MarsG76

Oh ow.... here come sleepless night, lethargic days, empty bank accounts but rewarding images you can call your own.... you own images that will cost a lot.... but you'll never regret it.

Great pictures you've already imaged..... 

Currently clouds.... so 0 ?

20 hours ago, benzomobile said:

What camera do you want to chill?

A Canon 40D, the one pictured. 

On 16/03/2019 at 20:16, benzomobile said:

IMHO, I suppose your goal is very difficult to realise ...  I've done many experiments in this area and  32° C  is my the maximum deltaT° .  No more chilling I ever reached, neither with more accurate insulation, nor with bigger Peltiers too ...

Beppe

 

A 32° on the sensor would have me very joyous... even might be too much, since a hot night is around 30°C here but generally summer is around the 22-24°C and winter 9-15°C at night, so either way I'd have to use the temperature controller to keep the sensor at the 5°C for which I'm aiming. Don't want to fight frosted sensors.

9 hours ago, Gina said:

Very true.  Only if you have a lot of heat to get rid of, such as in a refrigerator, do you need larger TECs.  For cooling image sensors you only want a small TEC.

I guess that is the case if the peltier is sitting directly on the sensor... but what about when the heat extraction is by using a copper plate which constantly fights outside heating sources such as the circuitry.. I'm guessing that the colder the better... currently I have a 20 degree heat up simply to the copper passing through the camera in a 24/25 degree ambient temp... hence my aim for -20 at the peltier side.

3 hours ago, Stub Mandrel said:

Don't trust the internal temperature given in the exif data, it isn't measured at the sensor, or is this from the sensor mounted thermistor?

I don't trust the EXIF, I have a thin NTC between the sensor and the cold finger giving me an actual reading.

Hello all....

Good news is that after I filled the inside of the camera with foam strips, sealed the cold finger entrance with expanding foam and placed silica gel into the card compartment, I could image all night for quite a few nights with no condensation problems....

With the on going quest to freeze my 40D sensor while exposing subs, I constructed the same setup to measure the temperature of the peltiers with the surrent setup... the result is that it did not got below 4.2 degrees (from an ambience of around 20), and that temperature was hit in about 20 minutes... the dilemma is that with the distance between the cooling system and the sensor, the cold finger passing through the inside of the camera, seems to heat up by around 20 degrees.. resulting in the actual sensor temperature hovering at around ambient temperature while exposing (give or take a degree or two).

Even though the system currently is a noticeable improvement over when the camera was uncooled, ultimately I would like to have the sensor at around 5 degrees.

As the next stage in experimenting, I got a bigger heat sink, the CoolerMaster 212X, added a second fan to it and attached two peltier stacked, the primary TEC powered with 12V and the secondary with 5V... the temperature drop was 44 degrees.. went down to -19 degrees from and ambiance of 25... at this level, even if the cold finger would heat up by 20 degrees, the sensor would still be at 4.. so my target would be reached....

The other problem is implementation since thee 212X is a lot bigger, but weighs the same as my current setup with the two TECs and two P4 heatsinks, so it seems like the heatsink might be unbalanced unless I stabilize it.

To be continued...

19 hours ago, Stub Mandrel said:

Thickness of the plate is virtually irrelevant with copper.

I did some calculations as I was sceptical and the limts are teh thermal bond between copper and sensor and copper and peltier.

You may be losing some cooling if you have a thermistor between copper and sensor, you need them, to be a close as possible.

Neil

 

Everything is tight with heat transfer paste... the NTC should report the accurate temperature since its tight against the copper plate, even if the sensor might not be to the temperature.. but believe me, its all very tight... I'm missing something.....

23 hours ago, Thalestris24 said:

Hiya

Dunno if this will help or if you've seen it before anyway:

https://nightskyinfocus.com/diyprojects/canon-450d-dslr-modification/

He doesn't do setpoint cooling but claims to get sub-zero temperatures at 30 deg C ambient. He also manages a ~1.8mm thick plate. Not sure if he had any probs with condensation.

