We had our kitbot running today and did a speed test.
Our configuration was using 6" wheels.
also used the 21 tooth for the counter shaft
and the 28 tooth for the wheels
we measured it’s speed to be 5’/sec
If you figure those cims are running at 5300 free speed
reduced 12.75:1 (gear) and 1.333:1 (sprocket)
we should have been moving 8.15’/sec
There wasn’t much if any efficiency loss in the way we had it assembled, and the speed test was fairly accurate +/- .5 sec
I’m thinking those cims are turning more like 3300 rpm
Has anyone else experienced this?
Originally I thought everyone using the kit chassis/trans. and 8" wheels would be up in the range of 10.5’/sec but after seeing this, it’s looking like they’ll be traveling more like 6.5’/sec
One thing I didn’t check (as I’m thinking of it now) is that we may have not been applying full power to the speed controllers, as we just grabbed an old RC and ran it. But I doubt the code wouldn’t be full power.
Yah, of coarse there’s going to be some loss, but over a 1/3?
and this test was giving it a good 10’ of room before the start line with no load other than the chassis, and the gear reduction as you said. I could understand 1/10 loss but not 1/3.
Even under a normal driving load, those speeds wouldn’t make any sense.
Assuming you don’t have a MAJOR parasitic load somewhere (which you probably don’t) this is totally crazy.
Dan, please check your voltage and get back to us.
My prediction, you’re not giving those CIMs full power.
5 ft per sec /(6"*pi/12)*60sec/min = rpm of wheels = 191
191 rpm * 2.63 * 12.75 = 6404 rpm at the CIM
I’m not sure what you mean by counter shaft, but assuming you mean the pinion sprocket (is that the proper term?) the sprocket ratio is 28/21, you have
I haven’t put your numbers in, but I designed a spreadsheet based off of JVNs’ motor calc sheets in the whitepapers, and from that I get an approximate speed of of 6.5 ft/sec for a full weight (120 lb) robot using the transimission as given. While the 8.15 ft/sec may be accurate with no load, you have to account for all of your weight as well.
there should be no (or almost no) loss in top speed because of weight. The only tiny bit of loss from weight would be the added resistance it puts on the bearings, but in every FIRST drivetrain i have seen, the bearings are well well below there ratings.
Dan,
Allow me to perform an analysis for “our viewers at home”.
My calculations match your analysis:
The weight will have minimal affect on the top speed. The only difference will be an increased normal force on all wheel/wheelshaft bearing surfaces, which will increase friction slightly.
This increase in parasitic loss is negligible.
Therefore, we can simply eliminate weight from our analysis.
Please refer to the attached spreadsheet.
This sheet shows the speed of the robot at 8.47fps (assuming 100% efficiency). From here, we can increase the “speed loss constant” to 60% which then shows an output speed closer to what Dan is experiencing.
This means, (as Dan stated) the CIM motors are either spinning slower than expected, or there is a large parasitic loss in the gearbox.
If we assume no parasitic loss, we need to find out reasons for a slower spinning CIM.
Some reasons:
-Voltage provided to the motor is closer to 7.2V
-Motor is wound differently, and top speed is limited.
I would almost immediately rule out a motor flaw.
Dan, please recheck your voltage. If that fails, replace the Chip motors and see if the problem persists.
For what its worth, I have compared free speed extrapolations to ground speed measurements for all the robots presently in our “garage.” The trend is pretty obvious.
2001: two 3/8" drill motors
32% 6ft/s real, 18ft/sec projected
2003: two 1/2" drill motors in high
38% 10.5ft/s real, 27ft/sec projected
(resulted in intermittent breaker and controller resets!! )
2003: two 1/2" drill motors in low
68% 5.6ft/s real, 8.2ft/sec projected
2004: two 1/2" drill and two CIM
78% 7ft/sec real, 9ft/sec projected
You can see that our “speed efficiency” rose dramatically once my team started choosing reasonable gear ratios. If you were to take these speeds and work backwards to the assumed load on the motor to get that speed, and the resulting current draw… the currents are astronomical, obviously. I hate to think what was happening while we were turning. I’ve always wanted to use a clamp on meter to find out if all this empirical data is self-consistent.
Oh, and for all of these tests I also did a “wheels off the floor” rpm measurement, which without fail was very very close to the free speed expected. Take this measurement, and just see if robot weight on those axles doesn’t somehow make a difference.
As far as our measurement methods, we timed the robots on the ground over a 50 foot window following a 30 foot accelleration zone. I imagine that was fairly accurate.
Given this history, my adjusted calculation to estimate the speed of our kit-built competitors was 6.3 ft/sec.
I tested the living daylights out of that gearbox. I tested the 4wd configuration with 8 inch wheels and it behaved as expected (around 9.5 ft/sec). I use a closed-form non-linear differential equation to solve for position, speed, and acceleration over time and it tracked exactly with my testing of the frame.
For the past several years, Team 33 has been able to get our robots to reach over 90% of the theoretical free speed. If you are only getting 60%, something is wrong. Easiest way to find it: remove the chip motors and check for friction. Without the motors in the system, everything should spin freely when you turn the wheels by hand. If it doesn’t, figure out why.
6" skyway wheel chair wheels (usual kit stuff). Chain had appropriate slack. Also the bot didn’t drift to either side when full power - it went straight. Indicating the fact that if there were something wrong it would have had to happen in both gearboxes, or at least one motor in each box. So I’m almost certain there wasn’t a mechanical problem. I also don’t think we happened to receive (4) defective cims. So if everyone is positive these cims perform 5300 rpm free speed. Then it really only leaves one conclusion, that the RC wasn’t providing full power. We also had the old victors (883’s?). Anyhow, I’ll measure the voltage tomorrow and let you know.
Well I couldn’t get there today, but I had Amanda check the voltage on the output speed controllers, she told me they were reading 12v on the output at full power. I also had her spin the drive wheels - spun freely. Gearbox is in good shape. So I’m out of ideas… I prefer my dew cold please…
I’ll have to make a little video to prove it all, since I doubt anyone will believe me now. I guess I’ll try a different method of testing the speed, but the last was pretty good, and the chassis is visibly slow.
