775 Pro active cooling

Technical Motors
1

We’ve been seeing lots of applications for 775 pro motors on drivetrains especially.

I’m aware they tend to fry themselves after about 2 seconds of stall. My understanding is this occurs in part due to the lack of airflow (no motion = internal fans not moving). The other part is also the higher current draw makes for more heat building up (I would suspect at least).

Question for ya’ll:

Has anyone run an experiment to see how much forcing air through a 775 Pro increases its life at stall? I’m thinking something like this with a 3D printed attachment to direct the airflow through the electrical input end of the motor.

I have a suspicion it wouldn’t make a huge amount of difference, but I’d be curious if it’s been tried…

Edit: some math

At stall, the motor will be consuming 134*12~=1600W, presumably all becoming heat.

from http://www.nmbtc.com/fans/engineering/temperature-increases-graph/, if I’m interpreting their graph properly, you’d want something like a 40 or 50 CFM fan to keep the temperature rise reasonable? Seems like it would be a big fan to keep an $18 motor happy. Plus, more space claim means we’re chewing up the advantage of having smaller motors… I think I’m slowly answering my own question… any thoughts would be appreciated though!

2

Can’t smoke a 775 is you don’t stall it. Realistically it only needs to survive 2:30 and if you make them easy to change the issue kinda goes away, unless you have a design where on motor is doing everything.

Teams I talked to at champs this year said the motors only ever got slightly warm after a match.

Extended play will probably be an issue

3

I love the 775Pro more than most but actually, this isn’t true. Vex/WCP provided the data to show it is not true.

In general temperature is what kills these motors.

Temperature can be thought of as the water level as you fill a leaky bucket. You can keep the water level down by putting in less water (which in this analogy is current that turns into heat) or you can keep the water level down by increasing the size of the leaks in the bucket.

Forces air is an example of the latter. I believe that it is more effective to work on the former.

What does that mean? It means working in the more efficient portions of the motor curve.

For example, if you are trying to do a task that requires 220W of power. You can gear the 775Pro such that it is running at 80% of free speed and 20% of stall or you can go with 80% stall torque and 20% of free speed. Both give you 220W of mechanical power. BUT the first one has the motor working ~68% efficient (generating ~100W of heat) and while the second has the motor working at ~18% efficient (generating ~1000W of heat). A thousand watts makes a pretty reasonable blowdryer. That’s a lot of heat to get rid of.

As a bonus, faster motor speeds not only mean less heat to dissipate (i.e. it turns down the spiget on the top end) it also forces more air through the motor effectively making larger hole in your “leaky bucket.”

It is a win-win.

Dr. Joe J.

P.S. looking at the locked rotor tests and doing some rough calculations, it looks like these particular 775Pro motors start to fall apart somewhere around 250-300C which is pretty much the melting point of high temp solder so it makes sense… It actually isn’t that hard to write code to estimate the temperature of the armature windings given Voltage, Speed, and Current. Maybe something the WPI folk might consider baking into their libraries…

1 Like
4

With regard to Dr. Joe’s suggestion

I think this is a great idea. We’ve been pretty aggressively discussing the 8X 775Pro drive for the upcoming season.

I agree whole-heartedly this is a temperature problem and we’re wrestling with how to address that. His suggestion is definitely on the table for us. The only tip I would add is to consider both the rear and side vents and remember you’re pulling air from the front, so you need to leave those vent’s open as well. The 40mm fans provided in some KOP by EBM Papst would likely do the job if you wanted to do it on a simpler, one fan to one motor basis. A more complicated iteration could be achieved with a larger offering previously available in the KOP also from EBM Papst. My gut feeling is that it’s more of a static pressure problem than a flow problem when you start considering fan ratings, because of the fairly minimalistic size of the vents on the front and the restriction to flow those provide. We’ve also discussed incorporating a thermocouple or thermistor in the air stream of an active cooling device, one per gearbox in order to provide temperature feedback, which could also be used to impose a current limit.

The other solutions we’ve discussed are as follows:

-One consideration is programmatically limiting current, A simple current ceiling for the drive is pretty easy to achieve with current control mode on the Talon SRX motor controllers. The more sophisticated option for programmatic protection is looking at a rolling average of how many amp-hours (or other appropriately scaled unit) we’ve used in the drive to impose a floating limit adjusted in some way by that average.

-One consideration is mechanically limiting current, by gearing to provide a traction limited drive a a much lower percentage of rated stall torque than we’d historically designed for, which is a practical example of what was described by the 80/20 situation in Dr.Joe’s post.

-The final consideration we’ve looked at is electrical limits, but we found some really bizarre data on examination here. Ultimately this prompted some really interesting research on the whole FRC power distribution system and what limits exist there. Couple of points on this one.

