave load) and move the filter capacitor because the ripple voltage seen is really low. A capacitor across an 1.2V opto LED has to be quite big to keep it lit, compared to if done ahead of the dropping resistor. Don't be afraid to add an extra series LED for quick visual.

 

Here is your circuit modified with a led for visual purpose also, bridge rectifier can be any 100V 1A rated, the series resistor can be from 4K7 to 10K.

 

Depending on the current level in the thermostat wire when operated, you could use a current transformer to give isolation, and not having to disturbed the existing furnace wiring.
Then, feed the current transformer output into a circuit like above.

 

I did a quick sim and the ripple current is up there with 100uF ~120mApk, 10uF would be about half but more than a 10uF 50V electrolytic is rated for. Another approach is moving the cap to the opto output side.

OP could use a scrap bin GFIC current-transformer, they are 1:1000 and with 30mA in the thermostat loop it's enough to work with, but AC output and I haven't seen any clamp-on probes for that low current.

 

Hi, All
Thanks for all the suggestions. I'll try some things based on this info and let you know what worked. For a test I put a 1.2K resistor across the heat and power connections to see if it would turn the furnace on. It did not, so I'm pretty sure any load I put on the system won't false trigger the control board input. But I didn't rule out the possibility of it holding it.

Thanks,
Jim.

 

Why don't you use an AC input optoisolator? Then put small capacitor across the output transistor, so that it doesn't switch off at every zero-crossing.

 

Hello,

The FOD814 or LDA110 are AC optocouplers as @Ian0 proposed.

Bertus

 

My furnace has a design flaw that I am trying to work around and in order to do that I need to be able to tap into the thermostat connection to see when the heat is being called for. The thermostat closes a relay in the thermostat that then closes a 24 V AC circuit in the furnace. I'm looking to sense that switch closure by measuring the voltage across the heat wire (white) and the 24 v source (red wire). I do know I should make sure I've got isolation.

Here's a first pass at a schematic. I have a pc817 opto-isolator that I want to put into service and have calcuated that a 1.2 K resistor will probably limit the current to a reasonable limit. I don't need for this to be a pure DC output. I can always loop on reading the pin for a period of time.

View attachment 252980

Any suggestion on alternatives? Or corrections?

Thanks,
Jim.

What sort of control arrangement does the furnace have? Is it just a gas valve, or is there a computer board between the thermostat and the rest of the hardware? And what is the purpose of the addition?
AND what is the claimed design flaw of the furnace? I am wondering about that part.
If you measure between the red and the white wire the voltage will be "high"until the thermostat demands heat, at which point the voltage will be close to zero.

 

What sort of control arrangement does the furnace have? Is it just a gas valve, or is there a computer board between the thermostat and the rest of the hardware? And what is the purpose of the addition?
AND what is the claimed design flaw of the furnace? I am wondering about that part.
If you measure between the red and the white wire the voltage will be "high"until the thermostat demands heat, at which point the voltage will be close to zero.

I suspect a computer board since it's a fairly modern 90% furnace. In any case you are correct that the voltage does what you say. I do not know the exact problem. It appears that a sensor in the area where the condensate collects is finding something that it doesn't like. The whole of that area has been replaced but it still faults if it runs for more than 30-40 minutes continuously. The only cure (AFAIK) is to power cycle the furnace. I don't know if it's design flaw or just a broken part that no one seems to be able to find. In any case a 8266, a relay and some sensors can perform the reset task when necessary.

 

Why don't you use an AC input optoisolator? Then put small capacitor across the output transistor, so that it doesn't switch off at every zero-crossing.

I like that. I didn't know they existed. One of the devices that Bertus mentioned has a 74 week lead time, but Digikey has the other one in stock.
Thanks to you both,
Jim.

 

I suspect a computer board since it's a fairly modern 90% furnace. In any case you are correct that the voltage does what you say. I do not know the exact problem. It appears that a sensor in the area where the condensate collects is finding something that it doesn't like. The whole of that area has been replaced but it still faults if it runs for more than 30-40 minutes continuously. The only cure (AFAIK) is to power cycle the furnace. I don't know if it's design flaw or just a broken part that no one seems to be able to find. In any case a 8266, a relay and some sensors can perform the reset task when necessary.

That condensate is rather nasty stuff! My Dad had one of those high efficiency furnaces and after a while the condensate destroyed the rubber? drain tube so that it leaked all over. Since the exact replacement part would also have failed in just 2 years I replaced the drain fitting with one that I bent up out of PE plastic water pipe. That stuff was far more chemical resistant. At least it resisted the condensate very well. Also it was on hand and cost about $25 less than the OEM part.

Perhaps you should investigate the condensate area and see if that sensor needs additional protection, such as a silicone sealer coating. OR it may already have been dissolved.