New Gadget project started. Fridge fan controller. (1 Viewer)

[wizzer59]

Your not from Yorkshire are you? I don't seem to understand a word your saying but words and pics are good and I'm sure the story will only get better

 

[wizzer59]

Jeez I thought you said simple?? are you gonna use all those resister thingys??

 

[Gromett]

Jeez I thought you said simple?? are you gonna use all those resister thingys??

Not on this one project no. Each resistor costs a fraction of a penny. So rather than just ordering each value as I need them I ordered a bulk load of 1460 of them in one go.
That is 73 different values, 20 of each value.

As I won't know what values I need up front I would have had to order a few of a range of values anyway.

It was only £7 for the lot

^{_{Subscribers  do not see these advertisements}}

 

[Gromett]

I couldn't wait to get started on this. So I looked at what I had and realised that the digispark has an LED built into it and it is connected to an IO port.

2 minutes later I had it flashing morse code for SOS... Not much else I can do with it, but at least I know the development IDE work

 

[pyro]

Good luck Gromett, the fun begins.

An interesting aside, that temperature controlled fan http://www.ebay.co.uk/itm/Arctic-F12-TC-120mm-PC-Case-Cooling-Fan-Temperature-Controlled-Silent-Quiet/182257356550?_trksid=p2047675.c100008.m2219&_trkparms=aid=444000&algo=SOI.DEFAULT&ao=1&asc=39060&meid=4bb7b834cdd94a64806f87bf39b374b2&pid=100008&rk=2&rkt=5&sd=182288768003&clk_rvr_id=1107436610887&afsrc=1&rmvSB=true could always be fitted with a series variable resistor in the sensor (thermistor) and thus change its range

Of the two solutions, I far prefer your project, which is well considered, well specified and will be gratefully implemented in our own M/H

 

[Gromett]

Good luck Gromett, the fun begins.

An interesting aside, that temperature controlled fan http://www.ebay.co.uk/itm/Arctic-F12-TC-120mm-PC-Case-Cooling-Fan-Temperature-Controlled-Silent-Quiet/182257356550?_trksid=p2047675.c100008.m2219&_trkparms=aid=444000&algo=SOI.DEFAULT&ao=1&asc=39060&meid=4bb7b834cdd94a64806f87bf39b374b2&pid=100008&rk=2&rkt=5&sd=182288768003&clk_rvr_id=1107436610887&afsrc=1&rmvSB=true could always be fitted with a series variable resistor in the sensor (thermistor) and thus change its range

Of the two solutions, I far prefer your project, which is well considered, well specified and will be gratefully implemented in our own M/H

Thank you...

I had looked at your link before. The problem is at an ambient temp of 40C the fan will spin full speed all the time, even if the fridge is down to temperature and no heat is being generated at the back of it. That is why I am using two temp probes to work on the differential.

^{_{Subscribers  do not see these advertisements}}

 

[pyro]

I understand, no real point that "solution" then.

I suppose a var resistor in series with the probe wouldn't help?

The measurement of the temperature differential is key to the success of this project. As I say, I know which solution I prefer (yours!) so I shall be watching progress with great interst

 

[Gromett]

I suppose a var resistor in series with the probe wouldn't help?

Not for me and not in this particular use case. The fan needs to be off below a certain temperature and off if the ambient is above a (variable) temperature if it wont' have any effect. For the range in between it needs to run at a speed low enough to minimise power usage and noise issues but also have the maximum effect. It is an unusual use case for fans in that an increase in temperature doesn't always result in a higher fan speed.

 

[Gromett]

Just tripped over this when looking for A chip's datasheet.

Got one on order. Looks like it will make my job much much easier.

http://www.awesomepcb.com/ardushield/

^{_{Subscribers  do not see these advertisements}}

 

[stevec]

Thought you might have used the Nano, Grommet.

 

[Gromett]

Thought you might have used the Nano, Grommet.

Why? I only need 2 inputs and 1 output. The ATTiny85 is only 97p each. The Nano is a couple of £ each and I wouldn't use half the features on it. Plus the nano is bigger and draws more current.... The 328 is overkill for this I think.

