EOS 1000D extreme modding

[yesyes]

Time for a bit of an update I think. The thread was already on page 6. Lots of DIY must be going on in this awful weather...

I finished the electronics and "firmware" for the Arduino. (1st picture)

On the display, the number on the left is the temperature of the far side of the cold finger. 2nd number from left is target temperature or cooling mode (I have programmed 4 modes: Off, cool at full power, cool to target temperature and cool to 1°C above dew point). The 3rd number is the heatsink temperature and the rightmost number is the ambient dew point.

The 2nd picture shows a test run. It was a bit disappointing to be honest. The far side of the cold finger did not go below +6 °C although there was ice forming around the peltier. You can just about make out some ice coming round the first bend in the cold finger.

I'm not sure why the far side did not get proper cold. Is this system too open, i.e. drawing in ambient heat?

I also had some weird effect when I switched off power to the peltier the cold finger temperature dropped by about 1.5°C rapidly, then it started rising again up to ambient. This makes me wonder whether there was some heat from the peltier hot side leaking to the cold side where the temperature sensor was mounted.

More testing and tinkering is required. I'm watching Gina's thread with interest as she seems to have similarly odd results.

[Gina]

Yes, Chris, I do indeed. I've sometimes found the temperature drop slightly after switching off the cooling. Most times it stays put for several tens of seconds before rising. This is EXIF T rather than a separate sensor. I do have a 1-wire digital thermometer attached to the cold finger where it disappears behind the sensor PCB and onto the sensor, but not connected it up as yet, or indeed got my Linux laptop working in the warm room. The 1-wire software I'm using is Linux based.

[yesyes]

Gina, when you run your tests is the camera casing closed? In other words, is the cold side mainly enclosed?

I'm asking because I wonder if at least part of my problem is that the cold finger is "out in the open".

[rfdesigner]

Chris

looking at what you're doing can I point out one potential problem

sensor failure.

You are building a system (cooling) with more than one sensor. If a sensor fails what happens?.. will it fail high or low or will it be detectable?, what if a wire just drops off will it look like a high or low temperature. You will be doing rapid cooling and moderate heating, you should anticipate mechanical stresses which could lead to sensor failure in time.

it's worth thinking though, fail safe is what you want.

I might want to use dual sensors, maybe just on the heatsink, such that if the sensors disagree then the system just shuts down.

RE: fan failure. I would certianly use a temperature sensor (or dual sensors) as you will pick up blocked vents that way.

just my 2 penneth

Derek

[Gina]

Gina, when you run your tests is the camera casing closed? In other words, is the cold side mainly enclosed?

I'm asking because I wonder if at least part of my problem is that the cold finger is "out in the open".

Yes, the camera is closed and as well as I can, sealed. The cold finger is thermally insulated inside the camera except where it goes onto the sensor. The other side of the imaging board isn't insulated but I think it should be really.

Outside the camera, the side of the cold finger at the back of the Peltier TEC is open though and ices up. Once I've decided on which Peltier(s) to use I shall enclose it in insulating foam. The icing up must be putting heat into the cold finger.

[rfdesigner]

Gina

When I was playing peltiers some years ago I came across the tip when using them in sealed chambers: leave an exposed metal surface that will be colder (fractionally) than the device you're trying to cool. That way any moisture will frost onto the coldest surface in preference to the delicate CCD.

just one piece of advice I was given.. not a golden rule.

Derek

[Gina]

Gina

When I was playing peltiers some years ago I came across the tip when using them in sealed chambers: leave an exposed metal surface that will be colder (fractionally) than the device you're trying to cool. That way any moisture will frost onto the coldest surface in preference to the delicate CCD.

just one piece of advice I was given.. not a golden rule.

Derek

Yes, I have heard this before. One reason it's not all fully insulated.

[yesyes]

Chris

looking at what you're doing can I point out one potential problem

sensor failure.

You are building a system (cooling) with more than one sensor. If a sensor fails what happens?.. will it fail high or low or will it be detectable?, what if a wire just drops off will it look like a high or low temperature. You will be doing rapid cooling and moderate heating, you should anticipate mechanical stresses which could lead to sensor failure in time.

it's worth thinking though, fail safe is what you want.

I might want to use dual sensors, maybe just on the heatsink, such that if the sensors disagree then the system just shuts down.

RE: fan failure. I would certianly use a temperature sensor (or dual sensors) as you will pick up blocked vents that way.

just my 2 penneth

Derek

Thanks for that, Derek. I always welcome these comments/suggestions.

The sensors I used are DS18B20's. So they are digital and will not fail to high or low temperatures. If they fail, they will just not send back data which can be detected by the Arduino library for these sensors. If that happens, the "firmware" I wrote will shut down the peltier and display an error message.