Louise

That is one neat and tidy project.. BUT... look at the tiny heat sink with the tiny fan he is using.. HOW does ho go from 30 deg to 0 with that??? Got me thinking.. he stated that he is powering the kit with a 4A PSU.... I did some measurements and I'm running 8Amps on the two TECs, fans, and temperatures controller, but still stopping around 10 degrees below ambient when the camera is OFF... when on it goes down from 25 to 18 as before so I thought I'll setup other TECs on a massive heat sink and fan as well as measure the temperature of my very first cooling project...

The Big heatsink has the TEC sandwitched between it and a small heat sink, the NTC in the middle grill and the whole thing sitting on a Styrofoam block..... the resulting lowest te mperature is 18.2... thats when the TEC is hooked up to a 50A 12V PSU and the heat sink is not even getting warm, so heat dissipation is definitely effective.

The second setup, the first project, draws 9 amps for the two peltiers and fans, the heat sinks get quite warm, and the result lowest temperature is... 13.1 

Both of these cooled items, small heat sink or the case of the DMK41 are WET with dew.... Whats going on? I cant get below 11.7 after about 60 minutes of cooling on the two TECs on the DMK but thats nowhere near freezing.... it reached 13.1 in 20 minutes than slowly went to 11.7... also the DMK is not powered...

The fan on the big heatsink has a speed control pot and the funny thing I noticed is that when I had it running max speed the temperature didn't drop below 18, but when I turned it to minimum speed, quite quickly, the speed dropped to 16 and stopped there... so I disconnected one of the extra fans on the DSLR, the bottom one of which extracted air from the fins and seeing if this will make any difference to the lowest temperature.

What am I missing?

20 hours ago, Thalestris24 said:

Fingers crossed! Though you should be able to do a lot better in terms of Delta T - I used to run mine at about +4 deg C which was probably about 15 deg below ambient (though it was a few years ago and I'm not certain now)

Louise

18 hours ago, Stub Mandrel said:

Hmm! I don't think the 450-D records actual sensor temperature. Imaging at -2.5C with the peltier flat out the Exif reads 2C.

I get ice on the outer part of the finger at room temperature, obviously that may not all get to the sensor even wit the insulation on, but until I sussed the heater element I suffered persistent icing up of the sensor.

Here's an image from 20 past midnight on 10 August 2017, during the period when I was getting mine sorted out.

Camera temperature in the Exif is 8C. The night-time minimum temperature that night was 6C, probably a couple of hours later.

To get that level of ice on the sensor glass (which is separated from the sensor by a decent air gap) suggests cooling well below zero. It is ice not dew as I looked at the sensor and sawice on it.

Very interesting... initially I thought that perhaps I'm not dissipating enough of the peltier heat, but it is just warm to touch so I don't think that heat bleeding on the peltiers is the problem... BUT I'm juggling now whether I should re-wire the two pletiers in series, this way they'll both run at half voltage and perhaps increase the efficiency of the cooling with less heating.... and OR my copper plate it too thin, (1.3mm thick) to effectively extract the heat from the sensor?? Which would be a problem because there is no way I can fit a thicker plate inside the camera, already I had to squeeze the unit shut, so I'd have to live with the 32 degrees plus down to ambient (which does improve slightly).

I'm using ebay 12706 peltiers from China, so perhaps they're not as good as they should be??? Get what you pay for?? 

My other option is to wait until winter and see how far the camera is cooled to when the climate is not so hot... night time 25 degrees C but feels more like 30... and so the cooling is struggling.... Either way, so far I do have an improvement so it's a start.... and good news is that the camera survived another night...

Hi All,

I modded another 40D... this time I squeezed very think foam sheets on the cold finger and into gaps where I could, I sprayed expanding foam into the gap where the copper plate goes into the camera, inserted silica gel packets into the card compartment, I added a thin NTC sensor between the sensor and the cold finger for accurate sensor temperature reading and added another fan to help dissipate the heat off the heat sinks. 

I had the camera running for 2 nights at 80% humidity and a dew point temp of around 20 degrees.. roughly the same conditions as during the nights when I lost my camera before... and this time the camera survived, so perhaps the foam did the trick, thanks for the suggestion @Stub Mandrel and @Thalestris24.... time will tell if my camera can survive the dew....

I'll tell you that with the foam inside the whole assembly is TIGHT.... very tight....

