Several 'remedies' were suggested. One was shielding the relay with aluminum foil which seemed to help on some mounts. Another was adding a flyback diode on the coil of the relay which was a good plan but, again, didn't cure the problem. Changing out the R8 resistor to put the voltage to the coil into the datasheet's specified range was tried which also did little good. Lastly we tried using an external pull-up resistor on the reset pin of the CPU and a small value capacitor to filter any noise (see my page on filtering the CPU reset line). This had somewhat better results but the resets were still there.
Finally Doug Dieter put an oscilloscope on the mount and voltage spikes were seen in the mount casting during relay operation. This is the 'noise' and appeared to be caused by the relay leaving the stepper motor's power bus 'open' for a short time while it changed state. Stepper motors have coils in them which can, at times act like the ignition coil on your automobile. The coil is energized to move the motor; if the power is cut the magnetic field in the coil collapses which can cause a back EMF spike of several hundred volts; which needs a place to go. Where it should go is back into the motor's power bus through a clamping diode. Unfortunately, when the relay is 1/2 way in between one state and the next there is, in effect, no power bus and the spikes, anywhere from 50-200v, ended up in the casting. This wasn't good so he bridged out the relay... EUREKA!
To see the page on the early attempts click here
The Final Solution
You can do one of two things. You can pull the relay from it's socket and wedge in some wire or resistor leads to bridge the lines on the relay or you can unsolder the whole damned thing (don't throw it out.. you never know when a 5v relay will come in handy) then bridge the lines. However you go you pull the relay and connect pins 3-4 and 7-8.
This does cause your mount to use more power. Measurements on an EQ6 showed about a 20% increase in current used during tracking and on my HEQ5 it's about a 33% increase. During higher speed movements the current flow is a bit less. The mounts use more power during tracking/guiding than they do slewing. With higher current comes higher transistor temperatures. Measurements Doug made on his Atlas showed that the transistors got to about 37-40C (case temperature); they're rated at 150C (junction temperature). On my HEQ5 the current is slightly greater than Doug measured (??) so they'll get slightly warmer but still well within the specs of the parts (BD233/234).
That's about it. I haven't looked too hard at the rest of the Synta board, but I'm told (by Doug Dieter and others) that the voltages are filtered where needed and that the relay was the only weak spot. The first couple of fixes outlined above may take care of what the Synta engineers should have designed in the first place. Use a scope to check for transient voltage spikes in the body of the mount. The Final Solution will eradicate these completely at the expense of milliamps.
Back to the HEQ5 Project page
Back to Ozzzy's Astromomy page