problems with CHI030 at low ambient temperatures

[Pat13]

Hello all
haven't posted in a while. What is the ambient operating temperatures for the Picaxe and the CHI030 boards?
I use CHI030 with 18M2+ chips for a variety of Halloween props, and I am having all sorts of issues with programs not running properly once they get into cool temperatures.
I will program a CHI030 board (I have several that I use for different props and applications) and everything will run perfectly on the bench, at room temperature. But when that board goes outside (in a protective container to keep out moisture, of course), and where the ambient temp is between 0 and 8 degrees Celsius, the program does not behave properly at all. Pins will be turned on that shouldn't be, pins that should be on are not, programs that should be turning pins on and off for a certain amount of time will only run for a couple of seconds. I tried switching out the darlington drivers, with success on one board but not on the others. I tried downloading the same programs into the respective boards, doing hard resets, all to no avail. I also tried adding hand warmers to try and get the temp up, but I don't think they got warm enough to make a difference. Has anybody else had these issues and is there a way to resolve them?
The CHI030 boards a running either 12VDC pneumatic solenoids (which have a very small current draw), are switching 12VDC relays, or are running LEDs ( again very low current).
Just to be clear, I have several props that use picaxe chips that are NOT in CHI030 boards, and they don't seem to have the same problems at the low temps. it seems to be specific to the CHI030 boards.
Thanks in advance for your help.

 

[hippy]

Most components will be rated down to at least 0C and often -20C. The datasheet I have suggests a ULN2803A is rated down to -40C but it probably varies between manufacturers. The PICAXE chip itself also seems to be rated to -40C. There is nothing I am aware of on a CHI030 that would seem to be particularly problematic.

Could it be a mechanical or electrical connection issue ?

 

[rq3]

I've never used the CHI030 boards, but I do have a surface mounted Picaxe 20M2 controlling my refrigerator at 1 degree C. As a suggestion, I would look very closely at the power supply connection to the board, and reflow the solder if possible. Sounds like the picaxe may be resetting from a bad solder joint opening and closing under load.

 

[Technical]

Are you using batteries or a power supply? Cold batteries could probably give more of an issue than the circuit itself.

 

[Pat13]

Thanks for the replies.

I am going to inspect and re flow all of the solder joints. As far as power supply, all of the boards are being powered by computer power supplies, both for the 5VDC and the 12VDC, so there is plenty of amps.

 

[rq3]

The computer supply may be the issue. I believe that most modern supplies need to have at least some minimum load to stay in regulation, otherwise they shut down and then try a re-start. Although I can't think of a reason why this would only show up at colder temps, but...

 

[Pat13]

I am using these boards to switch 12VDC relays, 12VDC pneumatic solenoids and 12VDC vent motors. I know that the ULN2803 has internal diodes. Could it be that the inductive voltage is too much for those diodes, Should I be adding diodes to the relays and solenoids? I am concerned that if I hook them up incorrectly, I will release the magic blue smoke, lol. Or should I double up the outputs, as suggested in manual 1? Is that even possible on the CHI030?

 

[inglewoodpete]

I am using these boards to switch 12VDC relays, 12VDC pneumatic solenoids and 12VDC vent motors. I know that the ULN2803 has internal diodes. Could it be that the inductive voltage is too much for those diodes, Should I be adding diodes to the relays and solenoids? I am concerned that if I hook them up incorrectly, I will release the magic blue smoke, lol. Or should I double up the outputs, as suggested in manual 1? Is that even possible on the CHI030?

How much current do the solenoids and motors take each? Ie. If you connect them to your 12v supply via a meter, how much current flows? The ULN2803 is limited to a few hundred milliamps per outlet. Do they operate/run when in a "normal" temperature (it's 37C here today - not "normal" for me either!!).

The darlingtons should not need additional diodes. They would have already failed if they needed external diodes, anyway.

I think you need to post your circuit and code so that we can check your set up.

 

[hippy]

I know that the ULN2803 has internal diodes. Could it be that the inductive voltage is too much for those diodes, Should I be adding diodes to the relays and solenoids?

That would not seem to explain why it all works in warmer temperatures but not colder, why other units you have built do not suffer similar problems.

If the issue is purely temperature related I would say it is unlikely down to any electrical effects unless they can be shown to present at colder temperatures but not higher ones.

It could be there is some EMI effect which can vary depending upon temperature and how the medium that EMI travels through is affected by temperature. I don't know enough about atmospheric physics to say.

It's very hard to guess what the problem may be. Replacing one Darlington with another and having one work may be indicative of where the problem lies, but may just be coincidence and now leading the eye away from where the actual issue lies.

Apart from throwing things at it in the hope of fixing it the only way to really determine what is going on is to get it to fail in an environment which allows you to analyse what is going on. You could add SERTXD to at least determine if the unit is resetting or passing through the points where things should be happening. That will at least determine if the software is working as it should or not, will help identify what the issue could be.

 

[grim_reaper]

Should I be adding diodes to the relays and solenoids?

I'm no hardware engineer, but I thought you should ALWAYS have a back-EMF / voltage suppression diode across a relay/solenoid/coil device. Are you saying there are no protection diodes at all?
Our offshore engineer went through 10 very expensive circuit boards last month when the supplier of some fire dampers (shut-off vent things driven by solenoids) neglected to mention they their new devices didn't have diodes/suppressors installed!

I agree with hippy though - this doesn't really relate to the temperature problem. I think we all need to see a full circuit diagram to ascertain what's going on?

