Peltier cooling a DSI II

[kinetic (Steve)]

Originally Posted by peter_4059

What period is the peltier switching in and out over? Does adding PWM help? I thought this would just increase the time to cool down leading to a bigger variation in the temperature?

you are dead right here Pete...I can have it STAY at 5 C but it isn't
reaching the cutout temp. That is about 1.5-2C.
Because the PWM gives finer control over the 12v rather than brute force
it never gets to cutout. Of course, once ambient drops then the cutoff
would be reached. When I ramped up the PWM to let it reach cutout the
variance was 1.5C to 8 C...so yes a greater range....no good.

Originally Posted by peter_4059

The other variable is temperature setpoint. I decided to target 10 degC as I've heard the noise is pretty low on the DSI II once you get below 15 degC and the peltier will always be able to achieve 10 regardless of the ambinet temperature here so I'll only need darks at one temperature. Have you tried the setpoint at a higher temperature to see if you get a more stable control?

Yep I agree...10c is a good value to nominate.
My better results with the DSI have been with the CCD below 15C and
best at 6-7C. (coldest night I have ever imaged)

Originally Posted by peter_4059

Final question - I was trying to work out a way to make the temperature adjustment pot a panel mount that could be adjusted without a screwdriver - any ideas?
Peter

Yes Pete, as you can see from the circuit...the pot is just a voltage divider with a very fine (20turn) control over a small fraction of
that voltage.
As you have decided that 10 C is your nominated value, then all you
have to do is measure what the pot reads at your chosen temp.
Then it would be just a matter of substituting a fixed resistor
value in it's place (but halving it) and adding the same value as a
1-turn panel mount pot....if you get what I mean.
This means that you are then only limited to fine control around 10C
though from that point onwards.

Steve

[peter_4059 (Peter)]

Originally Posted by kinetic

you are dead right here Pete...I can have it STAY at 5 C but it isn't
reaching the cutout temp. That is about 1.5-2C.
Because the PWM gives finer control over the 12v rather than brute force
it never gets to cutout. Of course, once ambient drops then the cutoff
would be reached. When I ramped up the PWM to let it reach cutout the
variance was 1.5C to 8 C...so yes a greater range....no good.

Steve,

Have you tried putting the RTD in a bigger piece of aluminium rather than in the cold finger - I'm not sure how accurate the RTD in that kit is (It seems to be designed to operate over a huge temperature range). The small mass of the cold finger probably means it can change temperature more rapidly than the larger block behind it.

Here's where I'm leaning towards for moisture control in the camera...

http://www.pwtec.com/srp.htm

Peter

[peter_4059 (Peter)]

Originally Posted by kinetic

all you have to do is measure what the pot reads at your chosen temp. Then it would be just a matter of substituting a fixed resistor value in it's place (but halving it) and adding the same value as a 1-turn panel mount pot....if you get what I mean.
This means that you are then only limited to fine control around 10C
though from that point onwards.

Why didn't I think of that? Thanks.

[kinetic (Steve)]

Originally Posted by peter_4059

Steve,

Have you tried putting the RTD in a bigger piece of aluminium rather than in the cold finger - I'm not sure how accurate the RTD in that kit is (It seems to be designed to operate over a huge temperature range). The small mass of the cold finger probably means it can change temperature more rapidly than the larger block behind it.

Here's where I'm leaning towards for moisture control in the camera...

http://www.pwtec.com/srp.htm

Peter

Peter,

But wouldn't that mean the CCD itself (backed directly on my block
which cycles presently between about 1.5C-5C (minimum hysteresis))
would have during the course of an imaging session, different noise
and hot pixel levels anyway?
At present, no non cooled session has ever had an ambient change of more than
a few degrees. This peltier cycling might blow that out to a 5 C differential
and show up in processing as terribly mismatched darks to lights etc.

My signature says it all....more research needed

Steve

[peter_4059 (Peter)]

I'm not seeing that on mine with the same control circuit - I'm wondering if sloppy response of the RTD in a block that is rapidly changing temperature is causing the variation in temperature you are seeing. I'm measuring the temperature on the end (surface) of the cold finger but controlling the temperature inside the larger block of aluminium behind the cold finger. Also I've encased the RTD in heatshrink and used thermal transfer paste in the hole the RTD is inserted into.

[kinetic (Steve)]

Originally Posted by peter_4059

I'm not seeing that on mine with the same control circuit - I'm wondering if sloppy response of the RTD in a block that is rapidly changing temperature is causing the variation in temperature you are seeing. I'm measuring the temperature on the end (surface) of the cold finger but controlling the temperature inside the larger block of aluminium behind the cold finger. Also I've encased the RTD in heatshrink and used thermal transfer paste in the hole the RTD is inserted into.

