DSLR Active cooling MOD process - Part 2

[Stub Mandrel]

5 hours ago, Thalestris24 said:

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.

I have a very similar loop of nichrome to the one he shows.

[Thalestris24]

37 minutes ago, Stub Mandrel said:

I have a very similar loop of nichrome to the one he shows.

I think the guy who modded my 550d did something similar. He said "I don't have condensation issues in the sensor filter because I found a way to include a heater just for the optical window of the sensor that maintains it a few degrees above dew point with no impact on sensor cooling efficiency, I also vacuum seal the sensor chamber to keep it dry inside." Even so, he recommended putting the camera in a container with some silica gel after use.

Louise

[MarsG76]

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?

[MarsG76]

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.....

Edited February 4, 2019 by MarsG76

[Thalestris24]

40 minutes ago, MarsG76 said:

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?

Perhaps to do with distance of TEC from sensor? I'm not sure your setup looks right (though I'm no expert!). Have a look at Gina's:

I would test things without the camera itself being involved. Then, only when everything is cooling as expected, would I connect the cold finger to the camera/sensor. I think my 550d one works fine if powered from just a lithium battery which only supplies maybe 2A. I'll check it now  

Louise

[Thalestris24]

31 minutes ago, Thalestris24 said:

I think my 550d one works fine if powered from just a lithium battery which only supplies maybe 2A. I'll check it now  

Louise

I just did a quick test. Powered from my blue DC 1298A 12V LiPo, from an ambient of 16.8 deg it cools to about 2.7 deg in about 5 mins.

I don't think the LiPo delivers more than a couple of Amps. The TEC is supposed to cool down to 20 deg below ambient which it probably can with more current. Even from just the battery, the heatsink gets noticeably warm. I imagine it relies on the fan to take the heat away. Mine is a Sunon maglev 12V 1.44W, 60mm. The direction of flow arrow on the fan points inward so it blows ambient/cool air through the heatsink fins.

Louise

 

[benzomobile]

Hi.

Only one Peltier of 45 Watt is enough to perform >30° C  deltaT.  Sub zero temperature doesn't  make smaller thermal noise in Canon cameras:

https://sites.google.com/site/beppelastrofilo/home/canon-1100d-mod-filter-removal-direct-cooling-electrical-dew-control

Beppe

[Thalestris24]

I just watched this on youtube and thought it was quite informative without being technical. I think I'll get a tec1 12706 myself and have a play just for fun  

Louise

[Thalestris24]

12 hours ago, Thalestris24 said:

I just watched this on youtube and thought it was quite informative without being technical. I think I'll get a tec1 12706 myself and have a play just for fun  

Louise

Of course, it's Chinese New Year so there will be delays for some things... Have ordered a 12V thermostat which looks the same as the one in my 550D. Also ordered a tec1-12706 with uk despatch but I also want a pwm controller which I've not ordered yet. Once upon a time I'd have built the electronics from components on stripboard but can't be bothered now and Chinese prices take away the incentive. I'm not going to be cooling any cameras or sensors - just having a play  

[Gina]

I've read that Peltier TECs don't like being run from PWM.

[Thalestris24]

15 minutes ago, Gina said:

I've read that Peltier TECs don't like being run from PWM.

I think you're right, Gina! It seemed like a good idea at the time ha ha
https://www.meerstetter.ch/compendium/peltier-element-efficiency#DC

Edit: I'll have to get a linear controller instead

Louise

Edited February 5, 2019 by Thalestris24

[Gina]

Useful link Louise, thank you. 

In my all sky camera system I used just two preset levels of cooling with a Peltier TEC.  A low level for daytime use just to take away the thermal heating and a high level for night-time to cool the camera and reduce noise.  These are simply 5v and 12v on a 15v TEC.  Instead of PWM I simply switch one or other supply to the TEC using power MOSFETs.  These are controlled by an output data line from a Raspberry Pi.

Edited February 5, 2019 by Gina

[Thalestris24]

1 minute ago, Gina said:

In my all sky camera system I used just two preset levels of cooling with a Peltier TEC.  A low level for daytime use just to take away the thermal heating and a high level for night-time to cool the camera and reduce noise.  These are simply 5v and 12v on a 15v TEC.  Instead of PWM I simply switch one or other supply to the TEC using power MOSFETs.  These are controlled by an output data line from a Raspberry Pi.

I want to keep it simple and just use dumb electronics. It's purely educational for me as I've never built anything with a peltier before. Also, I mustn't hijack this thread!

Louise

[Stub Mandrel]

What is rarely understood is that the heatsink has to dump the heat removed from the camera PLUS the heat used to power the peltier.

A peltier is 10-15% efficient, I think in our rather peculiar setups we need to assume 10%.

In which case a 40W peltier running would provide 44W of cooling and have to dump 44 watts into the surrounding air. Think how warm a 40W bulb gets or even a 10W LED lamp, that's a surprising amount of heat.

If you don't get that all out the heatsink the peltier will warm, not cool, the cold finger.

This is why (a) fan cooled heatsinks are essential and (b) why an underpowered peltier can work better (if the heatsinking is ineffective).

 

As for PWM and Peltiers, why not try fitting fit a decent sized capacitor across the peltier?

