Modeling the Kit Motors with USB-DAQ

Dave Lavery’s request during kickoff for the Pilot Program teams to model the characteristics of the kit motors gives us a chance to showcase DAQ and demonstrate a little bit of what we can do with the USB-6009 DAQ device. I will be working with team 418 to model the kit motors within the next week - I was just wondering how many other Pilot Program teams wanted to do this alongside us.

The software we will use is the one written by Russ Beavis for motor control and testing. We will be using an Allegro current sensor (the spec sheet is in the motor control software zip file) and a Hall-effect sensor for RPM detection (just slap a sprocket to the end of the motor shaft and read tooth info). The wiring schematic for the entire setup is also included in the motor control software zip file.

We will potentially be doing something a little more advanced in addition to the DAQ testing to double-check our work, but I’ll release those details once we’re sure we’ll have the hardware available to us when we need it.

-Danny

We were working on this today in a lunch meeting, and will be figuring out how to do it. Currently we think we are going to take the simple route and build a pulley on an axle that we can couple to the motors and load using a belt. Its funny, we had the “how to measure it” down in seconds, but as we are all electrical/computer engineers, the how to mechanically load the motor was the hardest part!

Luckily one of our mentors is a test engineer and uses Labview every day here at work, so that part is easy, the mechanics are the challenge!

Team 1676 wants to be involved.

Just like Kims Robot, we are still struggling with making a test stand, aka Dynamometer. Dumb stuff, like shaft couplers, pillow blocks and a way of measuring the linear force on the ‘brake’ belt are what’s slowing us down. I may have to just pay a visit to home Depot and whip out the credit card…

The DAQ and electronics seem trivial in comparison - at least at this point. We hope to have a dyno design ready to post this weekend.

I will point the students on the Test & Integration team to this posting, and they will make contact and see how they can assist team 418 in the task.

Don

Measuring the motor performance unloaded is easy as Danny suggested. But measuring the torque properties is completely different as you need to have a way to measure and control torque.

Any idea out there on how to do this? :confused:

I’m working on a dyno design that uses an extra CIM from a previous years kit as the load. With it configured as a generator, varying the electrical load on the output will vary the load on the motor being tested. The motor shaft is mounted in a pillow block using a couple of rollerblade bearings (which just happen to be 8mm ID, and the shaft is 8mm OD), and a lever arm is attached to a scale to measure the resulting torque. I figure this will work OK on the smaller motors, I probably can’t load the CIMs all the way up to stall with it, but I should be able to get enough points for a good curve.

I still have some work to finish up on the pillow block design, and I hope to get it all built this weekend. Now if I can just find the high-power electronic load we used to have at work, and a scale I can read with the PC, it should be fully automatic. I’ll post an Inventor pic when I get it all designed.

Some days I sure miss having access to a motor lab :frowning:

One of the guys in the National Instruments Motion group has a very crude dynamometer setup on his desk. All it is composed of is 2 motors (in his example they were the same type of motor) with their shafts connected, we’ll call them the “load motor” (LM) and the “motor under test” (MUT). The MUT is powered in the usual way (for FIRST it would be via a Victor) and a simple current sensor measures the current draw of the motor. The MUT drives the LM, causing the LM to generate a Back-EMF voltage. The Back-EMF voltage is a constant voltage generated by the motor per thousand RPM of the motor, and is linear across the RPM range of the motor; with the Back-EMF voltage it is possible to get a really good idea of the speed of the motor by measuring the voltage alone! The LM is then connected to a high-watt 10 Ohm resistor (like 300 Watts, it’s almost the size of the motors!) purchased from Ohmite. The resistor allows the voltage from the LM to produce a current, which then powers the LM in the opposite direction. By measuring the current coming off the resistor you can calculate the torque of the LM by using a torque constant (which gives you however much torque per Amp) found in the spec sheet of the motor. So, by measuring the currents and voltages on the LM and MUT you can generate some fairly accurate torque curves with a minimal hardware set.

One thing to note, however, is that in doing this you need to make sure the LM and MUT are aligned PERFECTLY so that there is no stress on the shaft coupling (and placing them in an enclosure would be wise, too). From what I hear the “first run” of the simple dynamometer wasn’t lined up as well as it should have been, and the snapping of the shaft coupler wasn’t a pretty sound (could have damaged the motors too!).

-Danny

PS: LOL, I guess this is the same setup as Jeff’s above!

So guys, hows it going with this? Anyone have any updates?

-Joe

The resistor I was trying to get from Ohmite was $335 with a lead-time of 8 weeks (took FOREVER to get that info). Riiiiiiight. I called PowerOhm in Katy, Texas and they’re sending me a $34 resistor (model WR25 with the B1 mount) early next week. So, we Team 418] should be able to have torque curves to the masses by hopefully next weekend if all goes well.

-Danny

Not as well as we would like, and now it’s getting towards crunch time. The key difficulty we are facing is a variable torque load, which can be measured with the USB DAQ. Sure, a spring scale and friction belt works, but taking measurements by hand is a drag, not to mention inaccurate.

Our current incarnation has a CIM from last year, controlled by a Victor, to counter-rotate against the motor under test (actively driven against the MUT as opposed to useing it like a generator against a fixed or variable resistor). Problem is, the victor is non-linear and even at the lowest setting sometimes is a bit too much for the MUT. So, we’re going to try either a fixed resistor to tone down the CIM for the smaller motors, or use one of the smaller motors (or a Non-FIRST motor?) to test, uh, the smaller motors.

Much of the testing we’re doing on the ‘final’ hardware is also taking away from the Integration & Test team’s availability for motor testing. The kids are surprised what they are learning from seeing data instead of 'gee, that looks about right…"

Don