Makings of a cooling module for DSLR

[mmalik]

I have developed this cooler that mounts to almost any DSLR [...that has a tripod socket hole]; will share more details soon. I think astro imaging can greatly benefit from such a cooling device in reducing thermal noise in the images. Regards

[mmalik]

Not all DSLRs have flip open LCDs, plus securely mounting to the LCD recess without intervention is almost impossible. Base provides easy non-invasive mounting and heat conduction paths for most DSLRs. Regards


Note: Following are a7S heat conduction points for the sensor and logical heat conduction paths for the base; point with red arrow specifically comes in contact with the base with a heat conducting pad installed by default.

[janoskiss]

Looking at it naively, I'd worry about temperature gradient across the sensor with a setup like that. :shrug:

[ZeroID]

Hi Mike,
How do you control & measure the effect of your cooler on the camera ?
Myself and a few others on here having gone down extensive experimental paths with cooling DSLRs and have found any external cooling systems are minimally effective at best and the only consistent solution is the cold finger behind the sensor with TEC and temp management.
I have a modified Canon 450D which gets a 20* delta on ambient as measured at the sensor ( BYE reports the CPU board temp ).
Our experiments with cold boxes and thermal transfer points certainly took internal temps down 5-10* but as soon as power was applied the temp rose faster than the system could remove and near ambient temps were the best acheived.
To be useful you need to reliably and consistently achieve deltas of 15-20* below ambient at the sensor. The general rule being that every 5* drop halves the noise at the sensor.

EDIT: Yes, I was going to mention gradient as well

[glend]

Be very careful of what this guy is promoting. He has a very long thread (2000 posts) on Cloudy Nights on the ongoing development of an external cooler attached to a DSLR base (a highly inefficient and as yet entirely unproved design). Even the old style cooler boxes are probably more efficient and they are long out of production.

As most well researched DSLR modifiers know, only cold fingers directly cooling the sensor produce the results needed to null noise. The arguement that if you buy a low noise modern DSLR camera (which make cold finger placement very difficult or impossible (but no one has tried it)), then cooling only the base and relying on plastic and frame conduction to get some temperature reduction at the sensor - is frankly wishful thinking.

His designs are entirely unproven on any camera, he has no way of actually measuring sensor temperature (relying on EXIF processor data), and he has avoided any field testing and Dark vs Bias/offset file comparison, which would validate or at least indicate what if any noise reduction was being created. Be aware that he has a commercial interest in selling kits or parts, and camera mods.

It is disappointing that he has now popped up here.

I have reported this thread.

[mmalik]

I don't think temp. gradient is an issue; I have done some testing and will be doing more. Will share some test results soon. Regards

[mmalik]

It is an Arduino based PWM (Pulse-width Modulation) system that uses analog pot (potentiometer) and reports real-time coldplate temp. Measuring effects on the camera of course is EXIF based. Regards

[mmalik]

If there was any goal in developing such a system, it was one... bolt-on cooling of the base of any camera that has a tripod socket hole.


Not all cameras have flip open LCDs; ones with flip LCD don't have secure means of mounting a cooler. In short, cooling of the back was definitely not a consideration.


Idea was to NOT have to open the camera; cold finger of course was NOT a consideration either.


Camera bottom cooling results are 'not' going to be uniform for all cameras; cameras like Sony alpha that have conducting structures inside may benefit the most; most Canons' internal structures at the base are not so conducive, but decent results can still be obtained regardless with adequate insulating post module installation. Note: Module does have coldplate side and TEC side insulation; it is the exposed camera side coldplate insulation that I am referring to. Regards

[ZeroID]

EXIF temperature results are measured at the processor and have no relevance to sensor temperature. Long exposure times will heat the sensor and neither be measured accurately nor controlled. Also the gradient on the temp through the conduction from the base WILL have an effect, there will be a noise gradient across the sensor. Not a cooling solution at all.
The only place I have heard of this working at all is on a high end metal bodied camera and even then it was not able to be accurately managed as the temp was still measured at the processor.

EDIT: All cameras have a tripod mount hole .. even my wee Fujifilm.

EDIT Again: And showing a complex piece of electronic equipment showing the number 6 and not connected to anything visible proves absolutely nothing. How does the user control and measure the unit and camera at the scope ? I don't think I'd want to be powering that piece of gear out in the field.

[mmalik]

I have a different philosophical perspective... there is something that always gets missed in such technical discussions is that a cooling device in itself that is designed and operates optimally comes first and foremost.


Cooling modules abound, big bulky heavy ones, cold finger ones, boxy ones, etc. This was an effort to design a compact yet efficient cooler that could be effortlessly mounted to the base of any camera. Lot of thought went into the type of power connectivity; RJ45 was selected for easy connection/disconnection after module gets mounted to a camera (so one doesn't have to contend with cables while installing).