-I was surprised to learn CCA for a typical 18Ah 12V SLA battery is about 300amps, but that doesn’t accurately describe the peak momentary current the battery is capable of providing. Cold cranking amps described here https://www.optimabatteries.com/en-us/experience/2014/01/what-does-cold-cranking-amps-cca-mean

-Then we start to look at spec sheets for our inline electromechanical current limiting devices and start finding all manner of fun things, I won’t spoil the fun here. The data sheets are all on product pages at AM if you’re interested

-Then there are speed controllers and their maximum capacity to feed motors, which becomes quite large when you have 8 channels to play with.

-Then there is the issue of the data available for the motors themselves. We’re staring down a boatload of wonderful testing data from VexPro and now struggling with how to get it to spit out a concrete current limit for the motors.

It’s complicated. This is exciting, but I’m afraid it’s exciting because it’s dangerous.

Sorry for the novel. Just thought I’d share where we were at in our process.

5

Awesome stuff to all! I love to hear about progress and bleeding edge thoughts! Thanks for taking the time to type it up for the community here!

6

I don’t currently have much time to devote to this, but it’s (theoretically) entirely possible to eek a little more power out of a 775pro though using active cooling, specifically by applying some variant of thermoelectric cooling.

What I’m imagining is something akin to a Peltier cooler being used to keep a copper vapor pipe wrapped around the motor at reasonably low temperature.

The issues I’d like to research are:

  • Legality of using a Peltier cooler in a Custom Circuit
  • Thermal Conductivity of a cooling pipe wrapped on the motor, and the effect on thermal mass of the motor
  • Whether or not a Peltier may be more useful on a forced air system to cool the motor

I’ll probably look into this once I have more time.

7

Those are issues I’d like to see you research too. Here’s my initial feel for them:

  1. A heat pipe (“copper vapor pipe”) probably isn’t useful here–they mostly are purposed to move the heat from the hot item to a heat sink or cooling system. They’re not really a cooling method in themselves. The hot-item in a 775Pro isn’t the body of the motor as much as wire in the armature. Not really able to heat-sink that. These might work with a CIM, where the heat goes into the motor frame.

  2. forced air cooling is only limited in it’s heat removal by the CFM of air you can get through the motor. It will take about 90 cfm (a five-inch blower) to remove 500 watts of heat at a 10C delta. Can you blow that much through? I know that the motor is producing 1600 watts at stall, but 500w removal might be enough to keep the smoke in long enough.??

  3. Peltier coolers are a great technology and I can’t think of any reason they’d be not-legal, but their efficiency isn’t suited to a motor. This 5-amp one from Adafruit can remove 60 watts, not 1600w.
    [Peltier Thermo-Electric Cooler Module+Heatsink Assembly - 12V 5A : ID 1335 : $34.95 : Adafruit Industries, Unique & fun DIY electronics and kits

SDvdiNsUoSJACIlTqlJur1sNHe6HP6L6DUiRDpdhxANgMxFysCa-scjuhVMxoChf7w_wcB](Peltier Thermo-Electric Cooler Module+Heatsink Assembly - 12V 5A : ID 1335 : $34.95 : Adafruit Industries, Unique & fun DIY electronics and kits)

If you pre-cooled a chunk of metal with a peltier, it might be some benefit, but I think that will be better suited to cooling the main breaker than the motor and we still have the problem that the 775 isn’t designed to dump the heat to metal, but to air instead. And the peltier will still need a fan on its heat sink unless you want to just heat up a chunk of metal during a match.

8

Not sure where those numbers are coming from, but when a motor is close to smoking, the delta is a lot bigger than 10 degrees Celsius.

9

Numbers from chart posted by gerthworm, but verifiable on motor data sheets.

There Is a temperature where ten degrees of cooling will prevent smoking, but as long as the motor is adding heat at a rate of 1600w and we’re only removing it at 500w, it will be getting hotter fast.

How fast? Probably really fast. Blow through the not-burn temperature in a fraction of a second? I don’t know.

10

This is the problem you will have to overcome. All of the KOP fans I have seen are axial fans. They are good for applications with low pressure build-up and high flow. Centrifugal blowers work better than axial fans when the application has a high pressure build-up (like blowing air through a 775 motor). I don’t know if centrifugal blowers are legal in any of the recent years.

I have also not seen an application where the air flow is so restrictive as trying to blow air through a 775. The air speed required to remove several 100’s of watts would be very high. When working on cooling systems for large industrial motor controllers, my co-workers found that higher air speeds increased the effectiveness of heat removal only up to some point. Using air speeds higher than that actually decreased the effectiveness of heat removal since the air is not in contact with the hot object for long enough to absorb the heat.

