Reliable Embedded CPU

[Alti]

One of the things I'd like to do down the road is to add a relay to control a 240V device such as a water heater or an EV.

This is all fine as long as you do not trade safety for convenience. Just be aware that this thing might come on or off and on again at any time. If this leads to inconvenience and the EV is damaged by thousands of on/off commands during one night then that is ok. But if someone unaware approaches an appliance that can unexpectedly turn on by itself, make sure this is not endangering anyone first.

[gnuarm]

Interesting, but that device is only rated for 10 amps and my water heater circuit breaker is 30 amps and the present heater (30 years old) has been measured at 18 amps.  Ideally it would be a 50 amp device so that it could be connected to control any device in the panel.  That's why I'm thinking a home brew device at one level or another, would be good.  50 amp relays are a lot more money than 30 amp units for some reason.  That's probably why it's hard to find higher current switches.  But I can always find a relay. 

Did your Arduino keep good time?  How did you deal with power fluctuations?  Did you need to reset the clock on power glitches and outages?

Use a Sonoff Basic to control a contactor, then you can control as big a load as you want, limited only by the rating of the contactor you select. Furnace contactors are relatively cheap and readily available, they're used for controlling the heating elements in electric furnaces and can be had with 240VAC coils.

Yes, that's why I was talking about a relay.  Contactor is another name for a relay. 

My concern is the utility of the software.  So I guess I'll have to play with the Tasmota software. 

[james_s]

A contactor is a specific type of relay, with large heavy duty contacts designed for repeated switching of large loads. I know from experience that the small cube relays used in the Sonoff devices do not last long if you try to use them to control higher powered loads. Especially reactive loads, they may be rated for 10A but if you look up the datasheet for a 10A relay, it's probably less than 1A switching an inductive load.

[gnuarm]

A contactor is a specific type of relay, with large heavy duty contacts designed for repeated switching of large loads. I know from experience that the small cube relays used in the Sonoff devices do not last long if you try to use them to control higher powered loads. Especially reactive loads, they may be rated for 10A but if you look up the datasheet for a 10A relay, it's probably less than 1A switching an inductive load.

"Probably"  So you have not looked?  The only issue with switching an inductive load is the need for snubbing.  Water heaters are not very inductive and EVs are also not very inductive.  A relay is a relay.  They are called "contactors" out of habit much like the old timers call capacitors "condensers" and use μμF.  Condenser is still the term used for the capacitor across points in auto ignitions. 

The term contactor is used in the power industry for high power handling relays.  What is important is the spec on the relay ADDED to the Sonoff, not the relay inside the Sonoff.  BTW, the relay inside the S31 is 15A, same as the rating on the electrical circuit as a whole.

[james_s]

"Probably"  So you have not looked?  The only issue with switching an inductive load is the need for snubbing.  Water heaters are not very inductive and EVs are also not very inductive. 

I did look, I investigated quite thoroughly after trying to use a Sonoff Pow to control power to my air compressor, it worked great for about a week and then burned up the relay (and melted the terminals where the wires connect) despite being rated for 16A and the compressor draws only 10A. I said "probably" because I don't remember the precise value which varies a bit from one relay to another and can't be bothered to look it up again. Snubbing does not solve the problem of inrush current, LRA on an electric motor can be 5x-10x the rated current and transformers and other inductive loads draw a significant surge as well. I bought a 30A furnace contactor on ebay for around 10 bucks and that does the job.

Water heaters, yes they're as close to an ideal resistive load as you're going to find. EVs I assume are power factor corrected. Still it's worth being aware that inexpensive devices like the Sonoff switches are probably rated in Chinese amps, ie take the printed specs with a grain of salt, personally I would not expect them to hold up to more than about half the rated load long term, but used with this in mind they work really well.

[gnuarm]

"Probably"  So you have not looked?  The only issue with switching an inductive load is the need for snubbing.  Water heaters are not very inductive and EVs are also not very inductive. 

I did look, I investigated quite thoroughly after trying to use a Sonoff Pow to control power to my air compressor, it worked great for about a week and then burned up the relay (and melted the terminals where the wires connect) despite being rated for 16A and the compressor draws only 10A. I said "probably" because I don't remember the precise value which varies a bit from one relay to another and can't be bothered to look it up again. Snubbing does not solve the problem of inrush current, LRA on an electric motor can be 5x-10x the rated current and transformers and other inductive loads draw a significant surge as well. I bought a 30A furnace contactor on ebay for around 10 bucks and that does the job.

