kens

Hmmm. OK, I think what I have to do no matter what is get my hands on my friend's telescope so that I can get a really good alignment set, and try other solutions in conjunction from there too. Thanks for this! Actually, you may be better off just skipping over the alignment steps when they come up on the controller. For the long time periods you are using you will never get a good enough alignment to keep the readings correct. So you just set up with a manual alignment of the mount then switch on the controller. When it asks "Begin Alignment 1) YES 2) NO" press "2". The mount will assume it is perfectly aligned so the readings wont change as it rotates.

I think what is happening is that the star alignment causes the handset to make a sky model. The sky model tells it that your alignment is not perfect so as it tracks in RA it "knows" that your Dec is also changing, as it is uncorrected. You need to clear any trace of the model from the handset.

The most scientific write up I've seen on the topic is this one: http://www.sbig.com/site/assets/files/1359/autoguidingversion3_mb.pdf although the data is somewhat specific to the SBIG guiders. There are several factors involved, including focal ratio, that determine how accurately the guider can determine the centroid of the guide star.

Latest activity was to test the noise levels on the various sensors and the likely real resolution. To do this I captured the readings in a file which I then loaded into a spreadsheet to do the calculations. Results are as follows: Accelerometer: Without any filtering (other than on the chip) I can get resolution to about 13 arc minutes. If I apply a moving average filter over 16 samples (about 1 seconds worth) this comes down to about 3 arc minutes Magnetometer: Without filtering I get about 13 arc minutes on this one also. With the same moving average filter I get about 4 arc minutes. Gyro: Unfiltered the noise level is down around 0.75 arc seconds per second but I need to adjust for bias. Not bad seeing as the rotational speed for tracking the sky is 15 arc seconds per second so well above the noise level. This suggests a use for the gyro in correcting for noise in the other sensors.

Sorted out the timing conflict and can now get readings from both GPS and IMU. Not as stable as I'd like and it's still early days yet. Next step is to hook up the LCD screen. To do that I need to get an Arduino prototyping board which will let me get at the pinouts while the display is plugged in. I'm also ripping out unwanted code as I'm getting perilously close to the 32k limit. So far I can see that I could at least get a handy set of digital setting circles. I was going to post a pic but it came out fuzzier than any of my astrophotography attempts. I'll try again in daylight

The GPS and IMU arrived. Wired up each one separately and they worked just as expected. The GPS got a fix and from the IMU I could get roll, pitch and yaw readings which were fairly stable. Hit a bit of a hitch when I wired up the two simultaneously. The program does strange things which seem to be related to a conflict when reading from the two devices. Once I sort that out (any help is welcome) I will look into some calibration code I found that helps with accuracy and also hooking up the LCD screen to display results. A minor hitch there in that I bought the display as a shield and it occupies all the headers without using them all. So I need a work around to access the relevant pins for the GPS and IMU.

I agree. The resolution is really theoretical based on the number of bits but the only way to find out for sure one way or the other is to build it. I'm still learning about the gyro.One blurb says "the gyro can be used to very accurately track rotation on a short timescale, while the accelerometer and compass can help compensate for gyro drift over time by providing an absolute frame of reference." But the data sheet implies that the gyro can only give the rate of rotation, not the absolute orientation. The magnetometer accuracy may not be a big deal as it is really only needed in the initial orientation of the scope whereas the gyro and accelerometer can be used to determine (theoretically) the orientation of the scope when it is in use.

I wasn't really planning to do anything with the gyro unless some bright idea presents itself. It just came as part of the integrated unit along with the magnetometer and accelerometer. I chose those based on their resolution which is way better than anything I had seen before. Largely because they are 16bit rather than 10bit or 13bit. My scope is an old Vixen (c 1983) on a Super Polaris mount. I've got motor drives on each axis but they are different to later models. They are unipolar steppers with 1:300 gearing. I've worked out how to drive them with the Arduino and can now control their speed through the whole range versus the 1x, 2x and 8x options on the DMD-1 controller. I think where I'll head is to see if I can get reasonable tracking using the sensors but I haven't thought it through yet. Maybe just apply periodic corrections to adjust for any alignment errors or something like that. My rough calculations say that the inclinometer can resolve down to 0.003 arc-seconds but the magnetometer can resolve only to about 40 arc seconds. Getting the code to do what I want is what I do best. That's what I like about Arduino as it makes the interface to the hardware so easy. Other ideas pretty much revolve around making things so that I don't have to bend so much - I'm not as young as I was when I first bought the scope.

When I thought I had fried my DMD-1 controller I decided to try to make a replacement with Arduino. As it turns out the controller was fine but my interest in Arduino was piqued. Having a pretty basic scope I thought something to help with alignment or even tracking would be useful. It would need a 3-axis accelerometer and magnetometer, plus ideally a GPS. At first all the devices I could find had insufficient resolution but I think I've now found one that could do the job. Its a breakout board based on a LSM9DS0 which has a resolution of 0.061 milli-g/LSB and 0.08 milli-gauss per LSB. Another breakout board has a GPS unit based on the MTK3339 chipset (not that it means anything to me - but just in case anyone is interested) I had to order the bits from the States as there was no local supplier so hoping it will arrive in a couple of weeks. Not sure whether it will work or not but should be fun to build. I also ordered a 2.8" TFT display to show the outputs and if I start to look at using it for tracking I want to use an IR remote control. Has anyone else tried anything like this?