zernikepolynomial

On 16/07/2023 at 01:20, AstroPCB said:

This might not be relevant but what version of Sharpcap are you using? I read somewhere that you need to use the Beta version with the Player One cameras?  I am not sure if this would make any difference to the issues raised in this thread?

It was already tried, made no difference.

On 08/07/2023 at 20:55, Luke Newbould said:

Interesting thread! - Only just saw I got tagged in this so I'm sorry about the late response here.
I've not had the chance to run sensor analysis on the cameras I'm currently testing unfortunately, they just got unboxed, briefly tested then mounted on my scopes so I can't offer anything for reference I'm afraid - first light tests went great though at least!

Regarding the variance observed between P1's graphs and the ones in this thread, my gut feeling is that there's a few things at play:

 

First is likely a 'silicon lottery' issue - probably safe to say if you got 100 of these cameras on the bench to test, you'd get 100 very slightly different end results! (same issue as in other fields, PC overclockers run into this issue all the time for example)

Second would be the test-to-test output variance in Sharpcap's sensor analysis - be it either through slight variances in the test environment (unavoidable without a laboratory grade test bed I'd guess!) or just because of the general shot-to-shot 'noise' in measurements from each frame, something makes it so that SharpCap sensor analysis numbers don't ever truly match for 5 tests in a row - I reckon the graphs published from P1 could just be from a great sensor and then taking the best graph out of 10 for example?

 

Thirdly, environmental variables could be at play with equipment this sensitive and numbers so small! - testing on a day where you can practically feel the static crackling in the air? using a regulated, well earthed power supply? the test computer's earth connection? strong sources of unshielded magnetism nearby? - the list goes on, but there's probably something at play that's near impossible to anticipate too - worth consideration imo!

 

Hope that helps! 🙂

The tests I conducted were pretty rigorous and well thought out. Natural stable sunlight, flat ROI, grounded separately powered USB hub (also tested with direct computer connection, no difference), paper diffuser. Another user on CN conducted his own tests and actually came out with a very similar result to mine.

The smoking gun on why it has to do with LRN engaging properly or not is test 5. Test 5 followed its own complete curve, well outside the std dev of the other tests. Sean from Player One stated that if it was an instability, the curve would be unstable, with points at each gain having a wild deviation; this was not the case, and it looks like a true systematic change in test 5. The silicon can not become better on one set of gains in a row, but then go back for the other same sets of gain ranges. Again, if it was instability in the system, we would expect wild results at individual gains. 

This is why we need more tests from other people, because if many or some of the cameras are having issues engaging LRN properly, it might be a fixable thing. Might even be sharp cap issue with the camera, resolved with an update. As far as hardware, read noise tends to be related to the amplifier and ADC. One interesting thing to note, on mine and other peoples tests with the camera, engaging the cooler results in a very slight bump in read noise. There could be MANY reasons for that behavior, and so I am not worried about that unless it was more noticeable like this general LRN issue. 

I just want to let everyone know, that this is not a Player One bashing thread, and I just heard about someone on the cloudy nights forum having Player One do a very nice thing for them and offered a replacement as well. They do take the time to try and solve these problems. Its just mine is not bad enough to pay for shipping. The other person had a minor dust issue, but they still offered a replacement. 

45 minutes ago, Stuart1971 said:

I think the take from this thread and other posts I have read is that Player One over promise and under deliver on there camera data, this is not the first time I have seen this issue with there cameras, which is odd as Player one is owned by one of the original partners in ZWO, they split after a disagreement and one stayed at ZWO the other formed Player One, so would think the cameras would be on par with the ASI counterparts….but that does not seem to be the case….🤔

@Luke Newbould may well have some insight as he is a big advocate for these cameras….

To be fair, I generally see the Uranus-c perform better than the ZWO counterpart, and to be fair my Ares-c is performing better at certain gain settings than the ZWO counterpart. For example, as Adam pointed out, my camera performs better at 7k full well than a standard ZWO.