 

[OldAgeTravellers]

Hi Gromett,
Great project that I will follow with interest. My two fans are switched by a thermostat and so of course will only run at full speed although I can't hear them inside the van or on the sitting side because my fridge is on the other side to the door. But I had to fit a switch otherwise they will be running all the time the sun is on the fridge vents even when the fridge is off. (We live in the South of France)
So your fit and forget system will be very elegant and can be left on permanently. Or perhaps there is a terminal on the fridge where we can get an ON signal to power up the unit. I will have a search but I suspect it may be buried in the electronics.
I can handle the soldering and testing but it is not worth all the other kit for programming the processor, would you be able to do that for the few of us that would really like to duplicate your project? If you are still prepared to share your circuit that is.
I will watch with interest.

Steve

^{_{Subscribers  do not see these advertisements}}

 

[Gromett]

The software and the hardware will be open source so anyone can replicate the project. It will include full instructions on how to make it, the circuit diagram and software source.

I have some concerns about building them for other people. I will need to think about it.

I am not sure if it will be classed as a commercial project and if so do I need a CE certification for it.
The other issue is that although the device itself is intrinsically safe. It will require people to source 12V safely and I am not convinced everyone can do this. If the 12v supply causes a fire in somebodies van I am of course concerned not only about my reputation but my liability.
finally, I have enough work on with my existing company and I don't want to have to start doing support for this project as well

If I have to do insurance, CE testing etc etc. The price will end up being a lot more than the £3-4 it will cost me to make them and probably in the low volume I can make myself would mean charging a lot more than they are worth.

I will think about it a bit more once I have the finished item.

 

[OldAgeTravellers]

I have some concerns about building them for other people

NO, NO , NO, I appreciate that it is not worth you making them for people.
What I asked is, would you be able to program the microprocessors for a few of us for a fee. Perhaps supply the processors to us at cost plus time plus postage and a bit to contribute to the cost of the development gear you have bought, if that is realistic. There would then be no comeback on you at all. We could pay you through PayPal if you have an account.
I would enjoy making up the circuit myself and installing it. Which of course would be totally my responsibility. But programming the processor would not be practical.
I have a breadboard and a large bag of resistors, voltage regulators, transistors and diodes from my late brother so may already have a great deal of what I need.
Steve

 

[Gromett]

NO, NO , NO, I appreciate that it is not worth you making them for people.
What I asked is, would you be able to program the microprocessors for a few of us for a fee. Perhaps supply the processors to us at cost plus time plus postage and a bit to contribute to the cost of the development gear you have bought, if that is realistic. There would then be no comeback on you at all. We could pay you through PayPal if you have an account.
I would enjoy making up the circuit myself and installing it. Which of course would be totally my responsibility. But programming the processor would not be practical.
I have a breadboard and a large bag of resistors, voltage regulators, transistors and diodes from my late brother so may already have a great deal of what I need.
Steve

I could probably do something along those lines. Although I am a fulltimer and I tend not to park near a post office most of the time so I would need to do them in batches.

^{_{Subscribers  do not see these advertisements}}

 

[OldAgeTravellers]

I could probably do something along those lines. Although I am a fulltimer and I tend not to park near a post office most of the time so I would need to do them in batches.

Thanks Gromett, please put me down for a programmed processor when you are happy with the result. My current set-up will be fine for the winter trip to Spain as long as I remember to switch it on.
Steve

 

[Robin McHood]

I could probably do something along those lines. Although I am a fulltimer and I tend not to park near a post office most of the time so I would need to do them in batches.

Me too please Grommet, I'd be grateful if you'd add me to the list as and when. Cracking idea.

 

[Gromett]

I would enjoy making up the circuit myself and installing it. Which of course would be totally my responsibility. But programming the processor would not be practical.
I have a breadboard and a large bag of resistors, voltage regulators, transistors and diodes from my late brother so may already have a great deal of what I need.

