2016 Bosch motor 6004 RA3 194-06

[kprzewodek]

Seems like broaching a 3/8" hex still leaves some meat on the gear. Think I'll try and see how it holds up.

[HumblePie]

I laid out one concept for packaging this motor that would fully support the load by sandwiching 2 mounting plates around (1 or 2) piece(s) of 1" square tube. The motor has only one (slotted) mounting hole, so I added a bore to locate on one of the motor bosses and provide some "clocking" of the motor. The drive shaft is a piece of 1/2" hex milled down to a 6mm square on one end. Another option would be to use a 3/8" hex all the way, so I'm curious to see Kevin's results on broaching the motor out to a 3/8" hex.

If there's interest, I'd happily share the CAD file. I got one quote online for waterjet cutting the plates, and it looks like about $100 to make 10 pieces.

[topgun]

I am interested in the CAD file. SolidWorks format hopefully.

[kprzewodek]

That's a nice robust design. I'll experiment over the weekend with broaching a hex and see how it goes.

For a less robust way of doing it I've had success 3D printing a hub inside two plates that you can mount an arm. Simplifies a bit in that you can eliminate bearings.

[HumblePie]

Originally Posted by topgun

I am interested in the CAD file. SolidWorks format hopefully.

Well, let's give it a shot.....

I modified the original plate to better serve as a corner gusset if you should choose to use it that way.

I then saved the Solidworks file as a Parasolid and zipped it. The link (to get around the 5MB limit) is here:

https://app.box.com/s/lcv76duqngz2pukh9qdnlgqz7i5hlqq1

Let me know what you think, and Merry Xmas to all......

[kprzewodek]

I was able to broach a 3/8" hex and it appears to work quite well. I haven't put the motor through FRC type durability testing yet but so far that seems to be a good option.

Actually the rotary broach that I have goes about 3/4 of the way through the gear however I was able to carefully tap a hex shaft the remaining distance by making sure the gear was supported.

For those on the fence about investing in a linear or rotary broach I would highly recommend as we have used several times over the years. The nice thing about a rotary broach is that the main investment is the tool holder (~$300) and then you can add a wide range of cutting inserts for around $40 a piece.

[Dad1279]

The square 'neck' of a 1/4" carriage bolt fits.

[Ryan_Todd]

Hey all, following up on Kevin's behalf here.

(Full disclosure: like Kevin, I am also both a Bosch associate and a longtime member of Bosch-sponsored FRC Team 862.)


Intro

On January 14th, Kevin asked me to spend some time torture testing one of these motors. I was able to jury-rig a fairly decent test setup, and gathered a pretty good chunk of data; at the end of the evening, I explored to see what it would take to break one of these motors altogether. Here's a summary of the test rig:


(click to zoom)

• Motor output gear broached to 3/8" hex, as described by Kevin
• Standard 3/8" hex shaft used as output
• Shaft supported by 3/8" hex bearings on either side of the motor
• Bearings, in turn, supported by a pair of old 4" VexPro wheels clamped in a vise
• Motor allowed to float freely in the space between the wheels
• Appropriate spacing ensured with polypropylene spacers
• PWM control through custom "electrical test bech" rig
  • Power supplied by fully-charged 2015 battery (swapped halfway through the evening)
  • 2 ft wire from battery to PD board
  • 4 ft 8 inches of wire from Victor to motor
• Assorted methods used for getting torque off the shaft, described below

Stall Torque Test


(click to zoom)

• Sought to determine what kind of torque a typical FRC team could reasonably expect to get out of this motor in real-world conditions (bearing friction, side loads, etc.)
• Plotted real-world output torque at stall against PWM control percentage
• Motor was driven "clockwise" (as defined on the motor data sheet); resistive force was applied in a "counterclockwise" direction until a stall condition was achieved
• Used small wrench on output shaft to transfer torque
• Used spring force gauge to measure output force
• Multiple data points taken & averaged for each stated value
• Reported torque = (measured force) x (3.75" lever arm)
  • 3.2 ft-lbs to back-drive without any power applied (not brake mode)
  • 2.0 ft-lbs to stall at 20% pwr
  • 6.6 ft-lbs to stall at 40% pwr
  • 10.2 ft-lbs to stall at 60% pwr
  • 13.3 ft-lbs to stall at 80% pwr
  • 13.6 ft-lbs to stall at 100% pwr

Motor Overheating Test (part 1)


(click to zoom)

• Used the FRC-standard "anecdotal touch test" to estimate how much abuse it would take to begin to damage one of these motors via overheating
  • This is a fully sealed motor like the CIM and mini-CIM, so it is necessarily designed to be far more tolerant of stalling out than (for example) a Fisher-Price motor
  • This category of motor will still suffer degraded performance if you let them get too hot
  • Rule of thumb:
    • If a motor is too hot to hold, then its performance is temporarily degraded; it needs to cool down before it will regain full performance
    • If even a quick tap of the finger on the surface of the motor is painful, then its performance is permanently degraded
• Motor was driven "clockwise" (as defined on the motor data sheet); resistive force was applied in a "counterclockwise" direction
• Test setup featured adjustable resistance
  • Hex shaft runs through hole drilled in wood block
  • C-clamp to squeeze wood against shaft
  • Torque measured using spring gauge at fixed distance along length of board
  • Wood wore away over time, so the compression needed frequent adjustment
• Full pwr, no load
  • 5 mins: "Pleasantly warm to the touch"
  • 10 mins: "Very warm"
  • Allowed to cool (10 mins)
• Full pwr, 4 ft-lb average load
  • 5 mins: "Very warm"
  • 10 mins: "Hot, but not yet painful to touch"
  • Allowed to cool (10 mins)
• Full pwr, 8 ft-lb average load
  • 5 mins: "Hot, but not yet painful"
  • 10 mins: Test aborted, resistance board fully eaten through
  • Allowed to cool (15 mins)

