CCD Sensor Cool Down Times

[Dogs857]

Ahoy GetDpi friends.

I recently bought back into MFD with a Pentax 645D, which is a fantastic camera. I have been running around taking photos of this and that, testing different lenses and doing all the usual stuff you do when you get a new camera. Whilst this is not my first MFD device, it's the first one that is attached to a camera, as I was using tech cameras exclusively last time. This brings me to the crux of my question.

Whilst I know that CCD seems to have become a bit of a dirty word here lately (all hail king Sony CMOS, long may he reign) I do know that noise can become an issue. With a tech camera it never bothered me as between shots I was usually fussing around with something or other. But now things are different. Last night I was rattling off multiple shots in a row of some bats flying around my island to test out some later evening type exposures. I have noticed how the noise does build up as the sensor is continually used. It's not excessive, but it is noticeable, especially if you try and recover some shadow detail.

So is there a rule of thumb at all for allowing the sensor to cool down sufficiently between shots to ensure you don't get excessive noise for longer exposures, or multiple exposures over a long time.

I know I can work all this out with some testing time and dedication, but surely someone has done this before. And please people, I don't really want to know how much better another camera could be for this kind of thing. I do own a Df as well, but I want to get the best out of the 645D.

Thank you.

 

[ErikKaffehr]

Hi,

You may get better answers if you mention the Pentax 645D in the thread title. There are some P645 users on the forums shooting long exposures. Ed Hurst...

Best regards
Erik

Ahoy GetDpi friends.

I recently bought back into MFD with a Pentax 645D, which is a fantastic camera. I have been running around taking photos of this and that, testing different lenses and doing all the usual stuff you do when you get a new camera. Whilst this is not my first MFD device, it's the first one that is attached to a camera, as I was using tech cameras exclusively last time. This brings me to the crux of my question.

Whilst I know that CCD seems to have become a bit of a dirty word here lately (all hail king Sony CMOS, long may he reign) I do know that noise can become an issue. With a tech camera it never bothered me as between shots I was usually fussing around with something or other. But now things are different. Last night I was rattling off multiple shots in a row of some bats flying around my island to test out some later evening type exposures. I have noticed how the noise does build up as the sensor is continually used. It's not excessive, but it is noticeable, especially if you try and recover some shadow detail.

So is there a rule of thumb at all for allowing the sensor to cool down sufficiently between shots to ensure you don't get excessive noise for longer exposures, or multiple exposures over a long time.

I know I can work all this out with some testing time and dedication, but surely someone has done this before. And please people, I don't really want to know how much better another camera could be for this kind of thing. I do own a Df as well, but I want to get the best out of the 645D.

Thank you.

 

[tjv]

Were you making long exposures, one after the other? I know it's a different sensor and back (not integrated like the 645D) but I've never seen noise increase when using my Credo 60 on a tech camera, even on zero latency mode in approx 22 degrees celsius ambient temp, even when keeping the back on for long periods of time. This is true even when I'm making long exposures (although exposure time certainly increases noise, etc., so maybe that is masking it.) The back does get warm though.

Can you post some examples? I'd be interested to see what you're talking about, even if just to inform me of some things I can look out for in my files to try reassess things...

One other thing is that I expect Capture One 9 does a great job with noise, particularly with long exposures, even up to 2mins. I've never used older versions of C1, but I hear people are reporting improvments in v9 over v8. Are you using Lightroom to process your files, and have you compared the noise levels if you convert in different software? I know C1 is not good for the Pentax, but what about Torger's bespoke software?

 

[algrove]

OP

Interesting question as the new PO IQ 100MP has a measurement built-in to show a graph of the temp build up. Perhaps Phase can answer your question moe precisely or someone who currently has that back in use and does long exposures. Just a thought.

 

[Paul2660]

Hi Dogs,

Actually CMOS can get hot too, note the issues currently with the Sony A7rII, with both long exposures (excessive noise) and using it for 4K video (camera shuts off due to overheating). Also it's interesting to to note Fuji with the X-Pro2 added a temp warning on their camera, which starts to blink if the camera detects too much heat.

