View Full Version : Gear ratios for launcher
swwrobotics
05-02-2012, 13:13
Has anyone found that magic gear ratio that works best for the launcher?
Our team tried doing a direct drive with the CIM motor but the ball did not go very far. I remember reading that a 2.7 to 1 gear ratio works best but I can't be sure. Any ideas?
We've been running 2 BaneBot 550s with 20:1's and they work beautifully. We may just be able to pull off our master strategy after all :D
JamesCH95
05-02-2012, 13:19
Has anyone found that magic gear ratio that works best for the launcher?
Our team tried doing a direct drive with the CIM motor but the ball did not go very far. I remember reading that a 2.7 to 1 gear ratio works best but I can't be sure. Any ideas?
We've been running 2 BaneBot 550s with 20:1's and they work beautifully. We may just be able to pull off our master strategy after all :D
The gear ratio depends heavily on the diameter of your end effector, assuming you're doing a roller/wheel shooter. Direct-driving with a CIM to a 1" roller is very different from an 8" wheel.
OP, care to provide any details? My team's experiments show that you should find the exit speed you need for your range, assume the ball will go 60%-70% of your theoretical tangential free speed, and gear accordingly.
Has anyone found that magic gear ratio that works best for the launcher?It depends on the motor(s) you are using, the diameter of the wheel(s), whether you are using only one wheel or top&bottom, the moment of inertia of the wheel(s), and other factors.
This post may be of interest:
http://www.chiefdelphi.com/forums/showpost.php?p=1118835&postcount=20
The gear ratio depends heavily on the diameter of your end effector, assuming you're doing a roller/wheel shooter.
Our team is using 6in or 8in wheels for our shooter, I can't remember which one. (I'm not actually on the shooting team)
swwrobotics
05-02-2012, 13:50
The gear ratio depends heavily on the diameter of your end effector, assuming you're doing a roller/wheel shooter. Direct-driving with a CIM to a 1" roller is very different from an 8" wheel.
OP, care to provide any details? My team's experiments show that you should find the exit speed you need for your range, assume the ball will go 60%-70% of your theoretical tangential free speed, and gear accordingly.
We are doing a one-wheel design that uses the new six inch wheels that came in the kit of parts and a flap on the top to adjust the angle the ball comes out of.
We were just worried that a high gear ratio wouldn't have enough force to launch several balls and it would be hard to get back up to speed.
MrForbes
05-02-2012, 13:59
The new kit wheels tend to tear up the balls, according to a couple of reports from teams who tried them. We're using a single 8" kit wheel from a year or two ago, driven by last year's FP motor with it's 19 tooth pinion, which drives an older FP plastic first stage gear from the FP gearbox. The ratio is 19:72, I believe. This gives us a pretty good range....
http://www.youtube.com/watch?v=81vOT2KY2ig
But the gear ratio you want will depend on a lot of factors, as others have mentioned.
artdutra04
05-02-2012, 14:54
We are doing a one-wheel design that uses the new six inch wheels that came in the kit of parts and a flap on the top to adjust the angle the ball comes out of.
We were just worried that a high gear ratio wouldn't have enough force to launch several balls and it would be hard to get back up to speed.How much compression were you putting on the ball?
Other than the fact that the tread pattern on the 2012 KOP wheels chews up the foam basketballs, any single-wheeled shooter with six inch wheels (when directly driven from a CIM) should easily be able to shoot the balls a more than adequate distance.
Simply changing the gear ratio won't solve this problem; the real problem is probably related to ball compression, shooter hood design, and/or misalignment/binding/friction in the moving parts of the shooter.
FYI: The 2.7:1 gear ratio is what is inside the CIMulator from Banebots. There will likely be a fair amount of teams using this gearbox this year, as this gearbox makes a 775 or 500-series motor "emulate" the output characteristics of a CIM motor. This allows teams without a lot of machining resources to keep their CIM motors in their drive train and use "emulated" CIM motors elsewhere on their robot.
swwrobotics
05-02-2012, 15:21
How much compression were you putting on the ball?
Other than the fact that the tread pattern on the 2012 KOP wheels chews up the foam basketballs, any single-wheeled shooter with six inch wheels (when directly driven from a CIM) should easily be able to shoot the balls a more than adequate distance.
