Power supplies for a Bag Motor

Hello CD,

I’m working on a research project where I build a low-cost dynamometer that runs off an arduino. My original plan for the system was to buy a few BAG motors and test my findings against the publicly available motor curves on motors.vex.com to evaluate the efficacy of the system.

The problem that I’m running into is that, due to coronavirus, I’m no longer able to access my school’s power supplies, so I need to buy a power supply to run the motor off of.

To my understanding, a standard FRC-battery type of setup won’t have a stable voltage, which could lead to bad data. Additionally, I want to be able to perform a locked-rotor stall test so the power supply would need to be able to provide at least 53 amps at 12v.

Any advice on which models to buy, or at least where to look for something like this would be greatly appreciated.

You’re gonna have a hard time finding a bench top power supply that goes to 53 amps. The one I have at home only goes to 5 amps. I got it on ebay used which is generally a good place to browse. Something like this may work but is super pricey and non regulateable. I bet if you do more than the cursory internet search that I did you may be able to find something better.

If I recall there was a Mac Pro power supply that could do quite a few amps at 12V. You might look into that further.

You could use a smart motor controller in voltage compensation mode with a fully charged battery in order to provide a stable 12V. I would use either a Spark Max (heavily preferred) or a Jaguar since they’re easy to interface with a computer without setting up a RIO.

I would try a car battery…

That would have the same problems as the FRC battery, they are basically the same thing. Charging and discharge inconsistency could skew the data.

This power supply claims to be able to supply 700W (58.3A @ 12V). I’ve never used it so I can’t verify that though. Keep in mind if you do use a power supply like that you’ll need to wire the AC side yourself too

Yes, it would skew the data. But the interesting thing is that it would skew the data in a realistic way.

“To my understanding, a standard FRC-battery type of setup won’t have a stable voltage, which could lead to bad data.”

jijiglobe mentioned that they didn’t wish to use an FRC battery due to this, so a car battery wouldn’t be much different.

What you want is a deep cycle marine battery. Think trolling motors. Trouble is that they are in the $100 to $200 range. Test at 10 volts (motor) rather than 12 to have a flat voltage curve. Assume the Vex curves are linear and adjust them down. Another option is to use throttle body motors. You can get as many as want by asking for them on CD.

Check Digikey. If in doubt use their technical support chat, which is awesome.

OP: Can you provide some description indicating how you plan to use the dynamometer? It might help to be able to understand the power requirements.

Most dynamometers I have seen utilize a speed controller to operate the driving motor at a known speed. With a properly tuned speed controller, supply Voltage variations should not impact the shaft speed of the dynamometer. The Voltage variations may impose limits on the torque that the dyne is able to transfer to the load device as the Voltage droop will limit the ability to induce current in the motor.

Some of the smart controllers can handle 24V inputs - use two batteries in series, then use voltage compensation on the controller to get a pretty consistent 12V.

1 Like

I think this is what I’m going to do. I have some batteries lying around already so this will probably be the cheapest way to pull this off with the fewest number of things I’ll need to order

So it’s a bit of a weird system, but basically the idea is that the motor rotates an torsion spring, and by characterizing the rotational inertia of the system, as well as the spring coefficient of the torsion spring, I can calculate the torque of the system at any given time.

Well… in theory… We’ll see how it goes.

Actually, you are on the right track!
I used precisely this approach to calculate motor inertia values.

Suggest that you look at this video:

You can certainly do the same thing - I’d be interested in comparing results!
If this is applied to characterizing mechanism inertia, this may tend to produce some significant vibration in the mechanism.

My inertia report is attached (because I couldn’t sort the link!). It provides a link to the servo-amplifier I used. The hardware chosen allowed me to inject an analog rate signal, which keeps me away from digital command bandwidth limitations.

Motor Inertia Identification 190106.pdf (433.1 KB)

1 Like

There are used server rack power supplies available at 1100W, 1200W and even a few at 2000W for relatively low cost. The challenge with many of the switching power supplies is that they may trip out on the starting surge of the motor. The fans in them are apparently very noisy. It would probably be best to use a motor controller and ramp the speed of the motor. Some of them may trip out on over-voltage when stopping the motor and it regenerates. I am having a problem like this with a high power lab bench supply in a test system at work now.

I would have imagined that you’d be fine with the starting current as long as that is still lower than the max current the power supply is rated for. I used something similar for a lab project with 4x 100W motors in parallel.

The over-voltage may be a problem though. I don’t know how these power supplies work well enough to know if they have a way of dissipating that.

What I have seen in the past is use of a blocking diode to prevent the regeneration from charging into the supply. That approach leads to a dc link capacitor to provide a sink for the regen energy … which can also lead to a pre-charge circuit for the capacitor.
Even with this approach, you can also run into bandwidth issues in the power supply current loop - how quickly does the power supply off-load current and then transition to carrying a load.
Make sense?

The challenge is that the starting surge current is typically something like 5-6 X the rated full load current leading very large, heavy and expensive power supplies if the rated load is anything over a few 100 W.

Understood. Large D.C. link capacitors and braking resistors were used in the industrial AC motor controllers I used to work on at previous jobs. In the current application, the whole system, with the controller circuitry fits inside a 2 1/2" diameter enclosure and has to operate at up to 175 C so large D.C. link capacitors and braking resistors are not viable solutions. Since our input voltage is at most 600 V, we are going to 1200V MOSFETs. The power supply tripping really only occurs in lab testing and doesn’t always happen. When the system is in use, the motor is pretty heavily loaded and cannot regenerate very much.