Here are my findings... when I'm exposing my subs at ISO1600, a 30 minute (15 minute also) settles at about 1 degree below ambient according to the NTC between the sensor and the cold plate.... when not exposing but the camera is on than it falls only about 8-9 degrees below ambient... when the camera is off than I get to about 10-12 degrees below ambient. the ambient temperature is 25 degrees C at midnight being summer time.... Now the EXIF temperature reported by APT coincides with the NTC reading... so the EXIF temperature on cameras looks to be accurate.

When exposing the same ISO and duration subs with the cooling turned off the sensor temperature rises to 32-34 degrees C (actually seen it go up to 38 once)... so I guess its an improvement of around 8-10 degrees...  

So my question is do these numbers look right, considering the ambient temperature and considering that I do have two peltiers running and cooling the copper plate back to back which are quite well insulated and the heatsinks are not too warm at all with all of the fans blowing and sucking air through them? I read about how some of you get ice forming and reach -5 degrees etc, is that because you're cooling your camera at a much lower ambient temperature to begin with?

3 minutes ago, Stub Mandrel said:

I wouldn't attempt to fully seal the camera - if any moisture does get in, it will be trapped.I'll explain how I sealed/ insulated my camera:

I used hotmelt on the back of the circuit board behind the sensor. At the left you can just see part of the thin layer of foam I used between the cold finger and the front of the board. There is more foam underneath the cold finger. Everything else is untouched to maximise the flow of air around the rest of the camera.

The 3D printed black object fills the gap between the back of the cold finger and the camera and the sides of the cold finger. It is hollow with foam fitted inside and against eth back of teh cold finger opposite the peltier.

This leaves one final length of exposed cold finger - that showing in the first photo. This is covered by an oversize piece of foam pushed into place as below.

I won't claim this as the 'best' solution but it has worked reliably for a year or two.

Thank you for they reply... I do have questions.. 

Adding silica gel inside wouldn't hurt?. I sprayed my main PCB with Polyurethane conform so I might not need the hot glue?

My other question is, if there is foam up against the copper plate inside and blocking the entrance into the body, won't it just get saturated with moisture that currently seems to accumulate and drip down inside the camera causing the havoc and cause the same effect just later in the night?

On 17/01/2019 at 01:17, Stub Mandrel said:

That 'analysis' fails to account for the higher voltages needed!

Taking the two comparisons he makes smaller one:

16 volts: 80 watts (cooling); 7.95 amps - power consumption 127.2W, efficiency  63%

21 volts: 86.2 watts (cooling); 6.74 amps - power consumption 141.5W, efficiency 61%

Any questions?

 

On 17/01/2019 at 00:56, Marci said:

Sorry - I worded badly... have a look at https://www.overclock.net/forum/62-peltiers-tec/121488-testing-undervolt-tec-air-cooled-3.html#post1319096 

The original post they refer to there (which originally contained much more data for a wide range of manufacturers, and the validated test rig etc) over on HardwareAsylum looks to have been lost (altho I’m 90% certain the original thread was actually on procooling.com a few years prior - it _was_ 20yrs ago... my memory gets hazy).

On 16/01/2019 at 22:16, Gina said:

TEC efficiency varies enormously with make.  But a very common mistake is to use a more powerful TEC than required - this just reduces efficiency.

On 16/01/2019 at 05:55, Marci said:

Insulating all exposed coldplate will also increase efficiency - you'll be losing a good chunk of cooling power to the air. Can improve efficiency further by finding a higher wattage/voltage etc and under powering it - look to PC TEC cooling forum archives for spreadsheets of make/model testing - getting you same wattage but for lower ampage at 12v = same cooling for less power consumption... amp-load saving can be dramatic

In this version of the project, I think I'll stick to my current peltiers, efficient or not it is cooling... once I have my current problems sorted out, perhaps for the next version of the cooling kit i'll concentrate to get most cooling for least amount of current drain.

On 15/01/2019 at 21:25, Thalestris24 said:

Just wondered if you insulate the copper? I imagine you only want it bare where it makes contact at either end?

Louise

On 15/01/2019 at 21:45, Thalestris24 said:

Here's a pic of mine. It shows the copper is insulated with a rubbers sleeve. It carries on inside and only has an opening to allow the connection to the sensor:

Louise

On 16/01/2019 at 05:39, Stub Mandrel said:

Mine's got foam over the copper and stuffed into the hole the copper goes into.

