Peltier cooling a DSI II

[kinetic (Steve)]

Originally Posted by peter_4059

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

Yes!...I did see that. In the dark I thought I had auto-level clicked on
and that makes the histo dance around as you describe.
But it wasn't.
I'll post two extremes of histo...one at 6 deg and one at 8 deg to show
that something is different in both, and therefore makes it hard
to stack. I think normalising fixes this but it could be a downside to
the whole peltier cycle changing for a few degrees around the setpoint.

If this is what causes it then a better feedback system is needed to get
the pelitier to stay at 0.1 degree from the setpoint.
Simple relay on/off even with hysteresis just isn't going to cut it.
Even driving it with PWM and tweaking it so the temp doesn't
vary has problems. When ambient changes I have to keep tweaking.

More dabbling needed

Steve

[kinetic (Steve)]

All set to go on the mono camera, hopefully tonight.
Luckily the Colour II and the Mono II cameras have
an identical casting for the back. A swap was done
in about 10 mins. Both warranties I think would now
be well and truly void

Dessicant has been placed into a small circular cavity
moulded into the foam shroud that encases the cold
finger. Another capsule is placed ahead of the PCB
in the void of the front casting.
If I get the thing just hovering above dew point all
should be good. If I decide to go under zero all should be
good as well. Here's hoping!

Steve

[peter_4059 (Peter)]

That looks the goods Steve. I'm looking forward to seeing how it turns out.

[kinetic (Steve)]

OK Peter, first reasonably successful night out testing.
Seeing is very good right as I post and getting better.
Edit 2:50am: clouded in!

First issue is dew!
Ambient started at 22C but the DSI chip reported 26, falling
to 24 during the 2 hours I have spent testing.

I then did a set at 16 C (DSI sensor) to stay above the dew point.
(Bureau says tonight that was 8C.)

Then I took it down as low as delta would let me and immediately
dew formed on the CCD coverglass (but not on the Baader IR in
the nosepiece). Got down to 4.5C

So a heater is needed in the nosepiece, or at least as close as
I can get one to the chip.

Attached: Darks for 24C (ambient), 16C and 4.5C (identical curves)
Lights for same. (note the blob of fog over Miaplacidus on the 4.5c )

and finally the processed results with as close a curve and final
histo as I can estimate.

Steve

[kinetic (Steve)]

Part of the Eta Car complex at 12°C cooled with
good seeing and another try for the halo around NGC 2822
close to Miaplacidus (cooled to 14°C).
Ambient was between 22-25°C.

Steve

[hikerbob (Bob)]

Steve I've located a cheaper temp/humidity sensor than the one I used on my dew heater controller ($15 US rather than $46 US). Slightly less accurate but in the same series http://www.techtoys.com.hk/Components/SHT1x/SHT1x.htm

I have one on order to try out. The one I've already used was the SHT15 http://www.sparkfun.com/commerce/pro...oducts_id=8257

Hooked up to a small micro (I use an Arduino) you can do dew point calcs. I use the dew point calcs to manage the heating cycle on the dew controller but it might also be useful to manage cooling. The code I use was posted some time ago on the PWM Dew heater control thread.

I'm in the process of doing a rebuild of my dew controller so have been looking at options to reduce the cost/board complexity and will hopefully be ready in the near future to post more on that.

Bob

[kinetic (Steve)]

Bob,

cheers for the reply..and thanks for the Ard heads up, I have read
the dew heater thread...your solution is a very ingenous way to tackle
the dew/pelt control for sure.
When I get time I will get to it.
I replied last night but in retrospect the posts always seem self indulgent
so I'm quick with the self-moderation. You would have the reply in email.
Thanks mate .

Tonight I set the controller to 10°C and the seeing was/is exceptional.
I plan to get a few objects in the bag even with near full moon.
Again a NGC 2822 field near Miaplacidus and the stars survive a quite
aggressive sharpening.

I think the cooling definitely helps with the faint stuff because the halo around 2822 is easily seen in an identical curve to all previous nights.
As Pete4059 has mentioned, 10°C seems to be a good cool point for
this DSI II. No prbs with dew tonight...dew point was 4C and the capture showed no signs of moisture with the dessicant system working.
I think this setup is a goer.

Steve

Steve

[hikerbob (Bob)]

Steve I've not seen your email. Gone missing somewhere.

Thanks again for all the info you post on your various projects, it provides a lot of inspiration for myself and others.

Bob

[kinetic (Steve)]

Originally Posted by hikerbob

Steve I've not seen your email. Gone missing somewhere.

Thanks again for all the info you post on your various projects, it provides a lot of inspiration for myself and others.

Bob

Sorry Bob, I meant unread posts re-directed to your email.
If you have that activated you can see posts, even my self
moderated ones.
Thanks for the kind words Bob. The boss doesn't quite share your
views...but that's all down to an error of judgement on my part.

Steve

[kinetic (Steve)]

A few days ago I put the whole control arrangement in a box
with some visual feedback and adjustment.
Also to save tripping over all of the wiring in the dark!
I got a 3.5 digit LED module from Jaycar and set it up to
measure 3 things via a 3 position switch on the front.
Excuse the rough Brother labeller titles on the front.

Pos 1 reads the 12V Battery that powers the Peltier.(Pic1)
Pos 2 reads the offset voltage (the temp adjustment)
(with a bit more refining this can be further op-amped/biased to read
a meaningful temp in °C)(Pic2)
Pos 3 reads the 5V regulator voltage (the LED module required 5V).

The 20 turn 1K potentiometer for fine temp control has been desoldered
and brought out to the front panel so adjustment is easier in the dark.

I would like to thank Peter_4059 for putting me on to this kit
from Jaycar. Peter, this circuit is doing exactly what I wanted it to
do. Cheers for the advice!

Steve

[peter_4059 (Peter)]

Looks very nice Steve - what size pot/combination of resistors did you end up using to replace the 20 turn trim pot?

Have you noticed the same thing I did with the histogram being shifted right? Have you needed to reduce the offset?

Also how is the dew control going?

Regards,

Peter

[kinetic (Steve)]

Originally Posted by peter_4059

Looks very nice Steve - what size pot/combination of resistors did you end up using to replace the 20 turn trim pot?

Have you noticed the same thing I did with the histogram being shifted right? Have you needed to reduce the offset?

Also how is the dew control going?

Regards,

Peter

I simply de-soldered the pot and araldited it to the front panel.
There is a clearance hole just big enough for the screwdriver to
sit and find it's way into the slot of the pot in the red-light of the dome.
I figured that I'd probably only adjust for a certain setpoint once
per night depending on what ambient was and what the dew point
was expected to be. A visual check of the offset voltage therefore
was needed too I reckon.
As I said, a direct temp readout in LED for would be achievable with
a comparator and some more thought but for now it will do.

(Edit: Pete: if I work out what adjustments/voltage divider/op amp
setup adds a direct temp (200mv scale DMM) readout I will let you
know btw)

As for the histogram shift, yes , see post #41...I did see it.
Posts 44, 45 and 47 have all been different results/setpoints and
ambients, and it seems (at least for my capture method) that LOTS
of frames and LOTS of darks, and the setpoint only varying by 2°C
max, I get a smooth result. Seeing obviously helps too...see post #47

Also, as I have just posted...10°C seems to be a sweet spot for
the DSI with ambient above 25-ish.
And no problem with dew on the night of the #47 post....10C/ DP 4C

Steve