Device to Provide Dry Contact on Power-Down (Normally Closed)

[EvilGenius]

Hello Friends
Its that time of the year when I will have a bit of time to work on a new project.
I am designing a PCB that will switch HVAC Thermostat wires on and off.
Board is powered by 120VAC, reduced to 12VDC or 7.5VDC for electronics (using VREG).
When no power is applied to the board (default), I need the two contacts to be closed. (normally closed, open on power-up).
I did a design based on relays but trying to avoid mechanical switches if possible.
The switch will be driven with a PIC (5V) and the output is a simple two pin dry contact.
Any suggestions...

Regards,
Rom

 

[MaxHeadRoom78]

You can get very small relays, some in DIP package such as Clare PRMA1B05 etc.
Difficult to do in S.S. with no power applied.
Max.

 

[EvilGenius]

You can get very small relays, some in DIP package such as Clare PRMA1B05 etc.
Difficult to do in S.S. with no power applied.
Max.

Max
I am looking into DIP packages for optocouplers and/or small solid state relays if that is where you are heading with your thoughts. I have to have the contacts closed with no power to the PCB.
Cheers,
Rom

Edit: Added a Product PDF

 

[MaxHeadRoom78]

The Clare relay quoted is a N.C. conventional relay in a DIP package.
Can be driven direct from Pic etc.
There are other makes also.
Max.

 

[EvilGenius]

The Clare relay quoted is a N.C. conventional relay in a DIP package.
Can be driven direct from Pic etc.
There are other makes also.
Max.

See edit (attachment on my previous post). Its a Vishay product based on CMOS (logic level).

 

[MaxHeadRoom78]

So is there voltage available on the N.C. side when board is powered off, if so what is the nature of the O.C. voltage?
Max.

 

[ronsimpson]

What is the voltage/current on the N.C. "relay" contacts? Is it just the 24VAC for the HVAC?

 

[dr pepper]

If you use a big power supply cap you can detect loss of power with a pic and have time to do something, just detect the ac going into the rectifier.
Then you could trigger a latching relay to a known state on power down.
Depending on the application all fairly pointless as you could just use a simple relay to do this too.
Dry contacts to me are 'volt free contacts', I spose we all have our own terms.

 

[MaxHeadRoom78]

Volt free or contacts that do not directly switch a primary current, IOW they enable any main switching device or circuit.
The Vishay S.S. device has quite a high on resistance, 20 ohms!
Max..

 

[dr pepper]

What do you mean by primary current, did I miss something.

 

[MaxHeadRoom78]

I mean what is the nature of the voltage that the dry contract is to switch?
Max.

 

[EvilGenius]

What is the voltage/current on the N.C. "relay" contacts? Is it just the 24VAC for the HVAC?

Volt free or contacts that do not directly switch a primary current, IOW they enable any main switching device or circuit.
The Vishay S.S. device has quite a high on resistance, 20 ohms!
Max..

This is answer to Max and Ron...
The contacts for HVAC Thermostat are dry contacts (no volts, or very low), used primarily for enabling the thermostat of the HVAC.
I was under the impression that datasheet was telling me my contacts are normally closed (which is perfect). Not sure about On-Resistor reference!? Could it be for the LED side of it perhaps?
During ON condition the resistance is in Giga (again perfect).

 

[EvilGenius]

If you use a big power supply cap you can detect loss of power with a pic and have time to do something, just detect the ac going into the rectifier.
Then you could trigger a latching relay to a known state on power down.
Depending on the application all fairly pointless as you could just use a simple relay to do this too.
Dry contacts to me are 'volt free contacts', I spose we all have our own terms.

DrP
I think you are overthinking it a bit.
It is a device to help the standby generator select one of multiple HVAC units by use of their enable wiring of the thermostat (not high power, not high volts, we are talking telephone wires 22AWG).
The sense or 120VAC supply voltage comes from the generator (line to neutral). Inside my board, it will have a 120VAC/12VAC transformer, voltage conditioner, and 5VDC voltage regulator to supply the PIC and opto.
The PIC sets time delays between these multiple sets of contacts to turn each A-C on for a defined duration and switch next and so on.
When Utility power is on, the generator is turned off, and my board goes powerless by design which creates my default condition for all A-C units to be enabled and function normally.
When Utility goes out, the generator runs and powers my board, then PIC starts the timing sequence for the individual A-C's.
Hope this makes sense....?
Regards,
Rom

 

[EvilGenius]

The more I think about this (thanks for making me think by the way), the more I think an SSR will not work.
They are probably using CMOS technology and a latching mechanism to create the normally closed contacts (output). What is powering their electronics inside the SSR? I think they are using the device connected to the contacts as the voltage source. So if the contacts are truly dry (no volts), I don't think the SSR will operate as expected (normally closed).
I might have to go back to old school "ice-cubes". Any thoughts or past experience is welcomed...

