Observer's report - Joe Cali


Total Solar Eclipse, August 1st 2008

Observers report from centre line NE of Jiayuguan

by Joe Cali

   See individual exposures of the corona at many different shutter speeds
   See composite images where all the individual exposures have been stiched into one long continuous image.
  See fascinating time-lapse photography of the Moons shadow sweeping across the sky
   Return to "Totality in the Gobi Desert" Home Page


 On July 20th at 1:30am, I arrived in Beijing having flown from Australia.  After a short hold up at the airport while customs decided if they would let me import the telescope in my suitcase, they said it was ok and let me go on my way.  My friend Bengt Alfredsson was on his way from Sweden and arrived in Beijing 12 hours later.  We spent ten days blistering our feet,  pounding the pavements in Beijing, Xi An and Shanghai travelling under our own steam. On July 31st,  Bengt and I joined up with the Eclipse City / Eclipse-Reisen 48 hour eclipse program including a charter flight to Jiayuguan ex-Shanghai.  


 Forbidden City


 Terracotta Warriors

 Forbidden City Beijing  

 Park of peoples Beijing

 Terracotta Warriors, Xian

 Beijing  Beijing  Beijing  Beijing
 CCTV Tower Beijing    Beijing Chengxiang Huamao Shopping Centre    Shopping Centre downtown Xian  Zhongda International Xian


We rented a pair of suites in Shanghai & spent five days looking around.  On the first night we walked along the street near our apartment building. Bengt looked at a pizza joint. I said, "Don't even think about it, we are eating local tonight!" We walked into a hotpot restaurant. Neither of us spoke a word of Mandarin or Shanghaiese.  It was a typical Shanghai hotpot restaurant -  a small stove in the middle of the table -  order the ingredients and a pot of stock and cook it yourself.  The menu had no English translation and no pictures.  So I said to Bengt, "What if I just tick a few things at random?  How bad could it be?" Bengt just raised his eyebrows.  After 15 mins a waiter arrived carrying a large stainless steel pot about 40cm diameter.  He set it down and we looked in.  To our horror, we saw a bowl full of fish heads floating in grey coloured water sloshing around in what must have been about 40 chopped red hot chilies. Bengt didn't say a word, he just looked at me, raised his eyebrows smiled and nodded. After the initial panic, I tasted some of the vegetables and seaweed that were lurking beneath the surface.  It was really very tasty.  Later, a waiter brought us more of the ingredients I'd randomly selected. These included wonton noodles, brocolli mushrooms root vegetables and seaweed.  It looked horrendous and it was very very hot but it was really very tasty.  Next night however, we tried the pizza restaurant!


 Pudong Riverfront

 Nanjing Rd Shanghai

 New World SHopping Centre  Nanjing Rd

 Pudong Riverfront from the Bund  

 Nanjing Rd, downtown Shanghai

 New World Shopping Centre, Nanjing Rd



After a few days in Shanghai, we made our own way to the airport to meet the Eclipse Reisen group at midday on July 31st.  I recall that at Shanghai airport they didn't have a conveyor belt to load luggage into the plane.  I watched as two baggage handlers threw my suitcase from the ground level up into the plane. Great! Thanks guys. Air China's bag handlers had already been busy progressively destroying Bengt's travel pack during the trip. In Xian, we resorted to sewing up the damaged section of his bag with a bootlace.  My suitcase contained an 80mm refractor OTA and an equatorial mount, mounting tripod and camera tripod.  We arrived in Jiayuguan at 9:00pm and went straight to a downtown hotel to an EC welcome cocktail & briefing.  After the briefing we went out to the square to watch a cultural stage show put on to celebrate the eclipse. We returned to the hotel tired and after a high tech shower I went straight to bed. At 1:30am hotel reception called to remind me that they would be giving me a wake up call at 6:30am.  Yep, thanks for that, I said wondering why, in the name of holy snapping fluffy ducks, they'd wake you up at 1:30 am to remind you they'd be waking you up 5 hrs later.  Bengt told me they did the same to him.