Point being, cooling device in itself comes first. Once you have such a device [...which we do now], then comes the camera, dark frames, EXIF temps., etc. [which I did test, will share results soon, and will do more of...]


Idea here wasn't to develop the biggest/bad-est cooler that could freeze the heck out of a DSLR, rather develop a smaller efficient cooler for moderate cooling of a DSLR.


Lastly, when it comes to DSLRs, there are variables in terms of camera design. Some camera designs will conduct heat better than the others. Does that negate the efficiencies achieved in cooling design itself... no. Regards



Note: Efficiencies achieved are numerous... TEC type/design [TEC is USA made], power used (volt/amps), [relatively lower powered design to produce less heat and more cooling], how TEC gets driven (linear vs. pulse modulated), fan used to dampen vibrations, fan power/cfm, heatsink used for heat dissipation, coldplate size/type, coldplate insulation materials/specs, power connectivity (type/ease of use), AC vs. Powertank flexibility, etc. I'll elaborate on in coming posts.

[mmalik]

I have done extensive testing of the TEC both with linear power and pulse modulated (PWM) power; I can conclusively say that PWM drives and produces cooling to the max. Oscilloscope traces of UNO PWM. Regards

[glend]

At the risk of giving this thread more oxygen; PWMs have been used to control TECs for years now, and do a good job when coupled with the right camera, simple control circuits and cold finger design. I would suggest that those interested in workable DSLR cooling read rcheshire's excellent thread on this subject in the DIY forum here. Rowland conducted sufficient field testing and Dark to Bias sub comparisons to validate his design , which was also peer reviewed by other builders, myself included.

Effective DSLR cooling is all about cooling placement, and there is no substitute for direct sensor cooling, although there have been many attempts by well respected DSLR modder's like Gary Honis, and others. I would remind readers of the hype surrounding the various 'cooler box' designs of some years ago, which were even sold as commercial products such as the Orion version; these are all gone now as users soon found that external cooling was not giving them the performance expected and fell far short of successful cold finger designs.

It would be useful if mmalik were to spell out exactly what his objectives are for this thread - which seems to be an offshoot of his Cloudy Nights thread. Is he promoting a commercial interest in the supply of componentry and kits?

[ZeroID]

Still no results to see for it. Yet to see a working system and quantified photograhic results from it.
He is selling a beer fridge .... probably expensive.

[mmalik]

Following test results are from early on testing on a Sony a7S during the development process; more to be presented in due time. Regards



Baseline 3min dark taken at 76F ambient with cooling module OFF; baseline dark RAW available here... (https://onedrive.live.com/?id=6A83BA2A53E2A28B%21818&cid=6A83BA2A53E2A28B&group=0)
3min dark taken after 30min of cooling module start; dark RAW available here... (https://onedrive.live.com/?id=6A83BA2A53E2A28B%21820&cid=6A83BA2A53E2A28B&group=0)


Note: Small (identical) sections of corresponding in-camera JPGs shown, zoomed-in at 150%

[mmalik]

Following test results are from early on testing on a Canon 7DII during the development process; more to be presented in due time. Regards


TEST METHOD:


A lamp was used to gradually warm up the camera (for ~2hours)
A baseline reading was taken at this time with cooling module off
Cooling module was started and a test reading was taken (after ~2 hours)
Heat source was present throughout the test
3 meter RJ45/Ethernet cable was used for cooling module/power/PWM interconnectivity
Baseline RAW... (https://onedrive.live.com/?id=6A83BA2A53E2A28B!823&cid=6A83BA2A53E2A28B)
Test RAW... (https://onedrive.live.com/?id=6A83BA2A53E2A28B!825&cid=6A83BA2A53E2A28B)

Note: Small (identical) sections of corresponding in-camera JPGs shown, zoomed-in at 400%

[mmalik]

While no one is arguing the end result of reduced noise in the images, I am at the last stretch of enhancing the inherent performance characteristics of the cooling module itself. I'll chronicle all that went into the development of the module in due time but let me jump to one of the last characteristics of the module being upgraded, and that's coldplate insulation.


Here is a test user leaving the module running indoors at room temperature with 1/16" insulation which proved to be insufficient. Upgrading insulation considerations in coming posts. Regards


Note: Images courtesy of Steve

[mmalik]

One enhancement that's being made is an upgrade of coldplate (closed cell) neoprene insulation from 1/16" to 1/8", here is why...

[mmalik]

A comparative analysis...

[mmalik]

You may think a lot is being made of (coldplate) insulation but if you knew all that went into development of the rest of the module it is worth the extra effort.


So here are my thoughts...


