, , , , , , , , , , ,

A few blog entries ago, I thought I had made great progress in getting the old telescope drives for Hopewell Observatory’s venerable Ealing Mount to work again. Unfortunately, it became clear that one had to adjust the amount of friction in the clutches very, very accurately, and I saw no way to fine tune it.

So I bit the bullet and decided to convert the mount over to an inexpensive system, at least partly DIY, that uses very inexpensive solid-state printed circuit boards and Android phones to control stepper motors that make the telescope point in the directions desired. (Instead of spending many thousands on a Sidereal Technologies rebuild.)

This system is called OnStep and is spearheaded by a number of very generous volunteers: Howard Dutton, who basically invented the system and wrote all the original code, along with Ken Hunter, George Cushing, and Khaled Bahayeldin, and a number of others whose names I don’t recall. It uses off-the-shelf components, chips and sub-boards, that cost very, very little; these are put on one of a slew of different possible 3D-printed circuit boards. There is even a Wiki that could use a bit of editing. It’s got a ton of information but when I was starting out, I found it extremely confusing, and I am not alone. I promised to try to improve it when I get the Ealing telescope working properly.

After getting the software to work, then you arrange the connections to your telescope’s gears, power supply, and communications inside your own mount.

I am immeasurably aided in this conversion effort by Alan Tarica, who is the co-leader of the Washington, DC-area’s Telescope Making, Maintenance, and Modification Workshop (which has been going on for about 80 years) and by Prasad Agrahar, who made a remarkable telescope in our TMMMW several years ago and went on to build his own OnStep conversion of an existing commercial telescope. Prasad’s example showed me that if our old Ealing drive died, we should try OnStep.

Well, the Ealing drive did finally die. (It had presented problems ever since it was first delivered to the University of Maryland Observatory nearly 50 years ago.)

Michael Chesnes and Bill Rohrer of Hopewell helped materially with removing the old components of the scope and with then trying to debug the electrical problem that has now sprang up with our roll-off roof.

Ken Hunter made for us, and debugged, an entire OnStep board and refused to take any money for it. Prasad Agrahar gave us some NEMA17 stepper motors and some wires and likewise refused to let us pay. Prasad drove all the way from Philadelphia to help Alan and me figure this stuff out in person, both at the workshop and out at the observatory. Ken has spent hours, remotely from Yuma AZ, walking me through the various steps in managing the many settings that need to be uploaded and adjusted in order to get things to work. Ken told me he used to run the ATMFREE list-serve, but retired from that after an injury, and he remembered meeting me once at Stellafane. He also very kindly sent us an antenna for the system so that it can run WiFi or BlueTooth more efficiently from inside our massive metal mount.

Alan and I are fairly far along in the conversion, thanks to all this help. I had to learn some of the basics of the Arduino operating environment, which one uses to set all the many, many parameters needed to get the system running. And had to improve my soldering techniques! Fortunately, all the heavy lifting of getting all of the many lines of code working together has been done by Howard, Ken, and the others, so all I had to do was set things up to fit our particular set of choices for the board, the stepper driver, the sub-boards, the gears, and the motors.

Here is our current setup: we have two (now three) MaxESP boards running OnStep version 2.04 (iirc). (Multiple boards because they are cheap and in case one gets fried by a lightning strike or stupidity. It happens!)

They have TMC5160 stepper drivers, connected to two rather beefy NEMA23 stepper motors (200 steps per turn), which I arranged to fit exactly in-line with the worm gear that we will later put back into the mount. We have tweaked the ‘CONFIG.H’ file settings the best we could, and with an enormous amount of help, I think I’ve set the speeds of the stepper motors correctly. The worm gear turns another gear with 20 teeth, which turns another one with 359 teeth. (All made by Byers, and made very, very well.)

(We had NEMA17’s run by the TMC2130 stepper drivers, but we didn’t think they were beefy enough to rotate the very large mount we have, even if we balance it perfectly.)

It’s been a very interesting learning expedition. It’s taken quite a bit of time, but not really very much money. With mass production, the components (screws, capacitors, diodes, resistors, and so on) if purchased in medium quantities, are really very inexpensive.

However, the stepper motors are still not behaving properly. They scream instead of moving, as you can see in this video. I will post the current parameters on the OnStep wiki, where I said. You can see and hear the action in this little video. When I try to slew to any random, dummy target, the steppers will start rotating and also start making a deafening squeal that gets higher in pitch and volume. However, after a little while, both rotors stop turning either completely or almost completely. The smart hand controller pretends that the mount is moving in both axes, but it’s not true.

Right now, I don’t know what is causing this problem.

Anybody have suggestions?