I was imaging last night after moon set, so from 1:30am onward. At that time it was still 23C ambient and 77% humidity (typical Aussie summer night on the coast). The cold finger system on my 450D pulled the sensor down from the internal temp of 24C to 0C within 5 minutes of startup and held it there for the next two hours while running 5 minute subs, with only ten second intervals between subs. The point here is that I don't know of another current technical solution that can produce such efficient, reliable, controllable, cooling of the sensor for a reasonable investment (and I have looked at this technology for the past couple of years). Alternative solutions have been promoted and trialed over the past few years: cooler boxes, gas change of state systems, external camera body cooling, etc. IMHO these designs are all flawed in that they do not address the main criteria required for astro dslr cooling: the ability to efficiently cool the sensor to a setpoint that is consistent with optimal sensor noise performance. The main issue that cold finger implementations bring is condensation, but that is an issue for any of the alternatives as well. Dessicants, body sealing (or bagging), dry air or gas purging, all can be made to work. It is worth noting that the professional Astro CCD camera builders still rely on sealed sensor chambers and direct cold finger cooling.
Agreed, you just have to work up the nerve to open the camera up. And of course a minor workshop and a need to tinker with stuff. Plus no money to buy expensive gear ...
It all helps ...
To me it's all part of the fun and challenge, one of the things I love about this obsess.... hobby.
I'm reasonably convinced my 1100D malfunctions are moisture related. Had another go over the last couple of nights and only once I attempted to drop the EXIF temp below 12C did I start having issues - this time an intermittent vertical line pattern in the image. It's working perfectly again this morning.
It has been a couple of warm humid QLD nights, so I'm not discouraged I just need to work on my moisture sealing...
If you can lay your hands on a can of PCB spray protector from JayCar or similar, this will solve most of the issues. A squirt on the PCBs close to the cooling source.
A simple way to reduce condensation is to leave some bare metal exposed to the air outside the camera. Strangely, this works quite well. I leave the back of the Aluminium mounting bracket uncovered; that is, the bracket that holds the TEC, copper finger and heatsink in place, all clamped together. Despite the rubber sheet insulation it gets quite cold.
So rather than seal in, leave them exposed to the air.
If you have done all of this the spray can is your friend.
Good suggestion Rowland. When I get my sensor back for the mono build I will be spraying (not the sensor but everything else). Sensor will be masked off.
This brings up the subject of the sensor cover glass fogging. Even if you remove both the stock filters the sensor coverglass is still there. There is a good utube video showing how the stock coverglass reflects light compared to an MC clear coverglass. CentralDS has to remove the coverglass to debayer the sensor so I have asked them to leave it off (they wanted another $150 to replace it with a MC clear one). I will deal with that when I get it back. There is a cheaper clip-in clear I can use. But the point being I need to engineer a sealed chamber, or dessicant chamber in front of the sensor surface.
I have thought of it Glen, but it's a little beyond my needs. Interested to see how you go about achieving it.
EDIT: I found an interesting write up on powering TECs and it seems that 1kHz is preferred to avoid flow reversal reducing cooling effectiveness. However, I don't think it makes a lot of difference where the PWM duty cycle is maintained near 100%. The switch off times are very low and a reduction in effectiveness works in favour of temperature control bleeding off as the TEC attempts to drive temperature below set point - which is pretty much the set up I use. Really just widening the OFF time to back away from overshooting set point. However, in hindsight, the effect was evident while profiling the cooling curve using long pauses between pulses.
The external end of my cold finger is slightly over sized so there is exposed copper, and I've had icebergs form on it!
I haven't considered the permanent fixing yet. It's currently held in place by a slot I cut in the case, and the tight bend of the cold finger on the outside of the case results in no wiggle. I'm considering an alternative, water proof case for v2
I'm pretty stoked that I've actually been able to collect some subs now though I'll post a crop later. With an EXIF temp of 12C they decent enough for me to work with and reaching my target of combatting the warm summer nights.
Re the desiccant chamber idea I'm part way through my IR and Coldfinger mod and have thought about going back to the cold box idea but filling it with argon instead. Or Silica gel packs to dry the air.
Using a T-Ring and an Ha filter or plain glass on the front effectively gives you a sealed system and OK, have to work out a system for the finger to get outside the box but it's one way to get below zero and nice turn into a popsicle.
I am posting this here in the hope that other camera builders might have some ideas of what is creating this problem. Note I have posted this in the Computer and Software section as well (because it involves BYEOS) but I suspect it's more likely to be DSLR cooling controller related.
Here is the problem:
I am getting what appears to be electrical interference (moving horizontal flickering lines) showing up on my laptop screen while running BYEOS and on the Focusing and Framing screen. It seems to be coming from the power supply for my DSLR cooler as when I turn it off, or the relay trips out, the interference disappears. It only shows up when the temperature controller switches on the PWM to feed the power to the TEC cooling the camera. This is happening with my new Mono Cold Finger Canon 450D. Strangely this has not been a problem in the past with my OSC 450D (but I only recently started using BYEOS), and until recently when I installed a powered USB3 hub on my pier. I have tried moving cables around to prevent any power supply cable parallel runs with USB cables, but it doesn't seem to work. I also ran the camera cooler off of 12V direct from a battery and I still gets this interference showing up when the TEC is swtiched on by the controller. So it would appear to be unrelated to the AC power pack for the camera controller. This leads me to assume its coming from the camera itself, or related to the shielding in the USB cables or the Hub as they are now located on the pier instead of being separate from the scope/camera direct to the laptop.
Any ideas? It's ruining my imaging.
Please ignore the sensor fogging and artifacts in the attached shot - I am still working out the sensor chamber isolation and it needs to be cleaned.
