|
PyMovie
Anyone have expertise with PyMovie and untracked asteroid data?? Since the Astrid records in .ravf format that Tangra doesn't handle, I'm forced to use PyMovie.? I've used it before for tracked recordings without issue, but I'm clearly missing something for these untracked files.? When I try to make a finder image, I just end up with unaligned images (stars are lines).? There likely is just one checkbox or similar I'm missing (likely has a big, flashing red arrow pointing to it that I'm not seeing ? ).? Unfortunately, the YouTube tutorials are 4 years old, and the software looks quite different now.
?
?
Steve
?
|
|
Re: Occult4 301
Steve:
I re-checked and the path shown on OW Cloud for (2845) and the one generated via Occult4 are essentially a precise match - down to the size of an individual house. But then they should be... Both most likely originated from the same program - Occult4. Yes, I now get a 97% chance for you using your updated position. Too bad you will probably be clouded out (I know that feeling well). George |
|
Re: Occult4 301
George,
?
My observatory is at 41.74685 N, 77.31678 W.? There was an error on an old post-it I had on my monitor saying the latitude was 41.81352, is that what I sent previously?? If so, sorry!
?
?
Steve
?
|
|
Re: Occult4 301
Steve:
Thanks. I decided to do a cross-check of this event comparing the "initial" Occult4 MPC-based path with the "refined" Occult4 JPL-based path, and also the path shown in OccultWatcher Cloud (these are attached). Conclusion... The path (and 1-sigma bounds) for this event on OWC is in very close (if not exact) agreement with the path derived after calling up JPL elements in Occult4. The "initial" MPC-based path produced by Occult4 is notably south of you and has much larger 1-sigma bounds. BTW: Maybe I input the latitude & longitude of your observatory wrong, but when I enter those numbers in Google Earth, I get that your observatory is located in a forest somewhat north of Wellsboro, PA. Thus the close "Miss" of the path on the JPL-based Occult4 output - and the resulting 0% chance for you (partially due to a very small 1-sigma). As for the compiled document on the Occult4 discussion...... I'm not sure I'd like for this to be spread too far - it was a bit hastily written and I'd rather not invite arguments by folks who want to nit-pick at it.? I wrote it for our small NEUS group. That said, I DO stand by the methods outlined and results - especially when the steps outlined in "301" are done. Way too many cloudy nights, George |
|
Re: Occult4 301
George,
?
Thanks for this--all of these have been very useful to me.? I've compiled together most of your "Occult class" emails into the attached file (did a few very minor edits)--if it doesn't bounce for file size issues.??
?
I did a quick test run using the Horizons method outlined in your last one, and an excellent event was found for just after midnight Mon/Tue night.? (2845) Franklinken occults a 9.9 mag star with an 8 mag drop for as long as 0.7 seconds--centerline basically right over my observatory (96%).? Weather forecast is dicey, but not without hope.? Unfortunately this one doesn't go anywhere near anyone else on this list with the possible exception of Kia (Kia--not sure where you are in Cleveland area, but this one is straight west-east, centered on Willoughby).? It would be a no brainer for a remote deployment for me a couple of km away if it wasn't a worknight (plus the dicey weather).? Instead, I'll try to run both the Astrid set-up and observatory to show I can get identical results for confidence building (my own confidence!).
?
?
Steve C
?
|
|
Weather
Since my monthly astronomy column for our local paper (the Wellsboro Gazette) for December is about astronomy weather, I went to weatherspark.com to look at historical weather--specifically cloud cover--for my area (historical data from Elmira-Corning airport).? I've used that site for years, since cloud cover impacted some aspects of my day job when we did field tests.? Winter is by far our most cloudy time here in the PA Wilds region--generally from early December until late February is the bad time, with the worst in mid-January.? We have clear skies (<20% cloud cover) less than 10% of the time during this period, and mostly clear (<40% cloud cover) less than 20% of the time.? At our peak in late August, those numbers go to 35% and 50%--a huge difference.? It looks like George has me beat, with a few percent even less of each in his area.? Those of you in western NY are about the same as me, while the rest of you seem to be much better--especially in winter.
?
If I add up the percentages and ignore time of day, I find that the chances of my weather causing me to not capture an occultation event are "only" 45% in August versus 68% in January.
?
Weatherspark now has year by year cloud cover data going back many years for my area that also shows cloud cover by time of day.? The one pattern for my area that popped right out to me while doing that was in late summer and early fall, we very very often had near complete cloud cover in the 3 AM to 6 AM time period.? I'm guessing this is our fall morning fog we get here.