Water heaters, yes they're as close to an ideal resistive load as you're going to find. EVs I assume are power factor corrected. Still it's worth being aware that inexpensive devices like the Sonoff switches are probably rated in Chinese amps, ie take the printed specs with a grain of salt, personally I would not expect them to hold up to more than about half the rated load long term, but used with this in mind they work really well.

There's not much in low price devices that isn't designed and made in China.  I've been using an Ankuoo switch for some years now to control the times my EV charges with no signs of wear with 12 amp loads.  There aren't many devices that are permitted to draw more than 12 amps from a 15 amp circuit.   Most devices have to be derated to 80% of the circuit capacity.

Part of the reason I want a timer to control a relay driving the load is to use it with 240V higher current loads like the water heater and a 240V connection to the EV.  Charging a car on a US 120V outlet suffices for many needs, but only 95% of the time.  Other times you want a faster charge so it is fully up in the morning after coming home with a low state of charge the night before.  Even a 40 amp circuit at 240V will bring an EV up to full charge overnight.

As to your relay problem, the start up current is not a big deal as it is over quickly.  Current is a thermal issue that has to be present for some time to damage the contacts.  Opening the contacts interrupts the current flow resulting in a very high kickback voltage from the motor inductance with a substantial arc across the contacts.  This is very damaging and will burn out the relay contacts regardless of the rating, it just takes a bit longer.  Electric motors are typically not snubbed and so the relay contacts suffer the full brunt.  Add a snubber and your relays will hold up much better. 

[PlainName]

Opening the contacts interrupts the current flow resulting in a very high kickback voltage from the motor inductance with a substantial arc across the contacts.  This is very damaging and will burn out the relay contacts regardless of the rating, it just takes a bit longer.

One trick I have heard of is to put a MOSFET in parallel with the relay contacts. If the MOSFET is turned on first and switched off last, it deals with the contacts arcing whilst the relay deals with the current.

[james_s]

Part of the reason I want a timer to control a relay driving the load is to use it with 240V higher current loads like the water heater and a 240V connection to the EV.  Charging a car on a US 120V outlet suffices for many needs, but only 95% of the time.  Other times you want a faster charge so it is fully up in the morning after coming home with a low state of charge the night before.  Even a 40 amp circuit at 240V will bring an EV up to full charge overnight.

As to your relay problem, the start up current is not a big deal as it is over quickly.  Current is a thermal issue that has to be present for some time to damage the contacts.  Opening the contacts interrupts the current flow resulting in a very high kickback voltage from the motor inductance with a substantial arc across the contacts.  This is very damaging and will burn out the relay contacts regardless of the rating, it just takes a bit longer.  Electric motors are typically not snubbed and so the relay contacts suffer the full brunt.  Add a snubber and your relays will hold up much better.

The only EV I've ever charged myself was a Tesla and it had the ability to have the charging scheduled and the current could be set to anything up to the maximum 32A supported by the included home charging cable depending on the plug adapter that was connected. Do others not have the ability to schedule or manually control charging?

In my case the relay was never opened while the motor was running, the existing pressure switch on the compressor did that. What I have is a Sonoff controlling the power so I can turn it on from in the house and let the compressor charge up before I go out to the garage to use it, it's normally empty when I switch it on due to slow leaks at some connectors so the motor starts up as soon as the relay contacts close, then when I shut it off the tank is normally full and the motor has already stopped running. Like I said, these Sonoff things are fine, I have loads of them controlling various things but don't expect them to survive being used at the full rated current printed on them.

[gnuarm]

Part of the reason I want a timer to control a relay driving the load is to use it with 240V higher current loads like the water heater and a 240V connection to the EV.  Charging a car on a US 120V outlet suffices for many needs, but only 95% of the time.  Other times you want a faster charge so it is fully up in the morning after coming home with a low state of charge the night before.  Even a 40 amp circuit at 240V will bring an EV up to full charge overnight.