I'm guessing that's good enough but will listen to any suggestions.. :-)

[Gina]

Precisely what I'm planning for my cooling except that I shall be using software in a laptop rather than Arduino. I'm using the same digital temperature sensors.

[rfdesigner]

Thanks for that, Derek. I always welcome these comments/suggestions.

The sensors I used are DS18B20's. So they are digital and will not fail to high or low temperatures. If they fail, they will just not send back data which can be detected by the Arduino library for these sensors. If that happens, the "firmware" I wrote will shut down the peltier and display an error message.

I'm guessing that's good enough but will listen to any suggestions.. :-)

Of course they are..Doh!.. I've been playing too many games with thermistors at work, I forgot you were using chips.

Anyway, yes it looks like you'll have it failing safe.

Another thought.. are your peltiers sealed around the edge to stop water getting between the plates?

Derek

[Gina]

Chris, I have a bit more info on the temperature dip on switching off the cooling. I got the same thing when testing this morning. I found the radiator cooling fan was rubbing on its casing due to not being upright, so I switched it off together with the Peltier supply and re-fixed it in an upright position. When the cooling was switched off the EXIF T dropped by 2C then went back to what is was before rising a degree or so while the cooling was off. When I switched the cooling back on the temperature continued to rise for a minute or more reaching +4C before gradually going back down as the cooling took effect. So there seems to be a time lag between the cooling input and what APT says happens in the imaging chip. I'm planning to get the 1-wire system going again shortly and then with temperature sensors at various places, I might get some more ideas of what's happening.

[rcheshire]

When I was playing peltiers some years ago I came across the tip when using them in sealed chambers: leave an exposed metal surface that will be colder (fractionally) than the device you're trying to cool. That way any moisture will frost onto the coldest surface in preference to the delicate CCD.

Unintentionally, the copper heatsink mounting plate on my system does just that. Perhaps replacing it is not a good idea - also a useful tactile indicator.

Love the system you've built Chris. Given more time I would consider doing something similar. For now, I'm happy with LED indicators for monitoring temp control and error detection.

[yesyes]

Not much progress on the actual cooling side of things...

Still can't get the far side of the cold finger below zero degrees.

Ordered a few lower power TECs as Gina suggested in her thread.

But I thought I'd post a screenshot of the Windows application that I wrote to configure and monitor the cooling controller. I can read all values here and set cooling mode and target temperature. The Arduino based cooling controller does remember the settings when it gets powered down and will resume operation with previous settings when powered up again even without the Windows app. I will need to use the app if I want to change any settings though.

[rfdesigner]

The Arduino based cooling controller does remember the settings when it gets powered down and will resume operation with previous settings when powered up again even without the Windows app.

excellent.

is there a way of getting it to report the target temperature without the app?

With setpoint temps you can use a set of darks for quite some time, and may have more than one set (-25c & -5C for instance)

it could be useful to know at a glance what the setpoint is, or are you planning to program it up at the beginning of each imaging session?

Derek

[yesyes]

excellent.

is there a way of getting it to report the target temperature without the app?

yes, see the images at the top of this page.

I have added 16 digits of 7-segment displays showing 4 values with 4 digits each. Leftmost is the current cold finger/CCD temperature. The next is the target temperature or the cooling mode if there is no specific target for this mode (OFF, FULL, specific PWM value). The next 2 are heatsink temperature and ambient dew point temperature. These always show values whether the Windows app is connected or not.

[Gina]

I'm looking into using an Arduino like you Chris and have just reread this thread Been looking at Arduinos and associated bits on eBay - cheap aren't they? Two reasons really - I think it would be more efficient than my way, by way of space taken up and I like the idea of displaying temperatures separately from the computer plus there's really no point in "reinventing the wheel" as long as you don't mind me filching your programming work I can also use the Arduino to control other things. I'm thinking of using a separate unit rather than putting everything in the camera box. The power MOSFETs I'm using now are logic level devices with a low RDS(on) (from eBay cost 99p) - data sheet here :- http://www.irf.com/product-info/datasheets/data/irfz44n.pdf

Also been thinking about your problems of getting enough cooling. As you know from my threads I have all my electrical cooling inside the camera box with the absolute minimal length of cold finger - I'm sure this helps as it's very much better than the longer one with external cooling I had before. The Peltier TEC I'm currently using on my wdiefield camera is a 12v 37W one from a UK supplier :- http://www.ebay.co.uk/itm/360468741168?ssPageName=STRK:MEWNX:IT&_trksid=p3984.m1439.l2649 With this and water cooling of the hot side I can get -12C EXIF T with long exposures.

I'm also looking at using air cooling rather than having all the complications of water cooling. Now I've reduced the power input to the TEC I believe I should be able to dissipate the resultant heat without resorting to water cooling. This is work-in-progress.