1. I think 1/8" now with some experimental experience and improved application (corner coverage) will do a great job on the TEC side


2. On the coldplate perimeter, I feel 1/8" will also do the job and will look nice as well; there is a possibility of considering 3/16" for the perimeter


3. The camera side of the coldplate is something users will need to improvise given the foot print of the their camera; on areas extending beyond the camera, of course folks can go as thick as they wish. This is where 1/4" can be a possibility but in limited application. For most part I think 1/8" here will do the job as well. Regards


Note: Neoprene used is closed-cell and is of really good quality with strong native adhesive that bonds very strongly to the coldplate, and was imported form EU

[mmalik]

3/16" is an odd one; not may made in that thickness specifically. I tested the strength of the native adhesive for bonding 1/8" & 1/16" together, and it looks good and very strong. So yes, 3/16" can be an option for the (coldplate) perimeter and where needed. Regards

[mmalik]

The gap (arrow) is the particular reason why TEC side can't have thicker than 1/8" insulation [...as it will start touching the heatsink otherwise]. Regards


Note: Coldplate as it appears before most of the neoprene is applied

[mmalik]

Improved neoprene insulation specs are as follows; thicker insulation and better edge coverage will prevent edge frosting seen earlier. Regards

[mmalik]

Improved RJ45 connectivity between PSU (power supply unit) and the cooling module was one of the remaining items and is going to insure clean power and signal integrity.


A lot of thought went into selecting following RJ45 cable...



Length (3 meters for optimal distance between components)
Shielding (for signal integrity, clean power/signal and interference prevention)
Gold contacts (for improved connectivity)
Color (for visibility in the dark)
Flat (to prevent tangling)
Etc.

[mmalik]

Following desiccant is what I am considering as dew-prevention and anti-fog measure for high humidity and even for low humidity environments where cooling might encroach upon dew point. These are designed for cameras and I don't see improvising or fitting to a DSLR an issue. These inserts are also re-chargeable. Reflections should not be a concern if properly shielded/placed. Regards

[mmalik]

As a result of some testing, new and improved insulation assembly in progress...


Perimeter layer one 1/8"...

[mmalik]

Perimeter layer two 1/6"...


Note: Combined perimeter insulation is now 3/16"

[mmalik]

Following is the continuation of new and improved insulation being applied...

Note: Spring was added have a firm thermistor contact with the coldplate

[mmalik]

TEC placement with zinc oxide paste

Note: Thermistor wire isolation/insulation from the TEC; arrow points to buried thermistor

[mmalik]

Ant-vibration measures... [Maglev fan is at the core of it]


Note: I'll elaborate on zero vibration guarantee in coming posts since there is lot more to it

[mmalik]

Here is the final status of the (camera-side) of the coldplate with new and improved (thicker) insulation. Regards

Note: White bolts have been buried as well...

[mmalik]

Pics of the fully assembled module with updated insulation...

[mmalik]

Insulation around the port...


Note: Additional measures (blue arrow) taken to prevent bolt icing/frosting getting into the port

[janoskiss]

@mmalik (https://www.youtube.com/watch?v=m12-vW3OquI) ;)

[mmalik]

Bolt is probably the only point on the TEC side the is prone to heavy frosting; will need to be insulated after installing...

[mmalik]

States of the coldplate (off and on...)

[mmalik]

A quick tabletop test of the updated insulation...

[mmalik]

A look at the power supply (PSU) for the system...

[mmalik]

PSU assembly.


Note: It is an Arduino Mega system for PWM

[mmalik]

PSU assembly..

[mmalik]

PSU assembly quick test with production cable lengths, insulation, etc. White cable is Cat7a shielded (SSTP) gold (3 meters).


Some camera tests soon. Other testing participants welcome!


Note: All readings are actual; camera didn't capture red displays properly which are quite clear and discrete in real!

[mmalik]

Some real-world tests are in; more details & analysis soon...


Note: Small sections of in-camera JPG (3min, ISO2000) zoomed-in at 400%; object is Bubble Nebula

[mmalik]

Sonly (a7S) noise evaluation [3min x36, ISO2000, Bubble Nebula]; RAWs here... (https://onedrive.live.com/?id=6A83BA2A53E2A28B%21891&cid=6A83BA2A53E2A28B)


Note: Baseline in my previous post is with the fan running but pot turned down all the way

[mmalik]

Notes from the field...


Operationalizing the cooler, in the filed that is, and integrating into my routine were major goals of mine. What I meant is I wanted to develop a cooler that was no additional burden than what it takes to handle the camera, setup and image otherwise. Good news is it was a pleasure to use; fan wasn't noisy at all. There were no vibrations as I have elaborated many posts back how those were mitigated. There was no internal condensation; I had descant solution in place as well which I have also elaborated in my previous posts. Having somewhat of an enclosed sensor chamber with round LPS filter attached to the far end of the camera adapter and desiccant inside the chamber is a good setup in my opinion to prevent any un-necessary condensation in high humidity environments.


