I need to replace relays on a Thermostat with Mosfet

[igeorge]

I want to replace the latching relays in a thermostat with mosfet
I found on the web a sugestion, but i am not sure if it is OK
At the furnace it is a 24VAC transformer
One wire come to the thermostat and return by closing the contact in one of the 3 relay
I am not sure if it is correct to connect the CPU ground with the wire coming from the thermostat
Please help

 

[NorthGuy]

When COM goes negative, you get it in the following order: Negative(COM) - Diode - GND - Mosfet - Diode - Load - Positive

I guess during this phase they want the capacitor to charge.

When COM goes positive, you get it in the following order: Negative - Load - Diode - GND - Mosfet - Diode - Positive (COM)

I guess at this stage they want the capacitor to stay charged and keep driving the mosfet.

If that's how it's designed, you only need to drive mosfet when COM is negative. At which point GND is a diode drop above COM. So, you can use another diode to connect your controller's ground to COM, and VCC to the other AC line. However, this way you only get power when COM is negative, so you probably will need a big capacitor to keep the power on during the half cycle when COM goes positive. Most likely, instead of connecting controller directly to lines, you'll need a voltage regulator for the controller.

I would use relays instead of mosfets.

 

[igeorge]

When COM goes negative, you get it in the following order: Negative(COM) - Diode - GND - Mosfet - Diode - Load - Positive

I guess during this phase they want the capacitor to charge.

When COM goes positive, you get it in the following order: Negative - Load - Diode - GND - Mosfet - Diode - Positive (COM)

I guess at this stage they want the capacitor to stay charged and keep driving the mosfet.

If that's how it's designed, you only need to drive mosfet when COM is negative. At which point GND is a diode drop above COM. So, you can use another diode to connect your controller's ground to COM, and VCC to the other AC line. However, this way you only get power when COM is negative, so you probably will need a big capacitor to keep the power on during the half cycle when COM goes positive. Most likely, instead of connecting controller directly to lines, you'll need a voltage regulator for the controller.

I would use relays instead of mosfets.

Thank you , but i cannot have a separate supply
The whole system gets power from 3 AAA battery pack
No AC line

 

[NorthGuy]

Thank you , but i cannot have a separate supply
The whole system gets power from 3 AAA battery pack
No AC line

I thought that COM in schematics was a 24VAC line to be used as a power supply.

If you want to use an isolated battery power, you could connect two diodes between COM and other AC line in series (looking in opposite directions of course), and reference your ground to the point between diodes.

When COM is negative, your ground will be the same as mosfets.

When COM is positive, mosfet ground will be at your ground when mosfet is off, but it'll go 33V higher if mosfet is on. To prevent the capacitor from discharging through your controller, you may install a diode on each of the mosfet control lines.

 

[KeepItSimpleStupid]

It's a very strange design, but it should work. Basically there will be a missing 2.4 V piece of the 24 VAC line centered around the zero crossing.

Note that there is a GND, and a COM. The u in uHEAT means the microprocessor side. I would have labeled GND uGND.

I too, think it's unclear. Somehow, I think that R should be connected to GND and you need to ignore COM.

Adding a fuse, would be a good idea.

In reality, you could breadboard one of the circuits with a small fuse (0.5A). like FAN with a battery pack and go. The furnace should be fused at a little over 40 VA/24 or usually 3A or use an energy limiting transformer. It's not guaranteed though. They usually put the fuse in the R lead before it leaves the furnace. An automotive 3A blade fuse is common.

FWIW:
A connection from R to G (Fan) turns on the fan
A connection to R to W turns on the heater. Warning, the furnace controls the fan, so it's not instantaneous on or off.
A connection from R to Y just turns on the outside unit. The thermostat controls the Fan. Turning on and off the outside unit under head pressure may blow the fusebox/breaker box AC fuse. The T-stat usually incorporates a 3-5 minute delay between off and on. So, the thermostat controls the fan. So, for AC, both Y and G are connected to R.

Battery powered stats are usually used because there is not an extra wire for COM to be sent to the thermostat. There are devices that can electrically add a wire, so you can put in a newer stat with a COM terminal. This way your not out of heat/cool when the batteries die.

