Testing out new EEVA rig ...

[vlaiv]

Well, if you can call it that way, but yes, I believe it will be interesting EEVA setup.

First introduction to setup. I don't say this often, but this is one sexy looking setup:

Scope is SW Mak102mm mounted on AzGTI head and pillar that sits on Berlebach adapter (3/8" to SW mount tripods) and HEQ5 heavy duty tripod. Camera is ASI178mcc (cooled color version) mounted at prime focus.

Due to poor weather, this is only daytime testing so no conclusive results (well, some are - ones that will not work ). First baseline shot - sensor mounted at prime focus:

Target is roof of a house that is about 315m away (according to google maps / satellite view, and measure tool). Roof is indeed green (dark green - image is not properly color balanced).

This image shows full FOV, debayered using super pixel mode for effective resolution of 1548x1040 and scaled down x3 (to 33% size).

At full zoom (1:1 pixel) it looks like this:

It is rather cold outside and I was shooting from balcony and yes, there were thermals both from my own house and houses in between, but I think this image is pretty much indicative of sharpness so we can take it to be baseline.

First test was simple x0.5 GSO focal reducer in 1.25" variety. With this sensor, my aim was to get something like F/4.5. That would be reduction factor of about x0.346, so more than "prescribed" x0.5. I eyeballed distance to sensor so no precise reduction factor (one of things that I wanted to test out was reduction factor at certain spacers combination) and here is full FOV:

Pardon the orientation - I was to excited that it worked at all on that distance and that I had no issues with focus so I completely forgot to properly orient the sensor.

It is hard to judge sharpness here because depth of field is evident because speed of setup is indeed increased. Here is section of the roof at 1:1:

Now this is not smack in the center of FOV and roof is tilted and for that reason I believe that it is depth of field rather than off axis aberrations.

Length of tiles in both images are: 165px in prime focus:

and 65px with reducer:

Ratio of the two is ~x2.54 or as reduction x0.394 - very close to target reduction. With this distance setup is at F/5.12 and additional spacer of 5mm should push it down to F/4.5. Nice.

Next thing that I wanted to test out was eyepiece projection and I decided to try out "reducer mode" eyepiece projection method (as opposed to "proper mode"). Unfortunately, achieved reduction is way too large. Eyepiece used was 32mm and sensor was placed at about 27-29mm - that creates somewhere between F/2 and F/1.2 system (probably around F/1.6). This creates all sorts of havoc on optical performance of system (illuminated field is only about half to third of sensor size, and sharpness is well, very non sharp ).

Here is full FOV at this setup:

I did not debayer this image - just made it appropriate resolution (bin x2 of raw data). That is true mess. Let me see if I can pull out reduction factor and 1:1 image.

Actually it's not that bad:

Contrast is suffering, but that is because this is very aggressive reduction and this scope is not properly baffled! It really needs to have rather long dew shield to remove issues with baffling. You can see that there is bright outer part of the illuminated circle - that is due to unfocused light going thru front corrector plate and ending up on focal plane (very far off axis for normal use) without hitting mirrors.

Anyway, tile length here is about 27px, so reduction factor is about x0.164 and that applied to F/13 system gives F/2.12 (so distance is about 27mm). To be usable configuration, I would need to bring it forward 6mm (to about 21mm) and that is simply impossible with eyepiece projection adapter that I have and this camera model.

In the end, there are couple more things to try out for results - proper mode eyepiece projection (with distance to eyepiece of about 43mm - so T2 extension of 16mm will be needed - I can get 15mm, that is 10+5mm) and afocal method with CS lens for this camera (about 12mm lens will be just right).

And night time trial as well - that one is probably the most important

If everything checks out - we will have very interesting beginner setup that can do it all - visual for planets and DSO and planetary imaging and DSO imaging (in form of EEVA rather than proper imaging, but I also plan to test this scope with a bit more serious sensor - ASI1600 mono + filters - to see what it can deliver if data is processed accordingly).

 

[geeklee]

Nice write up @vlaiv As always, look forward to hearing about your experiences and output on this setup when it's used 🙂

Edited December 29, 2019 by geeklee

[vlaiv]

Small update on this thread.