As to your relay problem, the start up current is not a big deal as it is over quickly.  Current is a thermal issue that has to be present for some time to damage the contacts.  Opening the contacts interrupts the current flow resulting in a very high kickback voltage from the motor inductance with a substantial arc across the contacts.  This is very damaging and will burn out the relay contacts regardless of the rating, it just takes a bit longer.  Electric motors are typically not snubbed and so the relay contacts suffer the full brunt.  Add a snubber and your relays will hold up much better.

The only EV I've ever charged myself was a Tesla and it had the ability to have the charging scheduled and the current could be set to anything up to the maximum 32A supported by the included home charging cable depending on the plug adapter that was connected. Do others not have the ability to schedule or manually control charging?

The Tesla control over charging times is extremely crude and not suitable for my situation.  I'm not sure why you mention the current level.  My car has the ability to charge at up to 72 amps from an appropriate connector.  What it can not do is charge only at the times of my off peak electric rates. 

In my case the relay was never opened while the motor was running, the existing pressure switch on the compressor did that. What I have is a Sonoff controlling the power so I can turn it on from in the house and let the compressor charge up before I go out to the garage to use it, it's normally empty when I switch it on due to slow leaks at some connectors so the motor starts up as soon as the relay contacts close, then when I shut it off the tank is normally full and the motor has already stopped running. Like I said, these Sonoff things are fine, I have loads of them controlling various things but don't expect them to survive being used at the full rated current printed on them.

I suppose they are not really capable of their rating then.  The Ankuoo I use has worked fine for a couple of years now, other than it glitching the program settings eventually.  Which model Sonoff did you have?  They have different ratings, so they don't all use the same relay. 

[gnuarm]

Opening the contacts interrupts the current flow resulting in a very high kickback voltage from the motor inductance with a substantial arc across the contacts.  This is very damaging and will burn out the relay contacts regardless of the rating, it just takes a bit longer.

One trick I have heard of is to put a MOSFET in parallel with the relay contacts. If the MOSFET is turned on first and switched off last, it deals with the contacts arcing whilst the relay deals with the current.

One of the issues of charging an EV at high rates is to not overload the main feed to the house if you use truly high currents and other appliances are on at the same time.  In the US 200 amp feeds to the house are common, but that can disappear when the hot water heater comes on, the stove is on, the clothes dryer is on, the furnace is on and potentially others along with a 40 or 50 amp feed to the EV.  So I have thought of wiring the EV outlet with an A/B switch to allow a single circuit to be shared with another load such as a hot tub that is not essential (not that I have a hot tub, lol).  I'd like this A/B switch to be controlled like one of the Sonoff switches. 

My original thinking in starting this thread was to use an embedded CPU with wifi to roll my own.  Maybe I should check out the Tasmota software to see what it is designed to work on.  I looked into some open source software once that was intended to simply control various existing devices like the Ankuoo, but trying to learn how it worked was such a chore I gave up.  Very convoluted. 

[alepape]

At first I hacked the supplied remote to fake pressing the buttons, but after a couple of months I reversed engineered the protocol and switched to using a $3 433 MHz transmitter board from dealextreme

would you happen to have the protocol documented somewhere? trying to use my raspberry pi to do the same...

[brucehoult]

At first I hacked the supplied remote to fake pressing the buttons, but after a couple of months I reversed engineered the protocol and switched to using a $3 433 MHz transmitter board from dealextreme

would you happen to have the protocol documented somewhere? trying to use my raspberry pi to do the same...

That's gnuarm quoting *me*.

I might have, but in the eight years since I last used that I've lived in eight houses or apartments in four countries.

[brucehoult]

At first I hacked the supplied remote to fake pressing the buttons, but after a couple of months I reversed engineered the protocol and switched to using a $3 433 MHz transmitter board from dealextreme

would you happen to have the protocol documented somewhere? trying to use my raspberry pi to do the same...

That's gnuarm quoting *me*.

I might have, but in the eight years since I last used that I've lived in eight houses or apartments in four countries.

I can't immediately find my updated code to do the 433 MHz control directly, only the remote control hacking version, but I found some stuff I got ideas from:

https://hoult.org/2262_datasheet.pdf

https://hoult.org/SC2262_emulator.cpp

One thing I had to do, in addition to that documentation, was trace how the remote's buttons were connected to the chip, and also I think some pins tied hi or lo to make the codes transmitted different from other similar remotes.