soft start unit for old fridge

[Jbates58]

hello all

i am having some trouble with my brew pi controller and fridge.

i originally had it all setup with SSR's to switch the fridge compressor on and off, and discovered that it only worked once. then every time after that, as soon as it tries to kick in the fridge, its asthough there is a spike and it resets the arduino.

i have tried to isolate it as far as using a battery to power the arduino so that it has no interferance from the fridge, and i have also tried a few different arduino boards.

after getting sick of this, i decided that perhaps the SSR's couldn't handle the draw of the fridge. they are 240V 3A rated, and the fridge draws about 1.8A. i also thought that perhaps the inrus current was the issue, but the highest i can meter if i bypass the controller is 5A, the SSR's are rated to 80A inrush acording to the data sheet. so i thought that ruled them out.


then i thought, stuff it. i will swap it over to mechanical relays. so i got some and re-jigged everything to get it going, and still nothing. it is exibiting the same symptoms. so to isolate it further i wrote my own code on the duino to turn the fridge on for 1min, and off for 1min. it functioned perfectly for 1 hour, after which time i ruled out hardware been the issue and went on from there.

but i am wondering if i use this http://www.jaycar.com.au/soft-start-kit-for-power-tools/p/KC5511 on the fridge compressor if that will solve the issue? or am i barking up the wrong tree?

thanks

Jason

[Siwastaja]

You have something horribly, horribly wrong. Any "soft start" is not the wayto go, and in worst case, may damage the compressor if the motor stalls from lacking torque. The only way I see is to actually find and fix the problem. It's most probably something very "stupid" once you find it. Please provide a schematic and maybe a few photos of the setup.

[Monkeh]

turn the fridge on for 1min, and off for 1min.

Don't do that. Compressors do not appreciate it.

it functioned perfectly for 1 hour, after which time i ruled out hardware been the issue and went on from there.

but i am wondering if i use this http://www.jaycar.com.au/soft-start-kit-for-power-tools/p/KC5511 on the fridge compressor if that will solve the issue? or am i barking up the wrong tree?

If it worked fine and you ruled out hardware.. why are you wondering if you need a hardware fix?

[Seekonk]

I have a UNO running my fridge at camp.  I use a hot start delay in software of 5 minutes.  Otherwise the compressor will be starting under pressure.  Most fridges now have a thermal resistor to pull in the start winding.  If that doesn't cool down the start winding will not come in.  I had an issue with mine early in the morning when batteries were weak.  It would momentarily start and shut down. due to a reset.  Have to agree.  Something is crap in your software.  Do you have any temperature filtering?  Inrush currents are quite large.  Your SSR is not going to last too long.

[Siwastaja]

turn the fridge on for 1min, and off for 1min.

Don't do that. Compressors do not appreciate it.

True. Which is why they come with some sort of safety delay circuit. Which I assumed is not bypassed, in which case the whole thing is simply not working (it shouldn't turn on after 1 min off time), but shouldn't give EMI problems either.

If the delay switch is bypassed, then this is really a good way to destroy the compressor by burning the windings by forcing the compressor on when it's stalled due to the pressure it can't start against due to the limited starting torque of the small 1-phase motor.

At least 15min on, 15min off would be recommendable to OP, IMO.

[Jbates58]

turn the fridge on for 1min, and off for 1min.

Don't do that. Compressors do not appreciate it.[/b]

i know, but it cost me nothing, and would otherwise be going in the bin.  i do know about it starting under loads and what not, im in A/C as a job, so i understand the loads against the compressor.

it functioned perfectly for 1 hour, after which time i ruled out hardware been the issue and went on from there.

but i am wondering if i use this http://www.jaycar.com.au/soft-start-kit-for-power-tools/p/KC5511 on the fridge compressor if that will solve the issue? or am i barking up the wrong tree?

If it worked fine and you ruled out hardware.. why are you wondering if you need a hardware fix?

just trying to eliminate possible causes.

