Having said that, I've attempted to power it up this morning for its first thermal test of the new cold finger and case mod and it's dead we went through a lot together. Not 100% sure I'm ready to crack open the known working 1100D so close to New Moon.
The 450D and 1100D appear to differ in that there is a chip on the sensor assembly circuit board whose pins protrude into the cold finger space, hence entry from the hand grip side (right) is simpler than from the left.
Bugga .. Thats why I bought the 1200D, to be left operational while I tinker.
The KM 7D had proruding pins as well, I trimmed them down to solder level. It's also near dead after many DIY attempts. This is an expensive hobby.
I've got my eye on another 450D advertised cheapish over here. A spare would be handy and I could always sell it on modded if I can get a process working.
I'd bought the sacrificial cam for spares so was surprised (and stoked) that it worked when I got it. It's a shame it made it this far only to fall at the final hurdle
But still...onwards and...I performed surgery on my remaining working 1100D, and here it is... it ain't pretty, but that's what prototypes are for
I did get the Peltier, heatsink and fan running on it earlier. My Peltier is the 30W version from Jaycar. At full power, 5 minute dark subs at ISO1600 were reporting a temperature of 14C in the EXIF tag. At approx. 2/3 power (20W) it was recording 17C. It's been another warm and humid day up here and the room temperature was reading 26-27C, and when I ran some subs to warm up the sensor before starting the cooler the EXIF tags were showing 32C. That gets me in the ball park of where I want to be. I can live without totally clean darks if it simplifies the dew situation, as the noise is very manageable with the EXIF temp reading around 10C.
There's some trimming to be done to the copper and some additional insulation to add, as I had an iceberg form on the very end of the copper (foreground in the third pic) after the Peltier had been running for a while. I also need to sort the power out as I'm using a DC coupler, and I need a neater USB cable solution...oh and a hole for the lens eject button
Dunk, those are darks right? Shoot some fast bias frames at your low temperature and compare them to some darks shot at your standard duration and your low temperature - once they look the same there is not much to be gained by going colder. You can't get rid of bad pixels with cold ( which is what bias frames are designed to highlight).
I've had been building up a dark library at -5C and stopped when I realised they all pretty much looked the same. Rowland has found the same thing with his.
Glen, yeah they're darks (cropped from the centre, but 100% resolution). More so than completely clean darks, I'm concerned with cooling effectiveness at a dark site with limited 12v power.
I figure if I can take 10-15C off the summer QLD night temps then it'll make the difference between being able to do some useful imaging and not.
Digressing slightly to a previous question about preferred fan airflow direction in one of the cooling posts, the answer is derived quite apart from any cooling consideration - wind velocity.
Depending on the fan type, typically the elevating type, designed to be vibration free, incident wind effect on the blades produces a very annoying harmonic which may not be audible, revealed only by touch.
The effect is eliminated by drawing air through the heatsink rather than blowing air into the heatsink. I think the latter is marginally more effective, but as I found, a significant problem in the field with any sort of breeze.
A crude analogy might be compressor blade stalling/surging in a gas turbine engine. This is controlled by the incidence of the airflow presented to the compressor by guide vanes - the airflow through the heatsink onto the fan guided by the fins is probably similar in effect.
Dunk. For our application finding a sweeter spot is just a matter of balancing supply to the TEC with the capacity of the active heatsink to remove the cooling plus supply energy.
Depending on the effectiveness of your heatsink a lower voltage/current may produce a higher differential. TECs are all about energy out extraction vs energy in.Too much in or insufficient out limits performance. A bigger TEC may confound your cooling efforts.
............ a previous question about preferred fan airflow direction in one of the cooling posts, the answer is derived quite apart from any cooling consideration - wind velocity.....
The effect is eliminated by drawing air through the heatsink rather than blowing air into the heatsink. I think the latter is marginally more effective, but as I found, a significant problem in the field with any sort of breeze.....
A crude analogy might be compressor blade stalling/surging in a gas turbine engine. This is controlled by the incidence of the airflow presented to the compressor by guide vanes - the airflow through the heatsink onto the fan guided by the fins is probably similar in effect.
Just wanted to make a few comments based on my testing and reading on this subject. First in relation to fan design, some fans are better at pushing rather than pulling air, and fan blade design is the determinant.
If your fan is pulling air through a heatsink, with many heatsink designs the path of least resistance is through the side of the heatsink vanes and thus airflow over the total vane area can be significantly reduced. With a pusher fan, with good pressure, I believe better heat extraction is possible, provided the air can flow through and is not blocked. There are a number of CPU heatsink fans on the market that push air down onto the CPU mount area and through the heatsink (the idea being that the air then spreads out across the motherboard and produces a general cooling effect). I am using this sort of architecture on my mono camera.
So how do we orient the fan to the cold finger, is there an optimal orientation? My past field experience would suggest that optimal cooling is derived from airflow which is perpendicular to the sensor cold finger. This keeps any warm air extracted from the heatsink away from the TEC and cold finger. However, on my new mono camera I am doing the opposite, with the warmed air exiting the heatsink and aimed at the TEC/cold finger (due to the folded nature of the heat pipe heatsink to achieve low profile). I believe that insulation of the cold finger and TEC area will prevent loss of cooling due to warm air passing, and I will be using a TEC/cold finger 'deflector shield' made of 2mm carbon fibre sheet, to angle the air flow away from the camera.
