The two jrk controllers are wonderful products with a superb utility. How about a third version with a higher current H bridge. The output structure of the 24V23CS added to the JRK 21V3 would be fine by me.
Personally I have gotten around the limitations of the current jrks by adding my own Nfet/Pfet H bridge but some of the motor control benefits like over current protection etc. are lost.
Thanks for sharing your thoughts. Can you tell us what kind of motors you are using, and in what kind of applications? The more common request I have heard is customers asking for a version with support for encoders. So, without much other feedback, our likely road map would be to make the encoder-supporting version first, with integrated drivers such as those on the existing Jrks, and then to move on to higher-power H-bridges.
We are using the jrk controllers in an az/el antenna pointer used to track low earth orbit satellites. At X band very high accuracy, <1/2 degree, is required. We use precision potentiometers, sometimes multi turn for FB. Not sure why one would use a shaft encoder for a rotary application when pots are so cheap and work so well, at least until they wear out. Motors used are often windshield wiper or window winder motors. 24 volts at 15 amps max likely needed. You might need more current momentarily to break free some ice, big bird or other minor obstruction.
My assumption is that jrk software would only need a minor tweak and the 4 FETs are a wash for the IC H driver. Straight forward PCB layout (just like the 24V23CS) No reason that a shaft encoder could not be a fourth input option, lotsa software required though to deal with setting up the encoder parameters.
Thanks again for a great product.
Hmm, it does sound pretty straightforward when you say it like that. I think maybe we didn’t have enough I/O on the microcontroller to support our preferred discrete MOSFET H-bridge, which is what we use on the simple motor controller 24v23 and which requires four PWM signals out of the MCU. Oh, and I just remembered that the MCU is so full that we already have to write some of the code in assembly, so it’s quite difficult to make fairly minimal modifications. But, we’ll reconsider it in light of your comments. By the way, is the current sensing feature important to you?
With external FETs it will be most cost effective and efficient if all four parts are N channel. You probably want to use a modern H bridge driver IC like the Motorola/Freescale MC33883. This will also have the benefit of reducing uP I/O. Left as an exercise is what to do about current sense.
You asked if I thought current sensing was useful. For circuit protection purposes it is obviously desirable. More subtle is the ability to monitor the motor current in real time via the USB. Monitoring motor current using the Plots function in the utility is very helpful in setting up the system to deal with a dynamic load. This is particularly true when the motor is remote from the PC when you can’t hear the motor. Using “motor current” in Plots is a good way to monitor the influence of external mechanical inputs, like the wind.
The jrk configuration utility is particularly inspired, don’t delete any functions.
My impression is that the more modern units are the ones that require more I/O since MCUs are getting more capable. From what I can tell, the MC33883 also needs one input per MOSFET, and it seems like it might be an obsolete part. I looked around Freescale’s site for a while, and the most recent version I can find is MCZ33883EG, but even that one says “No longer manufactured” and points to the MC33883HEG that doesn’t show up as on orderable part. Maybe they’re right in the middle of migrating and the web site is out of date.
Sorry about that! My mistake. Try the Intersil ISL83204. Having separate inputs, like the MC33883, for each FET allows shoot-thru control to be assumed by the uP. The intersil chip does this in hardware with two resistors. Has one input for output polarity and one for duty cycle.
There are lots of these drivers out there. One thing to keep an eye on for a motor control function, where the output may be DC for a considerable time, is that they have an internal charge pump and do not rely solely on a bootstrap for the top FETs. A lot of the earlier ones only had the bootstrap circuit.