Overall quite pleased with the ease of use and non-intrusive nature of the cooler; it let's you setup everything before cooling cable is attached between camera and the PSU. I have extensive inventory of parts if more folks would like to have a go at the testing.


I plan to do more testing in coming weeks as moon cycles and weather permit; I wouldn't call it testing anymore since it is now more of an ongoing use that I would like to get accustomed to so it no longer feels like something extra which it actually isn't. I keep cooler attached all the time so I don't have to worry about detaching/attaching routine and/or handling thermal conducting pads. Battery door is not affected so changing battery is easy. Cooler is not heavy so camera handling feels normal. I think you get the idea... it is the seamless use of the camera that was one of my pet peeves and I seemed to have managed it with your help.


If I have not been able to describe in words, following is what I mean...


Regards

[mmalik]

Cooling noise reduction is very significant. Here is a NO cooler scenario...


Note: Following are real values from an imaging session where no cooler was attached and ambient was same as above, with the same equipment

[mmalik]

Here is cooling and no cooling data superimposed.


Note: Shown data is from actual imaging sessions

[mmalik]

After all the experimentation throughout this thread it is quite apparent that there is a thermal run off (σK) on un-cooled cameras; what cooling does in essence is keep that in check, plus cool beyond the ambient stating point. Regards

[mmalik]

Here is the processed image from the same session...


•36x3min, ISO2000, FF, Dithered+Dk5, Guided, Filtered D1
•Sony a7S Cool*


Note: *It was not a full cool session since cooler was started after establishing a baseline with only the fan running; 100% cool sessions soon...

High Resolution... (https://onedrive.live.com/?cid=6A83BA2A53E2A28B&id=6A83BA2A53E2A28B%21928&parId=6A83BA2A53E2A28B%21245&o=OneUp)

[mmalik]

IBIS (In Body Image Stabilization) poses a unique cooling challenge since sensor is suspended in a magnetic field without any physical contacts for transferring heat. Camera I'll be profiling for 'Cold finger Inside' initiative is an IBIS based Sony a7RII


I have coined the term 'Cold finger Inside' that will become evident as you follow this discussion; what it basically means is putting a cold finger inside the camera in such a way that there is no permanent modification or alteration to the camera and can even be reversible if ever need. Plus it is a non-visible modification in itself.


Following is the starting profile of a7RII with Metabones adapter attached...

[mmalik]

A look at IBIS system (a7RII)

[mmalik]

Back panels/LCD removed... (a7RII)

[mmalik]

Panels relevant to 'Cold finger Inside',,,

[mmalik]

A sandwich view of a7RII...

[mmalik]

'Cold finger Inside' concept...


Core design idea is to connect sensor PCB window to internal base panel for heat transfer (...eventually via active cooler) and make this connection flexible enough so IBIS functionality stays intact.

[sil]

+1

and condensation promotion inside camera case

[glend]

I have yet to see a design that works as well as Rowland's cold finger. Many cameras are just plain unsuitable for a proper cold finger solution, and attempts are usually a mess with only marginal effect. If you want a properly cooled dslr then pick one of the ones that can support a cold finger attached to the rear of the sensor, with a proven history of delivering sensor temps of 0C. The cameras that are most suitable are Canon 450D, 650D, 1100 etc, basically any of the ones covered in Rowland's thread. Newer DSLRs are too tightly packed and produce more heat when used for astro work. A better investment might be an ASI1600MC-C if you want OSC, ASI1600MM-C if you want mono.

[JA]

I think there is scope for a carefully designed under camera bolt-on cooler, depending on the internal design of the camera in question. It's possible to get some idea of this by looking at the heat conduction paths to the sensor from either disassembling the camera oneself or carefully studying disassembled images/videos of one's camera on the net.

It certainly would make for a good DIY project, with far less complexity and downside risk than an invasive internally mounted cold-finger, as good a solution as that certainly potentially could be. There's no doubt that an effectively mounted internal cold finger would be better thermally, but an externally mounted unit, whilst it won't perform as well, due to thermal resistance at the various interfaces or through different materials, offers a simplicity and conversion back to normal, that is to be admired.

I can certainly see that on some metal bodied cameras (there are plenty of diecast magnesium bodied DSLRs )with good heat conduction paths to the sensor that there would be benefit. Of course the only proof is in the before and after images. As for concerns with internal condensation, one just needs to maintain surface temperatures above the dewpoint of the air inside/around the camera, which to be thorough could be somehow mapped with sensors and controlled electronically or more simply set by prior experiment/hand at some increased risk or performance disbenefit of reduced cooling.