Thanks Peter, I will give this a try, first chance I get

Steve

[kinetic (Steve)]

An update:
I made a different shaped cold finger last night after work on the
lathe wth a bit more thermal mass.
With the controller sensor in the thickest part of the cold finger
and the thermometer in the same spot I got the following behavior:

Ambient was 22°C
Setpoint was 5°C
Hysteresis was minimum

Fan was on full 12V
PWM was on about 50% duty cycle.

Temp dropped way below 5°C and went all the way to the
minimum temp cutout of 1.5°C approx, then continued to cycle
between 1.5°C and about 6°C (delta of 4.5°C)

On backing off the PWM I brought it back so it hovered around
the 5°C that I wanted it to.
Again it wanted to go lower.
Was this because ambient was 22°C this night instead of 30s the previous?
Probably.

So, again, I seem to have a very wide minimum hysteresis which
is frustrating because I would like to have it kick in and out within
a 0.5-1.0 range.
Maybe my thermal mass still isn't large enough yet.

I then cranked it all up to see exactly how low it would get.
I got it to -2.°2C.
From an ambient of 22°C thats a delta of 24.2°C.


Steve

[peter_4059 (Peter)]

Steve - have you got it set up to switch the peltier on on rising temperature or off on falling falling temperature?

[kinetic (Steve)]

Originally Posted by peter_4059

Steve - have you got it set up to switch the peltier on on rising temperature or off on falling falling temperature?

Hi Peter,

I set it up as default...as if it controls a cooling fan on a car.
This setup the circuit calls L/H and the diode and link are set
as it shows for L/H.

Hopefully that's right?

It cuts in (LED/peltier on) when temp gets above a set point and cuts
out (LED off) when it reaches the minimum setpoint.
Have I got that logic right?

Steve

[peter_4059 (Peter)]

Yes - that's what I did - link and diode in LH position and Peltier connections to the COM and NO terminals of the relay. My hysteresis adjustment POT is fully anticlockwise.

[kinetic (Steve)]

Originally Posted by peter_4059

Yes - that's what I did - link and diode in LH position and Peltier connections to the COM and NO terminals of the relay. My hysteresis adjustment POT is fully anticlockwise.

Ok...so you said your temp probe was at the coldest part of the
tip and the controller sensor is in the largest mass part?

Is your peltier cooling the entire back metal (minus fins)?
That's probably quite a larger thermal mass than my present finger
for sure.....

Maybe that's where I'm goin wrong...

Steve

[peter_4059 (Peter)]

Yes - my temp indication is a surface device I stick on the tip of the cold finger. The temperature control sensor is inserted into a hole that is mounted between the peltier and the flat back of the DSI (I cut the fins off the DSI and sanded it flat). I've used heat transfer paste between the peltier and the aluminium block and between the block and the back of the DSI.

The aluminium block is 65x45x15 mm. The temperature sensor (thermistor) is in a hole 25mm along the long side of the block and the hole extends about half way through the short side so 20mm deep. I covered the bare legs of the thermister with heat shrink then enclosed the entire thing in heat shrink. I put some heat transfer compound in the hole before inserting the thermistor.

The peltier is cooling the aluminium block and this in turn is cooling the back of the DSI so quite a bit of aluminium in total.

[kinetic (Steve)]

Originally Posted by peter_4059

Yes - my temp indication is a surface device I stick on the tip of the cold finger. The temperature control sensor is inserted into a hole that is mounted between the peltier and the flat back of the DSI (I cut the fins off the DSI and sanded it flat). I've used heat transfer paste between the peltier and the aluminium block and between the block and the back of the DSI.

The aluminium block is 65x45x15 mm. The temperature sensor (thermistor) is in a hole 25mm along the long side of the block and the hole extends about half way through the short side so 20mm deep. I covered the bare legs of the thermister with heat shrink then enclosed the entire thing in heat shrink. I put some heat transfer compound in the hole before inserting the thermistor.

The peltier is cooling the aluminium block and this in turn is cooling the back of the DSI so quite a bit of aluminium in total.

Ok thanks Peter. Definitely a case of bigger thermal mass and the position
of the temp and sensor probes.
If I had listened to you the first time you wrote that I wouldn't
have had so much hassle!

I changed the temp to almost the tip and now the best I get is
cut out 3.0°, cut in 5.0°. Minimum hysteresis, PWM 100%,Ambient 22.
The 2 input thermometer also shows approx 1° C differential between
the two positions on the cold finger....a few pics in a sec.

