Peltier cooling a DSI II

[kinetic (Steve)]

Many have probably tried this but I decided to dive inside
one of my DSI IIs and see how easy I could cool it.

Preliminary testing is very positive but no star shots yet.
As it's all being refined it all hangs precariously off the back
of the DSI housing until I've perfected the shape of the cold finger.

Two frames attached. First one is 5 mins after imaging when the
camera should have settled to ambient, which was 32 Deg C.

Second is half an hour later after the peltier had settled to a stable
output and the camera was cooling no more. CCD was at 4.8 deg C.

I'm not exactly sure where the DSI senses it's own CCD temp,
I have seen a post questioning where the SMD sensor is.
Maybe it's internal to the CCD....dunno, haven't read the ICX429 PDF yet.
Anyway, the CCD temps are shown here. Compared to a thermocouple/
DMM probe I had drilled into the cold finger the values are almost the
same. Probe said 5.2deg, CCD said 4.8deg.

The exact same stretch has been applied to both dark frames.

Steve

[allan gould]

Now thats a real difference I can appreciate. Well done Steve. Love to see the filal photos.

[White Rabbit]

Indeed, thats quite significant. I

[pjphilli (Peter)]

Hi Steve
How do you prevent fogging on the DSI2 chip when you cool below the ambient dew point?
Cheers Peter

[kinetic (Steve)]

Peter,

All benchtop testing so far. I've had the camera face down on
the bench with the nosepiece capped for darks, so I didn't
even check if the CCD had dew on it!

I could easily fry the camera this way of course....you don't
want condensation forming on an SMD circuit board!

The DSI actually lends itself well to making a cold chamber or
at least isolating it from the circuit board and still allowing it
to function. The CCD is soldered on to a rectangular section
of board that has fine gaps removed from the PCB material,
just leaving the tracks intact for operation.

I can't think why they have done this other than maybe to suppress
the electronic noise around the CCD area or for maybe a cooling
purpose. Dunno.

Either way, it would be quite easy to cut the whole CCD board
section out...extend the legs into a cold chamber and seal it up.
This could possibly add noise to the whole thing too and fail
miserably

Steve

[pjphilli (Peter)]

Hi Steve

Just remembering from my old Cookbook camera days which involved making the camera and interface electronics. It had the ccd chip mounted directly on the cold side of the Peltier. The aim was to get the chip down to about -30deg C to get decent noise suppression on the old fashioned Kodak ccd. Temperature was measured by a small temperature indicator transistor imbedded next to the ccd chip on its "cold finger". Dewing up was avoided by slowly increasing the Peltier current. As the cold finger got colder quicker than the ccd, any moisture in the sealed camera first formed on the cold finger and then froze leaving the ccd cold and dry!

I too have modified my DSI2 by slicing the cooling fins off the back heat sink, sanding the aluminium surface completely flat and then mounting a neat little fan/heatsink (to get rid of the Peltier heat)/ Peltier with the cold side on the flat DSI heat sink. A normal 40mm square Peltier neatly fits the fan/heatsink which I rescued from an old Pentium 2 computer where it was used to cool the CPU.

In summer my DSI2 can reach and indicated temperature of up to 35deg C. At this temperature, noise on long exposures is rather heavy. I usually cool the CCD down to an indicated temperature of about 19degrees C where the noise is acceptable. This only requires 5 volts at 0.5 amps into the Peltier which I derive from my astro PC.

I had thought about putting drying material in the camera so that I could take the CCD lower without fogging up but I have not gone to this trouble yet. I guess also from my Cookbook experience that if I did go well below the ambient dew point that any moisture would form on the DSI2's cold finger before forming on the ccd - must experiment with this one day!

Cheers Peter

[kinetic (Steve)]

Originally Posted by pjphilli

Hi Steve

Just remembering from my old Cookbook camera days which involved making the camera and interface electronics. It had the ccd chip mounted directly on the cold side of the Peltier. The aim was to get the chip down to about -30deg C to get decent noise suppression on the old fashioned Kodak ccd. Temperature was measured by a small temperature indicator transistor imbedded next to the ccd chip on its "cold finger". Dewing up was avoided by slowly increasing the Peltier current. As the cold finger got colder quicker than the ccd, any moisture in the sealed camera first formed on the cold finger and then froze leaving the ccd cold and dry!