?
?
Steve C
?
?
?
|
|
Occult4 301
This is the required final course in Occult4 prior to graduation.
Up to now (see prior topics "Occult4 101" and "Occult4 102"), what was stated about generating predictions using Occult4 reads like a mystery novel - without the concluding chapter. ???? This is that final chapter. (I had hoped to discuss this in a Zoom session prior to a written post, but my attempts in organizing a Zoom session... haven't worked out). In those 2 prior topics/posts, perhaps you noted 2 things: 1) The paths obtained as a result of your initial Occult4 search probably didn't match what you find in OW(C). 2) I mentioned a couple times that there is a final step you must do in order to get paths based on JPL elements The previous Occult4 101 / 102 topics covered how to get a listing of potential occultation events. In order to do this, Occult4 must have a source for asteroid orbital elements -- for "ALL" asteroids. JPL/Horizon doesn't publish such a catalog, but the MPC does -- MPCOrb.dat. Thus, our ""initial search"" was based on orbital elements from the MPC. Using the MPC catalog we were able to obtain an output in the "List and Display" of Occult4. But again, if you compared these predictions to what you see in OW(C) in most all cases you will find that the paths don't match. What we have to do now is to take the list of potential events that Occult4 produced (based on MPC elements), and obtain predictions for these same events, but based on JLP/Horizon elements. In short,... the resulting JPL-based paths are considered to be "more accurate". In most (nearly all) cases, the predictions you will then get should 'match' the prediction you see in OW(C) -- as in most cases the paths in OW(C) are derived from Occult4 (unless they are "special events" and/or Lucky Star predictions, etc...). [Note: I do all my searches for 1 day at a time - just my personal preference. The following assumes you have generated an initial output of potential occultation events - based on a search for 1 day - and this is showing in the "List and Display" section of Occult4]. Here's the steps you now need to do to obtain accurate path predictions (which are based on JPL-Horizon elements): * With a list of events showing in "List and Display"... Go to File, then select "Create 'UserAsteroid.csv' file from JPL-Horizons, for all listed asteroids" * 2 windows will pop up. Select "Yes" and "Yes" * Occult4 will then make an Internet connection to JPL-Horizons and download orbital elements for the same asteroids shown in your list, but the elements obtained will be JPL-Horizons based (and generated to the "epoch of date"). * As these are downloaded, Occult4 automatically saves them in a "User File" (the download is complete when the green progress bar disappears). * You can now "X Exit" the List and Display * You should now have the "Main Search Window" showing (if not, call it up) * As shown in the "Main Search Window", you now check the box that says "use USER file of elements" (see attached image) * Note that in the yellow box in the upper left corner, both dates displayed must be the same day you entered when you first generated your initial "List and Display" * Leave everything else the same as when you generated your initial "List and Display" * Click on "Search" * A Search will now be done using just the asteroids written to the User File. Since this is a small sub-set of asteroid elements, the search will go quick. * A new output will appear in "List and Display", and all the predictions will now be based on JPL-Horizon elements. These are the predictions you should use as they show the most accurate paths. (If you click on an event it should call up a path map and text in the map should say "JPL" and not "MPC"). * If you want Combined Magnitudes shown in the list, be sure to check the "Combined" box and then press "List events" again. * You can now save this list if you wish (File --> Save listed events) The "Graduate Course".......... There are some further "refinements" you can make in doing your predictions, but I've typed enough. What's stated above should yield accurate paths to plan your occultation attempts. I'd be up for discussing anything said prior, and going over some further details/refinements you can make in your predictions in a Zoom session --------- if anyone wants. Happy Shadows (IF it's ever clear), George |
|
Re: First "Real" Astrid Try (Long)
Including Mark's provided hardware, shipping from Canada, PayPal fee, and the 3 parts you buy yourself from Amazon or similar, I think I ended up spending about $480.? That doesn't include the focal reducer and spacers--I already had those.
?
?
Steve
?
|
|
Re: First "Real" Astrid Try (Long)
Steve:
Big Thanks. I saw that GitHub page and figured that might be it. Never worked with Raspberry Pi before so this will be "interesting". ? ...And I've put off delving into the world of WiFi configuring and setup (I've always MUCH preferred hard-wired connections), but, Alas..., I think ---? 'The Time Has Come'. ?????? I see some studying ahead......... I figured those chips had a 'window' over the actual sensor, but not so far above the sensor as to really diffuse any dust shadows. Ie., No real 'elevated' optical window like you find on the QHY or other expensive/cooled cameras. I do know that you work with spacecraft cameras, so I should acknowledge that. Thx! G |
|
Re: First "Real" Astrid Try (Long)
You are right about the marking, I hadn't noticed that.