Simply changing the gear ratio won't solve this problem; the real problem is probably related to ball compression, shooter hood design, and/or misalignment/binding/friction in the moving parts of the shooter.
FYI: The 2.7:1 gear ratio is what is inside the CIMulator from Banebots. There will likely be a fair amount of teams using this gearbox this year, as this gearbox makes a 775 or 500-series motor "emulate" the output characteristics of a CIM motor. This allows teams without a lot of machining resources to keep their CIM motors in their drive train and use "emulated" CIM motors elsewhere on their robot.
Our mentor said the we should have as little amount of contact with the ball as possible to pervent too much deformation and loss of accuracy. So you're right, that definitely could be the problem. But just out of curiosity, would it be worth more to have a high gear ratio or a larger wheel diameter?
JamesCH95
05-02-2012, 15:28
Our mentor said the we should have as little amount of contact with the ball as possible to pervent too much deformation and loss of accuracy. So you're right, that definitely could be the problem. But just out of curiosity, would it be worth more to have a high gear ratio or a larger wheel diameter?
It's true that you should have as little contact with the ball as possible, but you still need some minimum amount of contact, otherwise you won't impart much energy to it. Many teams seem to have found that 1-2" of pinch on the ball will work well.
With a 6" wheel you could be looking for 2,000-5,000rpm axle speed, depending on what range you want. the actual ratio will depend on what motor(s) you're using.
Has anybody posted test data anywhere that includes wheel rpm, wheel diameter, squeeze, launch angle, and distance?
and wheel moment of inertia, if available? (yeah, I know: dream on)
I've been keeping an eye out, but haven't seen anything yet. It would be most useful as a sanity check for some of the calculations.
artdutra04
05-02-2012, 15:40
Our mentor said the we should have as little amount of contact with the ball as possible to pervent too much deformation and loss of accuracy. So you're right, that definitely could be the problem. But just out of curiosity, would it be worth more to have a high gear ratio or a larger wheel diameter?From what it sounds like, your shooter barely compresses the ball. If so, the best thing to try out first would be to increase the compression between the shooter wheel and the ball; otherwise every motor and wheel combination you can try will likely have the same disappointing results.
The trick is to find a balance between not enough compression (not enough energy from the wheel gets transferred to the ball, leading to wimpy shooting range) and too much compression (too much energy is taken from the shooter wheel without any shooter range improvements). If you do some experimenting, you'll find an a good balance between the two.
Once you find the ideal level of compression, then you can look into substituting other motors, gear ratios, and wheel sizes.
wireties
05-02-2012, 15:50
FIRST Team 1296 is trying 2 FP 801 motors at 6:1 and 2 AM motors at 4.64:1 driving 2 8" wheels (both under the ball) - have not got it running yet, a proof of concept with a CIM at 3:2 worked nicely compressing the ball 1-1/2"
HTH
klmx30302
05-02-2012, 17:49
Team 166 is direct driving 2 axels with 2 8" wheels each. Each axel has a cimulator with 2 rs-550 motors each. We have not clocked its top rpm yet but the wheels spin quite fast as you can see in this video: http://www.youtube.com/watch?v=YtiK87pE_9U
We don't know the range of the final shooter yet, but the prototype could shoot 46-50 feet, the prototype was direct drive cims instead of cimulators.
pfreivald
05-02-2012, 23:10
We're making very consistent shots from key range with AM motors and planetary gearboxes, though the spin-up time between shots is too long. We're working on that problem now!
DampRobot
05-02-2012, 23:20
We're making very consistent shots from key range with AM motors and planetary gearboxes, though the spin-up time between shots is too long. We're working on that problem now!
My team is using a similar set up. Were getting about 16ft using belts driven by 1:1 sprockets off a AndyMark gearbox (the one for the Fisher Price). Accuracy is OK, but we've really been underwhelmed by its performance.
h1n1is4pigs
06-02-2012, 00:31
using RS-775s with 4:1 reduction direct driving to 8" wheels with about 3" of compression we are getting very consistently balls going well over 30 feet
the spin-up time between shots is too long.What voltage are you running the motor(s) at? i.e. what is the motor voltage when the wheel has come up to speed and no balls are being fed to it?