As a DIY project, I think the concept has merit, on a camera specific basis and could easily be housed in or mounted via a camera battery pack grip. As a commercial product ..... it would have some limitations, but I'm sure people would buy it if one could show a photographic benefit in before/after images.

Best
JA

[mmalik]

Thanks everyone; sorry for the delay. Here is what 'Cold finger Inside' looks like. It is a high quality, flexible copper tape (with heat conducting adhesive). Yellow tape is just a wrapping that serves two purposes; it thermally and electrically insulates the copper tape with the excretion of both ends where copper tape comes in direct contact with the sensor and the base panel. Regards

[mmalik]

Internal base panel connection of the 'Cold plate Inside'; area in direct contact with the plate is also marked. Regards

[mmalik]

Before and after view of the sensor connection of the 'Cold finger Inside'. Copper tape comes in direct contact with the sensor in the PCB window; back side is covered first with yellow tape and then with a7RII's native black tape material for insulation purposes. Regards

[mmalik]

Area highlighted that comes in direct contact with the sensor...

Note: There are additional measures in place (not shown) to thermally/phsically isolate copper tape from native ribbon cable.

[mmalik]

This is just a template that was created to cut 'Cold plate Inside' to spec; free hanging loop of cooper tape is about 2" long. Regards

[mmalik]

As a final step I added more insulation (arrow) at the sensor contact point and applied some di-electric grease for added protection from moisture. Regards


Note: Focus for di-electric were the ribbon cable connections mainly and some area around the center contact point

[mmalik]

This completes 'Cold finger Inside' installation.


Additional thermal separation of IBIS cover plate from the cold finger (arrows); cold finder in essence passes/hangs free except for two contact points, one to the sensor and one to the base panel. Regards

[mmalik]

Closure sequence and cold finger relationship to various layers as it gets buried...

[mmalik]

Sandwich view of the installed cold finger...


'Cold finger Inside' terminates at the internal base panel; external base panel 'not' yet attached. Regards

[mmalik]

All is well with cold finger in place...

[mmalik]

Relationship of the two base panels with each other and the cold finder...

[mmalik]

Zinc oxide thermal paste for mating the two base panels


Note: Both panels approximate 'very' well by default; this will improve conductance a great deal

[mmalik]

Final status after 'Cold finger Inside' mod...


Note: Same copper tape was used to line the external base panel (as shown) for active cooler attachment and heat transfer; plus it protects camera from any marring from the active cooler


Metabones adapter is there for a reason; it will provide alternate tripod mounting since original tripod mounting will get occupied by the cooler

[mmalik]

Coldplate module installed on a7RII (...will fit the same way to any DSLR/MILC). Regards

[mmalik]

Doesn't block battery door...


Note: Special attention paid to Metabones clearance while building the module

[mmalik]

LCD and ports are fully accessible

[mmalik]

MILC/DSLR use with coldplate module attached is quite utilitarian. Regards


Note: a7S shown; Cat7 can be connected once all setup to power the system

[mmalik]

System can be powered with universal AC or 12V DC via PSU (power supply unit) or a PowerTank/Car battery, respectively. Regards


Note: Real-time coldplate temp. sensing displayed on PSU; PWM (Pulse Width Modulation) and POT control of cooling is built-in

[mmalik]

a7RII Test Results...


30x3min, ISO2000, Darks
Cooler was attached and started at the start of the test
Test was conducted at room temperature
Initial coldplate temp reported by LED was about 72F
Coldplate temp. dropped to around 33F within first half hour
Coldplate temp at the end of the 1.5 hour long test was around 29F

[mmalik]

a7S Test Results...


30x3min, ISO2000, Darks
Cooler was attached and started at the start of the test
Test was conducted at room temperature

Note: Most tests I have conducted show that it takes roughly about 1/2 hour (half hour) for convergence, i.e., noise in images dropping and stabilizing for the rest of the session...

[mmalik]

Something I noticed, size of in-camera JPGs decreased as sensor got colder...


Note: File size dropped to less than half; a7RII files shown


This is an indirect testament to noise reduction; of course I have shown more scientific RAW noise evaluation before (a7S)

[mmalik]

I have been using base mounted cooler (with a7S) for a long time now; 'Cold finger Inside' was the next logical step which has been a great success with a7RII. a7RII was a unique challenge since it is an IBIS system.


I'll be showcasing a7S 'Cold finger Inside' next whose results I have already shown. Stay tuned. Regards

[mmalik]

While I have already shown a7S 'Cold finger Inside' test results (below), following is how the cooling mod was actually performed.

a7S is bit different than IBIS systems; even if cooling mod were not performed, it has some cooling paths built-in. Regards

[mmalik]

I added two new 'Cold finger Inside' paths in addition to the default (below).