I managed to test out proper eyepiece configuration, and I'm must say I'm rather underwhelmed by how it performed.

Just a bit of background first - I used TS eyepiece projection adapter:

It is standard item that has compression ring (and 4 screws) and it is quite wide inside. In fact there are three models I believe - for different eyepiece barrel diameters. I ordered one to fit my 32mm plossl (by GSO).

Adapter fits fine except - there is an issue with it. Eyepiece barrel has a slope at the top:

while adapter fits with flat surface. This makes EP hard to center. With adapter - you get 3 different self adhesive band - and no manual, so I just took one that fits best and inserted it into adapter. This makes EP body fit nicely within adapter but does not remove issue with tilt completely (although it is reduced because you don't need to slide adapter all the way down).

This creates repeatability problem - it is hard to position camera at proper / same distance each time. It is also a bit problematic to center eyepiece in adapter as you have 4 screws that you need to "balance" (pretty much like collimation of scope / adjusting finder or similar). In any case - here is EP with adapter attached:

Centering is done by making sure - there is same amount of space between eyepiece top and sides of adapter (eyeballing it).

Other than centering and positioning issues - adapter works quite well. It easily carries camera attached to eyepiece and everything feels sturdy enough.

Attached to scope it does create balance issues but for this test AzGti was not complaining:

I ran some indoor tests with this configuration - where distance between eyepiece and sensor is provided by extension T2 ring and at that time 15mm extension seemed to provide wanted reduction factor when used in "proper" EP projection configuration. In the meantime I forgot which extension I used, so I started first with 10mm.

This pushed system to the limit - I'm not sure if I was in fact able to properly focus as I reached end of focus range when I took test image:

At this close distance - reduction factor is rather severe.

I then remembered that I had actually determined that 15mm extension is proper one, so I switched it and here is what FOV looks like with that:

(images are binned x2 but not debayered - hence mono look. I plan on using super pixel mode so actual resolution will be same as this - although this image is reduced to 33% to show whole FOV and save on upload image size).

This is center crop. According to measured tiles I in fact got x0.25 reduction factor for F/3.3. I'm not sure why is this - either I was wrong in my indoor tests (although I tested it on a ruler and had 2.4cm map to 0.89cm which gives x0.37 reduction or F/4.8) or there is simply issue of repeatability of adapter. At these extreme reduction factors even 1mm of distance change can have dramatic effect on change of reduction factor.

I was also less then impressed with edge of the field in this setup. Maybe at F/4.5 - F/4.8 things would be different, but I'm not holding my breath for that. Here are roof tiles at the edge of the field (I just slewed scope so that tiles that were in focus in center of FOV become positioned in corners):

other corner

I shot two images in opposite corners just to avoid this being tilt issue due to adapter centering - but both corners show same thing - blurring is so severe that we can distinguish tiles any more.

This shows that proper ep projection configuration is also - "no go" in this case.

What remains is afocal with 12mm lens - I just need to wait for lens to test that out.

[vlaiv]

While testing proper EP projection today, I also managed to test x0.5 reductor again in more detail.

- I added 5mm extension ring to get closer to target F/4.5 instead of F/5.1 from first post.

- I also wanted to check level of vignetting in this configuration

- in first test, image seemed a bit blurry so I wanted to check sharpness of this setup - both on axis and in corners (because EP projection showed severe degradation in corners).

Here are results:

Reduced image shows nice definition across the field. Tile measurement shows reduction factor to be around - x0.36, or F/4.7 (closer to target F/4.5) - here is image of tiles cropped and 100% zoom (but super pixel debayer - not debayered but rather just binned):

I found setup a bit difficult to focus, and I don't think it is because it is fast. Maybe there is some issue with tilt - as nose piece with reductor can't be fully inserted into 1.25" receptacle on scope - there is baffle on tube that prevents this. I might try threaded connection in the future.

