Measuring optimal backfocus for EdgeHD OTAs

[randomic]

Celestron's EdgeHD product line features an integrated field flattener. As with all flatteners, these produce optimal results when the imaging plane is a specific distance from the flattener. In the EdgeHD whitepaper Celestron describe the optimal backfocus of 133.35mm (5.25") for the 8" model, and 146.05 (5.75") for the 9.25", 11" and 14" models. They suggest that the image plane should be placed within 0.5mm of this distance.

When putting together an imaging train it can be quite hard to determine the actual backfocus. You could add together all the optical lengths quoted by manufacturers, you could get calipers and actually measure each part or even try to measure the entire thing (although it can be quite hard to figure out where to measure from. At some point, you have to trust some manufacturer spec (unless you fancy risking your sensor).

Once all this is done you might, however, find that things vary ever so slightly; everything from the tightness of threads to the T-ring not quite giving exactly 55mm. How do you work out if you've done it all correctly?

In a table in the whitepaper (page 13), focal lengths are given for each OTA (for example 2125mm for the 8" model). Hypothetically then, it should be possible to measure whether or not you're at optimal backfocus by plate solving for your image scale. In the same table, an image scale is given for a sensor with pixel size 6.4 micron but you can use a calculator (such at the astronomy.tools one) to work out the expected image scale for your particular sensor. This does require that your image is as close to perfect focus as possible.

 

Putting all this into practice. I used my calipers to try to get the image train as close to 133.35mm as I could and then plate solved some resulting data taken with a DSLR with 5 micron pixels. From my calculations, if I'm in focus at the correct spacing, I should have an image scale of 0.485"/pixel.

However, my astrometry.net solves gave an image scale of 0.495"/pixel. Working backwards, this indicates that I was at a focal length of 2083mm, quite a way inside 2125mm. Although I can't find a reference I've read that, for an SCT, the focal length changes by approximately 3mm for each 1mm of backfocus, this implies that my sensor is 14mm too close!

Now, I'm no expert with calipers but I feel like I couldn't have been more than a few mm out, and if anything I thought I was too far. I suppose I could have been a bit out of focus but surely not ~10mm.

Is there a mistake in my logic of aiming for 2125mm focal length?

[globular]

I'm sure you know where to measure from... but just in case you don't it's from the red line in the below image:

With a visual back in place it can be hard to do this because it hides the exact point you need:

I suggest you measure the distance from the red to the blue line, add this to the 133.35 your'e after, then measure your actual back focus from the blue line (which is always visible):

I only mention this because it looks to be about the amount of back focus you're missing.

[Xplode]

I wouldn't trust the numbers given by any manufaturere to be 100% correct, FL varies a little, flattener-sensor spacing changes a little etc.
FL often seems to be just a nice "round" number, due to manufacturing differences it will of course vary a little, especially for mirrors.
For spacing i found it's best to always plan for having to adjust at least +/- 1mm from the manufacturers given specs.

Adapters/spacers should ALWAYS be measured without the male threads.

The correct spacing should not be judged by focal length or anything else than how the stars look.

[randomic]

2 hours ago, globular said:

I'm sure you know where to measure from... but just in case you don't it's from the red line in the below image:

[image]

With a visual back in place it can be hard to do this because it hides the exact point you need:

[image]

I suggest you measure the distance from the red to the blue line, add this to the 133.35 your'e after, then measure your actual back focus from the blue line (which is always visible):

[image]

I only mention this because it looks to be about the amount of back focus you're missing.

It's a good shout! I think one of the original Celestron documents had the distance indicated incorrectly in the diagram. I am indeed measuring from the red line. Maybe I just need to double check everything. I've got a ZWO camera on its way which should allow me to measure all the way to 6.5mm from the chip, it could be that my DSLR + T-ring is a ways off of 55mm.