Hi all,

Bad news is that the Cooled 40D with the clicky shutter fault and err 99 did not come back to life... so most likely the shutter on it has been condensed into the shutter grave.... but I did mod another 40D and am getting it ready for another cooling attempt.. I'll look deeper into the original 40D at a later date.

I did not insulate the copper at all since there is not much room to add any insulating material to it.. so my question is @Thalestris24 what is that black rubber you used to insulate called if I wanted to go to a hardware store and get some?

Right now I'm thinking of sticking some of the very thin foam you get products wrapped in around the cold finger to stop the water dripping off it, but also silicon and expanding foam seal the entrance of the cold finger which goes inside the camera, fill the inside gaps with silica/desiccant get packets and anti fog strips ... and remove the card storage housing and fill that area with silica gel also, giving me easy access to exchange spend silica gel packets. Hopefully this will keep the inside cold and dry....

But before doing this to the camera, I'll put the cold finger into a empty box (dummy camera setup) filled out with the above description, leave it running over night, checking every few hours the inside condensation status... if dryness is achieved, than the system will go live... this part of the year is a good time to check this since humidity is around the 85% mark and the dew point is around 20 degrees so a dewuey time of the year and so if it's dry now, it'll be dry all year. 

I honestly thought that dew will be the smallest issues in this project and it's actually the biggest.

If anyone who's done this sees any issues with my above plan, please do not be silent.

MG

I think I'll only sacrifice one more camera to the cooling camera project before either just reverting back to uncooled imaging or leave the cooling in place, but only use the heatsinks and fans without peltiers to avoid moisture buildup... and end up buying a cooled astro camera and feel like a failure....

@Stub Mandrel did you have this much of a problem with condensation? Honestly I didn't anticipate condensation to be such a problem.... I thought that this would have been the least of my issues, but as it turns out its by far the biggest problem.

Update 14 Jan:

I started tonight by comparing the actual difference cooled vs uncooled sensor temperature on a 15 minute ISO1600 sub. The resting temperature was 26.4° and it rose to 33.9° after the exposure when uncooled. When cooling was on it dropped to the 19° mark and stayed at this level. The actual temperature doesn't seem so great but it does make a considerable difference in the noise level on the subs.

Well that didn't last long... everything was going well until at around 02:50 I heard some clicks coming from the camera and after powering it on and off the camera malfunctioned with Error 99. The exposed copper plate was wet with dew and the 40D was wet with dew on card side which was facing down after the meridian flip. Looking at the orientation of the 40D, it seems as if the condensation which builds up on the exposed copper plate allowed the water to run down into the camera along the copper plate, and  looking at the angle which it was at, it looks like it would have dripped droplets on or near the area where the shutter mechanism is located, so I'm suspecting the Error99 to most likely be a shutter problem.

I'm starting to have doubts that I'll  beat this dew issue and have a reliable camera but as a last ditch effort I'll try to block the enterence from the peltier side, wrap the exposed copper with some material to stop droplets from forming, insert some desiccant packets inside the body and perhaps install some thin pipes which will gently blow warmish air inside from in between the heat sink fins and have the air inside the camera constantly moving.

The sensor temperature is reported at around 21°, which was 2-3° above the dew point so hopefully dew is not building up inside and stopping water from dripping in when the peltiers are orientated above will keep the inside dry.

The reason why I think that moisture is not building up inside the camera but dripping inside is because the camera failed both times when the peltiers were orientated up, allowing water to run inside the body along the copper plate. Both times the camera was ok for the few hours before the meridian flip and when the peltiers were oriented to the bottom.

Now just to see if the 40D will start to work again after drying out for a few days or whether I'll be using a different body.

Edit: On second thoughts I might ditch the idea of blowing air into the camera, potentially adding moist air inside... I'll just seal the camera at the entrance of the copper plate, with desiccant packets inside and foam where ever I can add to cover the copper plate to stop the water from forming droplets.

6 hours ago, Gina said:

You should never put anything solid in your ear... Pardon???

Especially anything with which you play pool or snooker... lucky that cotton isn't solid... ?

20 minutes ago, Stub Mandrel said:

Egad, he's speaking foreign again!

What do you call that ear cleaning cotton on a stick thingy?