Back to your question, IMO yes heat is the killer for either CCD or CMOS in regards to noise. But it's interesting, Phase One, with the P45+ was very clear in that they did not expect to see 1 hour exposures unless the outdoor temp was 69 degree F or less and the ambient humidity was less than 100%, more in the 60% range. In my use of the P45+, I took many 1 hour to 45 minute exposures in night work, (in Arkansas the summer nights are usually outside of the temp range for the P45+), and remember, the P45+ has to take a full dark frame after the first exposure of the same time, so a 45" shot, needs another 45" dark frame. But the P45+ never seemed to really get hot to the touch for me. A magic chip/controller combination that Phase One never really matched again, at least with CCD.

Move to the IQ160/260, and the all black digital back. In the summer I often felt the 160/260 getting hot, at times almost too hot to the touch, no kidding. And yes, the effect on the images, even at base ISO was noticeable. The increase in noise was huge. And it would take the back quite a while to cool off. In the summer here, I tend to carry a shield I made from a car sunshade to cover the back while it's not in use, i.e. while I am waiting for a particular shot.

It's interesting to note that Phase added a temp gauge to the IQ2/3 backs, where you can monitor the temp of the back, which I do use at times.

As a side note, I will be interested in seeing how well the new Hasselblad 100MB will handle the heat when being used for 4K capture, as so far most of the new camera platforms shooting to 4K, are having some heat issues. Note Canon added a heat pipe system to the 1DX MKII, which is unique, but also points out a known issue for heat build up. Fuji made it clear that the X-Pro2 could do 4K, but left it out again due to possible heat lock ups.

Paul C

 

[Dogs857]

Hi,

You may get better answers if you mention the Pentax 645D in the thread title. There are some P645 users on the forums shooting long exposures. Ed Hurst...

Best regards
Erik

Thanks Erik but I can't change the title after I post it can I?? I'll check.

Were you making long exposures, one after the other?

Can you post some examples? I'd be interested to see what you're talking about, even if just to inform me of some things I can look out for in my files to try reassess things...

One other thing is that I expect Capture One 9 does a great job with noise, particularly with long exposures, even up to 2mins. I've never used older versions of C1, but I hear people are reporting improvments in v9 over v8. Are you using Lightroom to process your files, and have you compared the noise levels if you convert in different software? I know C1 is not good for the Pentax, but what about Torger's bespoke software?

Good questions. Yes I was shooting longer exposures one after the other in rapid succession. I should qualify that the noise is not excessive, but certainly higher than was expected which led to the question. It was also getting quite dark so that didn't help. I was pushing the camera to see how it responded and maybe it was just too much for the 645D. It also might just be getting used to a CCD sensor again, which is also a possibility.
I don't have examples because I have an obsessive thing about deleting files that I am not going to use. It was only after everything was removed and cards reformatted that I thought examples could be useful. I'll do this again and see how things go.

OP

Interesting question as the new PO IQ 100MP has a measurement built-in to show a graph of the temp build up. Perhaps Phase can answer your question moe precisely or someone who currently has that back in use and does long exposures. Just a thought.

Intersting. I would like to see someone chart the temperature and how it affects exposures.

Hi Dogs,

Actually CMOS can get hot too, note the issues currently with the Sony A7rII, with both long exposures (excessive noise) and using it for 4K video (camera shuts off due to overheating). Also it's interesting to to note Fuji with the X-Pro2 added a temp warning on their camera, which starts to blink if the camera detects too much heat.

Back to your question, IMO yes heat is the killer for either CCD or CMOS in regards to noise. But it's interesting, Phase One, with the P45+ was very clear in that they did not expect to see 1 hour exposures unless the outdoor temp was 69 degree F or less and the ambient humidity was less than 100%, more in the 60% range. In my use of the P45+, I took many 1 hour to 45 minute exposures in night work, (in Arkansas the summer nights are usually outside of the temp range for the P45+), and remember, the P45+ has to take a full dark frame after the first exposure of the same time, so a 45" shot, needs another 45" dark frame. But the P45+ never seemed to really get hot to the touch for me. A magic chip/controller combination that Phase One never really matched again, at least with CCD.