Here is the schematic of the primary 'Cold finger Inside' path for a7S; I chose this location since it is where 'Cooler @ the base' TEC is located.

Note: I am coining the term 'Cooler @ the Base' for the active cooler to differentiate it from the 'Cold finger Inside' where 'Cold finger Inside' becomes an extension of the active cooler.

a7S is one of those cameras where 'Cooler @ the base' may be sufficient even if 'Cold finger Inside' were not installed given its compact body and well established default paths. Regards

[mmalik]

a7S 'Cold finger Inside' area of interest...

[gregbradley]

Do you have a link to threads by Rowland?
Who is he?

Greg.

[mmalik]

I have a different point of view; moderate cooling, 'not' deep cooling is the answer. Rowland's own words below, and I fully agree...

This project is all about moderate cooling with significant noise reduction, consistency, and preventing internal condensation without aggressive de-condensation measures. Regards

[glend]

Greg, Rowland is rcheshire here on IIS.

[mmalik]

Prepping the 'Cold finger Inside' for a7S...

[mmalik]

'Cold finger Inside' in place...

Note: Sensor end loops over and doubles up (end only); this is mainly to get the adhesive toward the sensor windows given the tape orientation; yellow tape covering is for thermal and electrical insulation

[mmalik]

A before look at a7S sensor...

Note: While a7RII, being IBIS, was bit more involved in terms of allowing for sensor motion, a7S sensor is quite similar otherwise

[mmalik]

Sensor bracket in place...


Note: Dip in the bracket keeps tape's looped end securely in contact with the sensor (window) by exerting positive pressure

[mmalik]

Added a secondary path since it felt logical and convenient given the footprint of the 'Cooler @ the Base'; this is to partially cool sensor bracket itself. TEC being toward the cold finger, makes 'Cold finger Inside' the primary path. Note: Yellow tapes are just for insulation

[mmalik]

With 'Cold finger Inside' (a7S), putting it all together...

[glend]

Some important considerations for anyone reading this thread:

1. With DSLR cooling, the only figure that means anything is the actual measured sensor temperature, not the cold plate, not the processor, not internal case. RAW file header temperture data is actually processor temperature.
2. Newer model DSLRs produce more heat than the common models with a proven history of modification. This is due to increased processor workload, in some cases mulitple processors, closer board stacking and battery heat.
3. Consider which DSLRs lend themselves to cooling mods. Rcheshire's excellent thread on camera cooling, here on IIS, is the bible on the process and fully tested, and produces consistent results with sensors at or below 0C.
4. Rcheshire's cooling solution development had no commerical goal and could be built by anyone capabkevof following his well documented instructions.
5. Mr Malik runs a business and is advertising an unproven product solution here on IIS.
6. Mr Malik has a history of this sort of thing on Cloudy Nights, his externally attached cooler is a good example.

If your looking for a properly cooled recent model DSLR, then CentralDS of South Korea has proven products. If you want a very good diy cooled DSLR then read rcheshire's thread. If you want a cooling optimised colour cmos camera solution, the ZWO ASI1600MC-C is more cost effective than almost new model DSLR, and has a proven track record - and can run its sensor down to a delta T of -45C.

[mmalik]

This is no competition with CCD or race to the bottom. DSLRs/MILCs don't need to be cooled that deep to reduce noise; a different animal altogether. Moderate cooling a modern DSLR/MILC will reap the most benefit without the overhead of condensation (and means to mitigate it) for most environments. [There is an option of chamber desiccant for extremely humid conditions; more in coming posts]

Experts (like RCheshire) agree there is NO need for very low temperatures to reduce most noise in DSLRs/MILCs. The concept is simple yet hard to grasp in a (cooling) world dominated by CCD. Regards

[mmalik]

'Cold finger Inside' (a7S) areas at the base...

Note: There is a seam along the base that separates plastic vs. metallic frames; 'Cold finger Inside' areas reside on the metallic side of the frame

[mmalik]

Prep for 'Cooler @ the Base' (a7S)

IMPORTANT: Arrow points to a raised point that needs proper alignment/negotiation with hole in the coldplate for a flush contact

[mmalik]

'Cooler @ the Base' (a7S) and test setup...

[luka]

Mike, while I admire your attempt to make a "moderate cooling system" for DSLR, I am starting to wonder where this project is going. It has definitely lost it's simplicity and it has developed from a simple cooling to the camera mounting to something that requires camera opening and modification. It seems to be slowly approaching a full cold-finger mod. You are even mentioning a dew control now (more about that below).