Edited October 4, 2020 by randomic
snipped images for brevity

[randomic]

2 hours ago, Xplode said:

I wouldn't trust the numbers given by any manufaturere to be 100% correct, FL varies a little, flattener-sensor spacing changes a little etc.
FL often seems to be just a nice "round" number, due to manufacturing differences it will of course vary a little, especially for mirrors.
For spacing i found it's best to always plan for having to adjust at least +/- 1mm from the manufacturers given specs.

Adapters/spacers should ALWAYS be measured without the male threads.

The correct spacing should not be judged by focal length or anything else than how the stars look.

Yeah you're absolutely right, and if you look at any of Celestron's marketing material you'll see that all the focal lengths are quoted at 10x the aperture. Presumably because it's a nice round number to say it's f/10. In the whitepaper, however, it seems that they've actually done some measurements; the EdgeHD 800 for instance is listed at f/10.456

Fundamentally, as you say, the roundness of stars is all that matters!

[JamesF]

6 hours ago, randomic said:

Although I can't find a reference I've read that, for an SCT, the focal length changes by approximately 3mm for each 1mm of backfocus

I'd have thought that's dependent upon the focal ratios of the primary and secondary?

James

[randomic]

1 minute ago, JamesF said:

I'd have thought that's dependent upon the focal ratios of the primary and secondary?

James

Yeah I believe it's f/2 and f/5 respectively, I haven't looked in to the exact maths, maybe I'm closer than I thought!

[fredvanner]

As  relative novice about to start trying serious imaging with my EdgeHD 8:

This accurate position must assume a fixed position of the mirror - which I assume will be at or near one or other extreme of the mirror focus adjustment range.

I also assume some adjustment must be left for fine focus adjustment, which must be done with the mirror adjustment since there is nowhere to put another focuser in the fixed imaging optical train.

Where should the mirror adjustment be when setting up this imaging configuration (with the flattener in place)?

If I'm missing something, please explain what.

[randomic]

32 minutes ago, fredvanner said:

As  relative novice about to start trying serious imaging with my EdgeHD 8:

This accurate position must assume a fixed position of the mirror - which I assume will be at or near one or other extreme of the mirror focus adjustment range.

I also assume some adjustment must be left for fine focus adjustment, which must be done with the mirror adjustment since there is nowhere to put another focuser in the fixed imaging optical train.

Where should the mirror adjustment be when setting up this imaging configuration (with the flattener in place)?

If I'm missing something, please explain what.

The mirror should be in the position such that the in-focus image is formed 133.35mm behind the end of the OTA threads. Does that make sense?

SCTs can form an in-focus image at a wide range of image plane distances but, with EdgeHD 800, the image plane at which the integrated flattener performs best is the one at 133.35mm

Yeah there is some fine focus adjustment to compensate for things like the mirror shape changing due to temperature.

[randomic]

Just got my 462MC (just in time for good weather to grab some shots of Mars!). According to the diagrams it's 12.5 - 7.5 = 5mm from the chip to the flat surface above the chip.

So once I have everything put together, I've measured with calipers from the part of the T-adapter which stops against the end of the thread of the scope all the way down to that flat surface. To hit 133.35mm backspace, I should measure 128.35mm (if I add a 2mm thick filter then I can add about 0.67mm so can round to ~129mm).

If there's a gap in the clouds tonight I'll measure the pixel scale. If I'm still far off the numbers given in the whitepaper then I might shoot Celestron an email to see if they can explain why. Perhaps the design has changed slightly since the whitepaper was written.

[ArEnJee]

I had the same question in terms of expected focal length, as I don't think it's possible to rely fully on published optical paths or callipers, and I did put it to Celestron, but they wouldn't be drawn on a definitive answer, giving the 2125mm as a nominal figure only.   In the end I worked backwards from the published reference setup for the Celestron OAG, and put together an image train for my AA294C camera, which in theory should be close to the correct distance.  When I plate solve in APT I get a focal length of between 2087 and 2090mm for the 294C depending on conditions.  Oddly this is 10mm more than the focal length I used to get for my Nikon D7200, which in theory should be too long due to the Nikon camera body's larger back focus (46.5mm v 44mm for Canon).  This is what prompted my query to Celestron...