9 hours ago, ONIKKINEN said:

I think you are being a bit unfair here. There is nothing wrong with sharpcap itself and just advicing to "take some images with it" is just plain silly when OP has a genuine worry here. I would imagine manufacturers also measure camera performance using exposures and measuring the pixel values within (like sharpcap does). These technical specs cant be seen in a single sub frame, at least easily so i really dont think looking at a subframe has any meaningful value here. But the small technical differences can be measured and they can have a real effect in integrations you would not have been able to otherwise see. Im sure you know all this...

Definitely not a panic over nothing.

However @zernikepolynomial if you look at your graph and the published graph they do look similar in many parts. Sharpcap has skipped gain 125 and gone directly from 100 to 150 which is a problem if you wanted to know what the stats are for gain 125 but im not sure there is a way to force sharpcap to use specific gain values for the graph. Both 100 and 150 look like fairly accurate measurements compared to the published values so i think it may be possible the read noise drop actually does happen at 125 but its just not shown. The actual amount of read noise is slightly higher, which i think is a bit concerning but it doesn't look too bad. The "low read noise" modes are something i dont really trust myself. My Rising Cam branded IMX571 camera has a low read noise mode, but there is no documentation on what this actually means. I do notice that the framerate drops exactly by half when this is engaged, leading me to think that this is some sort of double exposure internal dark calibration thing, which would not be useful at all for long exposure imaging where this probably gets disengaged. This would not be apparent with sharpcap measurements since the exposures used are very short hence my distrust for it.

If we assume your camera is not using the "LRN" mode (whatever that is), then at gain 150 you only have a discrepancy of 1.49 to 1.417 which is not too much. Still a bummer that its more than it should.

My results are not similar to Player Ones LRN mode. My results are only similar to the Normal mode both on the Player One website and the results I just published, except at gain 0. I still believe something happens after it jumps from gain 0 and messes up the results, because it randomly started working properly in test 5 of the published results. The way LRN works is based around the way they integrated memory into the camera, and its an easy to miss bit published on the Player One product page. 

Anyway, I agree that the camera is still useable, just underperforming and disappointing. 

7 hours ago, Adam J said:

Yes his worry is genuine but that doesn't mean he should be worrying about 0.1e of read noise or less by your own estimate. The maths just doesn't support that, 0.1e is not going to have a real effect in integrations especially for long exposure DSO imaging. When I have time I will come back and prove to you with some modelling that 0.1e read noise simply has no effect as the other noise sources are more significant. As above this is certainly withing normal variation from sensor to sensor. 

But if you don't believe me then take a look at Doctor Robin Glover's (yes the man behind sharp cap) presentation on read noise from the 2019 practical astronomy show, it's on you tube. You would spot the back of my head in the audience if you knew where to look. One of his conclusions is that read noise is so low in general in modern sensors that it is no longer a significant contribute to overall signal to noise ratio, so long as your exposures are not crazy short like 10seconds for braud band. 

here is a link: 

39mins and 30 seconds shows the effect of read nose from different cameras on overall signal to noise. 46mins companies 2.5e read noise to 7e read noise and basically even for that huge difference you get the same signal to noise as long as you expose for 120s or more. 

In the end it's just not going to make much difference and it might not even be a real difference at all. Just a difference in how read noise is being calculated. For example outlier handling of hot pixels etc. 

For short exposure stuff like solar system imaging it will matter more. But the IMX533 is more often a DSO camera. 

Adam

Adam, I have watched that video multiple times in the past, and I have actually myself derived his equation for calculating optimal exposure time based on a desired relative noise level and the cameras noise characteristics, as well as light pollution. If you run the calculation yourself, you will see that going from 1.38e to 1.5e results in an increase of (1.5/1.38)^2 = 1.18 times the original exposure length to reach the same noise target (18% longer). That is significant, and it does warrant investigation. Now, will that be noticeable in the images? It depends. If your doing long exposure astrophotography (fewer exposures, but longer) then no.  If you are doing short exposure astrophotography, then it can make a significant difference.  Also, if your signal is very weak, it can be VERY noticeable. All that information is in the video that you yourself have referenced.