?
Operating System load instructions are here--you can do half of it ahead of time (the first 5 major bullets in his list) if you want to:
?
Build the SD Card following the instructions here:
using the image file here: ?
I was worried about these steps--I tend to be software challenged.? But it turned out to be easy--even for me.
?
Both the Astrid and Runcam CMOS devices have windows. Typically the windows are really thin (0.4 mm for some of the ones we use at work), and close to the CMOS active surface--so they really do cast intense shadows I know.
?
?
Steve
?
?
|
|
Re: First "Real" Astrid Try (Long)
Steve:
Thanks. Interesting. Pictures are worth 1000 words. It reassures me a bit in the size of the ASTRID sensor (that it has a "usable" FOV). BTW: The numbers "296" can (sort-of) be seen on the sensor board - so you do have the Sony IMX296 sensor. While I'm writing (and waiting to hear from Mark about an ASTRID for myself),.... I gather the "Operating System" for the camera is on the Micro-SD card, and nothing but images end up (or are on) the larger Flash Drive (thumb drive). How does one get the operating system on the Micro-SD? It's not clear in Mark's assembly info on GitHub? BTW: Be careful about dust on the exposed sensors. You may get an "unexpected" occultation - especially when doing a drift-thru! I think this is something few folks realize! (You can 'test' for dust on a sensor/camera that has no optical window a bit in front of the sensor by doing an occasional flat field). Attached is what a flat from my NE2 looks like. George |
|
Re: First "Real" Astrid Try (Long)
Here's a photo of my Runcam (I'm not sure what version I have--not the most recent as it does have the 3 levels of integration) and the Astrid.? By eye, I measure the sensor area of the Astrid at 5.5 x 4.2 mm and the Runcam at 7.1 x 4.0 (prior to ROI windowing).
?
?
Steve
?
|
|
Re: First "Real" Astrid Try (Long)
The size/pixels I gave were from a posting from Mark--but it was pretty early in his development process.? Perhaps it was updated?? I didn't order anything other than the standard.? The number of pixels you quote makes better sense (note my confusion about binning).
?
?
Steve
?
|
|
Re: First "Real" Astrid Try (Long)
Steve:
Thanks for some more info. Please post more on your experience with your ASTRID. Well I finally found a moment to do a little more research on the ASTRID. A major concern is still its sensitivity as compared to cameras we have used commonly in the past for occultations, but that will eventually be tested... Another concern is the sensor size and resulting FOV - and how that may compare to cameras I currently own. A specific concern is... "will the ASTRID make locating the FOV, and finding Target Stars difficult?" So I did a little more digging. I was wondering if I could get a rough idea of how the FOV might compare to my RunCam NE2AE and my QHY174M-GPS. It seems it might be good to start by comparing the sensor to the one in the RunCam NE2. In scouring the Net about the NE2's sensor, there isn't much info. It's very scarce!... but here's what I could find: * Mono sensor * Rolling shutter * Pixels are ?? x ???? (Can't find this info anywhere!!!!!!) * Sensor resolution is [reduced] to 640 x 480 in NTSC. (It's actually a 16:9 chip reduced to 4:3 aspect) * This results in 307K pixels? (~1/3 Mpixels) * Sensor Image Area is 7.18mm x 5.32mm (diagonal of 8.93mm) (4:3) * ""Optical Size"" is 1/1.8 * Binning... 1X (tho the camera has an electronic Zoom mode) * Exposures are ~1/30, 1/15, and 1/7.5 sec * Max fps is 30? (actually 29.97) From Mark's Info on the ASTRID (specifically about the camera sensor) on Github..... ???? https://github.com/ChasinSpin/astrid/blob/main/docs/Camera.md The info on the ASTRID's sensor is as follows: * Mono sensor (The data sheet says Color, but Mark supplies a Mono version - it's the Sony IMX296LLR-C (Mono)) * Global shutter * Pixels are 3.45um x 3.45um * Sensor resolution is 1456 x 1088 (4:3 aspect) * This results in 1.58 Megapixels * Sensor Image Area is 5.023mm x 3.754mm (diagonal of 6.3mm) * ""Optical Size"" is 1/2.9" * Binning is 1X only * Max exposure is 15.5 seconds * Max fps is 60. So, in 'short'...... the sensor on the ASTRID looks like it's about 70% the size of the RunCam's in width, height, and diagonal. The star chart program I use (and love) is Carte du Ciel. The program allows me to input multiple FOV's for several scopes and camera combinations I use. I already have selections for the NE2 and the QHY in it. I always use the QHY withOUT any focal reducer on my 14.5-inch f/6 Newtonian - as I also do astrometry with it and I don't want any lens distortion on the edges of the FOV (ie., coma). Not using a focal reducer also allows me to obtain finer measurements since the image scale is 'better'. (I can measure positions to about 0.15 arc-seconds accuracy routinely). For the NE2 I use the supplied 0.5X