Aren Siekmeier
06-02-2012, 01:02
Has anybody posted test data anywhere that includes wheel rpm, wheel diameter, squeeze, launch angle, and distance?
I've been keeping an eye out, but haven't seen anything yet. It would be most useful as a sanity check for some of the calculations.
We took a decent amount of data with ball compression, I'll see if I can't dig that up and compile for public posting.
We're gearing for a top speed at the wheel edges of about 80 ft/s (54 mph!), since this is about double the speed of 40 ft/s that worked well for basically any shot in the near half of the field in our testing. So for a 4" wheel that's a 5:1 off of FPs (0673) and we have nearly the equivelant of 4 of them (2 FPs, 2 550s), so we're making up for lost torque (see here (http://www.chiefdelphi.com/forums/showpost.php?p=1120802&postcount=10) for more of the reasoning behind that). We're using the same 8" plaction wheels we tested with because that tread seemed to work better than others we tried briefly. I pushed for wheels smaller than the ridiculous 8 inches, but nobody ever tested anything smaller (we have screwed up ways of setting priorities, working on fixing that...) and people were worried about differences in traction with smaller wheels, we stuck with 8". The reasoning for smaller wheels was mostly weight, with a marginal benefit in how the frame could be layed out (more freedom), as well as just not having ridiculously huge wheels. But no one listens to me anyway... In any case, with our 8" wheels we have 10:1 total reduction off of 2 FPs and 2 550s in the works.
Garrett.d.w
06-02-2012, 01:16
Two weeks ago I made a pitcher using two rollers made out of 2.5" PVC covered in wedge top tread. This was powered by two 775s on banebot 46-1 gearboxes with no chain reduction. The compression of the balls was about 2 inches.
I would be comfortable shooting from half court with this thing. It is also good for rapid fire, the 775s have so much torque that the rollers will spin at a steady speed no matter what.
pfreivald
06-02-2012, 07:04
My team is using a similar set up. Were getting about 16ft using belts driven by 1:1 sprockets off a AndyMark gearbox (the one for the Fisher Price). Accuracy is OK, but we've really been underwhelmed by its performance.
Accuracy at key range will be more important than being able to chuck a ball 30+ feet, IMO.
What voltage are you running the motor(s) at? i.e. what is the motor voltage when the wheel has come up to speed and no balls are being fed to it?
They were using home-built speed controllers with a knob control -- I'm not sure what the duty cycle voltage was. (I wasn't there... had a funeral this weekend, so Saturday was the first FRC day ever without me there. Quite weird, I must admit!)
Saturday was the first FRC day ever without me there. Quite weird, I must admit!)
Not weird, but impressive. I hope your students recognize and appreciate your dedication!
pfreivald
06-02-2012, 16:51
Not weird, but impressive. I hope your students recognize and appreciate your dedication!
I believe they do!
mdiradoorian
06-02-2012, 17:50
It depends on the motor(s) you are using, the diameter of the wheel(s), whether you are using only one wheel or top&bottom, the moment of inertia of the wheel(s), and other factors.
That is very true because the bigger the wheel the faster the surface speed is. Our previous shooter prototype had 10" wheels and after using a cim motor we were using a 2 to 1 gear ratio, and the surface speed was abot a 130 miles per hour.
RRLedford
15-02-2012, 02:15
Team 3135 has a prototype shooter with two 8" AM-FIrst (AM-0420) wheels from 2010 mounted side-by-side on a 5/8" shaft. and temporarily driven by a single CIM with 1:1 ratio using dual 30 slot pulleys and the KOP timing belt.
Opposite the wheel pair is a 1/4" plywood sheet that the balls squeeze against it with ~2" peak compression. The plywood is covered with high friction shelf liner mesh. So, we are near near 5000 rpm on the 8" wheel pair.
We had to sand a radius on the outer corners of the wheels, since they were too sharp and were making lines on the balls. The wheels are somewhat lightweight, and it seems that the RPM drops a lot during the brief interval of ball compression while in contact with the wheel, so we are considering adding a large 60-tooth aluminum sprocket that has an outer hole circle where we can add several screws with steel washers for a better flywheel effect. This should reduce wheel slowdown and raise the exit velocity of the balls.