I too have modified my DSI2 by slicing the cooling fins off the back heat sink, sanding the aluminium surface completely flat and then mounting a neat little fan/heatsink (to get rid of the Peltier heat)/ Peltier with the cold side on the flat DSI heat sink. A normal 40mm square Peltier neatly fits the fan/heatsink which I rescued from an old Pentium 2 computer where it was used to cool the CPU.

In summer my DSI2 can reach and indicated temperature of up to 35deg C. At this temperature, noise on long exposures is rather heavy. I usually cool the CCD down to an indicated temperature of about 19degrees C where the noise is acceptable. This only requires 5 volts at 0.5 amps into the Peltier which I derive from my astro PC.

I had thought about putting drying material in the camera so that I could take the CCD lower without fogging up but I have not gone to this trouble yet. I guess also from my Cookbook experience that if I did go well below the ambient dew point that any moisture would form on the DSI2's cold finger before forming on the ccd - must experiment with this one day!

Cheers Peter

Peter,

I had thought of milling the heatsink face too but it commits me
to cooling the entire housing that way.
That would keep the entire CCD/ electronics/ housing cold for
sure but it might make the Peltier less effective.

I thought a better way to go would be as per the Cookbook.
I also have the book from ages ago.

I remember from making cold fingers for modded webcams that the
cold finger works just like a soldering iron in reverse.
As you want to transfer the heat to the very tip, you make the tip
the least cross sectional area.
The tip of my 'block' is less cross sectional area than the plate
that it faces on to, shown here in my test setup.
Also notice the drill holes into the 'block' and the plate.
That is where the temp was measured.
The block was slightly colder than the plate with the DMM probe.

The problem now is that dew will form on the narrowest part of the cold
finger.
One thing might make things easier though...the 429 CCD is sealed
with some sort of while sealant from allowing moisture through
into the camera. I'm not confusing it with transfer compound.
It is also sealed at the front window by a neoprene
gasket. This doesn't prevent it forming on the CCD legs however.

Dessicant and/ or sealing up any volume around the immediate
vicinity of the CCD may help this.

Steve

[pjphilli (Peter)]

Thanks for this further information Steve. It certainly is more of a Cookbook approach and more effective in achieving low temperatures for the CCD chip. I should imagine that the gradual cooling technique would also allow you to reach low temperatures a la Cookbook. In my approach, I only milled of the cooling fins off about half way so that the 40mm Peltier was centred directly over the internal ccd finger. However, this did involve butchery of the DSI2 rear, whereas with your method I guess you could put your DSI2 back to original if necessary. This has been a very interesting thread and I would be interested in reports of your further experiments/results in due course.
Cheers Peter

[peter_4059 (Peter)]

Hi Steve and Peter. I'm really interested in this thread as I too have built a peltier back for my DSI II. I was planning to fit it this summer. I am wondering if you have encountered any induced noise from the peltier device in the image?

Here's where my project has got to...

Peter

[kinetic (Steve)]

Hi Peter,

The dark at the start of the thread was approx 4-5C CCD
temp but only fed with DC straight from a 12v battery.
PWM would be the way to have fine temp control but as
you are probably aware, some PWM methods can be noisy.

I actually get induced noise from my Bartels stepper drive
which can be random and quite different on separate nights.
The good thing about noise though is it can be cancelled
out in the dark subtraction.
There is almost no evidence of noise in my best results to date,
even after some pretty extreme stretching, which I often
do because of the 5-10sec subs I regularly use in DS work.

Edit:
Thanks Peter for posting the details!
That seems like a great idea for the peltier control.
It even has some hysteresis adjustment. Cool!
cheers,

Steve

[peter_4059 (Peter)]

Steve,

I've edited my post with some pics of my new peltier back. I went for the shave off the fins and sand the surface flat approach. I've also built a temperature controller from a Jaycar kit that holds the temperature of the aluminium cold finger at 10 degC (+/-0.5 degC) and switches power to the peltier using a relay.

http://www.jaycar.com/productView.as...e&form=KEYWORD

[kinetic (Steve)]

To Peter (4059),

I went out and bought one of the kits you mention Peter and have
set it up running my DSI II cold finger arrangement as it currently
stands.

