Forgot Password
Pentax Camera Forums Home
 

Reply
Show Printable Version Search this Thread
08-19-2009, 05:31 PM   #331
Veteran Member
Torphoto's Avatar

Join Date: Mar 2008
Location: Trinidad W.I.
Photos: Gallery
Posts: 620
Well for fellow pentax users, the 5d mk2 gets some hot pixels after too much lv or video, difference is it goes away right after tho.

08-19-2009, 05:48 PM   #332
Pentaxian
falconeye's Avatar

Join Date: Jan 2008
Location: Munich, Alps, Germany
Photos: Gallery
Posts: 6,862
Well, the latest thing I heard (not officially from Pentax, though) is that Pentax really is working hard to do something about the green line issue. The issue got top management awareness.

The only thing is that we don't know how successful it will be. Let's keep our fingers crossed ...
08-19-2009, 05:50 PM   #333
Forum Member




Join Date: Jun 2009
Location: The Villages, FL
Posts: 70
QuoteOriginally posted by Torphoto Quote
I got a line, not to bad and shows up at 1600 iso, damn tropical heat

correction, shows up at 800 iso as well, so should I send this back?

Edit this line is at every iso, I'm thinking seriously of sending it back, tho wondering if I should wait to see if Pentax fixes it first.
Torphoto: If you saw that kind of line in a batch of film you purchased, would you send the film back? The answer should be "yes." The same should apply to your K-7. It sounds like the qustion is "when": Now, or when you are sure they acknowldege the problem and have a fix. The next question is "how long are you willing to wait?" A month? Six months? I guess you're ok as long as the warranty is in effect. I just hope those who might wait six months don't lose their credibility when they say the problem existed from the beginning. They might ask "Why didn't you send it in when the problem was first noted?"

And it seems the better place to send it would be Pentax - to do our part in helping them experience, diagnose, and fix the problem - even though it might take a couple of months longer to get the camera back. At least you would be more certain it was fixed, rather than chance another bum unit.
08-19-2009, 05:53 PM   #334
Pentaxian
falconeye's Avatar

Join Date: Jan 2008
Location: Munich, Alps, Germany
Photos: Gallery
Posts: 6,862
QuoteOriginally posted by gfmucci Quote
to do our part in helping them experience, diagnose, and fix the problem
This thread must have many more pages than you thought

08-19-2009, 06:45 PM   #335
Veteran Member
Torphoto's Avatar

Join Date: Mar 2008
Location: Trinidad W.I.
Photos: Gallery
Posts: 620
QuoteOriginally posted by gfmucci Quote
Torphoto: If you saw that kind of line in a batch of film you purchased, would you send the film back? The answer should be "yes." The same should apply to your K-7. It sounds like the qustion is "when": Now, or when you are sure they acknowldege the problem and have a fix. The next question is "how long are you willing to wait?" A month? Six months? I guess you're ok as long as the warranty is in effect. I just hope those who might wait six months don't lose their credibility when they say the problem existed from the beginning. They might ask "Why didn't you send it in when the problem was first noted?"

And it seems the better place to send it would be Pentax - to do our part in helping them experience, diagnose, and fix the problem - even though it might take a couple of months longer to get the camera back. At least you would be more certain it was fixed, rather than chance another bum unit.
Note my sig, no k7 and something new that's not pentax
08-20-2009, 07:09 AM   #336
Veteran Member




Join Date: Jun 2009
Photos: Albums
Posts: 1,679
QuoteOriginally posted by falconeye Quote
Well, the latest thing I heard (not officially from Pentax, though) is that Pentax really is working hard to do something about the green line issue. The issue got top management awareness.

The only thing is that we don't know how successful it will be. Let's keep our fingers crossed ...
Thanks for the update Falk. I sense your unofficial info tends to be more accurate than some others 'facts'.
08-20-2009, 07:39 AM   #337
Banned




Join Date: Sep 2006
Location: Millstone,NJ
Posts: 6,491
Someone from the UK posted this on DPreview Pentax SLR forum:
QuoteQuote:
I received my replacement K-7 today (#3384xxx c/w #3377xxx) and the problem is not apparent. I did an extended test on it and the dreaded green line didn't show. I'm cautiously optimistic
Re: Patheitic?: Pentax SLR Talk Forum: Digital Photography Review
08-20-2009, 03:23 PM   #338
Loyal Site Supporter




Join Date: Dec 2006
Location: Frazier Park, CA
Photos: Gallery
Posts: 835
I had the green line with the first K-7 body I had. I haven't had it on the replacement. I've done more video with it than I had with the first one, and in similar weather conditions. I'm not so sure that its because the sensor is any better, but it does seem much harder to get this camera to over-heat (can use video and live view longer) without the CameraTemperature indication going above 36 or 37 C. Granted, the weather has been a little cooler than it was the weekend I used the first camera body, but it hasn't been THAT much cooler.