This initial CIM driven setup has us making baskets on a real basketball court from half court into the 10" high basket.
We are replacing the CIM motor with either a pair of the Banebot CIM-U-Lator 775 (1-motor) gearboxes, each gearbox powering one end of the wheel shaft, or with a single Banebot CIM-U-lator 550-2 (dual-motor) gearbox. These gearboxes give a 2.7:1 reduction of these high RPM motors.
AN initial pair of CIM-U-lator=>775 gearboxes were ordered with a total of four 775 motors, because of the internal shorting problem. Expecting maybe one in four to test bad, our order arrived today with THREE SHORTED MOTORS out of four. We are bailing out on using the 775 motors for ANYTHING EVER AGAIN. I hope we can get a replacement pair of DUAL 550 CIM-U-lator gearboxes shipped ASAP.
-Dick Ledford
As a data point, we are using a 2011 FP motor on an AM planetary gearbox (3.67:1). The shaft is then connected to a 1:1 chain drive setup. We calibrated tonight and this setup at max speed through a Jaguar gives about 2200 to 2300 RPMs. Less than we expected, I guess there is more drag than we had figured.
As a data point, we are using a 2011 FP motor
When you say "2011 FP motor", what does that mean?
I looked at the 2011 KoP, and there's a FP 00801-0673 listed, and a FP gearbox with a 9015 motor. These motors are very different. Which one are you using?
We calibrated tonight and this setup at max speed through a Jaguar gives about 2200 to 2300 RPMs.
How did you measure the RPMs?
Less than we expected, I guess there is more drag than we had figured.
What were you expecting?
our order arrived today with THREE SHORTED MOTORS out of four.
For the record, did you call Banebots to report this, and what was their response?
RRLedford
17-02-2012, 12:07
For the record, did you call Banebots to report this, and what was their response?
Yes we did call. They found it hard to believe and wanted us to retest, but ZERO ohms is ZERO ohms. How many zero readings are we supposed to get?
They were very helpful though and sent us (2) gearbox face plates for converting both the 775 gearboxes over to dual 550 gearboxes via overnight ship.
So, we can finish our shooter now.
When you say "2011 FP motor", what does that mean?
I looked at the 2011 KoP, and there's a FP 00801-0673 listed, and a FP gearbox with a 9015 motor. These motors are very different. Which one are you using?
How did you measure the RPMs?
What were you expecting?
Sorry, we are using the 00801-0673 on the AM 3.67:1 planetary gearbox. The motor curve data shows a free speed of over 20,000 RPMs. Starting with the 3.67:1 ratio on the motor and 1:1 final chain drive, the theoretical no load/free speed RPMs are over 5000 RPMs. Certainly there is the mass of the wheel, axle, chain, drag, etc. In the end all this loads it down by more than 50%, for a final max speed of 2200 to 2300 RPMs.
To detect wheel speed we are using a custom Hall Effect Latch encoder we built based on the US1881, and verified the data via a Logic analyzer (clean and consistent square wave), frequency counter on a multimeter, and via counters in Java on a digital input. So I am confident with the actual axle speed we are getting.
Under load with all the variables I am not complaining about the 2200 RPMs, it was just a data point for the original poster on our FP 00801-0673 plus AM 3.67:1 geared shooter. :)
we are using the 00801-0673 on the AM 3.67:1 planetary gearbox ...and 1:1 final chain drive...final max speed of 2200 to 2300 RPMs.
2250 * 3.67 = 8258 motor RPM
0673 @ 12 volts @ 8258 RPM is pulling 68.5 amps and generating 512 watts of waste heat.
This seems like it would either trip a breaker or smoke the 0673.
Are you sure of your numbers?
2250 * 3.67 = 8258 motor RPM
0673 @ 12 volts @ 8258 RPM is pulling 68.5 amps and generating 512 watts of waste heat.
This seems like it would either trip a breaker or smoke the 0673.
Are you sure of your numbers?