Motor Overheating Test (part 2)


(click to zoom)

• With the adjustable-friction rig destroyed, any further tests could only be achieved with a longer lever arm
• No suitable supplies were readily available, so it was decided to skip straight ahead to the stall condition
• A large wrench was fitted to the shaft, long enough for the edges of the vise to act as end stops for the motor's rotation
• Motor was alternately driven "clockwise" and "counterclockwise" (as defined on the motor data sheet) at 100% power
  • Approximately 90 degrees of output shaft rotation before the end stops were reached
  • Whenever an end stop was reached, the motor was allowed to stall at full power for three seconds
  • After stalling for three seconds, the direction was immediately reversed
  • This process was repeated for 2 minutes and 30 seconds
• The "anecdotal touch test" was performed repeatedly upon the conclusion of this test sequence
  • A motor in stall tends to build up heat in its core faster than it can be transferred to the outside shell
  • First test, immediately upon the conclusion of the 2:30 test sequence: "Hot, but not yet painful"
  • Peak temperature of external shell of motor: "Just barely too hot to hold"
• Once the motor was allowed to cool back down, no difference was noted in the motor's performance vs. before the overheating test was performed

Death Test

(No picture for this one yet; sorry!)

• With all of that done, it was time for something a bit more... Aggressive.
• The large wrench from the previous configuration was left in place
• The spring force gauge was brought back and used to back-drive the unpowered motor
  • The motor was successfully back-driven in both directions with up to 21 ft-lbs of torque
  • No signs of damage thus far
• This clearly wasn't doing the trick, so low power was applied to the motor causing it to rotate "counterclockwise" (as defined on the motor data sheet)
  • Using the large wrench and force gauge again, an attempt was made to back-drive the motor again
  • No back-driving was possible against the powered motor
  • A peak of 27 ft-lbs was reached before something started snapping
    • With each snap, the output shaft rotated a short distance opposite to the motor's rotation
    • No loose teeth could be heard jostling around within the gearbox, however, and the motor ran perfectly normal (even at stall!) thereafter
    • Thus, it is reasonable to conclude that the observed failure mode was the result of the deformation of teeth within the motor's attached gearbox
    • Presumably, continued operation in the 20+ ft-lb range would eventually cause performance to degrade below the range of normal operation
  • Upon investigation, the hex shaft had rotated approximately 20 degrees relative to the hex broached output gear before the failure occurred
    • It is therefore possible that the resulting deformation of the output gear may have played a role in the observed failure
    • The use of the stock 8-pointed spline shaft may therefore stave off failure to an even higher torque level, but that is just speculation
• I hope to crack open the motor in question and confirm these hypotheses, but right now we're in the middle of building a robot!

Conclusion

This motor could handle a stall condition with the best of 'em, and withstood 2x its own stall torque before giving way. Even when it did "break", it still continued operating normally for applied torques up to and including its stall torque.

All in all, I have no reservations at all about recommending this motor for use in any application that warrants a motor with torque/speed ratings in this range:

• Deploying a collector
• Rotating a turret
• Winding up a catapult
• Manipulating the Portcullis, Drawbridge, Cheval de Frise, and/or Sally Port
• Et cetera

[MrForbes]

thanks!

[kprzewodek]

Thanks Ryan!

I should point out that there is a thermal switch built into this motor. If you are running close to or at stall for long periods of time you will most likely trip the thermal switch. It will reset over a period of time but you would notice intermittent operation. Not to worry its doing it job to prevent catastrophic meltdown.

[Ryan_Todd]

Originally Posted by kprzewodek

...there is a thermal switch built into this motor...

Well then, now I'm surprised that I never managed to trip it!

Perhaps a Round Two of testing is warranted, just keeping the motor in stall until the thermal switch trips and confirming that the motor retains its original performance characteristics after it has been allowed to cool back down.

Either way, that makes these motors seem practically rookie-proof...

[kprzewodek]

Originally Posted by Skywalkar

Well then, now I'm surprised that I never managed to trip it!

Hold it at stall for around 5s and it should trip. Less/more depending on existing temperature. This switch is based on internal temperature alone.

Originally Posted by Skywalkar

Either way, that pretty much settles the question that these motors are rookie-proof...

Careful. Don't think it's "anyone" proof. Mount correctly to keep side loading minimized and you should be ok.

[Dad1279]

Does anyone know where to buy more?

[kprzewodek]

Originally Posted by Dad1279

Does anyone know where to buy more?

Since we supply this motor "as required" by a particular Tier 1 we don't stock inventory and have it setup as a stock item in our system. We're trying to get a feel it teams find this useful this year. If it is we can look into something in future years. I'll do my best to donate replacement parts as needed but we cannot charge anything for it since we donate this to FIRST.

[cedwards]

Hello, our team has decided that they would like to use this on an elevator component. We have a few questions on how we can implement this motor to its fullest.

1) Wiring. We have the motor +/- and have been driving them. However, we are not sure how to wire the two addition wires (additional to the motor power) to gain access to the internal encoder (Hall effect sensor?).

2) If this is a hall effect sensor what part of the RoboRio do we plug it into? Is it digital or analog?

3) If the "encoder" that is in this motor is a hall effect sensor what component do our programmers use to get information from it to base position on. Is this a Counter object?

Thanks all, and happy last week of build!