Printable View
I am trying to replace an unreliable ULN2003 drive with a Gecko step/dir drive.
The controller has 2 PCB's. A CPU, and ULN2003 driver.
I have 6 signals from the CPU. The 4 lines are motor coil inputs to a standard ULN2003. Presumably the other 2 lines are an enable/disable latch, and a motor ground? Not sure. So basically I need a circuit that converts the 4 line coil input to step & direction signals. I imagine there are existing solutions, or if not perhaps a simple logic circuit will do it?
What is unreliable about it? Are you sure that replacing the ULN2003 will cure the unreliability?
I have not done such a thing, and I do not know offhand of any off-the-shelf solution, but my guess is that a simple logic solution will not be enough - presumably you will need some kind of "memory element", because I suspect that figuring out the step and direction outputs will require determining the difference between (at least) the prior state of the 4 output pins and the new state of those pins, and then deducing the needed step and direction signals from the change between the patterns. You will also need to "poll" the output lines to determine when they change, or else you will need some way to derive a clock pulse from the changed output pattern.
A microcontroller should be able to do the trick, however. (I am assuming that you do not have access to the Z-80 source code)
The darlingtons that the uln2003 drive, regularly blow. It may be a ground problem, or just poor heat management. There is also no low current idle feature. Nor is there room for better heat sinking, unless I swap in the Gecko driver. The driver portion of these controllers are notoriously unreliable. (I have two of them. ) The CPU side is reliable. (It is a rotary table controller) I am tired of repairing/debugging the 1970's tech. It is a 90:1 gear ratio, so 1 step behind is not going to impact the accuracy significantly, so even if a logic solution is used, once it detects the direction change, it will only be 1 or 2 steps behind, worse case. 99% of the time direction is always one way, once started.
Quote:
Originally Posted by doorknob
Z80 processor???.. i mean really...that processor was last seen in the 80's-
I would just upgrade everything- as it is much cheaper.
This Z80=zilog 80, was very popular when I was student 35 years ago. So you better replace it.
The CPU PCB & control is staying, What I need is a circuit to convert a simple 1/2, 2/3, 3/4, 4/1 motor coil signal, into a 2 wire step and direction signal.
I am fine with a micro solution so long as:
someone else is writing the code, and supplying the PCB, cause I am not doing that.
Preferably, it is a logic solution.
There is no previous known state of the motor... It is known at power up. 2 coils energized. I haven't looked at that, but most likely coil 1 & 2 everytime. The next state is either 2 & 3, or 1 & 4, etc..