So far this is the BOM for the build.
1 x ATTiny85 MCU
1 x 100R resistor
1 x 1K resistor
1 x IRL540N Mosfet
1 x 1N4004 or any higher specced.
2 x DS18B20 Temperature Sensor

I also need to find the best way to provide a stabilised and regulated 5v supply. Not keen on using a linear regulator due to losses/heat but a switched mode would put the cost of parts up too much. Still working on this bit.

^{_{Subscribers  do not see these advertisements}}

 

[Robin McHood]

What about a Mornsun dc/dc regulator or similar ? They have efficiencies around 85% and no load currents of a few mA?

 

[DBK]

Thank you...

I had looked at your link before. The problem is at an ambient temp of 40C the fan will spin full speed all the time, even if the fridge is down to temperature and no heat is being generated at the back of it. That is why I am using two temp probes to work on the differential.

If it's an absorption fridge at an ambient temperature of 40C its going to be running continuously I think. A compressor fridge might be different though.

Are you not measuring the absolute temperature anywhere and only the difference? If so is there not a danger the fan will be running when it doesn't really need to? For example, if the ambient temperature is 15C and the fridge starts cooling, conventional wisdom suggests a fan isn't required but if I understand your design the fan is likely to be running, albeit fairly slowly and not really doing a lot. It could also turn on if the ambient is even lower.

 

[Gromett]

If it's an absorption fridge at an ambient temperature of 40C its going to be running continuously I think. A compressor fridge might be different though.

Are you not measuring the absolute temperature anywhere and only the difference? If so is there not a danger the fan will be running when it doesn't really need to? For example, if the ambient temperature is 15C and the fridge starts cooling, conventional wisdom suggests a fan isn't required but if I understand your design the fan is likely to be running, albeit fairly slowly and not really doing a lot. It could also turn on if the ambient is even lower.

Subject to tests. But my thought is if it is cold then the heat will dissipate rapidly making the differential below the threshold. I will be using PID to set the fan speed however the inputs will be T2 - T1 - X. Where T2 is top temp, T1 is bottom temp and X is a preset variable. The result of this calculation if positive will be passed to the PID controller.

Lets say X is set at 2C, T1 is 15C in your example. The fridge turns on and generates some heat and the temp rises to 18C in the back of the fridge. T2 - T1 - 2 = 1 so the fan will run at speed 1. As soon as the temp falls back to 17C the fan will stop.

If during my tests this doesn't work then I will add a check for T1 <= Y where Y is a preset. So if the ambient falls below a fixed level it will skip the PID routines and set the fans to 0.

I have a rough idea in my head on how to do this. But until I have the electronics built and the basic code written I can't test to see what the issues are.

^{_{Subscribers  do not see these advertisements}}

 

[Gromett]

What about a Mornsun dc/dc regulator or similar ? They have efficiencies around 85% and no load currents of a few mA?

too expensive Got my eye on a few options. will post more later.

 

[Gromett]

What about a Mornsun dc/dc regulator or similar ? They have efficiencies around 85% and no load currents of a few mA?

I am currently looking at either buying a premade unit like this.
Link Removed

Or buying the chips in bulk like this;
http://www.ebay.co.uk/itm/141664251045

The latter would be my preferred option so I can integrate it onto one board. But I am struggling to find a bulk supplier of Inductors that aren't surface mount.
http://www.onsemi.com/pub_link/Collateral/MC34063A-D.PDF

 

[Gromett]

Found the inductors

Link Removed

^{_{Subscribers  do not see these advertisements}}

 

[stevec]

I also need to find the best way to provide a stabilised and regulated 5v supply

The Nano already has one on. No idea if it will provide enough current for your needs. Whats your power budget?

 

[Gromett]

The Nano already has one on. No idea if it will provide enough current for your needs. Whats your power budget?