I checked corners as well (slew mount - no change in focus until tiles that were in center of FOV are now in the corner), and they seem to be as sharp as center:

and other corner

Here is vignetting:

It shows that field is not quite centered on sensor. This is 5% steps - white is 100%, and corners are at 55%. Actually right side is at 45% because vignetting is slightly moved to the right on sensor. I was shooting at cloudy skies and some of center illumination is actually shadows on clouds that were in focus rather than true flat field - but this gives idea of what vignetting is - it indeed falls of to about 50% at the edges and flat fielding is necessary.

I wanted to get idea of how sharp this setup really is - so I turned the scope to very distant high contrast target:

This is top of TV broadcast tower - it is roughly 10.7km away. Here is what 100% center crop looks like:

Now this might seem like rather good image of something that is 10.7km away. It might even look like it is due to ground seeing and that you can't really get good sharp image of it.

That is in fact not true - this scope is very sharp. I tested it at night - it is collimated properly and it gave very sharp images of Moon at x236 power.

I decided to take again this tower in prime focus. Here is image at 100% without binning / debayering. Bayer matrix is still visible in this image, but presented like this - it is actually at critical sampling for this scope - best resolution that this scope can provide - and it shows

Yes - antennae elements can be clearly seen (and sharp? right?) all along the central column. In fact I think I can make out ladders used to climb to top of this tower at the back side. Again - this is 10.7 km away. This little scope is truly sharp

In any case - wanted to compare the two so I reduced image at prime focus to rough size of that taken with reducer:

Difference is clearly visible - left is prime focus reduced and right is image with reducer. Reducer introduces quite a bit of blur.

I wanted to see what sort of blur we are talking about here, so I tried gaussian blur with different sigma, and it looks like sigma of 1.9px is about right:

Now I know what sort of simulations to run in order to asses what sharpness will I get out of EEVA images.

I know - it is far easier to just go outside and do EEVA session instead of running sims - but one needs clear skies in order to do that

BTW, I'll try threaded connection to see if it will improve things and also reversing lens in the cell. I did that once before when I used this reducer with F/6 scope - and it helped. Maybe at F/13 better image will be from lens facing other way around?

 

[JBracegirdle]

Your test setup is really interesting, similar to what I'm trying to achieve.

I tried eyepiece projection on the moon using my Meade Zoom eyepiece as it already had a T thread, and would allow variable image scale. But I found it only  really had one really sweet spot, maybe it was a focus problem but it didn't seem to offer anything over prime photography.

I'm wondering why use eyepiece projection?

[vlaiv]

2 minutes ago, JBracegirdle said:

Your test setup is really interesting, similar to what I'm trying to achieve.

I tried eyepiece projection on the moon using my Meade Zoom eyepiece as it already had a T thread, and would allow variable image scale. But I found it only  really had one really sweet spot, maybe it was a focus problem but it didn't seem to offer anything over prime photography.

I'm wondering why use eyepiece projection?

Main idea is to create affordable setup that can "do it all".

I noticed that many people come and ask for advice for a scope that can basically do it all - as they often put it: "I want to observe both planets and deep sky objects, and I want to be able to record what I'm seeing - to take a picture of it". Some give a bit better explanation of what they want to image - but point is - they are limited in budget - usually up to 1000 of dollars/pounds/euros (I guess it is mental barrier of four figures spent on gear).

Of course - you can't accomplish that in said budget or with one scope, but I wondered if one is ready to cut some corners - what sort of setup it would require and what sort of budget would cover it.

Since I already own some bits and pieces of such setup, and I fancied idea of having a small Mak as grab&go lunar scope - I got myself new setup for both purposes - to test out "do it all" budget scope and to have lunar scope for quick peek.

In lunar role - this scope is beyond my expectations. Mount works good enough to track the target and hold the scope. Setup is less than a minute, and scope delivers sharp views beyond x200 power. What more does one need for grab&go lunar?

