Can someone explain to me why that combination is not going to work?
The two coils are connected at the center taps (the fifth lead). That will allow current to flow in undesired directions between the coils, which violates the design principle of the motor driver.
Try it out – I have done so with a six lead motor, connected as a five lead motor. The motor may still step/rotate but depending on the motor characteristics, the power supply and load, it will probably not perform as expected.
With a five-lead motor, it is rather difficult to figure out which pairs of the four leads correspond to the two properly phased coils. If you don’t get the combination correct, the motor may just execute random jumps in either direction.
As to your last remark… I’ve already been there… Now let’s hope I made a good note somewhere which wire is wich.
So… You’re saying: no smoke expected, but it just won’t work?
Maybe the “short” (others call it the fifth lead) interferes with the analog current modes? So maybe fullstep and half-step work?
In my tests with a six lead 12 V, 50 ohm/phase motor using full step operation, it did not seem to make a difference if I connected the two center taps together, converting it into a five lead motor. The motor worked normally, even under load. Indeed I used an oscilloscope to look at the voltage difference between the two center taps and although there were some switching transients, they were small and probably would not affect the performance. I next tried out a small five lead motor and observed normal behavior (once I figured out how to connect it). So at least for full stepping, there appears to be no problem using the A4988 with some five lead motors.
However, I did not experiment with either the current limiting or the microstepping feature of the A4988, so I reserved judgement.
When you think it through, during normal full step, full current operation, the center taps should be almost always at roughly ground potential due to voltage drops and that is what I saw with the oscilloscope. If five lead motors work in your case, go ahead and use them!
Perhaps the Pololu engineers could look at this issue more thoroughly, with a variety of motors and driver options, and come up with some guidelines based on experimental data. There are lots of cheap, surplus five lead steppers that are perhaps unfairly being marginalized!
I pretty much just assumed that a bipolar stepper motor would not work right with the two stepper motor coils shorted together, but your point about full-step mode is a good one. If the coils have the same properties and the taps really are in the center, then the tap points of the two coils will always be at the same potential when in full-step mode and no current should flow between them. It is not obvious to me what will happen when trying to microstep, however. We will try to do some tests soon and post again with whatever we find out.
Similarly, when in half-step mode, the half-steps are with 1 coil at full-power, and the other at idle (current = 0). That would mean that the driver need not drive the other coil at all. Leaving the ends of that coil undriven should similarly cause no current to flow between the centerpoints of the coils.
On the other hand, when the coils are NOT connected in the middle, you are /allowed/ to drive ONE side to either high or low, but that would mean trouble for the five-pin motors.