Differences in orientation may be hard to determine without actually trying them in operating conditions. Since most imaging does not take place during windy conditions (due to effects on telescope stability), I do not believe wind induced fan harmonics are a major risk.
Dunk. For our application finding a sweeter spot is just a matter of balancing supply to the TEC with the capacity of the active heatsink to remove the cooling plus supply energy.
Depending on the effectiveness of your heatsink a lower voltage/current may produce a higher differential. TECs are all about energy out extraction vs energy in.Too much in or insufficient out limits performance. A bigger TEC may confound your cooling efforts.
Thanks Rowland, I've been experimenting the past couple of nights and at about 1.5A (running from my 12v battery) the cooler dropped the camera temperature (as reported by the EXIF data) from 25C to 12C. In my previous experiments 12 is the borderline for acceptable noise, so it seems it can achieve my objective of countering the Queensland summer
I have yet to try one of my other heat sinks to see if it's more effective. I ran into an issue last night where it had cooled to 12C and then the shutter got stuck I power cycled the camera a couple of times and it isn't now working correctly. Puzzled by this one as it had been working perfectly up until the time I wanted to take a sub in anger hopefully I'll be able to resuscitate it with the shutter from my spare 1100D.
There is a question about where exactly the EXIF temp comes from. I prefer a temp probe on the cold finger near where It's enters the back of the sensor. Do you know where the file temp is captured?
I did attempt to fit a TMP36 but it was too big for the gap on the 1100D. No matter. Regardless of where the onboard one is, I had been taking darks and analysing the noise wrt the reported temperature. All that really matters is that I know what to dial up to get good subs. Having the actual temperature and a temperature-controlled solution will be the icing on the cake. Once I have a working camera again
I have to reassemble the camera around the temp sensor once it's fixed in place. 3 strands off an old ribbon connector makes a nice low profile lead to poke outside the camera body. Silicon roof and gutter sealer does a good job of holding it in place. Generally, placing the flat side down (before applying the silicon) with a touch of heat grease, and a small drop of super glue on one edge or tape it down to hold it hard against the copper until the silicon cures.
Here is the cooling profile of my 450D. Basically, a look up table in the Arduino code, rather than using stock standard micro controller PWM. Using a profile produces more stable, accurate set point control, with considerably less noise.
The chart illustrates diminishing returns, squeezing out that extra few degrees of cooling toward maximum differential.
Having done the same with the 1000D, the 450D is quite different and seems to require more energy.
Nominally, the profile starts with 10ms on and 50ms off.
Air temperature was 18C - I gave up at -7.5C due to the very slow decrease in cold finger temp - the system cools 26 - 27C below ambient air temp, but those last few degrees are not easily bought.
The graph shows the analog value of the temp sensor - reference voltage approx 3.3v.
System design as follows;
12V 6A (72watt) TEC
~120watt active heat sink
Sensor defogging ON
I tested the system while taking a very long exposure 20 minutes or so. The trend is an increase in cold finger temperature of ~1C over the range of the profile.
Now this might not meet everyone's specifications/expectations - here for its educational value and should give some idea of what to expect.
All other things being equal, cooling performance is dependent on the effectiveness of the active heat sink, which for me is a compromise - performance, balance and aesthetics.
Rowland there is a thread running over on CN (DSLR sub-fourm) talking about cooling modules. A recent post there suggested that simply toggling on and off a TEC via power on and off via the PWM, actually decreases the life-span of the TEC. The suggested solution was to install some electronic components, the quote from the thread is here:
"You can solve this problem electronically with an LC filter. The combination of inductor and capacitor stores energy that covers the "off" states and keeps constant power flowing to the peltier. This is recommended anyway with peltier control, as the instant on/off voltage change with PWM can shorten the life of the peltier if done continuously."
I'd be interested in your take on that comment. While the setpoint controller I am using does switch the PWM on and off, the PWM pot does allow control of the power to adjust it to reduced level to maintain almost steady state operation once you have achieved near setpoint temps. Sure this is a manual setting, and I guess it could be done electronically but I like having manual control over that. What are your views on adding LC filters to 'store' energy for off-states to smooth operation?
CN thread link is here, go to the end to find the post (#353) related to the LC.
Given all the threads I've read over there, there's a tendency to overstep the borderline called obsession
The real question is, if it shortens the life of the Peltier, by how much? Frankly, if I get one summers' use out of my $18 Peltier I won't be too upset if it needs replacing next year
I am always leary of posting image links on IIS because of the 'professional cadre' and standards expected. However, as a demonstration of cooled full spectrum images from these cameras, here are a couple, the first M42 from last night (sensor temp 0C):
we went through a lot together. Not 100% sure I'm ready to crack open the known working 1100D so close to New Moon.
I power cycled the camera a couple of times and it isn't now working correctly. Puzzled by this one as it had been working perfectly up until the time I wanted to take a sub in anger 
hopefully I'll be able to resuscitate it with the shutter from my spare 1100D.