I have a UNO running my fridge at camp.  I use a hot start delay in software of 5 minutes.  Otherwise the compressor will be starting under pressure.  Most fridges now have a thermal resistor to pull in the start winding.  If that doesn't cool down the start winding will not come in.  I had an issue with mine early in the morning when batteries were weak.  It would momentarily start and shut down. due to a reset.  Have to agree.  Something is crap in your software.  Do you have any temperature filtering?  Inrush currents are quite large.  Your SSR is not going to last too long.

the code is not mine, and there is plenty of people running the code on the same units that i am, as far as i can tell, this is isolated to me.

i am running the BrewPi system. based on an arduino and a raspberrypi, i have substituted the pi with an odroid, but that shouldnt matter.

the fridge was built in about 1985, so i doubt that it has any of that in it.

the ssr is capable of 80A inrush currents.

i have simply bypassed the thermostat unit, and replaced it with a relay.

here is a schematic of the setup. there is a capacitor on the +5v lines as i thought it might help, but made no difference.

Jason

[CJay]

Your relay driver circuit is, umm, odd, are you sure that's how you have it wired?

Are you also sure the fridge is isolated from the Arduino properly, I.E. no shared earth etc.?

Might be worth wiring in a couple of lamps in place of your fridge and seeing if it behaves then.

[Jbates58]

Your relay driver circuit is, umm, odd, are you sure that's how you have it wired?

Are you also sure the fridge is isolated from the Arduino properly, I.E. no shared earth etc.?

Might be worth wiring in a couple of lamps in place of your fridge and seeing if it behaves then.

how is it odd?

i am using digital 5 & 6 to power a transistor that is closing the relay.

its wired the same way that many relay driver boards are that you buy commercially. the power runs through the "active" led, through the current limiting resistor then through the gate of the transistor.

to me that makes sense. can you tell me how it is odd? im curious.

the lamp that is in the fridge is wired to the second relay pictured and is the heater unit to keep the temp stable. it functions perfectly. there is only an issue when the compressor kicks in.

as far as i can tell, it is as isolated as it can get. i have even tried to run the arduino off a battery and i still get the same issue.

Controlling a fridge at high repetition rate is inherently a bad idea. Get a proper VFD for it. The cheapest Chinese VFD I can find is only ~$110 brand new, and it talks with your MCU via isolated serial port, so essentially you only need to write program and wire the VFD up. No need to roll out any custom made circuits, and no exposed mains voltage hazard.

that just seems more complex and unnessceary than needed. i have made another one of these units for a mates fridge and it works perfectly. has no issues at all.

and to my understanding, all i am doing is replacing the mechanical thermostat with a relay that is controlled from an arduino.




Jason

[DenzilPenberthy]

Remember that the LEDs are diodes and will have a forward voltage drop when conducting.  So the 5v signal from D5 will drop 0.6v across the B-E junction of Q2, 3.3v ish across LED2, which only leaves 1.1v across R1. This only gives a base current of around 5mA. Is this what you intended?

 Q1 will get a base current of around 12mA due to the much smaller forward volt drop of the red LED.

Another point to mention is there are no freewheeling diodes on the relay coils. When there is current flowing in the coil and the circuit is interrupted by the transistor, a very large voltage will be developed across the coil. It is an inductor and you are forcing its di/dt to be very large. In effect the current 'wants' to keep flowing. This will damage your transistors and maybe injecting nasty spikes into places you don't want them.

Thirdly, If you are powering the arduino from 5v through its Vin pin, the board will probably not be getting it's full 5v. The arduino has a linear regulator which needs some drop-out voltage to operate correctly. Power the Vin pin from at least 7v.

[Jbates58]

Remember that the LEDs are diodes and will have a forward voltage drop when conducting.  So the 5v signal from D5 will drop 0.6v across the B-E junction of Q2, 3.3v ish across LED2, which only leaves 1.1v across R1. This only gives a base current of around 5mA. Is this what you intended?

 Q1 will get a base current of around 12mA due to the much smaller forward volt drop of the red LED.

Another point to mention is there are no freewheeling diodes on the relay coils. When there is current flowing in the coil and the circuit is interrupted by the transistor, a very large voltage will be developed across the coil. It is an inductor and you are forcing its di/dt to be very large. In effect the current 'wants' to keep flowing. This will damage your transistors and maybe injecting nasty spikes into places you don't want them.

Thirdly, If you are powering the arduino from 5v through its Vin pin, the board will probably not be getting it's full 5v. The arduino has a linear regulator which needs some drop-out voltage to operate correctly. Power the Vin pin from at least 7v.

oops, forgot to add some things to the schematic.

yes, there are flyback/suppression diodes on the relay. to eliminate the inductive spike.

the specs on the led's in the schematic may not be correct, nor the resistor values. i just drew that up in a rush thisafternoon. but i dont think they are an issue, as with the code on the arduino that cycles it on and off every min, it works perfectly.

the arduino 5v is supplied from usb from the odroid. but when driven off battery, it has an onboard 5v reg and that takes care of the voltage conversion to 5v.

i can measure the voltage on the transistor gate pin tomorrow afternoon when i get home. but it should be enough to keep it all going well, (unless there is more science in picking one than i thought).