[Thalestris24]

8 minutes ago, Stub Mandrel said:

What is rarely understood is that the heatsink has to dump the heat removed from the camera PLUS the heat used to power the peltier.

A peltier is 10-15% efficient, I think in our rather peculiar setups we need to assume 10%.

In which case a 40W peltier running would provide 44W of cooling and have to dump 44 watts into the surrounding air. Think how warm a 40W bulb gets or even a 10W LED lamp, that's a surprising amount of heat.

If you don't get that all out the heatsink the peltier will warm, not cool, the cold finger.

This is why (a) fan cooled heatsinks are essential and (b) why an underpowered peltier can work better (if the heatsinking is ineffective).

 

As for PWM and Peltiers, why not try fitting fit a decent sized capacitor across the peltier?

Well, I've ordered a variable buck converter from China - should turn up eventually! As I say, it's only to have a play, not for anything serious.

Louise

[Gina]

56 minutes ago, Stub Mandrel said:

As for PWM and Peltiers, why not try fitting fit a decent sized capacitor across the peltier?

Bang???

[Thalestris24]

7 minutes ago, Gina said:

Bang???

oops mis read your post, Gina. Yeah, I wouldn't put a capacitor across a high power pwm output.

* though I suppose you could put a few diodes in between  

Edited February 5, 2019 by Thalestris24

[Gina]

A suitable choke is what's needed to do that.  The result would be the same as you have on order ?

[Thalestris24]

2 minutes ago, Gina said:

A suitable choke is what's needed to do that.  The result would be the same as you have on order ?

Ah, ok. I'm just experiencing a visual migraine so not seeing or thinking straight  It will pass in a short while.

[Stub Mandrel]

I did think a choke might help, but to be honest a capacitor should work fine. Once it is charged (a few cycles) the current through the PWM should be no greater than without it.

[Thalestris24]

50 minutes ago, Stub Mandrel said:

I did think a choke might help, but to be honest a capacitor should work fine. Once it is charged (a few cycles) the current through the PWM should be no greater than without it.

Without an inductor wouldn't it be an effective initial short circuit? Anyway, hopefully the buck converter will do what I want.

Louise 

[Gina]

Yes!

[Stub Mandrel]

2 hours ago, Thalestris24 said:

Without an inductor wouldn't it be an effective initial short circuit? Anyway, hopefully the buck converter will do what I want.

Louise 

Only very briefly, the secret will be to match the capacitor to the load and the PWM frequency.

 

Less briefly:

40W @ 12V is about 3.6R. Power - V^2/R therefore R=V^2/P.

The  PWM frequency of the typical cheap eBay modules is 13KHz.

One cycle of power 1/13,000 = 0.00008 seconds.

To get smoothing the time constant RC should be >> than the supply period, 1 millisecond should be fine.

so if 0.001 = 3.6 * C, C= 0.001/3.6 = 0.00027 F = 270uF.

The energy in initially charging the capacitor (lets assume the power supply's internal resistance is negligible and it charges within one PWM on cycle) is 1/2*C*V^2

1/2 * 0.00027 * 12^2 = 0.039 joules, but this charging is typically at 50% efficiency so roughly enough to raise the temperature of a1/50 of a gram of water by one degree centigrade.

This might make a neat little blue spark, but it's nowhere near enough energy to fry your PWM module.

Science, don'tcha love it?

?

 

Edited February 5, 2019 by Stub Mandrel
Crap spelling

[Thalestris24]

1 hour ago, Stub Mandrel said:

Only very briefly, the secret will be to match the capacitor to the load and the PWM frequency.

 

Less briefly:

40W @ 12V is about 3.6R. Power - V^2/R therefore R=V^2/P.

The  PWM frequency of the typical cheap eBay modules is 13KHz.

One cycle of power 1/13,000 = 0.00008 seconds.

To get smoothing the time constant RC should be >> than the supply period, 1 millisecond should be fine.

so if 0.001 = 3.6 * C, C= 0.001/3.6 = 0.00027 F = 270uF.

The energy in initially charging the capacitor (lets assume the power supply's internal resistance is negligible and it charges within one PWM on cycle) is 1/2*C*V^2

1/2 * 0.00027 * 12^2 = 0.039 joules, but this charging is typically at 50% efficiency so roughly enough to raise the temperature of a1/50 of a gram of water by one degree centigrade.

This might make a neat little blue spark, but it's nowhere near enough energy to fry your PWM module.

Science, don'tcha love it?

?

 

It's academic - ha ha!

 

[benzomobile]

Hi all,

a suitable LC filter to transform PWM signal in  quasi continuous voltage (on power output) should have a frequency cut-off  of (circa) one

tenth of PWM freq, i.e., if PWM freq is 20 KHz, so the LC cut-off frequency will be 2000 Hz.

Too small filters don't perform to smooth PWM wave, while big ones do have bad effect on the feedback thermal control because they will

reduce  control readiness and  speediness, resulting in too large thermal hysteresis.

For this task, please  no high value caps across the power output.  Never ever!

I put here an example for a correct LC filter for a 20KHz PWM frequency:

L (inductance) will be a ferrite toroidal one, value of 1.4 mH

C (capacitance)  will be a 4,7 uF one.

 

Beppe

 

 

Edited February 6, 2019 by benzomobile