9 hours ago, Adam J said:

So first thing first. Are you aware that ZWO and Player One also ( unless I am mistaken) do not use Sharpcap for their chart generation. So what you have is three different measurement methods in play here. I think playerone have chosen a more generous method of measurement as your figures are not far from the ZWO figures for the ASI533mc pro. You say 1.38e at gain 200 7k full well. ZWO are saying about 1.3e at gain 200 and 5k full well. So a little lower but at what looks like slightly higher relative gain. Now these difference between Normal read noise and LRN mode are small as things go at unity gain you are talking about 0.11e between the two modes. I would expect the read noise variability from sensor to sensor to be larger than that tiny difference. If it was my camera the difference between 1.38 and 1.5 would not bother me at all and you won't visually detect it in your results that's for sure. 

But as above I think the bigger issue is you are comparing apples and oranges if you are looking at those charts and expecting them to perfectly match results from sharp cap. 

So I think you should stop testing the sensor in sharpcap and take some images with it. There are many other sources of noise beyond read noise that everyone forgets about and added together make your little differences in read noise insignificant in terms of total noise. Just expose very slightly longer if it is really eating away at you and your will have negated the effect anyway. 

This is a panic over nothing. Go take some images. 🙂

Adam

Adam, you could not be further from the truth. My discussion and tests with player one were long and well documented, all you had to do was ask. 

First thing first, Player One did also use sharp cap to compare their test camera to mine:

Player One's test camera in normal mode:

Player One's test camera in LRN mode:

Then I compared my results to theirs:

An example of my LRN mode without cooling:

An example of my LRN mode WITH cooling enabled:

I conducted many more tests, some were clearly erroneous results and not even consistent with the gain curve. I perfected my experimental environment and took several tests with the exact same conditions, I then compiled them into a graph comparing them to Player One's results (Sean [removed word]):

As you can see in the results, I was able to establish that LRN mode suddenly started working in one of the tests (test 5), but I was not able to reproduce it despite having the same exact conditions for each of the tests. In the tests on the above graph, I utilized stable natural sunlight and paper as a diffuser. I also checked to make sure the ROI was on a perfectly flat illuminated area, so systematic error was not having an effect. After this, Player One said the results were very strange via email, and offered to take the camera back. When I accepted the offer, they informed me I would have to pay shipping, but the costs were too high so I reconsidered. This went against their 30 day policy, but they claimed it was not a sufficient enough performance failure to warrant paying the shipping cost, and I found that to be understandable to some degree, but still a bit disappointing because Player One acknowledged the result was unusual and made me do extensive tests without coming to a conclusion.

On 24/06/2023 at 01:12, AstroFin said:

Following this thread with great interest since I also bought an Ares-C from Player One. Haven’t had a chance to test it yet though… 

While they did offer to take the camera back, Player One decided to go against their 30 day policy and somehow simultaneously claimed it was not a true "performance failure" to pay for shipping and still weird enough to take it back. Shipping costs back to their factory are way too much, and I would rather save for another camera than pay for that. Not to mention the risk of sending anything back to china just disappearing if it takes too long. 

You might have better luck, but anyone else should wait until you see more performance results before buying from them. My results may in fact be the true average, which is not much better than a typical 533mc camera from other manufacturers when taking into account how it responds to gain. Mine must be over 1.5e at unity gain (125), 16k full well. Can hit up to 1.5e at gain 150, 12k full well. My sweet spot will probably be at 1.38e at gain 200, 7k full well, or maybe even whatever I can get down to 4k full well. It will just be a balance of not overexposing with my fast newtonian operating at f3.75, while taking the shortest exposures possible due to its weight.