I awoke to my alarm at 6am. We went for an early breakfast. By this stage, I had not eaten for 24 hrs. I missed both lunch & dinner the day before. Just as I was leaving my room, the phone rang - my wake up call. Back to the room, grab the suitcase and camera bag. The suitcase contained my telescope, mount, my Manfrotto tripod Bengt's Manfrotto tripod and a couple of folding stools given to us by EC. It added about 6kg to the suitcase's usual 22kg weight. We were on the bus and on our way by 8am. I'll skip the fine details. During a busy morning we visited the Great Wall's first fire tower in Jiayuguan, the Jiayuguan Fortress and the Jiaquan wine spring. Following a late lunch we made our way to the eclipse site.

 first tower



 Jiayuguan First Tower  


 Jiayuguan Fortress


We drove north east from Jiuquan. While on the bus, our Chinese guide read out a statement from the Jiayuquan administration stating that we were not permitted to use GPS for the purposes of mapping. This received rousing applause from the passengers many of whom were using GPS to track the bus at the time.  Most didn't take it too seriously but approaching the checkpoint, there seemed to be much activity switching off bluetooth and secreting GPS units.  


The Police checkpoint

Eventually, we arrived at the police roadblock & check point. Authorities closed access to the centre line to anybody who didn't have an Eclipse City pass. There was some discussion on SEML that it sounded to some like dirty tricks by EC.  But in fact, the authorities just didn't want any people wandering around the area. There was a spaceport and military installations in the area. EC managed to acquire access to just one site in our area and another further west both were heavily guarded - guarded to keep us corralled in the one area. Anybody could obtain a pass from EC. At the check point, the police checked that we had our all-important Eclipse City passes.  After much fiddling around, videos & photos with the police and other activities we cleared the check point. I heard a rumour that someone managed to lose their pass and was put off the bus but I didn't see this myself.  It was made abundantly clear to all of us that without these passes we wouldn't clear the checkpoint.

After the checkpoint, the GPS units came back out. The ban notice that had been read out specifically mentioned a ban on using GPS for the purpose of mapping activities. Fortunately most people on the bus were using GPS units for the purpose of tracking our movements, for timing or to pinpoint their observing location. Nobody was making maps so most participants decided to leave their units running.

We drove 60km NE of Jiuquan. In the north, a large bank of stratocumulous clouds threatened and we were driving further and further under them.  The only hope was that they'd vanish or blow away as the temperature changed. I'd seen this happen before.

We finally arrived on site about 90 mins before first contact.  Eclipse City (EC) chose this location about 2km south of the centreline and 1300m altitude because of its scenic aspect. If you were just watching the eclipse, the site was fantastic.  But our group included many keen eclipse chasers using equatorially mounted telescopes.  Some carried heavy cases for about 500 metres over soft sand to get to the ridge that provided the most scenic view.  I don't want to be too critical because I know that EC had to do a lot of groundwork with Chinese authorities but it would have been nice not to have to carry the gear so far. Although Bengt and I shared the carrying duties, we were trying to hump a 28kg case over soft sand.

Bengt Alfredsson's GPS registered the following coordinates at our site
 Latitude  40 o; 23 mins 21.3s
 Longitude  99 o;5 mins 02.8s
 Altitude  1325 metres

As you will see from my wide field images, we stopped well short of the ridge. Basically we were about as close to the buses as we were allowed.  A number of factors affected this decision. I was concerned about how long it might take to walk up there and how long it would take to pack up and return.  The view of the landscape was better up there but the ridge lined with observers held its own attraction.  Our guide incorrectly told us we had to be back on the bus at 730pm local time - 15 mins after totality. I doubted that anyone would return to the bus by 7:30 but by choosing a closer location, I could wait until I saw others walking back and perhaps work that little bit longer.

I have been developing a small lightweight single arm fork mounting since TSE 2001.  The entire dec fork arm can move along slots to balance the load around the polar axis negating the need for counter weights. It fits in my suitcase with the 80mm f7.5 refractor OTA and all my clothes for a grand total of 21-22kg. I assembled it over a 20 minute period. The was a bit of breeze so I collected rocks in a large plastic bag to weigh down the tripod and stabilize the mount.