 

[MaxHeadRoom78]

Why go to ice cube relay when the Dip version should work? Also you can switch it direct with the Pic.
From reading the app sheet for the S.S. device, there needs to be a voltage present for it to work, just as the two wire hall sensors work in the N.C. mode. 20 ohms when conducting for the device quoted!
Max.

 

[tomizett]

I think we really need to be sure of the voltage and current that need to be switched before any better recomendations can be made. Maybe you could measure these?

I share Max's opinion that a small relay in a DIP package would be ideal - the term "reed relay" is what you need to look for. They're very relaible, fast, low-power etc, and do provide good isolation.

On the other hand, if you really do need a solid-state solution, the LH1501BAB looks about as good as you can get.
Regarding these matters -

They are probably using CMOS technology and a latching mechanism to create the normally closed contacts (output). What is powering their electronics inside the SSR?

From reading the app sheet for the S.S. device, there needs to be a voltage present for it to work

I think what's going on here is that they are using a pair of back-to-back depletion mode MOSFETs (which conduct with zero gate voltage applied), and biasing the gates off using a tiny photovoltaic battery (note it says "photodiode array"). This would account for the slow (by optocoupler standards) turn-on time of 2ms. If you're just switching a 5V TTL signal or something, the 20mOhm on-resistance probably won't be any problem. The datasheet says 20mOhm with 50mA of current through the "contacts", which is only 1mV of dropout, so it doesn't appear any voltage is needed to keep the ouput "closed".

For a discreet solution, depletion-mode FETs are the only semiconductor I know of which a "normally closed" action. You might be able to design your own solution using these, but I think in this application isolation between the different aircon units would be very desirable.

 

[EvilGenius]

I think we really need to be sure of the voltage and current that need to be switched before any better recomendations can be made. Maybe you could measure these?......

Reed relays might be what the doctor ordered. I take your word for it that they are fairly fast (ms is fast enough), and reliable (I don't plan to fast switch them or burning the contacts). Isolation is very important both for safety (preventing accidental bleeding or over-voltage surge)and proper functionality since I only want to turn one aircon ON each given time (whole idea of this circuit). Although I will test the thermostat wires for volts, doing it with a reed (a relay in general), I don't have to worry about interfering with function of the aircon!
Thanks for pointing me in the right direction.

Edit: did some search on Reed Relays (TH) normally closed. The cheapest I could find was $4.53 USD. Not Cheap!

 

[KeepItSimpleStupid]

See digikey and search for optomos or photomos. there should be hits for some photo led relays with NC and NO contacts.

Your lack of 120 VAC could likely be done better. I forget the part number.

basically, you want to do load sharing, so only one AC is on at a time. If you really wanted to be smarter, you could use deviation as a priority or some other method. In any event, you could use run time along with call for cool so if call for cool drops out, timing for that AC unit ceases and then on to the next one. So, you have a max run-time, but modify the max time based on call for cool dropping out.

I think it would be cool to base run-time on deviation of a baseline temperature independent of the stat. e.g. Loss of cooling means the setpoint becomes 80 degrees say for all zones. Now use delta T to do your time multiplexing. You also might give priority to certain zones. So if the zones are 1 deg and 2 and 5 deg hotter (deviations from 80 F), the 5 deg hotter zone gets to run longer.

 

[EvilGenius]

See digikey and search for optomos or photomos. there should be hits for some photo led relays with NC and NO contacts.

Your lack of 120 VAC could likely be done better. I forget the part number.

basically, you want to do load sharing, so only one AC is on at a time. If you really wanted to be smarter, you could use deviation as a priority or some other method. In any event, you could use run time along with call for cool so if call for cool drops out, timing for that AC unit ceases and then on to the next one. So, you have a max run-time, but modify the max time based on call for cool dropping out.

I think it would be cool to base run-time on deviation of a baseline temperature independent of the stat. e.g. Loss of cooling means the setpoint becomes 80 degrees say for all zones. Now use delta T to do your time multiplexing. You also might give priority to certain zones. So if the zones are 1 deg and 2 and 5 deg hotter (deviations from 80 F), the 5 deg hotter zone gets to run longer.

Impressive notes.
I am not an aircon guy nor want to play one on TV. Points well taken on your suggestions.
They make load shed modules for prioritizing 4 loads (thermostat dry contacts) but they allow more than one unit to come on based on the watt load on the generator. They also make Power Management module that does similar thing.

My design allows for only one aircon to come on with a manual select switch, or in auto position allow PIC to take over cycling thru each aircon for a specific duration. Each thermostat is already doing the work of sensing the temperature and queuing the individual aircon to come on. So I won't interfere with the function of the aircon voiding its warranty or damaging it.

The concept is to be able to use a lower power standby generator to do the similar work of a much larger genset to power up the entire load (house/business with multiple aircon). The board has to be user friendly, inexpensive and simple so that any technician out in the field can plug and play with minimum thinking and time consuming configuration.

 

[tomizett]

Ah... I'd missed the point that this was a commercial project. Had assumed that it was a one-off for your own home. Isolation would be mandatory, then, as there's no knowing the electrical environment it will end up in.