In the mean time - I'm fiddling with EEVA with this setup as that is the key for "do it all". It will aid observing in light polluted areas, but it will also provide base for "take image of DSOs" that I see (well they will see them when doing EEVA). In order for this scope to be good performer for EEVA - we either need large sensor ($$$) that will provide large FOV or if we go with small sensor CMOS camera that will be both for planetary imaging and EEVA/DSO imaging (and cheap) - we need means to exploit all corrected and illuminate field that scope is providing - that means some sort of focal reductor. We also need quite aggressive reduction if we want to get widest field possible out of this scope (and make a good match in resolution vs pixel size for available cameras that have pixels in range of 2.4um to 3.75um).

I've identified 4 different means to achieve focal length reduction:

- EP projection in "reducer" configuration ( focal plane of scope is at sensor and eyepiece lens acts as focal reducer - see first post - I could not achieve this successfully because I could not get sensor close enough to eye lens of eyepiece with this adapter).

- EP projection in "regular" configuration ( focal plane of scope is at focal plane of eyepiece and eyepiece is again bending - reimaging at sensor). I tried this approach today and while it works and you can "dial in" required/wanted reduction - edge correction is disastrous - there is so much blurring at the edge of the field.

- Regular reducer - very limited choices there given that most reducers are T2 or 2" and we have 1.25" option only here (maybe I could look for T2 reducer?). This seems to give most usable results so far.

- Afocal imaging. This uses eyepiece the way they are meant to be used - focal point of telescope is at focal point of eyepiece and eyepiece produces collimated beam at exit pupil. Lens is then used again to focus that light onto the sensor. This combination is the worst in terms of number of glass surfaces - but will potentially provide best correction / best sharpness because both eyepiece and lens will be doing what they are designed to do.

Problem with that configuration is - attaching everything together and matching eyepiece focal length and lens focal length and getting lens that is good enough (there are a lot of security camera / industrial type lenses out there that match sensor size - but most are basic and not well corrected lens - usually marked as 2MP - for this application we really need sharp lens - marked as at least 6MP or higher - 10MP, and of course lens performance needs to match the label   - and still be cheap / affordable).

That is left to be tested and I'll test that as soon as I get lens to test with.

This was rather long answer, short one is - we use EP projection as one way to get focal length reduction because we want wider field and better matching of pixel size to focal length with this scope and camera I'm using.

 

[JBracegirdle]

Thanks Valiv,

Your explanation is great. I did some experiments today with a 1.25" Antares focal reducer and my Orion StarMax 90. I tested it on a road 1km away. It worked really well at 0.65x to 0.43x at 0.35x it got a bit soft round the edges, but I thought still acceptable. I've not uploaded the images as they are the front of people's homes, even though they are on Google Street View.

I think your project is a great idea and hopefully will help many beginners.

[vlaiv]

1 hour ago, JBracegirdle said:

It worked really well at 0.65x to 0.43x at 0.35x it got a bit soft round the edges

This is really valuable. Next on my list was to test what sort of reduction can I get without sacrificing sharpness. Our setups are quite alike - only difference is that my scope is 102 and not 90mm (Antares reducer is same as GSO, just rebranded).

At about x0.45 reduction we are talking about F/5.85 - at one point I thought of getting it to F/6 and seeing how sharp it is - so that sort of matches. At 600mm it will be 1.65"/px after super pixel mode, and I'm afraid that is slow for 4" and OSC sensor - that means another bin x2 after for resulting image size of 750x500 and 3.3"/px. That is ok but a bit small image size - maybe if it is sharp enough it will be decent when resized to fit computer screen?

ASI183 type chip would be more suitable as it would provide different "magnifications" - but I feel it is too expensive for basic setup.

[vlaiv]

Here is another update.

I figured out that I have small CS lens that comes with guide camera - 2.5mm focal length. This lens is not suitable for my ASI178 camera in many ways (one of which I figured out during the test today) - first it is made for 1/2.5" sensor size, while ASI178 is larger - 1/1.8" format. This means that anything that lens picks up will be concentrated at center of the sensor with "extreme vignetting" (if it can be even called that).

Second thing that I just found out - due to mistake, is that ASI178 cooled version requires C mount lens rather than CS mount lens. This means that only C-mount lenses can be used on this camera model. I was under impression that ASI178 cooled also uses CS - but that was my mistake as I did not bother to find specs for cooled model and assumed that regular non cooled model has same housing - not true.

