PHD Guiding Problem

I just got PHD Guiding to work with my setup but noted a disturbing behaviour that I cannot really explain.

My setup:

Vixen GPD2 mount with Synscan Goto Kit
Refractor INED70 70/420 f/6
Guiding Cam Meade DSI II Pro monochrome on Off Axis Guider Lumicon 2″
Windows Vista 32bit
PHD Guiding v 1.11.0
ASCOM Platform 5.0b w/o the recent enhancement (because this does not work with the Celestron driver)
Celestron Unified Driver 5.0.14

Scope connected with USB-Serial-Adapter to Laptop

PHD Settings:

PHD Guide 1.11.0 — 5/17/2010 12:08:53 AM
Guiding begun
lock 172.3 302.5, star 172.3 302.5, Min Motion 0.15
RA angle -0.04, rate 0.0006, aggr 1.03, hyst=0.10
DEC angle -1.64, rate 0.0019, Dec mode 1, Algo 1, slopewt = 5.00

Dec mode is automatic, Dec algorithms is „resist switch“

Conditions: no wind, good balanced setup.

Problem:

The calibration process of PHD succeeds and I start autoguiding and start taking pictures. This works fine for a while then suddenly I hear an „eeeek!!“ sound from the mount and the PHD Guiding Graph (from „Enable Graph“) shows a huge spike in Declination correction. I saw this several times (perhaps 4 times) during the night, but I cannot reproduce it. The result of this sudden move of the mount is that the photo is garbage as all objects are duplicate.
My questiion is of course what has caused this and how to avoid it. I don’t think PHD is to blame in the first place because in my eyes it seems to react on the situation only. But I’m not sure. Without guiding enabled, I have never seen this happening before.

Here is an extract from the PHD Logfile during this time (the complete log file is here including a few more guiding sessions) :

#Frame Time dx dy Theta RADuration RADistance DECDuration DECDistance StarMass
253 507.749 0.12 -0.25 -1.1 0.0 0.10 85.0 0.24 10244.91
254 509.777 -0.01 -0.18 -1.6 0.0 0.01 85.0 0.18 10960.38
255 511.649 0.09 -0.28 -1.3 0.0 0.09 85.0 0.27 10297.99
256 513.552 0.04 -0.28 -1.4 0.0 0.05 85.0 0.28 10613.33
257 515.487 0.02 -0.27 -1.5 0.0 0.03 85.0 0.27 10359.79
258 517.406 0.01 -0.41 -1.5 0.0 0.03 85.0 0.41 10460.42
259 519.324 -0.14 -0.27 -2.1 0.0 -0.12 85.0 0.28 10722.88
260 521.228 -0.17 -0.35 -2.0 273.1 -0.16 85.0 0.36 10728.56
261 523.568 0.10 -0.36 -1.3 0.0 0.09 85.0 0.35 10279.31
262 525.471 0.14 -0.23 -1.0 0.0 0.14 85.0 0.22 9908.49
263 527.374 3.17 4.99 1.0 1000.0 2.67 0.0 -5.18 6541.70
264 530.244 -0.70 -0.48 -2.5 603.1 -0.34 85.0 0.53 10500.74
265 532.928 -0.47 -0.48 -2.3 778.1 -0.44 85.0 0.51 10985.85
266 535.767 0.06 -0.49 -1.4 0.0 0.03 85.0 0.49 10250.72
267 537.701 -0.07 -0.58 -1.7 0.0 -0.04 85.0 0.58 10072.85
268 539.636 -0.13 -0.51 -1.8 0.0 -0.10 85.0 0.52 10124.39
269 541.539 -0.11 -0.36 -1.9 0.0 -0.09 85.0 0.37 14931.81
270 543.473 -0.36 -0.33 -2.4 561.9 -0.32 85.0 0.35 10273.70
271 546.094 0.06 -0.26 -1.3 0.0 0.04 85.0 0.26 10256.53
272 548.013 -0.14 -0.52 -1.8 0.0 -0.10 85.0 0.53 10740.87
273 549.932 -0.05 -0.46 -1.7 0.0 -0.04 85.0 0.46 10331.17
274 551.866 -0.18 -0.55 -1.9 0.0 -0.14 85.0 0.56 9961.63
275 553.800 -0.28 -0.55 -2.0 430.9 -0.24 85.0 0.57 10006.91
276 556.296 -0.08 -0.63 -1.7 0.0 -0.08 85.0 0.63 9906.23

The deviation is in both RA and DEC and only because the Max Dec Duration step is 85 it seems to take longer to correct in Dec. The sound therefore seems to come from RA correction?! Now that both axes show the problem, I think this rules out any trouble with periodic errors, gears and worms. It looks more like someone tripped over the mount (but it was not me).

Here a few graphs that illustrate this behaviour better:

1) Here is an extended guiding sequence which shows some periodic behaviour in DEC (green) and one huge spike. This spike is the problem I try to understand.