Pic 1: T1 measured at tip, T2 measured at round base. Control sensor at
base too.

Pic 2: Cranked up max PWM and setpoint below zero. Got to -4 C
Pic 3: Ambient to finger differential (T1-T2 with T2 measuring ambient)

Steve

[peter_4059 (Peter)]

It looks like you've got better control of the temperature now. I think the Envisage software selects a dark from the library as long as it is within a few (5?) degC of the operating temperature. It looks like you have a fair bit of humidity to deal with for your next challenge - this is the bit that's got me worried the most as I bet the Meade electronics won't like getting wet - let me know how you deal with it. I'm thinking of adding a small metal capsule full of silica gell to the case.

Peter

[peter_4059 (Peter)]

Steve,

I finally got mine all assembled today and ran some test darks with the cooler switched on and off. Here's 60 secs without the cooler (29 degC) and 60 secs with the cooler set at 17 degC. I can report the temperature held stable on the DSI readout - not even 1/2 deg swing.

I started having trouble with the image going all the way to the right of the histogram with longer cooled exposures. I thought I'd done some damage to the camera so switched the peltier off and let it warm up and things returned to normal. I now think it was light leakage into the front of the camera around the scopestuff adaptor plate as the camera body cooled down but will have to run some more tests to verify this..

I also installed my desicant holder and this seemed to work ok (no free water droplets inside the camera when I opened it up).

Peter

[allan gould]

Peter
Have you considered dry ice on the back?

[kinetic (Steve)]

Originally Posted by peter_4059

I finally got mine all assembled today and ran some test darks with the cooler switched on and off. Here's 60 secs without the cooler (29 degC) and 60 secs with the cooler set at 17 degC. I can report the temperature held stable on the DSI readout - not even 1/2 deg swing.

Looks great Peter!
I think the onboard CCD temp in Envisage is pretty slow to respond. It certainly takes a while to show change when I turn on the peltier.

Originally Posted by peter_4059

I also installed my desicant holder and this seemed to work ok (no free water droplets inside the camera when I opened it up).

Sounds promising!
I'm leaning towards filling up the volume around the finger as the way to minimise condensation.
I figure that I just can't risk any moisture sloshing around in the camera so I tried another concept today.

I heated up the cold finger with a heatgun and pressed it into some polystyrene foam so it burned it's way down into it.
After some experiment I got the temp just right so that it wouldn't burn cavities away and have air pockets around the finger.
With some downward pressure on it as it melted into the foam I then poured cold water over the faceplate to quickly cool it and stop the melt.

A few tests with this and it held the same 2°C delta at a setpoint of 3°C. With the insulated shroud it also showed some thermal inertia as well. When the peltier cut out it continued to cool a few tenths of a degree before starting to warm up.

I then cranked it to see how low I could get it to go.
Ambient roughly 21°C...finger got to -5°C easily. A differential of roughly 26°C.
So , theoretically, I should get to below 10°C on a summer night of 35°C

Steve

[allan gould]

I know its not DIY but have a look at this site for a bolt on DSI cooler that really is nifty.
http://www.camerabug.com/
Light and easy to use. Better than the Aus version? Dont know but Im tempted. Have downloaded the pdf and the specs look very promising.
Allan

[kinetic (Steve)]

Thanks Allan, sure looks the goods.
I reckon the milled one piece back looks fantastic.
Bit pricey though.....I'd buy one in a heatbeat though
if cash was no issue.

I tried the cooler tonight on the DSI II colour.
Compared lights and darks at both ambient 22°C and
cooler at 6°C.

The first immediately obvious problem I had was the
frames at 8°C peltier cut-in are different levels to
frames at 6°C peltier cut-out.
This makes it very hard to stack a set with such a
difference. hot pixels are definitely dimmer though.
Here is a 22°C frame with four obvious hot pixels, alongside
a single 6°C frame.
The stars are constant brightness but the four hot pixels have
faded.

FWIW Peter, I found that the electronics in the fan doesn't like
being run on PWM. I had a lot of noise in my preview images.
When I ran the cooler and fan on direct DC (battery) the noise
is gone. The PWM kit does have an optional diode and capacitor to
suppress hash on the square wave output. I will try this another night.
That still doesn't explain your noise issue, you don't have the PWM
circuit.

Steve

[peter_4059 (Peter)]

Steve,

I'm not sure if it is noise I'm seeing - It looks like the cooling is causing the effect of offset to increase so the histogram is being shifted to the right. If I reduce the offset it comes back to the left. I'm not sure if there is an explanation for this - did you see the same thing?

Peter