How many stuck pixels are acceptable?

[stephen_usher]

Last week I received from FLO a Celestron Skyris 445C camera and I've discovered during testing that there are three stuck pixels which show up even at the lowest gain levels. One of these, a blue pixel, is close to the centre of the frame, which will make imaging problematic.

I was just wondering what would be the acceptable number of stuck pixels you would accept before sending such a camera back as defective.

Would having just one in a very awkward position be enough in your opinion?

I've attached a dark frame (30fps, minimum gain) to show the problem

[Oddsocks]

Stephen,

I have looked at your image at individual pixel level, heavily stretched, and can see that you have not only, single defective pixels, but defective clusters.

I have pasted four cropped images at the end of this reply.

These are RGB composites so suffer from chrominance noise, but at the individual colour channel level you can identify multiple adjacent pixel failures.

When clusters appear like this they do indicate a manufacturing defect and this would warrant replacing the camera while still under guarantee as substrate failures such as these may spread over time and become worse.

Individual hot or dark pixels while annoying don't have a great effect in the final image as they can be easily compensated for by dithering the camera between frames or allowing the planet to drift slowly across the sensor during AVI acquisition but cluster defects are a different story.

Here is a link to the Sony specification for allowable defects across different zones of the CCD, these specs are those that Sony's own QA dictates are acceptable at delivery to the device builder, Celestron don't publish their own tolerances but there is no reason that they should not be comparable.

http://www.eureca.de/datasheets/01.xx.xxxx/01.02.xxxx/01.02.0246/ICX445AQA-E(E)-1.pdf

You should contact FLO for further advice and possible return for exchange.

 

 

 

 

[stephen_usher]

Stephen,

I have looked at your image at individual pixel level, heavily stretched, and can see that you have not only, single defective pixels, but defective clusters.

I have pasted four cropped images at the end of this reply.

These are RGB composites so suffer from chrominance noise, but at the individual colour channel level you can identify multiple adjacent pixel failures.

When clusters appear like this they do indicate a manufacturing defect and this would warrant replacing the camera while still under guarantee as substrate failures such as these may spread over time and become worse.

Are you sure that the problems that you see aren't caused by the JPEG artefacts? I had to compress the original BMP file as it was too large for the forum to accept.

[stephen_usher]

Here's a stretched version of the original which has then been JPEG compressed.

No sign of clusters there, just CCD noise.

[Ant]

Either way you need to contact FLO for advice. I did try and find a data sheet for that chip, as they sometimes say what the level of hot pixels is acceptable for a particular chip... I couldn't find that info.

FLO may not see this thread, so please don't rely on this thread for an answer.

Cheers

Ant

[FLO]

Hi Stephen, 

I am at home juggling kids and packing the car for Kelling but, at first glance, it doesn't look too bad and it isn't unusual to find 2 or 3 dud pixels on a sensor. But, we want you to be happy with your purchase so if you decide you'd like to try another we'll happily send you one for comparison  

Coincidentally, one of my colleagues has taken a Skyris 445 (at least I think it's that one) to Kelling to play with assess so we'll have firsthand experience with the camera soon.

HTH, 

Steve 

[stephen_usher]

Thanks for the reply, Steve.

Given the awkward position of that rather bright blue pixel I think I might return the camera to you and take up your offer to try another. Do I need an RMA number?

I am surprised that this got through the quality control process at the factory though, especially given how good Imaging Source branded cameras are.

Thank-you again for your great service.

[stephen_usher]

I just have to say a big thank-you to the staff at FLO.

A replacement camera arrived this morning. This time only one stuck green pixel in an out-of the way place, so it's no problem at all.

[FLO]

We like happy endings  

[ollypenrice]

One pixel, wherever it is, isn't going to bother you. The whole point about fast frame imaging is that you get masses of natural dther so this pixel will be averaged out. It is never going to sample the same bit of planet on each frame! Good service from FLO but I think that if you try both cameras you'll find a similar result. You should see a dark frame from my 11 meg deep sky camera...

Olly

[Sp@ce_d]

Looking forward to seeing some results from this camera. I have to add, from experience, FLO have excellent customer service.

Sent from my iPhone using Tapatalk - now Free