The ultimate which ATIK ccd thread.

[RAC]

Just as a pointer to what I have found so far:

SBIG ST-8300M: Bias base level 740 (ADU)

QSI 683: Bias base level with high speed read: 1240

QSI 683: Bias base level with quality read: 240

This tells me that there is something to dig into here, and that it has to do with electronics implementation! Yes?

/per

My Atik 383L has a bias base level of 350adu. It takes about 10sec to download an image and this doesn't bother me at all. If I had a fast option I would still use the slow one.

[Gina]

For comparison, I'll check the bias ADUs and download times of my 460EX and 314L+. The latter I estimate as around 7 or 8 seconds for the 460 and 2 or 3 for the 314 but that's more of a guess than anything. I'll probably take darks too for the sake of completeness.

[perfrej]

Yes, I use the slow one for images and the fast one for automatic plate solves and automatic focusing. Way to go, really.

/p

[ncjunk]

I don't see an issue with download times the old art285 i have is 11 seconds theres no problem. For focusing focusmax takes a sub frame around the star and download times are around a second.

I'd prefer less noise any day over shorter downloads.

Am I the only one not bothered....do i look bovered? Is this face bovered...(sorry i'll stop now)

Sent from my GT-P5110 using Tapatalk 2

[Gina]

Quite agree. A download time of 11s or 20s or even a minute would be fine when taking subs 0r 5m, 10m, 20m, etc. And binned subs for plate solving or small area download are all very much faster

[NickK]

Earlier on today Neil and I had a PM conversation about doing a comparison of the 460 and the 383. At the moment I own both (long story!). This isn't scientific in that the conditions are certain to be different between two nights, but it might be interesting.

I'm attaching a single 300s sub taken with a Borg 77 with a 0.8fr and a Ha filter and 2x binned - for both the 460 and the 383. The 460 was the night before last, the 383 I've just taken. Ideally it would be unbinned, but the 460 run was binned. Later I'll add some dark frames for comparison too.

A few obvious issuesto take into account - I just substituted the camera, so the FR spacing isn't correct and the corners look bad!! and I think the SX OAG on the filter wheel might be intruding into the light path of the 383 a bit givenm the bigger chip, but it seems to have got stuck and so I couldn't move it.

The difference in field of view is apparent, I'll leave others to consider sensitivity/noise. What is also worth noting is that the 383 will cool to -45 rather than -25.

460 mono

pickeringstriangle-bin--003Ha.fit

383L mono

Pickerings triangle 383 - -003-Ha.fit

Helen

Was looking at this - something odd there - as you've got small images, 383L at native resolution is 17MB!

One option for faster download is use a smaller sub-image The camera will only transfer the part of the image you want instead which lowers download time

[Uranium235]

Ive been reading this with interest, but I do have to question the response curves posted by Per (great though it is). Im pretty sure the ICX285 has a QE about 10% higher at 650nm than the 44% quoted on the graph. Surely its more sensitive than the 8300 in Ha?

http://www.dangl.at/.../atik_314_e.htm

Out of interest, I was (and still am) considering the 383.... but its the shutter, it kinda puts me off as its a moving part (and therefore prone to eventual failure), and it will mean a different focusing routine to the one I use now (0.5s with subframe box at 1x1 bin). With a 383, to make up for the lost QE I would have to get a faster telescope - so thats more dosh and a leap into unknown optics. Even them im limited to F5, otherwise its a whole new set of 2" filters (even more dosh!).

But when you look at the FOV on offer for the cost.... its tempting.

[ncjunk]

Sums up my confusion, i like the extra sensitivity of the 428/460 but then the fov of the 383 would be lovely.

Sent from my GT-P5110 using Tapatalk 2

[perfrej]

I haven't heard of any shutter problems and they are designed for millions of operations. Focusing is a breeze if you adopt Focusmax. You achieve perfect focus in 30-60 seconds and it i salways just that; perfect!

As for general info on the subject of interline versus full frame, check out this blog entry: http://www.qsimaging.com/blog/?p=25

Getting a faster scope because lower QE sounds a bit of an over-reaction. I think you need to look at the bigger picture and factor in gain and noise as well as well depth.