28 minutes ago, tooth_dr said:

I think this is the best thread on SGL!

These adventures might help someone to skip the mistakes and show them what to expect...

Update: Since it is cloudy tonight and looking like it'll be clear skies tomorrow, (and this was on my mind all day)... I opened up the 40D to replace the faulty sensor.

When the back was off and before pulling out the bad sensor I wanted to try and see if I can somehow create a contact between the slightly damaged ribbon cable and/or re-seat the connectors to see if that would make any difference... thanks Louise @Thalestris24

The thing that was obvious when I pulled out the sensor bottom ribbon connector is that it's covered by some thick and obvious gunk, so I cleaned it (as well as the top connecting cable) with isopropyl alcohol on a q-tip.

I put the back on and held it in place as I inserted the battery and to my ecstasy the sensor was delivering images again...

I assembled the camera and once again it's ready and waiting for the next clear night.

Talk about an astronomical roller-coaster ride...

7 hours ago, Stub Mandrel said:

How did you protect the various connectors from the conformal spray?

Personally, I would be very worried about spoiling connections with it.

7 hours ago, Thalestris24 said:

 

7 hours ago, Gina said:

I wondered that too.

I covered all connectors with tape and insulated them from the conform spray... that is not an issue since the camera worked flawlessly for 3.5 hours....

I had a look this afternoon again and the live view is still just purple vertical lines changing and flickering, but otherwise all other aspects respond with  no freezing.

Another thing I did notice is that one of the ribbon connectors on the sensor was slightly torn, damaged by the despldering of the aluminum plate that covered the connectors where the sensor is plugged into, so I'm thinking that it was just barely contacting by microns and now let go.

The other thing is that I didn't bother replacing the aluminum plate on the PCB since it seems like it's for heating purposes, unnecessary now, or am I mistaken? It's only a plate that covers the sensor connectors and some ICs under it.

When I replace the sensor, I'll spray the conform on the exposed pins on the back top/bottom of the sensor...

...and so it continues... something is telling me that this project will exhaust all of my 40D spares and I'll end up reverting to the order original uncooled DSLR...

2 hours ago, Thalestris24 said:

Hope yours continues to work 

I was celebrating too early. At about 3:30 I was presented with vertical lines instead of the actual image after a 600 second exposure than just black after the following 900 second sub.

The camera responds and triggers the shutter and works properly except that the images are always clean black frames and live view only shows vertical lines!!! Did my sensor fail now?

3 minutes ago, Thalestris24 said:

Oh I have already. I just have to be in the mood to fiddle about with it!

Louise

Think of it as getting in the mood to see the results that that unit will deliver..

36 minutes ago, Thalestris24 said:

Good news! Gives me hope for mine though I still haven't got around to looking at it!

Hope yours continues to work

Louise

I think that you should definitely look into your 550D, look on eBay for a replacement main PCB, they don't seem to be that expensive when are from salvaged stock.. I ordered a main PCB for my 40D for $60, so when it arrives, providing that its not faulty, I'll have all of my 40Ds up and running.. talk about plenty of redundancy....

Thank you and all for the ongoing help and input...

Hello all... Today I powered on the 40D with the hope that it might come back to life after sitting and drying for a week but no such luck so I pulled apart the dead 40 and replaced the main PCB with the PCB from my other 40D. It was a pity to pull apart a perfectly good camera with the risk that it might not come back but I figured that this way (if successful) the hardware/camera will get more of a use than just sitting in the bag.

Before installing the PCB I sprayed it with the polyurethane electronics protection spray to coat it and protect it from the condensation.

Another mod I done to the original design is adding the thin NTC temperature probe between the sensor and the copper plate as I wrote about doing in a previous post, result is shown in the image. This will report the actual temperature of the sensor.

When I assembled the unit I was very happy that it still worked!!! WOO HOO... so now I'm more confident that I won't have another dead camera due to condensation! ...at least not for a long time.

After the preliminary check of the function of the camera, the sensor and the USB connection, I put the cooling controller back on the camera but instead of making the same rats nest of cable ties I had before, this time I put the controller into the 40D hot-shoe, looks much neater.

I have the system setup and doing some 600 and 900 second ISO1600 Halpha sub of the Rosette Nebula to test it out right now... so far so good...