focal reducer and 2 spacer rings (and only use the NE2 for occultations). My intent is to also use the same 0.5X focal reducer on the ASTRID - with a number of spacers that will give me the same amount of focal reduction as I get with the NE2. That said,.... Since the ASTRID's chip size is 70% of the NE2's (as best I can ascertain), I can input a setting into Carte du Ciel and see how it compares to the 2 other cameras. The result is in the attached image. Essentially, the ASTRID -seems- like it should give me very close to the FOV height of the QHY (NO focal reducer), tho the width will be reduced since the ASTRID is 4:3 and the QHY has a 16:9 imaging aspect. I can find Target Stars very easily with the QHY, so it seems I won't have an issue with the ASTRID. But this brings up another concern......... When I use the NE2 on my 14.5-inch, since I am using the 0.5X focal reducer, I must use the "PVC 1.25-to-2 inch" adapter (that IOTA came up with for the NE2) to reach focus. The lens of the focal reducer then falls 'below' my focuser - about even with the inside of my tube. If I use the ASTRID, I might expect that the lens of the focal reducer might fall in about the same location. I CAN place the focal reducer attached to the ASTRID in a 'similar' adapter, but it's questionable if the ASTRID (with it's large body) can slide far enough in the adapter to reach focus. So I might have to eliminate one of the spacer rings, which will reduce the focal reduction, which will also slightly reduce the FOV of the ASTRID. So in 'short'.... Reaching focus with an ASTRID """might"""(?) be an issue if you use a scope other than a refractor with a large focus travel range. 'Guess I'll find out. (Worse case, the mirror on a newtonian can be moved forward slightly toward the secondary). NOTE......... This is based on (often scattered/sparse) info I could find on the Net - however reliable. Please double check all data listed above before taking these rough calculations as fact. George |
|
Re: First "Real" Astrid Try (Long)
I have not done head-to-head sensitivity testing.? The pixel sizes are so different, with the Astrid's 1.55 um pixels, it really is better matched to short focal lengths.? I run the observatory C14 at about f/4, 1600 mm, which would be 0.2 arcsecond per pixel unbinned.? The overall chip size is 6.3 x 4.7 mm.? Looks like i have been running it binned 2x2, though I don't remember setting it up that way
?
I did do a quick timing check of it yesterday looking at my SEXTA test box.? Looks to be right on to the 2 msec resolution I can test to, and the duty cycle of the exposures is close to 100%.
?
I have done everything over WiFi with it, have not tried to do anything over the USB.? But you can look live, and it seemed to be broadcasting images the entire time it was counting down to the event recording.
?
BTW, the video file recording of .rasf type limit you to PyMovie--Tangra (and I assume LiMovie) can't handle that file type.
?
?
Steve?
|
|
Re: First "Real" Astrid Try (Long)
Hi Steve:
I was hoping I'd hear something about your experience with the ASTRID. Please post more 'adventures' (info about its performance). I have one on order, and Mark said he was testing one which he might get to me, but that was a week ago.... Haven't heard anything since. One MAJOR concern I have is SENSITIVITY.... what can it record as compared to a QHY174M-GPS and/or RunCam NE2AE -- same scope, similar exposure? ...And I'm wondering a little bit about chip size - how it might compare to a RunCam. (Haven't looked up chip size to compare yet). Yes, it does seem to be a great camera for unattended mobile work, and your test was a good simulation. Too bad you didn't stay up that extra 1.5 hours to just check on it (and adjust the focus). A general question.... If I'm using the ASTRID in my home observatory, I gather you can monitor what the ASTRID is doing (see the FOV) by having a laptop connected (via USB) to the ASTRID as the event is being recorded. Yes?/No??? ....Or is it simply better to connect and control via WiFi?? [I try to always use a hardwired connection if possible - just a habbit]. It will be interesting to see how it performs in quite cold weather - say, -10dF (or around 0dF)? -- or even colder. This, (and mobile work), is one of the main reasons I ordered one - I tend to temporarily "retire" the expensive QHY during the very cold winter months. During winter I mainly just use my RunCam(s). BTW: You once mentioned/asked about a QHY sequence to automate the recording. I'm attaching a screenshot of such a sequence I programmed. Works great for unattended captures. Just modify the event time and maybe the capture duration. Works in SharpCap 4.x onward. I went after 10 potential events last night (rare clear night). Got 10 Misses. Went to bed at 3am. Temp 11dF. Keep us posted on your ASTRID adventures! George |
|
First "Real" Astrid Try (Long)
?