/p

[Uranium235]

I haven't heard of any shutter problems and they are designed for millions of operations. Focusing is a breeze if you adopt Focusmax. You achieve perfect focus in 30-60 seconds and it i salways just that; perfect!

As for general info on the subject of interline versus full frame, check out this blog entry: http://www.qsimaging.com/blog/?p=25

Getting a faster scope because lower QE sounds a bit of an over-reaction. I think you need to look at the bigger picture and factor in gain and noise as well as well depth.

/p

Thanks, I'll take a look at Focusmax to see if I can better my current method. Oddly enough, I was reading about the differences between interline and full frame a couple of days ago (in regard to the way the chips are read out).

Looking at the "bigger picture" (quite literally), I can see the 383 being a timesaver if it avoids (or cuts down) on mosaics. What the 383 can do in one pane in 3 hours, the 314 would take nearly 6 panes and 12 hours.

[ollypenrice]

I'll measure the download time of the 11 meg around new year when the instrument of choice (a calendar) should be readily available...

lly

[perfrej]

Folks, I have completely forgotten about the effect of micro-lensing on QE. Truesense sells the monochrome version of the KAF-8300 (M) with and without micro-lenses. As these do focus the light better, they eliminate the sensitivity disadvantage that comes from anti-blooming gates. In addition, it comes in no glass, clear glass and AR coated glass versions.

QE 54% (micro-lens, clear glass)

QE 60% (micro-lens, no glass)

QE 56% (micro-lens, AR glass)

QE 37% (no micro-lens, clear glass)

All of these are specified at 540nm. Now, obviously SBIG makes use of the AR version since they specify a QE of 56% according to the graph below. Note that Ha has almost 50% QE.

KAI-11002, Olly's pet, comes in similar versions with and without micrlo-lenses. Which one is in your 11000-camera? The Atik ones use the AR glass versions but they use grade 2 sensors. Manually checked, yes, but it is not the grade one version without column defects. They are honest enough to tell us so and they state that the price premium is not worth it. I subscribe to that reasoning because it is, after all, not a top brand with top prices, so you - the customer - can choose something that doesn't cost an arm and a leg and still get going.

The thought of writing yet another "book" on how to choose your CCD comes to mind...

/per

[perfrej]

Ive been reading this with interest, but I do have to question the response curves posted by Per (great though it is). Im pretty sure the ICX285 has a QE about 10% higher at 650nm than the 44% quoted on the graph. Surely its more sensitive than the 8300 in Ha?

If the 8300 sensor in question is AR glass with micro-lenses it has almost 50% (like 48 point something) QE in Ha! See my post above.

/per

[Uranium235]

If the 8300 sensor in question is AR glass with micro-lenses it has almost 50% (like 48 point something) QE in Ha! See my post above.

/per

Thanks Per food for thought

It might be worth me asking Atik whether the 383 does indeed have microlenses and AR glass as its not mentioned in the camera specs. Cor blimey, its 700g - a bit on the heafty side when compared with my current camera(s).

[Gina]

Good heavens, yes Almost as heavy as my super-cooled DSLR

[Uranium235]

Just checked the 383 manual and the window is quartz glass with BARR (anti-reflective) coatings. MIcrolenses not mentioned (but I have a feeling it does/should have them).

Does the 383 respond to the subframe/ROI command in Artemis? I dont quite fancy having to wait for a 1x1 bin full frame image to download just for focusing.

[perfrej]

Most cameras support sub-framing and thus download them in a jiffy.

/p

[Uranium235]

Most cameras support sub-framing and thus download them in a jiffy.

/p

Lovely

And the good news is that its not out of range to get a new one if I flog both my two 314 cams.

I wonder if there is anyone else who made the leap from 314 to 383, did they regret it? Sounds like im trying to talk myself into getting one...lol

[jnc71106]

i do know someone that went from a 314 to 383 and he didn't regret it, from memory he said it wasn't as sensitive and was noisier but nothing you would call a deal breaker

[Uranium235]

i do know someone that went from a 314 to 383 and he didn't regret it, from memory he said it wasn't as sensitive and was noisier but nothing you would call a deal breaker

Hmmm... if its not a million miles off in noise/sesitivity (say for arguments sake, -10%), then in the grand scheme of things I could just expose for 10% longer. And if im already taking subs of 600s (my most often used sub length), 10% is just another minute. Cant see that being a big problem.

