it's not weird, my K20D is good regardless of the light... while there were others who complained

in reality a lot of people with issues go to the internet to try to find solutions and than see others like them so we get more than the fair share of complaining, while most others with well functioning cameras cannot be even bothered to post as they will become "fanboys" or whatever...

in this particular case, where new firmware fixed most other K-5's and not yours, you may actually have a faulty camera in some other way, than whatever was the error fixed by the firmware, so you may need to consider returing/replacing... not exactly good, but this is consumer electronics for all of us...

The tests were more than accurate to reproduce the front focussing fault of the K-5 versus the more accurate K20D

All subject distances were 30"

As I explained in a previous post : The shots shown are CROPS to show the pertinent parts of the frame to fit an 800pixel wide pic to upload to this and other forums. The centre focus point of the camera was aimed at the black centre focussing line of the chart in each case and on the Phones Jack 1 in the last test

I shall add this to my OP to avoid having to repeat

Smeggy, I'm sorry for your problems just get with your dealer and try another body. I have a friend here that just got his third D7K after repeated issues......this one works fine, all Mgf seem to be having some problems, but for the vast majority the Firmware appears to have solved the K5 problems. BTW, Charlie tried to trade his last D7K for a K5, but his dealer said it would take six weeks, so he didn't.

For me, I am just going to relax until I have my K5 on Monday, no need to stress out, if it will shoot Squirrels (and girls!) it will be just fine.
I know the entire Forum is waiting to see if it will shoot Squirrels.....is there anything in life more important than that? I didn't think so!!

Meanwhile, I will just relax and wait until Monday......


Best Regards
Rupert

Yup, I waited (mostly) patiently for the 1.03 update until making a decision to return the body.

There's something that has been implicitly covered, but perhaps not sufficiently emphasized.

Simple logic dictates that if the hardware and software are the same, then the cameras should behave the same. Since they do not, then we must assume that something is different. In this case, it's safe to assume that the software is the same. The hardware is SUPPOSED to be the same, but clearly, it is not, and hence should be considered "broken", in spite of the fact that it works in most respects.

Perhaps the discrepancies are merely a matter of manufacturing tolerances, and Smeggypants' (and others') K5 is at an extreme of one of these tolerances. As for my own K5, preliminary "tests" show improvement in FF, but also some inconsistencies. Also, the fact that the AF Assist beam appears to be behaving differently now, makes it more difficult to evaluate the degree of improvement (having not planned for comparisons in advance, before loading 1.03).

Granted, it's not difficult to reach an incorrect conclusion when one does not have knowledge of all the variables, common sense (logic) sometimes fails to take that into consideration. But nevertheless, the above conclusion is a possibility, at least.

Get a new one, Smeggy, and good luck.

Yes I've noticed the assist beam behaving differently. I had it turned off for my tests and frankly it would an annoyance in most of my shooting conditions.

One striking behaviour was in and around EV2 the assist beam would light up in Contrast LV AF, but not in phase AF. Didn't come on at all at EV2 in any mode mode before.

Thanks

Why can't Pentax have AF correction that can be calibrated so as to give the same focus as LV. i.e. in this new 'calibration' mode using a tripod, a shutter press, automatically and in sequence, AFs on a chart in LV mode, then AFs in PD mode and then the camera adjusts its AF internal settings (rather than moving the lens internals) to align the electronic PD signals. It could even have several settings for a zoom to maintain accuracy over its range of FLs and even subject distances etc.

Good idea sir!!!

If it did this for every shot, that new mode would be intensely slow.

I could see a type of 'green button' implementation of this, though - you hit a button and it calibrates the AF with LV, and then makes the adjustments to every PD shot after that until you either 'reset' that, or hit the button again to recalibrate for another lighting situation.

.

I don't think that's what he meant. I'm sure he meant it as a method of calibrating your lenses one time for various focal lengths and hopefully light levels.

I would be more than happy to calibrate the AF for each of my lenses for a range of EV levels

We have many FW 1.03 threads, so can't be sure where this would belong (perhaps nowhere). But anyway, it's pleasing to hear that so many people find 1.03 solves their problems, but just to add my own experience to that of Smeggy's: my K5 definitely still has FF with 1.03. I've just done some tests at EV 0, comparing Liveview AF to normal AF, and whereas LV gets it right every time, normal does not. Same behaviour with the Tamron 28-75 f2.8 and the Pentax 18-55 WR. Also, a quick test (i.e. not confirmed) seems to indicate that results are similar when using LV/Phase and LV/Contrast i.e. both seem to get it right. Interestingly, the AF assist light comes on under LV, but not otherwise. Bought my unit in a shop in Germany, and not sure how to handle it now.

Sorry, but I also have some concerns about alignment of the target. I use this same target and it is critical to have it aligned to the camera. If not, then the out-of-focus area on one side will be different from the other side. Judging by the angle of the lines, there seems to be some variability in your samples.

Which aperture did you use? Widest is best for focus testing. My K20D front-focusses in tungsten light, but only at apertures of 2.8 and wider.

From my experience, this isn't the best way to test focusing accuracy. The focusing sensor isn't a small 'point' but a little square 'area'. It's very hard to let it 'see' and 'focus' on a thin parallel line. Every time you think you focus on the same location, but actually the sensor might focus at a different 'point' within its 'area'. I've tried many times to focus on the same chart you used here, and I can achieve BF or FF or spot on focus very easily by tilting the lens just a little bit up or down, using the same camera/lens.

The best way to test is to focus on a flat target which is big enough for the focusing sensor to see and is parallel to the camera CCD/CMOS sensor, and start from there. You can easily make something like these:





Good luck testing

smeggy

for what its worth mate, Ive put the new firmware on my K5..and the FF is very much improved in Halogen light, I wont say its perfect, as contrast AF still does a better job. but its pretty near perfect, although it will still FF if I push it hard enough in very difficult very low light situations,

If you remember I did exchange a K5 that had a moderate amount of FF for another body, and in the tests of the second body I found it not to FF as bad as the first body...so Im convinced some K5's were better at AF in low light than some other were.....yours I feel is in the first set and not fixable with the 1.03 firmware....my advice is to return it for another unit, then do your testing again,

theres no doubt yours is buggered

Hi there. Please see this post for info. You may have missed it.

https://www.pentaxforums.com/forums/pentax-k-5-forum/136324-k-5-1-03-versus-k...ml#post1421891