The nano has a linear regulator on it not a dc/dc buck converter. It will operate off 12v but is very inefficient. It's recommended voltage is 7-12V but your battery may hit 14.5V which may cause the regulator to overheat in prolonged use. It uses an UA78MO5 according to the circuit diagram.
https://www.arduino.cc/en/uploads/Main/ArduinoNano30Schematic.pdf

If we put in 14.5v to this regulator and take out 5V that is 9.5V we have to dissipate. If we draw 500ma that is almost 5 watts of power we need to get rid of the heat for. A heatsink would be needed and we would be adding to the heat behind the fridge and wasting battery power. My plan is to make this ultra low power and not generate any additional heat. Even the power mosfet will operate without a heatsink. This is why I went with the IRL540N instead of the cheaper IRF540.

In addition it would mean having two boards still, one for the arduino and one for the actual circuit I am designing. I prefer to have a single board with everything on it.
The nano also has a USB port and FTDI pins adding to the size unnecessarily. The nano is ok for prototyping but not for the final product in my view. However I have an UNO for the development work.

I have looked at all the different Arduino and Naked Atmel chips and for this project the ATTiny85 is best suited. It has 5 IO ports in an 8 pin dip package which will keep the PCB/Veroboard design small and requires no external components to operate the chip itself.

 

[Gromett]

ok, time for a few more details on what I am using for this.
I looked at a number of temperature sensors and eventually decided on the DS18B20 which is a one wire temperature sensor.

The beauty of this device is that although it is 3 pins, it can operate using parasitic power on only 2 pins and it transmits the data as serial data. In addition each chip has a unique serial number so you can have multiple sensors on one wire. What I will be doing is connecting both the + and the - to ground.. Strange I know but this tells the chip to take it's power parasitically from the data line. I then connect the signal wires of both sensors to a single pin on the ATTiny directly with no interfacing needed. Using the onewire library I can then read the temp from each sensor in turn. This leaves me with 4 IO pins left on the ATTiny.

The ATTiny has 5 IO pins which can be all digital. Or each pin can operate in analogue mode with 2 PWM pins and 3 ADC pins. The temp sensor will go to one of the ADC pins but the pin will be set as Digital. I will then use one of the PWM pins to drive the MOSFET which will power the fans. So as you can see I am only using 2 of the 5 IO Pins.

In my unit, I will be adding a small bluetooth transmitter which will report on the fans RPM, and the temperature of the 2 sensors back to my main Arduino project. This will use pins 3-4 of 5. Then I may add an additional temp sensor to go inside the fridge which will use the final port.

The in fridge temp sensor is pointless for anyone else building this project as it doesn't actually do anything except report the temperature. But because I am adding bluetooth to mine I can report this temp back to my main project.

Yes, this is just a small break me in project to refresh my old electronics knowledge. I have 4+ other projects planned for once this is done hence my bulk ordering of components

^{_{Subscribers  do not see these advertisements}}

 

[stevec]

Good luck with the project. Nothing like a real world project to practice on. After developing embedded 8 bit micro projects for the last 30 odd years I've learnt to always budget for a couple of extra I/O pins. It almost invariably ends up that you need them for those little extras (or someone else....usually the marketing department.....asks for something else halfway through the project).

 

[Gromett]

Good luck with the project. Nothing like a real world project to practice on. After developing embedded 8 bit micro projects for the last 30 odd years I've learnt to always budget for a couple of extra I/O pins. It almost invariably ends up that you need them for those little extras (or someone else....usually the marketing department.....asks for something else halfway through the project).

totally agree about having a few extra pins. As the core of this project only uses 2 out of the 5 I think I have specced it about right. I couldn't find a 4 pin ATTiny with 2 IO ports so having 3 spare was kind of forced on me

 

[gus-lopez]

I think as @Birtha & @John Laidler have already attempted to point out it is the 30º ambient that is the factor. Above that you have loss of condensing on an absorption fridge. Even on domestic compressor fridges the condenser & fan have different ratings for the same unit depending on which country it is specified for. A Uk one used to be around 30º whereas a Southern Mediteranean one would have a figure of 37º at least. I have just checked mine & the fans are running at full speed with a 16º differential .. i look forward to seeing how well it performs.

^{_{Subscribers  do not see these advertisements}}