You mentioned that it can hold your C.F. at 10 deg C (+/-0.5).

After a bit of benchtesting I found that with my fan/peltier/finger
on a full 12v I can have the pelt run until it reaches a low setpoint
then it kicks out.
It rises a certain amount then kicks back in again. All good so far.

Then I decide on a setpoint...I started with 15 deg C.
Nicely cycles...as it should.
Then I set the hysteresis to about halfway. This gives about a 4 deg
delta.

This hysteresis is setable to from between 3 deg delta to about 10 deg
delta.
I could not get it to keep the finger between a degree of change
What setup allows you to have it cycling with only 0.5deg delta?
Have you got the 12v fan running on the full 12v or is it PWM
controlled?
The only scenario I can imagine where the peltier could hold the
temp so close would be the relay clicking in and out rapidly,
every second or so? You wouldn't want that?
OR....the fan speed lowered so the cooling effect is only just getting it
to 10 C and it stays there purely because of it's relationship to
ambient....you wouldn't want that either? Ambient could change over
the hour.

What I'm getting at is the hysteresis of the circuit only seems to let
you have a minimum delta of about 3 deg C.

Sorry for the questions Peter but I'm finding it difficult to duplicate your
exact delta of 0.5 deg C.

It certainly sounds like an ideal setup.

Steve

[peter_4059 (Peter)]

Steve,

I'll have to fire it up again (and open up the box to see where the hysteresis is set). I put it all together earlier in the year when the ambient temperature was lower so this could be playing a part in what you are seeing. Also - how are you measuring the temperature? I'm using one of the Aus Geographic inside/outside temp and humidity sensing devices - it has an external sensor for outside temp that I'm attaching to the end of the cold finger.

I've got the fan running with the full 12 V.

I'll go and set it up tonight and report back.


ps have you got it attached to the DSI and if so have you noticed any noise issues? Also have you sealed up the DSI case/provided dessicant to prevent dew condensing on the electronics?

Cheers,

Peter

[peter_4059 (Peter)]

Steve,

I had another thought - I've embeded the temperature sensing device for the control circuit in a big chunk of aluminium that sits between the peltier and the flattened back of the DSI. Perhaps this thermal mass is helping to hold the temperature more constant. The peltier, fan and the mounting was from a cheap in-car mini fridge.

Peter

[peter_4059 (Peter)]

Steve,

Just set it up again and ran a test. Ambient temp is 26 degC. Setpoint on the cold finger is 13 degC according to my TempTec device. The peltier is cooling for 15 secs on and 10 secs off and the temp is holding within 0.1 degC. Do you have concerns about the life of the peltier or the relay with this mode of operation? I haven't opened up the case yet but suspect I set it for minimum hysteresis.

Hope this helps.

Peter

[kinetic (Steve)]

Thanks Peter for the feedback, much appreciated!

Originally Posted by peter_4059

how are you measuring the temperature? I'm using one of the Aus Geographic inside/outside temp and humidity sensing devices - it has an external sensor for outside temp that I'm attaching to the end of the cold finger.

I'm using my Fluke DMM with JK thermocouple.
It agrees pretty well with the same inside/out thermometer you describe.
The DMM has a faster reaout sample rate though for fine readings.

In the pic attached I have shown where the thermistors are connected.
It's the coldest part of the block.

Originally Posted by peter_4059

have you got it attached to the DSI and if so have you noticed any noise issues? Also have you sealed up the DSI case/provided dessicant to prevent dew condensing on the electronics?

Not yet. Once I get this thing cycling nice and not too much
of a variance...as long as I can set any temp around 5-10C that
would be fine for summer imaging.
No need to go much lower because the 429 CCD is so clean a picture
anyway at below 20!
No thoughts yet on sealing it. As I said in an earlier post, the layout
and the neoprene grommet at the front lends itself well to volume
filling/ insulating etc.

