Boxdoerfer PowerFlex MTS-3SDI

Dual axis microstep telescope drive

Back to http://arnholm.org/
I received this telescope controller in April 2003. It was my third controller after the Vixen SD-1 and the Astromeccanica DA-1. The Vixen SD-1 was very nice and accurate, but controlled only one motor and had no PC-connection. The DA-1 suffered many technical problems and finally failed, so I decided to try the MTS-3SDI instead. It was very reasonably priced, controlled 2 motors and had a PC option, so it was a good alternative to the more expensive Vixen SkySensor 2000.

I compare the SD-1, DA-1 and MTS-3SDI controllers here. There is also a MTS-3SDI product website.


MTS-3SDI - Initial impression

I ordered and pre-paid the unit + PC cables in mid-March 2003 and it was sent 31. March. Sigurd Boxdoerfer was very clear and easy to communicate with. As I had Sanyo Denki motors from my previous DA-1 unit, I had to adapt them, and Sigurd Boxdoerfer was very helpful. More details on the motor adaption can be seen here. The postal "service" in Norway caused quite a few delays, however, so I did not receive the unit until April 16. When it finally arrived, this is what I saw:


MTS-3SDI controller with PC-cables, straight out of the box.

The black motor cables suitable for my motors are quite long (more than 2 meters) and quite flexible, so they will not cause problems when tracking. The plastic bag at right contains new connectors for my motors. I chose to solder these connectors to the motor wires myself, even though Sigurd Boxdoerfer offered to do it at no additional cost. I didn't want to trust the postal service, however, so I did it myself, and was successful.

The power cable is attached to the motor cable and comes with 2 "banana" type plugs, easy to adapt to a 12V DC power supply. At left you can see the coloured PC connection cable + a 5m serial extension cable. You can choose to use the extension cable or not.


MTS-3SDI connects to a PC using a 14-pin connector on the back side.

Protecting the PC connector

I found everything looked quite solid, except that I suspected the contact + cable on the back was a little fragile, with a possibility that the cable could be ripped out during an observing session. Also, I did not want snow or humidity to come near the contact.

When I observe, it is usually very cold and I want to ensure everything is easy and secure, so I designed a simple protection cover using a thin aluminium plate and some wood. I fastened it using the already existing machine screw holes on the back. The cable is now very secure and emerges conveniently at bottom. Very little weight is added, and the added thickness is only about 10 mm.

Configuration and initial tests

The controller came pre-configured for my non-standard combination of motors and mount. It proves the controller is very flexible.

I connected the motors while inside, and tested the tracking accuracy of the controller. The results were very favourable, the tracking was the most accurate of the controllers I have tried. Without any adjustments at all, I measured the tracking error to be only 0.007%. See the results here. Even more impressive is the fact that even this accuracy can be improved by adjusting the tracking frequency, something that can be done with floating point precision!

User Manual

A 43 page printed user manual in English was provided, and the manual is also available in PDF format. The manual contains a lot of information, but it could be better organised. I had to read it several times :-)

Essentially, I read through and ignored all the talk about programming the controller using the normal buttons and dip-switches. I find it better to use a PC program to upload such information. Later, I found that a simple program for this purpose already exists.

Of particular interest to me, was the RS232 serial programming protocol. I wanted to write my own PC program to communicate with the controller and thereby implement GOTO capabilities. I found that even though most details were available in the manual, some assistance was required to do the low level packing and unpacking of command parameters. I program in ISO C++ and the examples were in non-standard QBASIC, a rather big difference. Also, reading the EEPROM data is not straightforward, and the manual examples are not always sufficiently detailed. But after a couple of email exchanges I had it all working! More on this below.

Writing a PC GOTO program for MTS-3SDI

As noted above, a few very low level things must be mastered before you can communicate properly from a PC program. 32 bit long and float variables must be sent in a special way, and for EEPROM data a different floating point representation must be reconstructed. But when these things are mastered, the programming is quite easy, the MTS-3SDI behaves exactly as specified in the documentation. Quite interestingly, it seems the MTS-3SDI provides many more features when operated from a PC, compared to using only the hand controller buttons.

One feature available in the programming protocol, but not available directly from the buttons, is GOTO. The support for GOTO is quite rudimentary even from the programming protocol, but it does exist. You can ask the controller to move the motors to certain stepper motor step positions. By combining some of the EEPROM configuration data with the controller clock and stepper motor readout functions, plus some math, I wrote a GOTO program that worked!

Another thing worth noting is that if PC control is the main operation mode, you do not have to operate the hand control at all. Simply leave the MTS-3SDI in OFF position and turn it on from the PC!! This is very convenient under my conditions. If I want to, I can put the controller in a plastic bag and forget it, then do everything from my laptop.

MTSca program

More information about the MTSca program and software download is available from the MTSca software page.

Testing the MTSca program

A preliminary version of the program was tested April 21/22 2003. Several Messier objects not seen before by me was easily found using GOTO from various bright stars. The most prominent result was the 2h 20min exposure (composed of many shorter webcam exposures) of M51 seen at right. The image was captured with K3CCDTools.