All went smoothly until I attached the stepper motor and switched on the drive control. The indicator lights didn't light up the right way. I immediately started prodding & poking plugs and connectors. Then I saw it. I'd assembled and tested the drive at home in Australia and again in Shanghai but between Shanghai and the eclipse site, two motor wires were somehow torn out of the motor body. I thought everything was packed well enough but apparently not. I remembered seeing the suitcase thrown at the airport. Could that have been the fatal blow?  At the observing site with 20 mins until C1, there was nothing I could do but disconnect the motor. I have a "failsafe" slow motion knob on the other end of the worm so that, in case of drive failure, I can turn the worm by hand.

Some days before the eclipse, I calculated that at 9:58:30 UT(14.5 mins before totality) the azimuth of the Suns north limb would lie exactly west. I looked as the shadows grazed along the flat flange plate of the declination arm. This confirmed that the polar axis was indeed pointing north.  

The sky 15 mins before totality
The Sun played hide and seek with the clouds until 20 minutes before totality when the cloud
seen below and to the right of the Sun finally drifted away.  Looking at the direction of drift
I confidently said to Bengt that we were in the clear.  The other two clouds to the left, drifted
west & can be seen in the totality images beneath the Sun in the distance. This picture was taken
approximately 15 minutes before second contact. The Sun remained visible from this point until totality.

Observing program

80mm f7.5 ED refractor
The motor failure threw my detailed plans into some disarray as I quickly reassessed my photographic program. I decided to abandon planned time-lapse photography of the partial phases to concentrate on successful coronal photographs. The time saved would give me a chance to try a dry-run for the coronal photography sequence using manual operation of the worm drive. To operate slow motion, shutter speed dial and cable release requires 3 hands. So I needed to work out the best way to do it. I found that using the left hand for holding the shutter release and driving the worm and the right hand for changing shutter speeds was the way to go. The OTA is a Skywatcher ED80f7.5 refractor with a Pentax K10D 10MPx DLSR attached.  I hope to have completed an automated exposure control program and interface for this camera by next years eclipse but this year I was employing a capture protocol based around the camera's built-in 5 stop auto-bracketing.

A 600mm focal length and 15x23mm sensor meant that exposures longer than 2s were not necessary as the corona would be over exposed in the entire field. Exposures longer than 1s would blur due to the absence of a drive.

Task:  photograph the corona at all shutter speeds between 1/4000s and 2s.

Wide field cameras

Next I set up two wide field cameras on one Manfrotto tripod using a simple Aluminium RHS mounting bar.

Video camera
Panasonic VDR-D150 standard definition Video camera.  The camera has an excellent quality Panasonic 30x lens 1.9-57mm FL (35mm equivalent 35-900mm). The lens has low flare. On this occasion I added a Marexar Ultrawider wide-angle adaptor.  I bought this adaptor for my 35mm camera in 1979 when I was a struggling student and couldn't afford a real wide-angle lens. Although I have various wide-angle lenses, I continue to use this attachment for some applications.  Attached to the Videocamera  it creates a circular field, clipped top and bottom by the frame - 140 degrees wide and 90 degrees high.  

Task:  photograph the passage of the lunar shadow across the sky starting 2m before C2 ending ~2 mins after C3.

Still Camera (time lapse imagery)
The other camera mounted on the same tripod was a Pentax Z1 (35mm film camera) with a Pentax SMC-M 18mm f3.5 lens loaded with Fuji Pro H ISO 400 film.  This model camera has a built-in intervalometer. This is convenient for any constant exposure time series work.  I usually use it for automated photographs of partial phases before and after totality and for automated wide field capture during totality.  

Task : Take a time lapse, 1 exposure every 4 seconds starting 20 s before C2 and continuing for 36 exposures(144 sec).

Both wide field devices work unattended after I press the start buttons 2 mins before second contact.


Darkness Descends

Totality was approaching rapidly.  I looked at my "Atomic Clock."  It's a small white faced analogue travel alarm that I like to use for eclipses due to the ease of reading the face.   11:10:10.   Time flies.  I almost missed the start time for the wide field cameras.  I started the cameras on schedule but very nearly missed the window.