Also, you keep quoting Rowland (rcheshire) but there is a very important sentence that gets quoted and gets skipped over:
Firstly, your cooling is applied to the "wrong place". Your sensor temperature will also depend on the temperature of the camera body and hence on the ambient. Also temperature consistency can only be achieved if a cooling system has a temperature sensor at the imaging sensor (exif data is useless as we know). While your design is simpler than a full cold-finger mod, it lacks all these aspects. Also, the cooling system needs to have the extra cooling capacity to deal with temperature variations that happen during nights and between different nights (30C hot Australian nights come to mind...).

Regarding dew control, your system seems to be more complex than the fully-cold-finger modded DLSR. Instead of cooling the sensor, you will be also cooling the camera electronics and actually the whole camera. Condensation on the electronics can do serious damage unlike the simple fogging up of the sensor. Rowland used a simple resistive heating of the front glass of the sensor to solve the dew issues but you will have to address dew control of the whole camera body. You should have a look at Glen's great idea of packing the whole camera body inside a plastic bag containing silica gel and which is filled with argon gas. But as I already said, this is getting closer to a full cold-finger mod.

Also, I can see that in one of your tests you used ISO2000. You should probably stick to full-step ISO numbers (200, 400, 800, 1600, 3200 etc) as lots of DSLRs still take exposures at full-step ISOs and then use multiplication to get the ISO values in between.

[glend]

All good points Luka, but i wonder about giving this thread any oxygen at all. His Cloudy Nights thread covering the same ground, was started back in 2014, and he is up to 2757 posts now, and gradually people there have given up on trying to engage in any rational discussion, and walked away. He doesn't listen and the goal seems to be to trap unaware parties into buying from him.
This is simply an "infomercial" thread, much like those annoying TV channels that sream ads 24×7. He obviously has email alerts programmed so any response give him a chance to come in quickly over the top of it and post another photo with a few words, to push down any unwanted comments or concerns, out of sight of the next casual viewer lured into the trap.

All we can do is keep reporting this thread and hopefully IIS admin will do something about killing it.

[mmalik]

Thanks Luka; temperature consistency is at the heart of this effort. I'll show some lab and filed data in coming posts how consistent the temp. remains after initial convergence (i.e., time it takes for sensor to thermally stabilize in terms of noise reduction) which typically has been about 1/2 hour. Regards

[luka]

Glen, as this thread seems to be staying here the best may be to keep posting every now and then. Otherwise your comments will be buried and the newcomers will never see them.

Mike, can you please explain how the "temperature consistency is at the heart of this effort" as your temperature sensor is not "at the heart of the camera", i.e. the sensor. As far as I can see your solution can only keep the cold finger at the constant temperature but not the sensor. The ambient temperature may drop by 10 degrees C overnight and your sensor will do similarly as the camera body cools down although your cold finger may be at the constant temperature. And let's not talk about difference in between nights or summer/winter.

Furthermore, I am guessing that the lab data you want to show is similar to what you already have shown a few pages back. While it shows that the sensor temperature goes down and reaches equilibrium (as it has to), it is not what I mean as it shows nothing about longer term variations.

Also I am curious, have you ever tried measuring the cold finger temperature (or sensor noise) while camera exposing over, say 30min, but without the cooling.

[mmalik]

Luka, I have done some retrospective study of my RAW data going back to pre-cooler days and found that to be NOT the case. It may seem logical to think that as night temp. drops the sensor temp./noise would drop with it, but RAW data speaks otherwise. Yes camera body temp. may apparently drop, noise at the sensor behaves quite differently. And there is a reason for it... will explain as I present some field data in coming posts while folks ponder.

And I am not talking EXIF here [...plus some cameras like Sony alphas don't even log EXIF temp.]; I am talking actual noise evaluation of RAWs (as night goes by) and getting the real picture what's doing on at the sensor level in an un-cooled scenario.

Once I have elaborated that, it will be easier to explain what role cooling plays in terms of consistency. Regards

[mmalik]

Here is how noise in RAWs crescendo-es overnight in an actual imaging session in an UN-Cooled scenario...

[mmalik]

Here is how noise in RAWs stabilizes in an actual imaging session in a 'Cooler @ the Base' scenario...

Note: Comparison with un-cooled sessions shown

[mmalik]

Following is a7S 'Cooler @ the Base' sample for the data presented...


36x3min, ISO2000

[mmalik]

Luka, a7S is not a typical camera, and old rules don't apply. Here is why ISO2000 is recommended for a7S. There is ton of data out there to support this. Regards

[mmalik]

Here is how subs look at stages pointed out in the graph...

Note: Click the image below to look closer

[rcheshire]

Mike. Not sure how I am categorised an expert. All of the material in my publication is available to anyone with an inquisitive nature. It just so happens I decided to put it in one place.