COM may not be labeled in an old furnace, but it's the side of the transformer NOT connected to R.

The terminal designations come from Red, White Green and Yellow.

In fact, the only part you really need is the bridge and a 0.5A fuse in the R lead.. Just connect the fan circuit and substitute a switch where the MOSFET goes.

So, I really would like you to consider running a new wire to support C with a new stat or one of these add-a-wire gizmos and a new stat that doesn't use primary battieries. They still might use "non-replaceable" batteries for a calender and backup purposes.

 

[igeorge]

Thank you very much both of you
Let me make my self clear.
The principle of one single cell , heat or coll or fan, i found it on the web
For me, analog = Amazon's jungle
I have the basic of analog, but that it.
In exchange, i can write programs, which i enjoy.
In what concern this project, they are some restrictions.
I cannot run an extra wire, which was the best in my case. I need to make it universal so you can replace the old one with this ( in a 4 wires system)
I am limited to live with the battery which i have on the prototype
I cannot put latching relays as they are too big for my PCB ( even the SMD type from Panasonic, and they cost too much , 6 dollars a piece
My budget for the whole project is 20 dollars. I am at 10 now with the CPU, PCB, LCD, and few resistors and capacitors
With one CPU , i do touch buttons, LCD display, real time clock, temperature measurements, program 4 zone a day/ 7 days a week, extended absence, and RF communication with the vents ( they have a motor) on the room, and central control.
Not too much on the board, but plenty of features. I still need to connect to the 4 wires from old one, in case the RF goes down ( i do not want frozen pipes in the house)
Please , if i do not ask too much, can you provide a hand drawing to the proposed changes ?
Thank you

 

[kinarfi]

after a few assumptions of the t stat is battery powered and simply a switching mechanism, like mine, and the furnace has a 24vac transformer, like mine, and it has the standard 5 wires to it like mine, here's what i would try..
Because 24vac is 33vp-p and one leg is grounded, you will get a 33vcd on your cap, too much for a fet gate, so I added zeners to hold it down to 12.
Good luck, please let me know how it works out.
PS according to Spice, The zener will dissipate 120mw, the 2k resistor will dissipate 200 mw, and the 20k will dissipate 50 mw

 

[igeorge]

Thank you Kinarfi for spice and drawing
I cannot run the spice as when i stat it said with " missing symbol spst" the switch
But i am confused here
Which wire is the common from 24vac transformer and which are the return wires ?
I assume that the switch is the like simulatting the µprocessor outputs to command the gates
How i can interface the CPU and the furnace with your schematic ?
Also, do not forget that my system has only 4 wires like in the drawing attached
When i try to run the simulation it complai about the model for 1n4007 and models for fets
Will you be so kind and send to me the models and symbol for the witch please
Thank you

 

[KeepItSimpleStupid]

Your coming from a different angle, HOWEVER, I wanted to make you aware of this product: https://www.smarthome.com/304AAW/Ve...y-for-Programmable-Digital-Thermostats/p.aspx

 

[KeepItSimpleStupid]

In simulation, put an AC source between COM (What I believe to be R) and your furnace relays. Have those relays return to the other side of the AC source.

Your load could just be a resistor, say 24/0.030 ohms for simulation. So, tie one end of the resistors together and to one side of the 24 VAC simulated source.
Put the other side of the 24 VAC simulated source to COM. Put the other end of the resistor to uFAN, uHEAT and uCool.

This should verify operation. As I said earlier, I think the point labeled COM is actually R on the stat. To make matters wierd, R is a "common" terminal to each of the contacts.
R to G
R to Y
R to W

 

[igeorge]

thank you
i will try

 

[kinarfi]

Any diode and any fet that you have in your library will work, I just grabbed the first ones I found for the simulation, the spst switches just represent what ever your u Processor has for a switch. Some t stats have what's called a heat anticipater which is just a series resistor in line with the t stat switch, but it went away with the mercury bulb t stats. Others use power from the furnace transformer to operate. Mine, and I assume yours, run off batteries and there fore are just switching apparatuses that can be programmed to switch when we want them to. I don't think you need the full wave bridges either, a FET will pass current both ways as long as the gate is biased on and your fets are good for 60 v and 3 amps so you should be good.