The only other outstanding issue are the filters. There is no way I can afford a 2" filter set, so ive got to find a way of butting the camera up to the manual FW as much as possible. Either by way of a continuous male t-thread, or modifying a manual FW so it has the same short camera-side connection that the EFW2 offers.

[ncjunk]

I almost bought an 8300 but it went pear shaped.

Just for your amusement-

I saw the ad and sent an email asking about it and waited for a reply. I didn't get one after an hour and thought i better just send an email saying i want it anyway just to be sure. Well, two emails arrived after over 2 hours. 1 which the seller had sent 10 minutes after my original email saying it was still available then a second after my second email saying it was sold.

His first email got stuck somewhere between his server and mine as the header showed it left his server at 8 entered mine at 10:30 (half an hour after his email saying it was sold).

So although emails are great just remember they can go horribly wrong sometimes when you least expect it.

Feeling thoroughly cheesed off, as if i've just received socks in all my presents for christmas when i wanted an action man, i've gone and bought an atik 428ex instead from first light optics.

So I flip flopped right to the end!

Sent from my GT-P5110 using Tapatalk 2

[NickK]

Just checked the 383 manual and the window is quartz glass with BARR (anti-reflective) coatings. MIcrolenses not mentioned (but I have a feeling it does/should have them).

Does the 383 respond to the subframe/ROI command in Artemis? I dont quite fancy having to wait for a 1x1 bin full frame image to download just for focusing.

Yes - the 383L will only send/download the subframe requested. Additionally if the target is bright enough you could use preview mode for additional speed (at least till you get to the rough focus point.

My usual way of focusing is binning 3x3, 1/4 sized subframe, preview on - this means the downloaded image is very small.. then switch to 2x2 and 1x1 before going non-preview. It can do 255x255 binning .. but 8x8 is a good "find a target quickly" mode too.

I remember Billy indicated that improvements can be had by targeting the subject at the 1/3rd of the way down the sensor..

The 4xx series will also do binning/subframe/preview too.

[ollypenrice]

I presume I have the microlenses on the 11000. There is more to this that just the boost in QE. The downside can be artefacts arising from bright stars. Now I have no way of experimentally confirming that this is the source of the occasional artefact but I think it is. An example; same scope (TEC140) same chip (Atik 4000 but one OSC and one mono) and same target in the same week. The OSC threw up the odd diff spike when, on the same target, the mono didin't. So I presume that the Bayer Matrix was the culprit, not that it was a big, big deal. Or, shooting Lum on the HH region, the Atik 4000 produced no artefacts whereas the 11000 produced a semi circular halo on one part of the image.

The 11000 is an utterly enormous bucket and is remarkably fast. 2.5 hours per panel in Ha goes deep. And relatively speaking it ain't expensive.

After calibration there are no colmumns. I'm happy.

Olly

[Uranium235]

Yes - the 383L will only send/download the subframe requested. Additionally if the target is bright enough you could use preview mode for additional speed (at least till you get to the rough focus point.

My usual way of focusing is binning 3x3, 1/4 sized subframe, preview on - this means the downloaded image is very small.. then switch to 2x2 and 1x1 before going non-preview. It can do 255x255 binning .. but 8x8 is a good "find a target quickly" mode too.

I remember Billy indicated that improvements can be had by targeting the subject at the 1/3rd of the way down the sensor..

The 4xx series will also do binning/subframe/preview too.

Thanks Nick, thats cleared a few things up nicely. Good to see it can go up to 8x8 a well, 6x6 what I use for checking that im in the right place before framing (on Ha targets).

I managed to bodge together a zero distance male/male adapter for the camera > FW connection. I found an old male/male adapter with an unusually long thread, so I popped it in a vice and took a hacksaw to it. I now have a contiunous male thread 7mm in length which will allow the camera nosepiece to be directly connected to the FW with no space wasted.

So, now ive made the adapter I just need to get the camera to put it on.... easier said than done!

Neil, good luck with the 428 - it will bring you much joy

[IanL]

Just thought I'd throw this in for a bit of fun, hot of the press after a lot more head scratching and coding:

More on how you can roll your own comparisons over here: http://stargazerslounge.com/topic/191202-imaging-toolbox-updates-compare-cameras-scopes-mosaics-and-more/