[DenzilPenberthy]

I don't think anyone here is going to be able to help you with a hardware issue that you probably don't even have (as you say it works fine with one set of code but not another), using a schematic that bears very little relationship to what you have actually built.

i can measure the voltage on the transistor gate pin tomorrow afternoon when i get home. but it should be enough to keep it all going well, (unless there is more science in picking one than i thought).

You've drawn the schematic with a BJT. BJTs don't have gates, they have bases and they are not voltage operated, they pass an amount of current from C-E which is related to the amount of current flowing from B-E. the volt drop from B-E will always be a silicon P-N volt drop (0.6-0.7v ish).

[Jbates58]

I don't think anyone here is going to be able to help you with a hardware issue that you probably don't even have (as you say it works fine with one set of code but not another), using a schematic that bears very little relationship to what you have actually built.

i can measure the voltage on the transistor gate pin tomorrow afternoon when i get home. but it should be enough to keep it all going well, (unless there is more science in picking one than i thought).

You've drawn the schematic with a BJT. BJTs don't have gates, they have bases and they are not voltage operated, they pass an amount of current from C-E which is related to the amount of current flowing from B-E. the volt drop from B-E will always be a silicon P-N volt drop (0.6-0.7v ish).

hmmm ok, well still lots to learn.

i thought that there was perhaps some sort of inductive spike that may have been getting to the arduino through the relay, and that was perhaps causing the issue. and i thought that a soft start unit might solve the issue. i will try my unit on my mates fridge and see what happens from there.

thanks all.

Jason

[CJay]

Your relay driver circuit is, umm, odd, are you sure that's how you have it wired?

Are you also sure the fridge is isolated from the Arduino properly, I.E. no shared earth etc.?

Might be worth wiring in a couple of lamps in place of your fridge and seeing if it behaves then.

how is it odd?

i am using digital 5 & 6 to power a transistor that is closing the relay.

its wired the same way that many relay driver boards are that you buy commercially. the power runs through the "active" led, through the current limiting resistor then through the gate of the transistor.

to me that makes sense. can you tell me how it is odd? im curious.

OK, the base current running through the LED is odd, pretty much every schematic I've ever seen for controlling a relay ith a digital pin drives the base through a resistor only, any indicator LEDs are driven in other ways,

https://www.google.co.uk/search?q=drive+relay+arduino&safe=off&client=firefox-b&source=lnms&tbm=isch&sa=X&ved=0ahUKEwjSyqr4vJPOAhXCNJoKHc1PC_wQ_AUICCgB&biw=1366&bih=633

The Sainsmart relay board the BrewPi project Wikia shows is very different from your circuit as it uses an opto isolator but still doesn't put the LED in series with the base of the transistor.

http://www.sainsmart.com/arduino-pro-mini.html

Other than the connections from the isolated side of the relay are the *any* connections between the fridge and the Arduino? Do they share an earth perhaps?

[Seekonk]

"the code is not mine, and there is plenty of people running the code on the same units that i am, as far as i can tell, this is isolated to me."

There are a lot of schematics and code out there that works for most people that have errors in thought.  I've spent 3 years tweaking my system to account for real life situations that could never have been thought of.  Can you point us to this published article so we can see what variances you have made.

Running relays off the micros 5V is always a recipe for disaster.

[Jbates58]

its just the BrewPi software, its freely available.

i havent made any variences that i am aware of. when it first came out, it used the mechanical relays. now that it has evolved a bit more and become more mature; they have moved it to SSR's.

Jason

[Seekonk]

its just the BrewPi software, its freely available.

Jason

I make a simple request.  I google BrewPi refrigerator and there are a dozen.  You aren't worth my time.

[Jbates58]

its just the BrewPi software, its freely available.

Jason

I make a simple request.  I google BrewPi refrigerator and there are a dozen.  You aren't worth my time.

sorry mate, im not sure exactly what you are after? would you like to look at the hex file that has the code?

i have made no changes to the software, it auto uploads to the arduino from the brewpi github. i install the brewpi software on the odroid, and from there it does the rest. i tried both hardware approaches and am having the same issue on both.