2)  This is the same as above, but zoomed to better resolve the spike. You can see that both RA and DEC are affected by the spike.

3) This next graph shows the deviation in RA and the duration of the corrective pulse in RA given by PHD Guiding (value divided by 250 to fit in here). This illustrates that PHD did not send a pulse before the spike (so no overshoot reaction or so).

4) Same for DEC.

Here is another occurence from the same log file which looks slightly different. Here you see dx and dy deviation (Dec is green, RA is red) and a huge sudden peek in dx at frame 115.

Funny enough the dx value does not really change although PHD thinks it changed the DECDistance. Here is the extract from the log

PHD Guide 1.11.0 — 5/16/2010 10:49:21 PM
Guiding begun
lock 468.2 224.9, star 468.2 224.9, Min Motion 0.15
RA angle 3.02, rate 0.0009, aggr 1.00, hyst=0.10
DEC angle -1.66, rate 0.0020, Dec mode 1, Algo 1, slopewt = 5.00
Frame,Time,dx,dy,Theta,RADuration,RADistance,DECDuration,DECDistance,StarMass

109,252.446,0.17,0.74,1.3,0.0,-0.03,180.0,-0.75,4058.94
110,254.428,-0.09,0.68,1.7,0.0,0.15,180.0,-0.67,4285.21
111,256.456,-0.04,0.73,1.6,0.0,0.12,180.0,-0.72,4001.59
112,258.562,-0.02,0.72,1.6,0.0,0.11,180.0,-0.72,4002.96
113,260.683,-0.43,0.67,2.1,545.8,0.46,180.0,-0.63,3420.70
114,263.444,0.01,5.38,1.6,724.4,0.62,180.0,-5.36,3692.50
115,266.424,1.08,5.42,1.4,379.7,-0.32,180.0,-5.49,4184.83
116,269.014,0.72,5.74,1.4,0.0,-0.06,180.0,-5.78,4078.33
117,271.042,0.55,5.51,1.5,0.0,0.09,180.0,-5.54,4164.62
118,273.148,0.74,5.59,1.4,0.0,-0.05,180.0,-5.64,3735.08
119,275.160,0.61,5.59,1.5,0.0,0.05,180.0,-5.62,3551.44
120,277.235,0.62,5.64,1.5,0.0,0.06,180.0,-5.67,4135.35
121,279.294,0.75,5.67,1.4,0.0,-0.06,180.0,-5.71,3965.38
122,281.275,0.77,5.78,1.4,0.0,-0.07,180.0,-5.82,4096.16
123,283.210,1.04,5.67,1.4,390.4,-0.33,180.0,-5.74,3969.61
124,285.768,0.58,6.04,1.5,0.0,0.10,180.0,-6.07,5826.68
125,288.030,0.51,5.84,1.5,207.6,0.18,180.0,-5.87,3969.44
126,290.651,0.88,5.95,1.4,0.0,-0.13,180.0,-6.00,3843.11

Possible explanation

After some discussions with peers in the internet, I made some progress. I can now reproduce the behaviour by doing the following:

On the SynScan handcontroller I set the speed to very slow (Rate=1) and then press the declination button to move the mount to the South („up“ button). After a whiile I repeatedly press the declination North button very short a couple of times. This is to siimulate a number of guiding commands like PHD Guiding would issue them when moving the mount to follow the guiding star. After about twenty times pressing North, you hear a (comparable) loud sound and watching the gears I could see them jump in the opposite direction! This means although pressing North all the time, the gears suddenly jumped in South direction! I made a little movie of my DEC motor that illustrates the issue. The sudden jump is large enough that you can feel it on the complete mount and I’m confident that it would also cause the movement in RA.

THe problem does not show up in South direction.

A number of people offered their explanations. The most reasonable in my eyes are

  1.  there is either stiction in the motor that causes the tooth of the gears to bind and cause a sudden release of the tension. See the FAQs on the PHD Guiding page for an explanation how stiction can cause sudden movements.
  2. there is a problem in the internal gears of the motor that causes that one tooth is skipped.

I think I could rule out the mount itself because

  • the worm moves easily in both directions when being moved manually
  • the problem does not seem to occur periodically, e.g. after every complete rotation of a gear (as far as I can see them) or the worm
  • I could reproduce the problem by removing the DEC motor from the mount and carefully pressing my thumb on one of the gears and pressing the North button in short intervals at low speed.

I contacted the dealer where the SynScan kit was purchased from and asked for the motor to be exchanged, as in my eyes this is the most likely source of the problem, either due to internal stiction or a tooth being skipped.

I got a replacement DEC motor now and it shows the same problem! Now I need to find a way around this. I guess the most reasonable is to ensure that PHD Guiding only guides into North or South direction. Then I’m pretty sure to avoid the problem. THe other option is to use better fat that reduces stiction.

Kommentar hinterlassen

Antworten

Deine E-Mail-Adresse wird nicht veröffentlicht.


*