[yesyes]

Hi Gina. I absolutely don't mind sharing the Arduino sketch (that's what they call a program for an Arduino). I also have the circuit diagram of what I've built and I can also share the Windows app I wrote to control / monitor the cooling. If you'd rather use Linux, then I don't have an app for that (no Linux programming experience) but I can share info on the communication protocol I use between Arduino and PC. This is the same protocol I designed for my home automation project. It might be a little over the top for a single cooling controller (it can address 64k devices with 64k different values each). But it's quite easy to understand. It's basically 16 bytes data packets being sent back and forth over a (virtual) serial port.

Let me know what you would like me to provide. (everyone else is of course also welcome)

The Arduino sketch will still need a bit of work on my part. I haven't yet been able to test it fully (for example set-point cooling) because I still have issues with the actual cooling. I haven't had time to work on this recently but will resume as soon as I can. First thing will be to test with a single TEC. I also have 12V 35W TECs now (one 30x30mm and one 40x40mm). Will test with those next.

[Gina]

Windows software is fine - thank you ATM I'm running Windows only in the obsy - XP on my netbook. I'm also running a Windows desktop indoors. Win 7 Pro 64bit on an Intel Cor i5 2.8GHZ with 4GB RAM and 1TB HD space in 2 drives. I use this for running DSS and PS for image processing. I have Linux on 4 boxes, Mac OS X on 1 laptop and Win XP on 3 boxes inc. the netbook. Gosh - have I got that many computers?

[Gina]

Which version of the Arduino are you using Chris? I've been looking on eBay and there's a lot of choice. Do you have any particular recommendations? Could I use the same board to control 2 or 3 stepper motors as well? I mean running everything off the one chip. Or would it be better to use separate Arduinos. Think it's time I got going on this Manual temperature control is a bit hit and miss.

[yesyes]

I've been meaning to reply and prepare the software and schematics for you. But I've been rather busy these days... I should be able to get this ready this weekend.

I'm using all kinds of Arduinos, mostly the Mini version as it is the smallest. For the cooling controller though, I used an Arduino Nano because it has USB on board. The Mini only has a serial port and you would need a separate USB to TTL serial adapter.

Theoretically it should be possible to run a few steppers as well as controlling the TECs. However, bear in mind that an Arduino can only run one single program, no multitasking. So you would have to integrate the stepper control into the TEC controller sketch. It would be far simpler to have a separate Arduino; especially as they are quite cheap.

[rfdesigner]

Which version of the Arduino are you using Chris? I've been looking on eBay and there's a lot of choice. Do you have any particular recommendations? Could I use the same board to control 2 or 3 stepper motors as well? I mean running everything off the one chip. Or would it be better to use separate Arduinos. Think it's time I got going on this Manual temperature control is a bit hit and miss.

I use PICs which are similar.

My homebrew mount driver is based on two stepper motors, drivers and a PIC in the base and another in the hand controller and all works pretty well. The base PIC runs a real time clock and keeps track of where both motors should be whilst at the same time communicating with the other PIC over I2C. I manage all the different tasks by running interrupts fired off the real time clock. When I started running into problems with not enough processing power I upped the operating frequency from the internal 8MHz to an external 20MHz.

All of what I've described above should be achievable on an arduino.

Derek

[Gina]

I've been meaning to reply and prepare the software and schematics for you. But I've been rather busy these days... I should be able to get this ready this weekend.

No problem Chris - no hurry, I haven't even ordered any hardware yet

I'm using all kinds of Arduinos, mostly the Mini version as it is the smallest. For the cooling controller though, I used an Arduino Nano because it has USB on board. The Mini only has a serial port and you would need a separate USB to TTL serial adapter.

Ah right, I'll get a Nano for the cooling and get that going first. Sounds like a good bet for other things too but I'll give that some thought.

Theoretically it should be possible to run a few steppers as well as controlling the TECs. However, bear in mind that an Arduino can only run one single program, no multitasking. So you would have to integrate the stepper control into the TEC controller sketch. It would be far simpler to have a separate Arduino; especially as they are quite cheap.

OK I'll do that then Thanks

[Gina]

I use PICs which are similar.

My homebrew mount driver is based on two stepper motors, drivers and a PIC in the base and another in the hand controller and all works pretty well. The base PIC runs a real time clock and keeps track of where both motors should be whilst at the same time communicating with the other PIC over I2C. I manage all the different tasks by running interrupts fired off the real time clock. When I started running into problems with not enough processing power I upped the operating frequency from the internal 8MHz to an external 20MHz.

All of what I've described above should be achievable on an arduino.

Derek

Thanks

[Gina]

I've ordered a Nano :- http://www.ebay.co.uk/itm/320853660843?ssPageName=STRK:MEWNX:IT&_trksid=p3984.m1497.l2649

[ncjunk]

Carefull they are addictive, just dont look at the shields you can get or you'll start wanting to control the world with them