I tried the Astrid in prepoint mode last night on 2001XA226, a low probability but bright target at 02:14 local time.? This target likely would have just been in the trees for my observatory, so I used the portable Astrid setup on the back porch.? Set-up consists of:
?
- 5" Celestron SCT on a manual Voyager mount
- Meade 0.33x reducer adjusted to about f/4 with spacers
- Astrid
- Heated dew shield with Pegasus controller, running at about half power
- Celestron Power Tank Lithium Pro (158 W-hr), for both the Astrid and dew shield
- Galaxy S7 Tablet
?
I loaded the event into the Astrid from OWC in advance.? I left the plate solving search parameters wide with respect to focal length (I hadn't recalibrated the focal length since I changed the BFD spacing).?
?
I focused first on the moon, then Jupiter's moons.? I then pushed the mount to the general prepoint location.? I commanded the prepoint routine, and was somewhere in the 5 to 10 degree range off in each axis initially.? I followed Astrid's direction arrows and distances, using the knowledge that the Voyager hand wheels are about 3 degrees per turn.? In about 5 adjustment cycles, I was within 5% of the center--more than good enough.??
?
I set it for automated data collection.? Note that the suggested recording durations based on I assume estimated time error and event duration were narrower than my preference, so I manually edited them to be longer.? Took me three tries to get these new times to "stick", I think I wasn't updating and reloading them properly.? This was about 2 hours before the event.?
?
I waited about a half hour, did a quick recheck on the prepoint (that you can do anytime without any effort).? The pointing was still great, but the focus was poor as the telescope was cooling down to the 20F temp we had here last night.? I tweaked the focus to correct, and went to bed--this was about an hour and a half prior to the event time.? I also checked the box for Astrid to power down after the event.
?
This morning when I checked it, Astrid fan and the dew heater were still running.? However, I was not able to connected to Astrid over the WiFi.? Tried both my tablet and laptop without success.? I assume the fan continues running when power is applied, even if the Raspberry Pi has shut down.? So I know I can get at least 8 hours from the battery I used (most of the power goes to the dew heater, and the battery indicators were between quarter and half).
?
Took the USB stick out of the Astrid, plugged it into my laptop with the latest/greatest PyMovie installed.? File chosen, loaded up, times downloaded.? But the images initially looked blank.? Moonlight and/or clouds was my initial assumption.? But looking closer, I noticed there were hugely defocused star tracking across the screen.? They were large enough that you could see the SCT's secondary obstruction, at least 20 pixels in diameter.? They were so large/faint that I don't believe it would be worth the large effort to analyze the data.? There was a relatively bright star drifting through center at event time, likely the target.? It likely would have been a good data set if the focus held.
?
So I am assuming that the less than an hour of equalization time, plus the continued temperature drop overnight was the cause of the defocus.? Any two mirror telescope will be much more prone to this due to the secondary magnification, with is about 5 for the C5.? If I am remembering right, defocus due to change of distance between M1 and M2 goes as the square of the secondary magnification (I am ignoring the focal reducer).? Maybe this is a driver to use refractors for prepoints?? Even though the refractor tube is 2 to 5 x longer than the SCT's, that length just acts linearly--so I think the math would say a refractor will end up being 5 to 10 x less prone to this issue.? Anyway, next timeI will likely try to give it more equalization time, plus maybe insulate and heat the tube itself.
?
There is one other Astrid issue if you intend to use an Android tablet to control it (if you will use a laptop or IPad, no worries).? It requires a long description, if this applies to any of you, let me know and I'll fill you in.
?
Quick initial impression--the Astrid is an amazing device.? In a perfect world I'd want it to have a bigger sensor with larger pixels, but all-in-all you aren't making a mistake getting one if you want to do multiple stations.
?
?
Steve C
|
Steve C
George Viscome