It is most likely that your most effective means for preventing the 775’s from burning up is really to limit the input power in an appropriate way, possibly incorporating sensor feedback on the rotational speed of the motor.

11

Some other items to consider:
When pushing air through, you tend to compress the air and build up heat. When pulling air through, you tend to have an inlet temp more in line with ambient.
Some speed controller style fans with a shroud to help suck heat out of the motor are relatively low power (roughly 0.1 AMP) for 10 to 20 CFM (assuming no obstructions I imiagine). Putting such a system in place will not let you run at stall forever, but it might be worthwhile to help improve cooling and general motor temps as they would still be running even when the system is sitting still.

Anyone know what CFM the motor pulls at free speed or Peak efficiency?

12

Not to be a spoil sport, but…previous discussion in another thread brought a bru-ha-ha due to “compress the air”. All air compressors must be… A devise used for the purpose of compressing air that is not a legal air compressor could be cause for inspection hastle.

13

I doubt that Ike is advocating using a compressor. A fan will increase the air pressure but with commercially available fans, we have never noticed a measureable increase in air temperature due to the air pressure rise when working with our heat sinks i.e. fans running but no output load on equipment under test.

Having said that, it would probably take something like a compressor to get significant air flow through the 775 motor but then you run into other problems (see my previous post).

14

Ah, yeah so the delta that’s talking about is between the air coming out and the air going in (room temp). The air coming out isn’t going to be up to the motor temperature*. So 10 deg C may be in the right ballpark. It helps that the motors are probably designed to lose heat quickly.

*And in fact it’s quite difficult to simulate what it would be, testing would be the best method.

15

I think the most attainable goal isn’t to remove ALL the heat in the system, but rather to remove enough to enable the motor to operate at peak throughout 2:30. In order to dodge some of the cots fan requirements, I had also debated using an additional 775pro to drive an impeller/blower rig to handle the forced air setup. The current use of the motor at essentially free speed is entirely minimal, and may allow enough air movement to solve the issue.

I like the idea of pre-cooling. I think the benefit of using a Peltier would be not in removing ALL the heat, but simply removing ENOUGH to attain a slightly longer window of peak use.

Honestly I think all of this would only buy an extra ~5-10% power, which may or may not be worth the effort… But it’s certainly a fun problem to think about.

16

Or you could use a CIM? :slight_smile:

17

That mindset is parallel to the idea of just buying a COTS drivetrain.

Sure, it’s totally doable. But how much does it teach you?

18

I run into these type of heat transfer problems in my daily job. There’s more to cooling than just matching the total heat flux of the two systems.

The discussed scenarios are talking about stalled or locked rotor instances. When the rotor is locked, the air will not swirl through the motor and evenly distribute the cooling and therefore create local heat zones where the air flow is blocked. I’m certain that even with enough total flux to remove the total heat generated from the motor, the local areas will exceed the temperature limits and still smoke the motor.

If the rotor is still spinning, forced air injection would help remove more heat, but a locked rotor is still at the mercy of its material limitations.

19

So, the highest static pressure rating I’ve seen on a tube-axial 40mm fan is 27mm h2o, which is an astonishingly high static pressure rating for a 40mm fan. This fan isn’t legal (unless CTRE or somebody decided to sell a motor controller cooling accessory that uses it, if I’m reading that rule correctly <R32>)

Unless we’re considering a pretty novel approach here, I’m assuming we’re talking about pulling air through the motor, because a) air flows in the front of the motors and out the sides and rear and b) a solution to pressurize the motor vents on the front isn’t straight forward or really all that accessible in most cases, but it could be done.

In the case of the fan and air flow configuration described above we’re actually talking about decreasing the static pressure in the motor case by about 2 Torr with no flow, so we’d see a temperature decrease of an astonishing .056C. It’s really pretty negligible, per what I can only assume are my exceedingly inadequate calculations. (I used 760 Torr as standard pressure and 21C as room temperature, in case anyone is wondering)

Unless somebody wants to find a dimensionally accurate model of the internal geometry of the motor and beat it to death with some wicked CFD work the only conclusion I can come to is:

There is no question a duct and fan running constantly to move air through a 775pro over the course of the match will help with heating, but It’s not going to let you run them 1 for 1 in place of CIMs. Some other effort to reduce individual motor loading is likely necessary, be it using more motors, imposing current limits, or some combination of those and other ideas.

The potential exists for some super scary single speed as well as shifting drives next season.

The potential also exists for eight 775pros worth of smoke to come billowing simultaneously out of a robot, which could be kind of cool to see, but I’m sure would be a heartbreaking experience for the team involved.

20

Isn’t locked rotor the ultimate worst case scenario here? Brushes are first to die in this use case, right? I’m not sure I fully understand which pieces of that beautiful motor data provided by VexPro are actually applicable to this conversation.