[Siwastaja]

Sorry, but if you use random things made by random people randomly, without understanding any part of it, and add your own random things, the chances it would work are slim. But even worse, there is no way anyone could help you, especially on the forums. Someone might be able to help in real life. So, my suggestions are:

(1) continue doing random things until it works by luck. It may take a looooong time.
(2) start from the beginning, this time by learning and understanding. Do every step yourself. It may take as long as the "keep it random" way, or even longer, but the result is guaranteed to be a great success, and learning things helps you in the future projects.

[Jbates58]

Thanks. I do understand that.

But iam currently using the same components as my friends fridge, and his has no issues. Thebonly difference between our setups currently is the fridge itself. All the hardware is the same. I am yet to try swapping the compressor to the other relay to eliminate that as the issue. And on the weekend I will be taking my unit to his fridge, and his unit to my fridge to rule out something. Then I might know where the issue lies. I will report back with my findings.

Jason

[Jbates58]

ok, i have taken my unit over to my friends house and tried it on his fridge. works perfectly.

so that leads me to something in the fridge causing the issue. any ideas as to what? i am thinking some sort of inductive spike or something, but am unsure.

Jason

[Seekonk]

I don't understand what the two relays control.  They had a page indicating what type refrigerator configurations are out there.  You want us to give you answers like magic.  You so don't have a clue how to get help.

[Jbates58]

I don't understand what the two relays control.  They had a page indicating what type refrigerator configurations are out there.  You want us to give you answers like magic.  You so don't have a clue how to get help.

one relay controls the compressor, and the other the heater. i have the heater working perfectly, but when the compressor kicks in, it restarts the arduino.

yes, i'll admit that this sort of stuff is a little beyond my knowledge, but i have given it a go with what i know, and tried to peice it together. sure i might not use all the right parts, or quite the correct parts, and am happy to change them if need be. but the relay controls are the exact same as what came on the relay driver board, i un-soldered them and moved them to a bigger proto board to fit them both.

i have taken the unit to a friends fridge and it works, so that means that my fridge is the issue. i would like to find out why, as all i have done is removed the mechanical thermostat and replaced it with a relay. which is what you are meant to do.

i have tried to swap the relays over, and get the same results on both. so that eliminated them as the issue aswell.

and you are right, i dont know alot about this stuff. so my ability to ask for help isnt the greatest as i dont know exactly what im dealing with.

Sorry, but if you use random things made by random people randomly, without understanding any part of it, and add your own random things, the chances it would work are slim. But even worse, there is no way anyone could help you, especially on the forums. Someone might be able to help in real life. So, my suggestions are:

(1) continue doing random things until it works by luck. It may take a looooong time.
(2) start from the beginning, this time by learning and understanding. Do every step yourself. It may take as long as the "keep it random" way, or even longer, but the result is guaranteed to be a great success, and learning things helps you in the future projects.

i have not inserted my own random things. i used an arduino board, 2 relay driver modules and 3 ds18b20 probes. which is what it says to use.

i have removed all the componants off the relay boards and moved them to a bigger proto board where i can fit them both on the sheild, and wired them back up the same way.

[CJay]

Have you earthed the Arduino to the fridge?

Do you have a multimeter?

[Jbates58]

yes and yes.

it wasnt earthed when running it off battery to eliminate a power drop issue. but other than that, it has been.

Jason

[Jeroen3]

Switching off relays is fun  ! Especially with inductive loads. All sorts of funny business can occur.
Including but not limited to:
1. Dips when turning on the coil.
2. Spikes when turning off the coil.
3. Noise when closing high current path.
4. Spikes when opening inductive current path (like a motor or compressor "hint hint")
5. Spikes from 2 or 4 coupled onto nearby wires or boards.
6. Physical shock due to relay triggering a faulty solder joint.
7. Physical shock of compressor rattling the relay contacts.
Think of anything, relays are evil.

For 1. you must spice up you supply.
For 2. you can add snubbers on the coil. DC relays have diodes for this. If you have proper diodes (not from eBay) you'll probably be fine.
For 3, yeah... inrush current limiters or shielding perhaps. I'm not sure.
For 4, fit surge supressors on the load.
For 5, shielding shielding shielding. If all else fails, add a low impedance path to ground with some Y-caps somewhere. So that the spike does not have to go all the way over your board to the supply of the board.
For 6 and 7, relays are evil.