The shadow moved across from the northwest and pretty much right on time we were treated to a beautiful diamond ring. The clouds to the west acted as a wonderful projection medium for the celestial animation that was the approaching shadow. Watching this as I had in Libya, I mananged to repeat the error I made in Libya and didn't remove the solar filter.  

In Libya and again here I forgot to take off my solar filter and to add to my woes, didn't focus the telescope perfectly.  Watching the diamond ring naked eye and operating the camera by feel, I managed to do it again.  I didn't recheck the focus just prior to the eclipse and I didn't remove the filter.  I didn't look into the camera until 15 seconds had passed by which time I completed 40 exposures  between 1/4000 and 1/8 seconds. By the time I'd taken those 40 blank exposures, worked out what I'd done and rectified the problem, my first usable totality exposure was taken at 11:13:30 and I'd blown almost half of totality.  I'm going to invent a solar filter that takes itself off before totality ;-)

Despite making such an idiotic error, the auto-bracketing of the Pentax K10D is very efficient. I had rehearsed the exposure sequencing a number of times even with my eyes shut.  The value of doing this became very apparent.  In the remaining minute & a bit, I rattled off another 70 exposures. I completed my totality exposures at all shutter speeds from 1/4000 to 4s.  The reason I make errors like this is because I do look around during the eclipse. I operate the cameras by feel and don't stare at the camera the whole time. But this gives me the best of both worlds.  I do look through the camera to get a better look at totality. The camera has a right angle finder that gives me a great view.  The manual slow motion also didn't limit me much.  I allowed the camera to drift while I shot the inner corona up to the edge of the field and re-centred only for the outer coronal images where I needed the field positioning to be correct so as not to clip the corona. The longest (2s) exposure is blurred a little but this is only used for the very outer coronal streamers which are pretty diffuse anyway.
Bengt and I even had a short discussion during totality about whether we could see a "tunnel" or not.  We agreed that we could see an indistinct tunnel or coffee filter but it was very wide and low to the horizon.


Image taken at 11:13:45 close to mid-totality. In this image the coffee filter or tunnel can be seen wide and at a low angle to the horizon. The edges of the shadow were  about 100km away in each direction so it stands to reason that the angle would be low and wide.  By contrast, the observations and photographs of the tunnels taken in Australia in 2002 show the shadow edge much higher in the sky.  


In the comparison photograph below, taken near Cameron Corner by Bengt Alfredsson during the December 4th, 2002 eclipse, the edge of the shadow was only 10km away in each direction so its edge naturally appeared at a higher altitude.   The horizontal angle of view of both photographs is similar. Both were taken on 35mm film cameras using short focal length prime lenses.. Bengt used a Nikon 20mm While I used a Pentax 18mm Bengt's image has been vertically cropped. There was a whole lot of black below the horizon.


Totality & the tunnel, ©Bengt Alfredsson, 4th December, 2002.


Totality ended.  In my rush to complete the full shutter speed sequence, the diamond ring appeared while my camera shutter speed was set too fast but I was just happy to have rescued the situation.  

After some consideration, the appearance of the tunnel was probably to be expected.  The angle observed between the edges of the shadow and the horizon and the maximum height of the bright surrounding sky, can be calculated using simple spherical trig expressions. The height is directly related to the diameter of the eclipse path.  On August 1st, the edges of the lunar shadow were about 100km away in each direction and therefore lower to the horizon and less distinct due to diffusion by the air mass.  By comparison, the much narrower path observed from Cameron Corner in 2002 was only 20km diameter and only 10km in each direction.  As a result, the shadow edges were much higher in the sky and more distinct because they are not being diffused so much by the atmosphere. The observed angles of the shadow with the horizon were correspondingly greater. Therefore, the best observations of the spectacular tunnel apparition is best seen near the end of eclipse paths that have narrow path diameters.

This was my first attempt at video of an eclipse.  Auto exposure control worked very well. The wide angle view 140 x90 degrees is a good field of view. I however didn't think to turn off the camera's auto focus. The camera went hunting for focus at C2 and C3 and the focus it found each time became progressively worse.  Before totality the camera was in focus, during totality it was slightly out of focus.  After totality, it found the front element of the wide angle adaptor and focussed on that. Even with the slight out of focus during totality, if viewed on a small screen, the shadow progress has been recorded. The camera's audio recorded every exposure of my time lapse camera so that I could work out  the exposure times for those film images. On the track I can hear the shutter of my K10D on the telescope although with EXIF data, this is of less value. It also recorded our conversation and ambient sounds from the crowd.