 

[NorthGuy]

after a few assumptions of the t stat is battery powered and simply a switching mechanism, like mine, and the furnace has a 24vac transformer, like mine, and it has the standard 5 wires to it like mine, here's what i would try

Could you explain your design more? If you put AC on V1, then when voltage becomes negative on the top rail, the mosfets should start conducting (because they act as diodes when subject to reverse voltage). As a result the loads will be subject to half-wave power regardless at what's on mosfet gates. Is that how you want it to be? Or it is not really AC coming from the furnace, but something rectified?

 

[kinarfi]

Still thinking about your project, what type of furnace do you have, most fan switched just turn the relay of for the fan which is also turned on by a temperature sensor in the furnace to start the fan after the temp rises from the fire and turned off after the temp drops, if the fan switch is not activated, it is also turned on by the cooling switch and some times, you have one speed for heat and one speed for cooling. Unless you actually use the fan switch, I would just forget about it and leave it on automatic, which is no operation from the t stat

 

[kinarfi]

you are right, my mistake, use the bridge

 

[igeorge]

you are right, my mistake, use the bridge

my thermostat has 3 relays to close the common and return to furnace
My furnace is a lennox with a Honeywell control board
How and when any of the components are turned on, is not my problem
All i need that the contact is made to a set point and open to other
The fan, if in auto, will close the contact too with heat or cool and stop with them
The furnace will keep the fan longer, but this is not my problem
On manual, the fan run all the times

 

[NorthGuy]

Let me make my self clear.

If it is indeed AC over there, your design criteria are very tough.

Do you know how much load do you get on these lines? I wouldn't think 24VAC is used directly on the motor. It is most likely just a signal to the furnace's electronics. If it's in the range of mA, you can use an opto-isolator. It will take very little space. You can use two in series and they will conduct in opposite directions.

 

[igeorge]

on the installation drawing it shows a 24 vac transformer
One wire go to room and other to board
I know is just control signal because on the board they are some diodes on each return wire and a cap
But this is just mine
I like to make it universal and capable to drive direct the motor fan if necessary
Make it simple, i need the features on mine, without relays : contact 120 vac / 2 amps

 

[kinarfi]

I have a lennox also, here is a photo of my control package on the furnace, is it similar to yours?

 

[KeepItSimpleStupid]

Unless you actually use the fan switch, I would just forget about it and leave it on automatic, which is no operation from the t stat

This is NOT TRUE. In AC mode, the thermostat controls the fan.

In HEAT, the FURNACE controls the fan except for the ability for the thermostat to cause continuous operation.
FAN
ON - Fan is on continuously regardless of heat or cool
OFF - System is OFF. It does NOT mean that the fan will turn off immediately if in HEAT mode.
AUTO - The thermostat in AC mode or the the furnace in heat mode controls the fan,

Since you are designing a thermostat, it is important to add what's called an anti short-cycle timer whose operation is as follows. If the Y terminal turns off, it then can only turn on after it has been off for x minutes. 2 is usually a good number. In a thermostat that draws power from the furnace, that timer starts immediately on power up. It acts as a Y_enable and prevents the AC system from restarting immediately after a power failure. The Fan CAN start right away.

Let's explain the "Thermostat controls the fan a bit better"

Furnace gets a "Call for heat".
furnace says, the blower will blast perceivably cold air if I turn the blower on immediately, so the furnace can do two things.
1) Just delay the fan turn on or b) use plenum temp with a timeout.
2) Use Plenum temp (If Tp > x) or (Fan is in ON mode) then turn on blower.

When the "call for heat" is satisfied, then the flame turns off.
The fan turns off either :
1) after a set delay or
2) after (Tp<x).

The job of the "anticipator" in the old stats does create a sort of hysteresis and it heated the bi-metalic spring when the there was a call for heat. This had the effect of turning off the heat sooner,

My current stat does know the BTU's of the furnace and the cooling system, the outside temp and the time. I'll bet it uses this info to do a better job of not overshooting the setpoint when in setback mode. The stat is also smart and is able to report furnace errors and save history. It can report blower RPM and CFM values.