The canon 6d is a great way of tackling slower scopes with larger image circles. Scopes like the EdgeHD 800, while not able to entirely fill out a full frame, still benefits from the extra size of both the sensor and the pixels, and the relatively low noise. I only sold my canon 6d because I decided to move away from those kinds of scopes (think EdgeHD, Rumak, ect..).

It is not a driver issue. It appears to be a sensor issue, and I will wait to hear back from Player One on monday about what to do.

Player  one responded back and it looks like it definitely could be an issue engaging LRN. His results are identical to mine in normal mode (LRN turned off). So he wants me to try not using ascom.

I just got the new Ares-c from player one, and there is some kind of read noise issue going on. Its almost like the low read noise mode is not even working, and I have to set the gain to 200 just to get a comparable read noise to the graphs on the product page. 

The read noise at unity gain is supposed to be around 1.35-1.38e, but I only reach that at gain 200, otherwise I am well over 1.5e at unity gain which is way out of specification. 

Can anybody help me out? I have consistently gotten these similar values on 3 tests.

Most of it seems pretty good. The springs for primary collimation dont seem great, but I think most people know that. Its finder shoe requires a ton of tightening to make it secure, but it can accept a second screw, optionally. Spider looked good, and it was not bent. Secondary seemed robust, and it is adhered rather than enclosed, so I don't think mirror pinch will be as big of a problem. Bottom ring clamps on too tight, making loosening the rings to rotating the scope very difficult; this makes it poor for visual use.

But the biggest con/issue is the loose paint on the inside. I talked to Agena, who'm talked to GSO, and they say it is not a defect. GSO states it its overspray. So be warned when you get this scope, or get these scopes from Agena, because you will not be able to exchange or use the warranty if it has this paint issue. You may have to void your warranty to fix it. 

I cloned your project and decided to make a few adjustments to make it look like the first version. I think the latest version was not rendering correctly. Anyway, I am going to print out a PLA prototype at my local shop. 

What kind of screws for the standoffs did you use? And print method/material did you use? Did you use self tapping screws and FDM print method?

I was thinking maybe heat inserts could be put in the bottom holes. I think the top holes would be too delicate and it would ruin the standoffs.

1 minute ago, malc-c said:

Can't really help as this forum is predominately UK based, and the UK has different consumer laws than the US.  Here in the UK there is a binding act passed in law that gives consumers protection even if something is outside of the normal 12 months warranty period.   

I am so frustrated with iOptron, they gave me literal crap as a replacement to a scope that came with a faulty focuser, and spit in my face when I tell them its got obvious damage on multiple parts. I would accept this scope for a used price, not a new price. Anybody else have advice on what to do?

I live in the United States, and I received a clearly damaged iOptron Mak 150 from iOptron. It was a replacement scope for a scope that had a damaged focuser.  I still have the warranty on it. 

They are refusing to honor the warranty despite it having a chip in the secondary baffle shield/skirt, scratch on the secondary mirror, and leaking adhesive/scratch on the side of the secondary mirror. I am not too worried about the adhesive/scratch on the side of the secondary mirror, but everything else seems to be clearly against their warranty, and I just want a replacement without this kind of damage. They will not honor the warranty, and even acknowledged the chipped secondary baffle, but are claiming the scratch is not a scratch, when it clearly is. I asked other people if they noticed damage with their older versions of the same scope, and they said NO.

I am an physicist who specializes in adaptive optics, and I have joined to give you my experiences and expertise in telescopes. 

I am not sure what you did to the corrector plate around the crack. It looks like residue marks from something you put on it (maybe more cleaner?). Regardless, I think you should put some super glue right or glass adhesive onto the end of the crack (small amount), and if it sinks into the crack a bit, it might help it from getting bigger. Next, you should collimate it and see if the crack has permanently deformed the entire corrector, if it has, replace it. If its just causing a small amount of obstruction, then the scope will be fine because obstructions in the plane of the aperture (at the corrector and primary) become buried in the PSF (point spread function) due to the fourier effect of telescopes and lenses, and it will only reduce the contrast slightly.