Paul C

Paul, I forgot to take into account environmentals as well. It's always a balmy 30+ degrees Celsius here this time of year, and 70-80% humidity. I'm sure that doesn't help either.


I am going to re-shoot something similar again in the near future and see what happens. I'll take a base photo, then a number in succession and see if the noise increases with exposures or if it's just the CCD struggling in lower light. I'm guessing it's the latter.

It would still be interesting to know if there is a rule of thumb for CCD sensors and long exposure buildup though.

Thanks everyone who chimed in and to those who sent messages as well.

 

[Paul2660]

OP

Interesting question as the new PO IQ 100MP has a measurement built-in to show a graph of the temp build up. Perhaps Phase can answer your question moe precisely or someone who currently has that back in use and does long exposures. Just a thought.

With the latest firmware the temp gauge will also work with CCD backs for sure IQ 3 and 2. Not sure about the IQ1 series.

Neat feature to watch as you turn the back and to see how fast the temp goes up depending your shooting environment.

Paul C

 

[ondebanks]

Intersting. I would like to see someone chart the temperature and how it affects exposures.

That someone will be me

It would still be interesting to know if there is a rule of thumb for CCD sensors and long exposure buildup though.

There is.

Tech alert, people! :lecture: There will be graphs! There will be equations! Look if you dare! You might even learn something. That something might even improve your art!

----

The approximate rule of thumb is this: long exposure noise (dark current) in CCDs doubles, for every 6 degrees Celcius rise in temperature.

The exact relationship is this (from the observational astronomy module that I teach):

There are 5 variables in this equation. As always, if you know any 4 of them, you can calculate the 5th. So if you know 3 of them specific to the CCD design (the reference temperature T[SUB]o[/SUB], the dark current at that temperature D[SUB]T[SUB]o[/SUB][/SUB], and the temperature increment which doubles the level of dark current Delta-T[SUB]DBL[/SUB]), then you can pick the 4th (any sensor temperature T) and calculate the resulting 5th (the level of dark current at that temperature, D[SUB]T[/SUB])

Here's an example for a Kodak CCD, where the dark doubling increment was determined to be 6.4 degrees C:

The Pentax 645D uses a Kodak KAF-40000 CCD. Its datasheet includes the following specs:

This gives us the 3 design-specific numbers we need. Delta-T[SUB]DBL[/SUB] is 5.5 degrees C, T[SUB]o[/SUB] is 60 degrees C, and D[SUB]T[SUB]o[/SUB][/SUB] is 42 pA/cm[SUP]2[/SUP]. With a bit of unit conversion, using the pixel size (also given in the datasheet as 6 micrometers), we convert 42 pA/cm[SUP]2[/SUP] to 92 electrons/pixel/second. That is at 60 degrees C, remember. Plugging those numbers into the equation, we find the following for the 645D sensor:

O degrees C -> 0.05 electrons/pixel/second
20 degrees C -> 0.6 electrons/pixel/second
40 degrees C -> 7.5 electrons/pixel/second

Approximate reckoning with the "doubles/halves every 6 degrees change" rule would get you close to these numbers; you should really use 5.5 degrees though in this case, per the spec.

Plotting the 645D/KAF-40000 for a sweep through temperatures gives me this graph:

So, at what temperature does it start to be a problem for your photos?

To answer this, we should really consider the whole noise model; but a ready reckoner is that when the accumulated dark current/noise exceeds the readout noise, you become aware of the increasingly noisy shadow regions - and this is for a typical pixel: 'hot' pixels will be severely bright long before that. The KAF-40000 datasheet indicates a typical readout noise of 13 electrons at fast (photographic) readout speeds. From the graph, you get 13 'dark' electrons per pixel in a 1 second exposure at 45 degrees (the sensor may get much warmer than the ambient temperature), or a 10 second exposure at 25 degrees, or a 100 second exposure at 8 degrees.

There's just one other ingredient missing from this analysis: the rate at which the sensor is self-heating. This resembles Delta-T[SUB]DBL[/SUB] but instead of being the 5 or 6 degrees temperature change required to double the dark current, it's the continuous exposure time required to raise the sensor temperature by 5 or 6 degrees. We could establish this by experimentation...

Ray