 

[Pat13]

I'm no hardware engineer, but I thought you should ALWAYS have a back-EMF / voltage suppression diode across a relay/solenoid/coil device. Are you saying there are no protection diodes at all?
Our offshore engineer went through 10 very expensive circuit boards last month when the supplier of some fire dampers (shut-off vent things driven by solenoids) neglected to mention they their new devices didn't have diodes/suppressors installed!

I agree with hippy though - this doesn't really relate to the temperature problem. I think we all need to see a full circuit diagram to ascertain what's going on?

From Picaxe manual 3:

Standard Circuits 2 - Using a Darlington Driver IC
If a number of output devices are being controlled it may be necessary to use a
number of output transistors. In this case it will often be more convenient to use a
ULN2003 Darlington driver IC. This is simply a 16 pin ‘chip’ that contains 7
darlington transistors similar in value to the BCX38C. The ‘chip’ also contains internal
back emf suppression diodes and so no external 1N4001 diodes are required.

Even though this states that no back emf diodes are required, my suspicion is that the diodes in the darlington ic are failing. I am using the darlington side of a CHI030 board to switch six 12VDC relays. The relays are by Varley, and are discontinued stock. The darlington supplied with the board is actually a ULN2803. When I noticed the program was not running correctly, I took voltage readings across the outputs of the darlington and half the outputs were partially open, showing a range of 5 volts to almost 8. I replaced that darlington and ran the program again, but it failed again after approx. 30 minutes. Here is the code I ran. Be aware that I am a Halloween prop builder, not a computer programmer, so my code is rudimentary at best.

'#picaxe 18M2+ CHI030 board. The darlington side of the board is used to switch 12VDC relays.  Those relays switch 110VAC items.

symbol Coffin = B.0
symbol Fester = B.1
symbol GraveGrab = B.2
symbol Zombies = B.3
symbol Axworthy = B.4
symbol OneArmGG  = B.5

	
Start0:
	Init:
	Low Coffin 
	pause 200
	Low Fester
	pause 200
	Low GraveGrab
	pause 200
	Low Zombies
	pause 200
	Low Axworthy 
	Pause 200
	Low OneArmGG  
	pause 200
Main:
	let w0 = time
	random w0
	high Coffin
	pause 15000
	low Coffin  
	let w1 = w0//65000 + 17000
	pause w1 
	goto Main
	
Start1:
	Pause 10000
Main1:
	let w2 = time
	random w2
	high Fester
	pause 13000
	low Fester  
	let w3 = w2//65000 + 13000
	pause w3 
	goto Main1

Start2:
	Pause 20000
Main2:
	let w4 = time
	random w4
	high GraveGrab
	pause 13000
	low GraveGrab 
	let w5 = w4//65000 + 24780
	pause w5 
	goto Main2
Start3:
	Pause 30000
Main3:
	let w6 = time
	random w6
	high Zombies
	pause 11000
	low Zombies
	let w7 = w6//65000 + 19550
	pause w7 
	goto Main3
	
Start4:
	pause 40000
Main4:
	let w8 = time
	random w8
	high Axworthy
	pause 26000
	low Axworthy 
	let w9 = w8//65000 + 26770
	pause w9 
	goto Main4
	
Start5:
	pause 50000
Main5:
let w10 = time
	random w10
	high OneArmGG
	pause 13000
	low OneArmGG
	let w11 = w10//65000 + 26770
	pause w11 
	goto Main5

 

[inglewoodpete]

I replaced that darlington and ran the program again, but it failed again after approx. 30 minutes.

With that extra information that you have now provided, it is pretty clear that you need diodes across (at least some) of your coils. You may as well put them across all solenoids/relays/motors.

The darlingtons should not need additional diodes. They would have already failed if they needed external diodes, anyway.

That's what I originally said, without knowing that the chips had failed after a period. Perhaps I should have said "The darlingtons should not normally need additional diodes.". For me, it raises a concern that the current being drawn by the loads is possibly pushing the limits of the ULN2803's rating.

I have used ULN2803s with solenoids operating close to the chip's limits and not needed additional diodes. You haven't told us the current that the indictors draw, or the resistance of their coils. It's hard to help you further without this information.

 

[marks]

Hi Pat13,
you code just seems to make outputs high then low so pretty standard.

in some situations (rare) you do install back emf diodes at the relay
when a relay de-energises it can sometimes influence other circuits briefly.not really what you have described

reading 7v this seems to be the difference between the two supplies and probaly enuff to keep a relay energised.
perhaps check you have removed R9 resistor from your board. ch30 board when using different supplies.

common ground is still required.
if you test your output of the 2803 you can remove your 12v supply.
its an open colector output so high5v input > output 0.6v out ,low input > probaly 0v floating if 5v the jumper r9 needs removing.
with the 12v supply connected high5v input > output 0.6v out , low input > 12v

 

[fernando_g]

The ULN2803 actually have very good back-EMF diodes. Way back at another job, the component engineer characterized them and found them to be surprisingly good and well matched.

I'm going to ask therefore a silly question: To what supply voltage are you tying pin 10 to? Or are you leaving it open?
The reason I'm asking an acquaintance of mine tied it to the +5v rail. When the relays were de-energized, all of the relay's energy was dumped into +5v, causing the processor to reset.

Pin 10 should be tied to the same voltage that powers the relays.

 

[Buzby]

Pin 10 should be tied to the same voltage that powers the relays.

Yes, if pin 10 is left open, or connected to the digital supply (+5v or +3v ), then the built-in diodes can't work, and you will end up with a dead IC.

Pin 10 *must* be connected to the 12v, or 24v, or 48v, or whatever is the supply for the inductive loads.

( Also, the grounds *must* be connected together, or else it might/might not work, or might work sometimes, but never reliably. )

Cheers,

Buzby