Not sure if you've come across this, thought it might give a few ideas.
He's stripped a 450D and assembled it bare bone in a custom body for sealed cooling.
I particularly like the black silica gel ring idea and the electronic shutter simulator.
its in Croation, just let your browser translate it, if you're using Chrome.
I still think a sealed enclosure is the way to go. so if you can somehow seal the area from the main pcb to the nosepiece, I think it'll make condesation control a lot easier.
have you thought of using warm air from the warm side of the peltier to keep the sensor face warm?
I believe a well known ccd name uses this technique.
this video for the qhy8 illustrates this http://m.youtube.com/watch?v=x7iEa2i4KYs
Alistair. That was the original idea. I had high hopes, but space is very restricted. Like most of my projects, it's just a matter of time until I get a bright idea. If not, then I will redesign the whole thing and basically, start again. The concept is simple. It's the mechanics that are problematic...?
EDIT: Refering back to the 1000D cooling mod, I looked at the possibility of piping warm air, but was too impatient to get things up and running and tested.
Max differential is now -18 - camera off. So I am quite pleased with the range of control and margin to accommodate sensor warming during imaging. Tweaked the voltage on the internal TEC to improve the range.
EDIT: Bright idea! Will report back. Just needed to think about it... elementary dear Watson...
I think I have found a way to pipe warm air after all.... excited!
Not sure if you've come across this, thought it might give a few ideas.
He's stripped a 450D and assembled it bare bone in a custom body for sealed cooling.
I particularly like the black silica gel ring idea and the electronic shutter simulator.
its in Croation, just let your browser translate it, if you're using Chrome.
Hi Jo. Serious cooling of a DSLR... I used a 40 x 40mm 15V 8A TEC achieve a max differential of 34C - 38, if you were prepared to wait an hour or so, but -30 with relative ease and control.
The size of the TEC is not the whole story. Heatsink rating is very important. Basically, for any size TEC, performance is predicated on the ability of the heatsink to reduce the temperature differential between the hot and cold side - that is, maximum heat removal.
The heatsink must also dissipate the energy used to drive the TEC, otherwise it becomes less efficient. It's not a simple formula and is also dependent on heatsink design. Passive, forced, brazed or skived fins, copper aluminium and so on.
Thermal resistance of the heatsink is another factor and possibly the best indication of performance. Low is good. If in doubt double the TEC wattage to get the heatsink capacity.
It's more about heatsink than TEC module - choose carefully. I find that TECs have a sweet spot depending on the heatsink rating and there is a practical limit to the size of the heatsink - more power is not necessarily more cooling - heating instead.
If this hasn't put you off, good luck with the mod.
Inside the heat resistant sleeve is 10R dew heater. Take a length of nichrome and wind it tight around the leg of a 1W resistor. Slide the wound end off as you go and wind some more. This makes a nice little spiral heater element with a good range of voltage/heating control once stretched slightly to separate the coils.
Insert into sleeve and bend at the corners - solder a couple of thin wires capable of handling a few amps. Fixed as shown with the clear gel double sided tape from Jaycar. It's a bit tight - next version will be a tighter coil.
This works really well. A bit too good, so it needs to be run at a voltage adequate for the level of cooling.
Thanks heaps for the info Rowland, I'll start shopping around.
I was planing on mounting the TEC on the inside wall of my aluminium case and having the heat sink on the out side, do you think that would work?
The idea was to use the case as a bit of a heat sink as well and I was hoping it might send a bit of heat around to the glass window on the front.
Jo. I have not built a cooler box, using cold fingers instead. TEC modules, to be efficient, like minimal temp differential between hot and cold side, when powered. Not to be confused with the sensor ambient and cooled differential. Therefore, if you use your TEC as a heater, cooling will be limited, as a result. I don't know how effective defogging will be.
However, if you locate your heatsink near the area to be heated, or connected to it, you can take advantage of the natural inefficiency of the system and/or trade off a liitle in the way of cooling performance. Rather than using the entire box heated.
My latest effort limits the degree of cooling, to a out 20, in order to avoid damage to the camera internals. To do this, I use a smaller heatsink. I am happy to achieve a working differential of 12. Reducing dark noise by 4 - half for every 6 degrees or twice the noise /6C.
Inside the heat resistant sleeve is 10R dew heater. Take a length of nichrome and wind it tight around the leg of a 1W resistor. Slide the wound end off as you go and wind some more. This makes a nice little spiral heater element with a good range of voltage/heating control once stretched slightly to separate the coils.
Insert into sleeve and bend at the corners - solder a couple of thin wires capable of handling a few amps. Fixed as shown with the clear gel double sided tape from Jaycar. It's a bit tight - next version will be a tighter coil.
This works really well. A bit too good, so it needs to be run at a voltage adequate for the level of cooling.
Thats excellent work Rowland. very impressive.
so is the heat just radiative from the nichrome wire? or does it heat up the shield and that in turn heats the contact area on the glass?
when you say its too effective, what actually happens? does it completely prevent fogging now? so even if the sensor is exposed as it would normally be while exposing, does the heating element prevent fogging?
how about the space below the glass. won't the glass radiate heat onto the sensor face as well?
Hi Alistair. Heating is conductive and convective. I have purposefully ensured that the element is loose in the sleeve. However, the fit is too tight in the camera and I will have to make a narrower coil to keep it all loose and minimize firm contact with surfaces.
The filter glass gets warm, tested out of the camera, and yes there will be a radiative effect to the sensor face and conduction around the edge of the filter glass sensor interface. The tape is a barrier to some extent and is not deforming - so I guess OK.
