Bidirectional H-Bridge DC-Motor Motion Controller: Backstory
Back when Stephen and I were working for the University of North Carolina’s Chemistry Department, a researcher asked us to come up with a way to prolong the lifetime of an expensive lens used in an ultraviolet (UV) laser system. The laser in question was a noble gas eximer system, based on either argon fluoride or krypton fluoride, with high peak energy in the UV range, either 193 or 154 nm.
This wavelength is absorbed by almost all optical materials, including UV quartz, so calcium-fluoride optics were being used to focus very high power to a small spot. This sort of laser delivers its power in short pulses, nanoseconds in the time scale, so the peak power tends to be large—moreso when the energy is highly focused. Even calcium fluoride absorbs some of the energy and will develop color centers, places with high absorption that reduce the intensity of the light.
One characteristic of this sort of damage is that the lens will “heal” itself if it isn’t stressed to the point of fracture and some of the damage is due to heat from the absorbed UV. We reasoned that if the lens was constantly in motion, the laser would not damage the lens as quickly. This brought up its own set of problems. For example, the lens motion would have to be uniform and very linear, with no displacement except for straight-line motion.
Fighting the Budget
Our Instrument Facility came up with a crossed roller translation platform with a very rigid mounting in all but one axis. We were tasked with providing a method of providing a reciprocating force to the stage to provide a smooth passage of the lens without any motion either forward and backward or up and down.
Price was a big issue, and the budget did not allow us to use rotary feedback elements as would normally be used. Stephen thought a dc motor and gear head assembly, driving a split band over rollers as used in many disc drives, would be suitable if the problem of non-uniform speed could be addressed. The lens needed to be able to vary its speed over a broad range and the actual needs had not yet been determined, so whatever system that was used would have to accommodate a broad range of speed while providing an extremely stable and reproducible rotation from the dc motor. Friction of the translation stage would vary over time and temperature, and this system would need to run for a long time, with millions of shots from the laser.
We decided on a driving circuit that would sense the current through the motor using a low-value, low-tempco resistor and then using this voltage developed across the resistor to add an adjustable amount of positive feedback to the circuit that would provide the current to drive the motor. Using MOSFETs as pass elements in a H-bridge arrangement allowed the polarity of the voltage to the motor to be switched as the stage traveled to the end points. Opto interrupters were used to detect whenever the stage would travel to end points, and the position of the optical interrupters would provide the capability of setting the travel of the stage to suit any size of lens.
When the stage would get to the optical interrupter at either end of its travel, a flag on the stage would break the infrared beam between the optical interrupter, and the CMOS logic would change which pair of MOSFETs was conducting. One of the MOSFETs conducting at any particular moment would have its gate controlled by an operational amplifier, and the current through the current sensing resistor would provide a method of sensing and controlling the speed of the motor.
The voltage across the current sense resistor was compared with the voltage on the wiper of a potentiometer used as a speed control. The voltage across the potentiometer was precisely controlled by a voltage reference, and the position of the wiper provided a voltage that corresponded to the desired speed. If the voltage across the sensing resistor was too low compared to the voltage on the wiper, the op amp would increase the drive to the gate of the MOSFET that was conducting then. An op amp amplified the voltage across the sensing resistor, and a portion of this signal was summed with the voltage on the wiper of the control potentiometer and presented as one input to the current regulating op amp. The amount of positive feedback was adjustable and used to compensate for any increase in drag in the translation stage.
Success!
The controller worked well. The lens would last a long time in use as the damage, being largely thermal, was spread out over a large linear area instead of a tiny point. Despite the cost constraints that we had, this is still the most effective solution. Any method I know about today would involve some form of positional feedback, a tachometer, a lead tester dual voltage (LTDV), perhaps a ruled scale with optical sensing with some method of keeping track of the position, and some way of correcting for deviation. All of these methods involve transducers that cost as much as the entire drive system. As far as I know, the last time that I looked into this system, it was still being used but on a different laser and lens system.
This sort of problem arises often in instrument control and machining. Even something as simple a model train set would benefit from having uniform, controllable speed. This could be used in fiber drawing, a wheelchair drive, or any place where motion is desired that is both controllable and where the speed has to be the same even if the load increases or decreases. This speed controller allows for the voltage to be non-constant, such as from a battery power source, yet provides a controlled speed that does not vary.