It makes sense that consistent temperature across a data set is beneficial. CCDs demonstrate that clearly. My point is, that very low temperatures, while reducing dark current significantly, can be problematic technically, in DIY systems. Go for consistency and take darks. Temperature regulation makes it easier to compile libraries for long term use.

The biggest issue is the camera getting wet. Sort out the protection requirements and then work out the cooling system details. As yet no one has come up with a really efficient cooler box. But Glen's argon bag is a clue.

[mmalik]

Thanks Rowland; this is exactly the mastery that goes in 'Cold finger Inside' mods. Each camera is unique so I design internal insulation (electrical & moisture) accordingly.

To give you one example, look at the base of a7RII. First off it is all metal, no circuits. Second, I put in a thin layer of neoprene rubber insulation at this level to prevent any wetting.

Note: Neoprene goes in between the two layers shown; each image being a layer

This pretty much thermally isolates the whole bottom to prevent any inside wetting.

In short, each camera would have similar measures in place to prevent internal condensation. Regards

[mmalik]

Here is the neoprene at the base of a7RII which serves as thermal, as well as moisture insulation. Regards

[mmalik]

While one has camera open for 'Cold finger Inside' mod, it can be an opportune time to perform an astro mod as well.


Performing a 2-in-1 cooling and astro modification on a7RII next...

[mmalik]

a7RII LPF stack...

[mmalik]

Once LPF stack is removed, a full-spectrum look of a7RII...

[mmalik]

Houston, we have a problem...


IMPORTANT: A post Full-Spectrum a7RII dark shows severe light pollution from an internal light source. After much investigation, this has now been "confirmed" to be a 'known' problem with a7RII. There is a light source in/around the shutter mechanism that gets exposed after a full-spectrum mod. Stay turned for its remediation!

[mmalik]

a7R and a7RII are completely different animals; people have done a7R full-spectrum mods without such issues and the reason is that a7R had no LPFs; no LPFs meant there weren't IR sources inside to begin with, no IR sources meant full-spectrum had no issues. Sony had their head attached while they designed a7R.


a7RII 'has' LPFs as I show above; removing them in a full-spectrum mod exposes the internal light pollution which otherwise would have been blocked by the LPFs. This is a Sony design flaw [I am sure intentional] but I would call it a blunder; but we have some choices which I'll get into next. Regards

[mmalik]

This kind of leak (like the one in a7RII) is not amenable to any kind of external filtering since light source lies within or close to the shutter; so any blocking of such an IR source has to be on the sensor itself. So here is the solution I came up with...

[mmalik]

How Baader fits in a7RII's native bracket...


Note: Adequate insulation is needed to block light getting in from the edges of the Baader and the sides of the bracket; a job only a professional should attempt!

[mmalik]

a7RII dark with Baader installed...

[mmalik]

An actual image taken with a7RII Baader modified above...


Note: Image courtesy of a professional astro-photographer (DL); taken after 40min of astronomical darkness

[rcheshire]

Mike. I see what you are doing with the base of the camera. How have you addressed internal components. Anything within proximity of the conductor strip? Exclusion rather than insulation is going to protect in humid conditions.

[glend]

Haha, yeah, a professional like you, eh? Free advertising through forums. There are hundreds of filter mods that have been conducted by camera owners, its not hard to do and the step by step instructions for some cameras are available online for free. Folks in Australia, who may not wish to try it themselves, can usually get a camera repair shop to do the work for them at reasonable cost. I seem to recall a Sydney camera store was charging $150 for a full spectrum mod.
Don't get sucked into this guys web, if your considering DSLR mods ask about it on the IIS forum first.

[mmalik]

Rowland, you are correct and I have accounted for the exclusion. Not sure if you paid attention how I have excluded besides insulation in my older posts. Will see if I can elaborate the exclusion concept further but it is there and accounted for. Regards

[mmalik]

Can some of you mates help get this guy blocked for good? He has to be in violation of the rules. Regards

[mmalik]

Additional thermal separation of IBIS cover plate from the cold finger; cold finder in essence hangs free except for two contact points, one on the sensor and one on the base panel.


Even the loop of copper tape and the loop of (one of) the ribbon cables are separated by neoprene thermal insulation between the two.


I may have covered this earlier; here again since Rowland brought it up. Regards

[mmalik]

Third generation a7S, sort of, from Sony called S3CA [UMC-S3CA]


Price: ~$4,500
Shipping: Fall 2017


Note: Has no LCD on the back (shown below); may offer some cooling feasibilities?

[mmalik]

S3CA features:



12MP stills
4K/29.97P/25P video
4K video in XAVC S format at up to 100 Mbps
Up to 125min of 4K standard quality (29.97P/60 Mbps) video with optional onboard 64GB SD card
Standard ISO 100–102400, expandable to 50–409600



Note: Side by side look at S3CA and a7S