Time lapse imagery 35mm film
The results of this project exceeded my expectations. I used my trusty old Pentax Z1 35mm film body loaded with Fuji ProH ISO 400 film. The Z1 has a built in intervalometer and I fitted the camera with a Pentax SMC K18mmf3.5 (ultra wide lens circa 1975). It may be an old design but has low flare and is quite sharp wide open except at the corners.  

Background.  Bengt Alfredsson has for many years been taking wide angle images of totality.  He uses the Nikon 9 stop bracketing facility in the MC21 remote control unit. Bengt shoots 4 sets of 9 stop bracketed exposures per 36 exposure film with a 20mm lens.  He triggers one at C2, one at C3 and the other two spaced evenly between during totality.  Using 9 stops of bracketing he gets a lot of discards but the technique guarantees he always gets some good images.  In 2002, he took a stunning image of  the tunnel that was later published in my eclipse chasing feature in Australian Geographic(Issue 78, pp 86-99). Much to my horror and dismay, the photo was wrongly attributed to me. If you see the article, the closing photo is definitely Bengt's.    

I don't have the luxury of choice. The Z1's intervalometer doesn't allow bracketing to be combined with interval shooting. I've noticed at previous eclipses in 2002 & 2006 that the Z1's auto-exposure provides perfectly adequate exposures for very wide-angle imaging when metering mode is set to centre weighted averaging. In CWA mode, the meter won't pay too much attention to the tiny coronal patch of light. Some evaluative metering algorithms might weight it more heavily.  In 2006 with the luxury of 4 minutes of totality, I triggered the Z1 camera manually while attempting to shoot a horizon panorama. The resulting images were OK but in the light of the results obtained this eclipse, I think this approach of shooting a time lapse with a fixed camera aspect is superior to a panorama or changing aspect.  

During this eclipse a mere 113 s of totality demanded a hands off approach.  I loaded a roll of Fuji Pro H colour negative film with its generous exposure latitude, I decided to use a fixed camera direction time lapse approach to depict the shadow transit. With only 36 exposures to play with (35mm film camera) some triage was in order. I considered photographing in the anti-solar direction but decided on a westerly aspect with the corona more or less centred. With the camera pointed at the Sun, there was no point allocating too many exposures after C3.  The shadow would be out of field quite quickly. So I started the camera at 11:10:25UT (2m21s before totality). There was a 2 min delay until the intervalometer started so the imagery began at 11:12:25UT or 21 s before totality. My 36 exposures at 4 second intervals took me through to 11:14:45 just after 3rd contact.


Time of Exposure for each frame [±0.5s]

Exp #    TIME UT(H:M:S)            Exp #            TIME UT(H:M:S)

1            11:12:25                                         20                            11:13:41
2            11:12:29                                        MID-ECLIPSE        11:13:44
3            11:12:33                                         21                            11:13:45
4            11:12:37                                         22                            11:13:49
5            11:12:41                                         23                            11:13:53
6            11:12:45                                         24                            11:13:57
C2         11:12:46                                         25                            11:14:01
7            11:12:49                                         26                            11:14:05
8            11:12:53                                         27                            11:14:09
9            11:12:57                                         28                            11:14:13
10          11:13:01                                         29                            11:14:17
11          11:13:05                                         30                            11:14:21
12          11:13:09                                         31                            11:14:25
13          11:13:13                                         32                            11:14:29
14          11:13:17                                         33                            11:14:33
15          11:13:21                                         34                            11:14:37
16          11:13:25                                         C3                           11:14:39
17          11:13:29                                         35                            11:14:41
18          11:13:33                                         36                            11:14:45
19          11:13:37              

The best way to view the results is to open the first image.  Keep your mouse on the NEXT link and keep pressing to shuffle the images through as quickly as possible.  I tried setting this up in a flash gallery but this works better. The flash changeover time is too slow. If you have a slow internet connection, go through the images once to get them into your cache, then repeat & you should be able to shuffle quickly through the sequence.