Originally Posted by peter_4059

I had another thought - I've embeded the temperature sensing device for the control circuit in a big chunk of aluminium that sits between the peltier and the flattened back of the DSI. Perhaps this thermal mass is helping to hold the temperature more constant. The peltier, fan and the mounting was from a cheap in-car mini fridge.
Peter

I think you might be right here Pete.
In the next day or so I'll try feeding the pelt with a PWM 6-12V and
see if I can get a nice cyclic minimum delta.


cheers,

Steve

[peter_4059 (Peter)]

Steve,

The other thing I did was use both halves of the relay for switching the peltier (I think you'll find it has a pair of parallel contacts) to minimise the load on each set of contacts. From memory my peltier draws 3-5A and this was near the limit of each switch in the relay. It looks from your photo that you are only using one set however it is very easy to parallel up the connections on the blue terminal block.

Peter

[Scorpius51 (John)]

Hi guys

Thought you might like to look at this TEC controller chip. I have been involved with a Peltier controller for a photodetector (discussed in another thread) and have used a circuit based around this controller. It has given us control from +25C to -45C with ~+/- 0.1C variation.

Take a look at

http://www.analog.com/static/importe...ts/ADN8830.pdf

and

http://www.analog.com/static/importe...027AN705_0.pdf .

We made up a sealed enclosure to house the TEC/detector module and used a desiccant sachet. It was able to give us -45C with no significant frost.

Cheers
John

[kinetic (Steve)]

Originally Posted by peter_4059

The other thing I did was use both halves of the relay for switching the peltier (I think you'll find it has a pair of parallel contacts) to minimise the load on each set of contacts. From memory my peltier draws 3-5A and this was near the limit of each switch in the relay. It looks from your photo that you are only using one set however it is very easy to parallel up the connections on the blue terminal block.

Peter

Thanks for the tip Peter,
my 40x40mm pelt draws 2.8A @ 12 volt so I wasn't worried about that.
And as I want to drive it all with a milder control (see below) the
measured amps were about 1.5. Appreciate the advice. No sense in
making the relay contacts work hard!

Originally Posted by Scorpius51

Hi guys

Thought you might like to look at this TEC controller chip. I have been involved with a Peltier controller for a photodetector (discussed in another thread) and have used a circuit based around this controller. It has given us control from +25C to -45C with ~+/- 0.1C variation.

Take a look at

http://www.analog.com/static/importe...ts/ADN8830.pdf

and

http://www.analog.com/static/importe...027AN705_0.pdf .

We made up a sealed enclosure to house the TEC/detector module and used a desiccant sachet. It was able to give us -45C with no significant frost.

Cheers
John

Great advice John thanks. that is the sort of idea I wanted in the first
place rather than bodging up a heap of semi-applicable kits to do
a 3/4 functional job. (hehe see below...again!)
A very wise man once told me to get things working for you rather
than waste time the other way round (Rally )

Anyway.... going against that advice somewhat :
today after work I tried feeding the pelt/fan assembly
with some Pulse Width Modulated 12V.
I used the Jaycar KC5225 kit: A TA494 PWM IC driving a Mosfet.
Capable of 10A with one Mosfet fitted..and 20A with a second in parallel.

Results were that straight away the fan struggled sometimes to rotate
at lower duty cycles. Temp was very controllable though. I could
keep a setpoint almost stable at 5 deg C .
Because of the fan problem I decided to feed the fan with the full
12V and only the peltier got the PWM.
A slight improvement in setting a reliable setpoint. 5 deg C only drifted
from between 4.7-5.5 C.
The thing that made it drift was the dropping AMBIENT temp of course.
With just a slight tweak of the 5k Pot on the PWM board I could get it
back on 5 C straight away.
Ambient needs to be kept an eye on as well.
My Fluke DMM only has one K type thermocouple input.
Jaycar sell a great little 2 input meter with fairly fast update and max
and hold functions. In the last shot you can see the screen of this
meter holding 5 C. A quick press of the T2 button and I can read ambient
with the other K thermocouple (meter comes with 2 of them)
It also has a T1-T2 function where you can see the delta the peltier
is working with.

Steve

[peter_4059 (Peter)]

Looks like you are making progress Steve. 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?

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?

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