As with all this, I need to factor in operational warming of the sensor + filter glass temperature. Deicing capability will be the sum of these, rather than brute force heating applied directly to the filter glass.
It is quite brutal at 7V and defogs in an instant. At lower voltages there is a warm up time that I will need to observe. It evaporates fogging rapidly. I fogged the filter with heating off and timed the rate of clearing at room temperature and then incrementally increased the heater voltage, timing each phase. At room temperature the glass was very warm at 7v, but touchable.
Obviously, there is competition for cooling and if I could find a non-destructive way to pipe warm air, that would be kinder to the camera.
Jo. I have not built a cooler box, using cold fingers instead. TEC modules, to be efficient, like minimal temp differential between hot and cold side, when powered. Not to be confused with the sensor ambient and cooled differential. Therefore, if you use your TEC as a heater, cooling will be limited, as a result. I don't know how effective defogging will be.
However, if you locate your heatsink near the area to be heated, or connected to it, you can take advantage of the natural inefficiency of the system and/or trade off a liitle in the way of cooling performance. Rather than using the entire box heated.
My latest effort limits the degree of cooling, to a out 20, in order to avoid damage to the camera internals. To do this, I use a smaller heatsink. I am happy to achieve a working differential of 12. Reducing dark noise by 4 - half for every 6 degrees or twice the noise /6C.
Thanks Rowland, I drew up a quick plan to show you what I am thinking of doing.
By having the case air tight I hope to not have to much of a problem with condensation.
Hi Jo. Neat design. As you are using an airtight case can you build an evacuated chamber between the filter/cover glass and the sensor assembly. That will prevent fogging almost completely. Or just isolate the optical path and throw in a small bag of silica. I think this will be more effective than using heat from the TEC. Alternatively, duct air from the hot side of the TEC to the sensor face. Then again as the case is airtight, dry it out with silica. That should be effective, providing you can be assured of no leaks.
I guess the inside of the case will be insulated? The cold finger is a good option. Should work well.
Thanks Rowland, I drew up a quick plan to show you what I am thinking of doing.
By having the case air tight I hope to not have to much of a problem with condensation.
Hi Jo,
I'll be very interested to see how you'll get the camera working without the buttons and shutter and so on.
with that 450D mod I'd posted earlier, he added a shutter emulator. I believe certain sensors and switches need to remain connected? how would you know which ones you could do without?
if you can get it working barebone with just the pcb and the sensor, then its very easy to house it in a custom case and keep it sealed. That would be awesome.
Shutter emulator? Definitely interested. How? Thinking about dispensing with the shutter. How about a new thread?
Cable release? Intervalometer? Yeah Rowland this could be quite a long discussion and I don't want to take over your awesome thread so I'll start another one.
Using a spare 1000D sensor, I've fitted a nichrome element in the recess reserved for the AA/UV/IR filter - resistance = 7R7 as it happens.
Testing - I started by cooling without the dew heater running and predictably the sensor face fogged up badly.
Powered up the dew heater, starting at 1V, incrementing by 0.5V to 5V. Dew began to clear at 3.5 - 4V, but not completely. At 5V no dew.
The heater element needs to be aligned with the inside edge of the plastic frame, as shown to be effective. Made a small adjustment on one side to move element closer to edge, which had an immediate effect clearing that side of dew.
Second test - started cooling with dew heater 5V - no sign of dew. To be expected. At 5V the element is working overtime to clear an accumulation... etc etc
Third test - started cooling with dew heater at 4V - no sign of dew. Speaks for itself really.
Given test observations and the high thermal resistances involved because of the use of plastics, condensation prevention seems to be due to radiant heat and to a limited extent, conduction. The heater was only successful when an edge of the element had line-of-sight to the sensor face.
No doubt there is a trade off with cooling, but not equivalent to heater power. Yet to do a temperature test.
The best way to install the element is to attach it to a very thin, rigid plastic frame, in place of the AA filter. The element hangs below the frame and is removable. If the AA filter is retained stick the element to that. But you will most likely need to remove the piezo device to get things to fit. Wiring is very tight.
To fabricate the element, which is 3mm wide and 1mm deep, wind nichrome around a 2mm drill shank - tight - no gaps in the coil. Drop the coil into a bench vice - again, keep the coil tight and support in the vice with a 1mm straight edge. Tighten the vice to clamp the straight edge. When released from the vice the element will twist. This all but dissapears when stretched. Depending on requirements, stretch to say, 0.5mm or 1mm spacing, depending on sensor size and target resistance.
I have aimed to keep voltages low. It might be necessary to strip the supply leads back far enough, to clear the sensor assembly - insulate with heat shrink to keep as thin as possible.
Nichrome is difficult to solder. I run a fine grit over the ends, then wind onto the ends of the supply lead and solder- minimal.
144 turns of 0.32mm nichrome = 12R - element width = roughly circumference of drill/2 and make an allowance for wire gauge.
With this little test rig, the temp differential ran to ~20C with a 12V 6A TEC and 4.8V applied to the 7R7 sensor dew heater, which is hot to touch, but not burning. I can live with that. Cold finger temp was -2.15 ambient 17.7.
I dropped the dew heater voltage to 4V - at -1.5C the sensor started to fog up. Increased to 5V, clearing in 5 minutes, with an increase of 1.5C cold finger temp. 4.8V seems to be the spot, evaporating light introduced condensation in about 5 - 10 seconds.
Given that this was an open air test, defogging should be more efficient with the sensor in the confines of the mirror box/rear telephoto lens.
For me - Myth Busted. Let's see how it works in the field on a cold night - might need to wind the TEC voltage back a little. -5C is adequate. Watch this space.