View the Time Lapse Sequence

80mm f7.5 Refractor Photographs

As already mentioned, the images of totality are compromised by loss of focus.  My first reaction was to think that I'd forgotten to focus the refractor but examination of the handful of partial images indicated that the refractor lost focus between an image taken 10 mins before C2 and another taken 3mins before C2.  Alexander Birkener and Ralf Schaefer mentioned that they had measured a very large temperature drop of 27 Celsius.   The measurement was in sun not shade and taken 1 m above the ground.  It now seems that the temperature shift may have caused or at least contributed to the loss of focus.  

On the flight out of Jiayuguan, I completed a basic image stack. I aligned all the images and created a set of masks to radially blend the images. Joerg Ackermann was sitting in front of me and very kindly didn't recline his airline seat which allowed me to use the laptop during the flight. By the time we reached Xian airport, I had completed a basic image stitch.  The technique I apply is the same technique set out in Russell Browns eclipse tutorial available on the Adobe site.

After applying the unsharped mask, the poor focus became obvious. The unsharped mask did nothing to improve sharpness.  So I tried an unsharped mask filter. No matter where I applied it, it didn't improve things.  There is a technique I sometime use to rescue images my photography students have taken when they haven't focussed on their intended subject or have too much camera shake.  

Step 1     Make a copy of the layer
Step 2     Apply a high pass filter to it. I used a 3 pixel offset.
Step 3     Set layer blending to overlay

This sharpening technique improved the apparent focus dramatically and has lifted the quality of the resulting composites from garbage to mediocre.  

I then used combinations of other techniques including alpha channel contrast masking, 3d render scripts, unsharped masking ( filter and spin ) to create the set of final images.  With limited fine detail available to me, I have played around in processing some images to show large structures (low spatial resolution) at the expense of fine detail.  
View the individual exposures in increasing exposure order

View the image composites

Looking ahead
Next eclipse is of course July 2009.  My development of the mount is almost complete. The refractor mount and my clothes all pack into a suitcase that weighs in at 20kg.  I need to upgrade the mediocre worm drive with a more accurate drive. It may surprise some but I've set up a numerical model for a tangent drive and determined that for a solar eclipse instrument, a tangent drive has some advantages.  A bisymmetric tangent drive is one that starts with the arms apart , the arms become parallel halfway through its operation where the drive  performs very accurately then drives the arms apart again. The advantages of tangent arms are found in lighter weight higher torque for the total weight of the drive and much higher drive accuracy during totality when it's needed at the expense of less accuracy during partial phases when shutter speeds are fast and accuracy isn't needed. The problem is that for a custom mount like mine, you can't easily buy very small lightweight worm gears that are also highly accurate.

 drive error

Bisymmetric tangent drive error. During a 10 minute window either side of the centre position, the error in the drive rate is better than ±0.01 arc
sec per sec.  Provided the drive is well made that translates to a drift of 0.1 arc sec during a 10 second exposure, a typical maximum exposure
during totality.  It is relatively easy to manufacture a high precision tangent arm, compared to a worm drive.

Beam Splitter
I have acquired a 1/10 wave 50% beam splitter plate.  I will be building a beam splitter box for the back of the refractor. Similar to a flip mirror box, it gives me simultaneous optical viewing and photographic acquisition during totality for the loss of one stop of light.

Automated Photographic Acquisition
The Pentax K10D and K20D have an analogue bus between the body and the battery grip. Shutter speeds can be shunted up and down by closing a switch between a common pin and one of two other pins; one that shunts the shutter speed up, the other down.  I hope to have a program available by next year to automate multi-camera operation.  The current programs on offer don't do what I want to do- at least not for Pentax cameras.  I want to write a simple spreadsheet table of elapsed time, and shutter speed for an exposure at that time.  The program will look up this table and control the cameras shutter speed, trip the shutter and write to a nice fast SD card. The software will be available free to the SE community & it may also work on some Nikon cameras that have a similar bus to the Pentax.

See you next year in Shanghai!

Joseph Cali

Return to "Totality in the Gobi Desert" Home Page