I saw similar lines - only vertical - before I had insulated my cold finger. Once I taped that up, it seemed to go away. Appreciate this probably isn't helpful to you as you've insulated yours well previously.
Doubt that it's the USB hub or cables...if they were susceptible to any interference they wouldn't be able to transmit data very well.
Have you tried regulating your fan speed and seeing if it changes with varying voltage?
Thanks Dunk. In my setup the fan runs directly off the 12V supply produced by the controller power pack, and thus it is constant. The problem does not occur when the controller is not sending power to the TEC. As soon as the controller relay closes and sends power to the PWM the interference starts, and once the setpoint temperature is reached and the controller drops the PWM relay to shutdown the TEC, the interference stops. So I don't think it is the fan causing it. The problem seems to be TEC power related or lack of proper earthing of the cold finger (which I am leaning towards in the absence of anything else). I am thinking of running an earth wire from the controller negative to the cold finger to test, but if the cold finger is isolated from the camera earth/frame I am not sure It's will improve the situation. I need to check the camera frame to cold finger continuity. If, in putting the cold finger on the carbon fibre spacer, I have isolated it from the camera electrically perhaps it is acting as an antenna and inducting a signal into sensor image - if that's even possible.
As I am an RF engineer, I thought I would add my two cents.
I assume you are using well shielded cables for both USB AND power? And also large diameter cables for power in particular?
PWM controllers operate in the KHz band mostly. One thing, that could now be different when you moved things around, is the cable lengths used and the resistance in the power cables to the current surges that are seen each time the TEC is triggered by the PWM pulse train. Long cables could now be antennas, or simply transmitting the current switching back through the controller, into the laptop.
It could be a simple as grounding the cold finger, but if you didn't have the problem before you moved things around, and when you moved the USB hub and cables, if you changed any of the cables, I would concentrate on those areas first.
Last edited by sharptrack2; 14-12-2015 at 02:44 PM.
Reason: Clarification - meant to say TEC instead of fan.
Thanks Kevin, I think I might have found the problem.
I just did a bench test of the rig, without the powered Hub in the circuit. Sure enough the interference was there when the power was being applied to the TEC. I tried unplugging elements one at a time and it's clear that the TEC + is where the noise/interference is coming from. I ran an earth wire from the Camera hotshoe ground to the cold finger and Bingo the interference disappeared and stayed that way throughout the test down to 0C. So that appears to be the problem, lack of a good earth to the cold finger (one that is tied to the camera's earth). Rowland has mentioned this in the past and while I thought I had it connected obviously I have isolated the cold finger when I added the carbon fibre spacer. In my colour camera the cold finger touches the camera frame and the metal shield on the temp sensor is in between them as well so that's why it was never an issue with that camera.
So now I have to think about opening up the camera to add a wire between the cold finger and the camera frame or maybe just continue to use a wire to the hotshoe earth until I have to open the camera again for some reason. Photo of the test rig setup attached.
Thanks for your help.
My cold finger colour camera seems to be struggling to achieve the delta T performance that it was capable of last year. I had been running it with the Set Point at 2C, and this worked well until recently (it has been summer but the drop required from ambient was well within the original capability).
After I accidently reversed polarity on the TEC, and heated the interior of the camera, I swapped out the original TEC for a new Jaycar unit. The new unit, the same (51W 6A), seems to be unable to get down to that setpoint of 2C and gets stuck at about 6C. It seems to lack power. I did a test today, side by side with my new Mono 450D, same cold finger and fan/heatsink, and running the same Jaycar TEC (51W 6A); the Mono camera achieved a drop of -27C from ambient easily. The colour camera could only reach a drop of -14.2 C from ambient. It's not the fan/heatsink, and can't be the control box as it was used for both tests, thus its not the PWM.
My original testing on this colour camera was achieving the same temperature performance as the new mono camera. All connections seem ok. I suspect the new TEC that I put in the colour camera is defective, or a lower rated product labled as a 51W 6A. I have one spare TEC so I will swap it out when I can.
Any other ideas?
Good to know. My second buy was a dud, no action at all. Seller replaced it fortunately. I use it as a 'dummy' while mocking projects up now, kind of handy to have.
This thread is a great resource for people who want to attempt a cold finger mod.
My contribution may not be so useful as it is a digic II model without live view, but as the 400D's are becoming throw away items I considered it to be a good experimental platform. I currently use a 1100D in a cold box with gives me very consistent 15-16 EXIF temps, but happy to add a DSUSB and spend more time on focus and framing to experiment with the finger cooled 400.
Having got stuck into it, I have also found this model to be very accepting of a large chunk of copper inserted into it. The large CF slot makes a very handy pathway onto the back of the chip requiring no mods to the camera case, and having removed the pin holder, the remaining CF card infrastructure makes a nice framework to support the finger. It will be full spectrum with no filters or clears so in turn has only received a 4 resistor (4 x 22ohm) array packed into the old filter housing. As with my cold box setup, some hot-side air will be directed onto the field flattener body to take care of the glassware and filters leading into the camera body. I will try out how effective self amalgamating tape is as an insulator on the finger leading out of the body, and also my old friend 'shoe goo' to take care of the insulating of the area around the back of the sensor.
Should be fun
Last edited by KWR; 14-10-2016 at 12:25 PM.
Reason: Spelling
Cheers George, nice to hear some other intrepid DIY Astro Nutter is having a go. It's a great learning curve regardless of the usefulness of the camera afterwards, gives you an appreciation of what is necessary sometimes. I have yet to really use my 450D in anger as I have done a major rebuild of the pier and mount but I promise myself to get back into it when the clouds and time cooperate a bit.
I just need to work on my moisture sealing...