08-20-2009, 04:09 PM   #339
Senior Member
noVICE's Avatar

Join Date: Aug 2009
Location: Brisbane
Photos: Gallery
Posts: 165
as I said earlier, I took mine to 41C and I could not get the green line to show, but there was some colour banding. I was using ISO3200 and I was deliberately taking shots with dark areas in them. I am still hesitant to say that mine is fault-free, I think it will just be a matter of time. Another thing that I noticed while "heating it up" with LV was the noises it was making. You could clearly hear things expanding inside the body of the camera which makes me wonder if this issue is partly due to compact size of the body.
I think that getting an extended warranty with this camera is a must. For Aussies unfortunately the importer told me that they did not offer the three year warranty like Pentax USA, so you will have to get it through the retailer.
08-20-2009, 04:13 PM   #340
Pentaxian
falconeye's Avatar

Join Date: Jan 2008
Location: Munich, Alps, Germany
Photos: Gallery
Posts: 6,862
QuoteOriginally posted by mtngal Quote
it does seem much harder to get this camera to over-heat (can use video and live view longer) without the CameraTemperature indication going above 36 or 37 C.
That's one of my theories. That some units have more effective cooling than others.

My pre production K-7 reqired 15 minutes to display the red temperature alert sign when recording movie, at 27C ambient temperature. And it reqired another 15 minutes to reach 40C where it started to exhibit the green line.

Because of the SR mechanism, the sensor housing cannot be directly mounted to the metal part of the body. There somewhere must be contact points carrying the sensor heat to the body metal. Maybe, those points have too much variation.
08-20-2009, 09:43 PM   #341
Pentaxian




Join Date: Jun 2009
Location: GMT +10
Photos: Albums
Posts: 10,585
Perhaps the answer to my question lies buried further down in this thread, but why is a green vertical line the specific defect symptom?

I mean, random dead pixels appearing in an image I can understand (physically inert circuits on the sensor as a product of manufacturing defects) but why a green line? And one that runs the length of the image? Why isn't the it (for arguments sake) a horizontal green line, or a diagonal red line, or a green line that only appears for half of the vertical length of the image. Etc.

What physically is going on? How can the issue of heat impact sensor processing in such a specific manner?

I could certainly understand this if it were a mechanical issue with the sensor. Eg once the sensor itself gets hot, a gap opens up on the sensor surface - kind of like an expansion gap - and hence a line appears. Or the sensor has no expansion gap and at a specific line on the sensor as heat increases adjacent pixels are pushed together and short out or something.

However what I can't get my head around is how such a specific symptom could be produced by heat impacting the electronics. From my experience, when you over-heat electronic circuitry (eg an overheating video card or CPU on a PC), you don't normally get such surgically precise problem symptoms - errors are more erratic and 'fuzzy', crashes and BSOD's are random etc.

So why specifically a thin green line?
08-21-2009, 12:16 AM   #342
Inactive Account




Join Date: Jul 2009
Location: Oslo, Norway
Posts: 29
I have no reference for this, but having worked with designing integrated circuits for almost 9 years I have some theories.

The pixels on a sensor is probably read out either in rows at the time, or columns. My guess is that there are some enable circuits that enables a single rowof pixels, and then there is amplifier circuits for each column(could be opposite, but I'll stick to this for the arguments here). The enable circuits cycles through all the rows, and for each of them, the amplifier circuits are picking up the signals and routing them to the AD conversion. If there is a designweakness in either the enabling circuits or the amplifier circuirs, this would manifest in horizontal or vertical lines. So this may be the reason why there is always a vertical line showing up. Diagonals would make no sense at all.

Why is it always a vertical line showing up on approximately the same place. That's harder to tell. Stitching has been mentioned, which means assembling several dies into one large sensor, or there may be other design decisions like splitting the sensor into several power domains. It could also be that some of the support circuits are interfering with the sensor causing this. The transfer speeds needed to get the data out of the sensor is quite big, and it can cause all sorts of problems. But one thing is certain; due to the systematic nature of this thing, it is almost certainly a flaw in the design of either the sensor itself, or some support circuits.