Here is diagram for 178 from ZWO website (marked with red arrow is 12.5 distance that corresponds to CS flange distance):

Here is diagram for cooled version. My version is earlier than this one as well - it does not sport removable 2" nose piece like in this diagram, but distances should be the same:

Again - marked in red are important dimensions - 11mm + 6.5mm = 17.5mm - flange distance for C-mount.

This means that my camera / lens could not be focused to infinity, and was in fact focused to much closer distance (like macro mode with regular lens - if you move lens further from camera it can reach focus at closer distances). This in turn means that eyepiece was not operating as it should - properly focused - producing collimated beam.

In any case, here is result of experiment:

It of course has severe vignetting - both in illuminated zone (this is due to 32mm EP having 27mm field stop and scope having 21mm rear baffle) and huge area of sensor being completely in the dark - that is because of 1/2.5" lens used on 1/1.8 sensor - only 6.4mm diameter is going to be illuminated out of 8.9mm diagonal sensor - that is inner 72%.

Reduction in focal length is much less than I expected. Something like x0.8. I have no idea why is this. Maybe because lens was focused so close instead of at infinity? Another thing is that regular pixel per angle formulae probably don't give correct results for such short focal lengths as 2.5mm?

I was about to order 8mm fixed focal length lens for 1/1.8" and one varifocal lens 4-12mm for 1/2" sensor - I simply can't find suitable varifocal lens for 1/1.8" sensor in hope that one of them will be suitable, but now I'm having second thoughts about this approach. I'm not happy with results, but on the other hand - setup was not as intended - CS lens was used with C mount and possibly quality of result is hugely dependent on this.

I will also need to find some sort of "short" CS thread / T2 adapter in order to make everything work. I actually needed to push lens even further about 1-2mm in order to have enough T2 thread to make everything via threaded connection.

Back to so far preferred method - focal reductor. As luck would have it, I had a chance to test this setup with artificial star . Sun was in such position that something on TV tower gave reflection for brief period of time. This allowed me to evaluate "spot diagram" of center and edge of the field.

I was planning on doing two tests - threaded connection for focal reducer and reversing the lens (because I reversed it once before and I wondered if it would work better with original facing) - but as this artificial star opportunity presented itself - I took it to see what sort of aberrations I can expect.

First center of the field (well actually around 1/3 off axis):

This looks rather good - it is round and not overly large. I would be happy of having stars like that. Seeing was rather poor, and this is best frame out of 15 or so frames I took in fast succession.

Now edge of the field performance:

We no longer have single point but rather a streak in direction of center. This looks like tangential astigmatism, but as we will see later, it is very much to be expected in this setup.

Other corner:

Same thing, again - in direction of center of FOV.

I would be rather discouraged by this if it was not for the text I read few days ago:

https://www.telescope-optics.net/miscellaneous_optics.htm

(section on focal reducers)

here is spot diagram for cemented doublet focal reducer:

We are interested in FLAT FIELD part of this spot diagram as we are using sensor that is flat. And a small reminder - above edge of the field is at about 0.5 - 0.6 degrees. I think that above spot diagram almost perfectly matches what has been recorded.

For better results than this - one should have better corrected / matched reducer (or perhaps slower lens corrected at infinity?).

I'm about to test second proposal - using slower lens with larger clear aperture to see if it will make a difference. Reducer that I'm currently using is 1.25" reducer that has about 25mm clear aperture and about 102mm of focal length - making it F/4 lens. Since scope has T2 connection - reducer that has up to 40mm can be used, but there is no one on the market. However - we need doublet corrected at infinity? How about 30mm finder? That should have something like 120mm focal length and have 30mm (again F/4 lens) - but we could stop that one at 25mm and with 120mm we would have F/4.8 lens? Or maybe keep full aperture and see the impact on vignetting. In any case - I don't believe this option is in line with original idea - having do it all scope with easily accessible parts as it would require a bit of DIY.

 

 

[shirva]

Great write up,, even better reading about folk trying different approaches to the hobby. Nice scope set up,,