Why is it only a problem on some cameras, and why are they seemingly getting better? CMOS is manufactured to a pretty wide tolerance, which any good designer should take into account and make the design robust enough to tolerate. The temperature variations will stack up on the process variations, and it the latter is too big, the device gets out of its envelope and errors starts to occur (EDIT: errors first show up where the design is less robust, not completely at random). It could also be that variations in the sensor and variations in the support circuits some times both pull in the same direction, pushing the system outside its limits.

Why didn't Pentax or Samsung discover this before release. I'm not saying they didn't, but assuming it took them by surprise. Prototype production of CMOS is usually more tight in tolerance than full production runs. Also, if the problem is related to the interaction between devices, it requires a very large number of samples to find failing units, and even more to discover a trend.

Why does it seem to improve on later models? When a chip is put into production, the people at the fab starts to tweak the production parameters to increase the yield (the percentage of good devices), which in effect will tighten the tolerances. Also, the people at production test will start to adjust the test limits and modify the tests to better pick up bad samples. All this results in a more uniform quality of the devices over time.

Is it possible to fix in firmware? It depends, but I'm assuming that's what will happen. If the problem has to do with interference between the sensor and the ajacent circuits, then the pixels are probably good in the first place. Maybe something can be done with the readout scheme to make it better. If it is the sensor, then it may be interpolated, which the camera already does for dead pixels. It may sound disturbing to some, but I don't think it will be visible in a real picture. It wouldn't surprise me if some of the very sucesful, and highly acclaimed models from other brands do stuff like this without anyone knowing about it.

Anyway, this is not based on anything other than my general knowledge of semiconductor design and manufactoring. The problem with the K-7 may be something completely different, but I find this quite plausible.

Last edited by kjellhar; 08-21-2009 at 01:56 AM.
08-21-2009, 02:52 AM   #343
Pentaxian




Join Date: Jun 2009
Location: GMT +10
Photos: Albums
Posts: 10,585
Great information. Thanks kjellhar. Although I am woefully ignorant on circuit design, what you are saying makes a lot of sense.

It also suggests to me that the same issue may be lurking in the K20D sensor, but since it is rarely subject to the same heat and other stresses as the K7 sensor, it hasn't shown up in any real world use of the camera.
08-21-2009, 03:04 AM   #344
Inactive Account




Join Date: Jul 2009
Location: Oslo, Norway
Posts: 29
Well, I'm not suggesting anything about a specific camera. The K20D may be flawless (unlikely), or they have worked around all the serious bugs before release.

Many of the issues I describe, especially the problems with interaction between components, are very dependant on the design. You could have two versions of a design, each with the exact same components, but the components are only positioned differently on the circuit board. One may fail while the other may pass. It may sound like black magic, but once you get into it, it makes sense. It is very difficult to analyse though, so it's not surprising that quite many consumer electronics devices enters the marked with flaws.

Kjell
08-21-2009, 03:17 AM   #345
Pentaxian
Class A's Avatar

Join Date: Aug 2008
Location: Wellington, New Zealand
Posts: 8,937
QuoteOriginally posted by rawr Quote
why is a green vertical line the specific defect symptom?
GordonBGood is speculating about this at DPR. Don't think he's got it yet, though.
Reply

Bookmarks
  • Submit Thread to Facebook Facebook
  • Submit Thread to Twitter Twitter
  • Submit Thread to Digg Digg
Tags - Make this thread easier to find by adding keywords to it!
body, camera, dslr, iso, k7, line, photography, sensor
Thread Tools Search this Thread
Search this Thread:

Advanced Search


Similar Threads
Thread Thread Starter Forum Replies Last Post
Sony outsources some sensor production to Fujitsu deadwolfbones Pentax News and Rumors 10 10-02-2010 03:24 PM
Samsung's new BSI (backside illumination) CMOS sensor will be in production by 2010 yakiniku Pentax News and Rumors 7 10-15-2009 09:34 PM
HELP!!Busted my DA* 16-50 pentaxpaul Pentax SLR Lens Discussion 10 03-02-2009 10:21 AM
Busted! pingflood Post Your Photos! 7 01-03-2009 09:57 PM
ZX-10 Busted? ericc Pentax Film SLR Discussion 1 12-01-2006 11:41 AM



All times are GMT -7. The time now is 06:43 AM. | See also: NikonForums.com, part of our network of photo forums!
  • Red (Default)
  • Green
  • Gray
  • Dark
  • Dark Yellow
  • Dark Blue
  • Old Red
  • Old Green
  • Old Gray
  • Dial-Up Style
Hello! It's great to see you back on the forum! Have you considered joining the community?
register
Creating a FREE ACCOUNT takes under a minute, removes ads, and lets you post! [Dismiss]
Top