I was pretty certain until you put it that way. :) It doesn't appear to run very hot so far. I can certainly check the current today by measuring the voltage drop across a high wattage, low resistance power resistor in series.
I have seen plenty of shooters on CD this year with the same kind of setup - AM gearbox, FP motor, aiming for about 3 to 5,000 RPMs. If your concern is correct aren't we all in the same boat? If not, any idea where my assumptions have gone wrong?
Thanks!
I have seen plenty of shooters on CD this year with the same kind of setup - AM gearbox, FP motor, aiming for about 3 to 5,000 RPMs. If your concern is correct aren't we all in the same boat?
Yours is the first I've seen on these forums with an 0673 running at 39% speed at full voltage.
If not, any idea where my assumptions have gone wrong?
If you are sure of the motor part number and the gearing, then I'd suspect either the wheel speed measurement is incorrect, or else the 2250 rpm is not at full motor voltage.
So we just rechecked wheel speed, and it maxes out at about 3500 RPMs, or 12,991 RPMs motor speed. We measured the voltage across a .01 ohm power resistor and it was about 210 mv. So at max speed it looks like we are drawing about 21 Amps. In-rush current when the wheel starts up peaks out at about 50 Amps for a brief moment.
Now we need to check the motor voltage, maybe that is where the flaw in the applied math is. SW is sending a "1" to the Jaguar, but I guess there is more voltage drop that I expected.
Thanks again for the insight. Everything is working ok, at this point it's just a study of the real world numbers we are collecting compared to the expected data based on the motor curves.
So we just rechecked wheel speed, and it maxes out at about 3500 RPMs, or 12,991 RPMs motor speed. We measured the voltage across a .01 ohm power resistor and it was about 210 mv. So at max speed it looks like we are drawing about 21 Amps. In-rush current when the wheel starts up peaks out at about 50 Amps for a brief moment.
3500*3.67 is equal to 12845 motor rpm.
According to the motor curves,
An FP-0673 at 11.79 volts (= 12 - 0.210) and 12845 rpm should be drawing about 40 amps.
Since you measured the current to be 21 amps, let's use that data instead:
According to the motor curves,
An FP-0673 drawing 21 amps at 12845 rpm is getting about 9.7 volts. Add the 0.21 drop across the measurement resistor and your Jag is putting out about 9.9 volts. I'd expect it to be a bit higher than that, unless:
- your battery is weak
- you've got excessive resistance in your wiring or connections
- you're running a bunch of other heavy-demand stuff while you're running these tests
- the Jag is not putting out 100% PWM
- the motor curves are wrong or the motor you have differs substantially from them
Now we need to check the motor voltage, maybe that is where the flaw in the applied math is.
The math is sound, but the data or assumptions could be wrong. I agree it would be a good idea to check the voltage at the motor, to validate the above engineering calculations.
Everything is working ok, at this point it's just a study of the real world numbers we are collecting compared to the expected data based on the motor curves.
When the engineering calculations don't jibe with the observed test data, it could be a sign something is not right. I understand this doesn't necessarily apply to FRC, but in many engineering fields when the test results don't agree with the expected results based on calculations or simulations, you stop the presses and figure out why.
So we just rechecked wheel speed, and it maxes out at about 3500 RPMs, or 12,991 RPMs motor speed. We measured the voltage across a .01 ohm power resistor and it was about 210 mv. So at max speed it looks like we are drawing about 21 Amps. In-rush current when the wheel starts up peaks out at about 50 Amps for a brief moment.
Now we need to check the motor voltage, maybe that is where the flaw in the applied math is. SW is sending a "1" to the Jaguar, but I guess there is more voltage drop that I expected.
Thanks again for the insight. Everything is working ok, at this point it's just a study of the real world numbers we are collecting compared to the expected data based on the motor curves.
Ok here is a good reason it is good to compare real world measurements to the expected/theoretical numbers. After scratching our heads we asked SW one last time if they were sending '1' to the Jaguar. They said 'Yes. Wait - It's a Jaguar? We have it coded as a Victor'. Opps! Since Jaguars have a larger band of PWM Full/Reverse signal input, that means they really were not sending it a '1'. They changed to Jaguar and now it runs a good amount faster at hight input values and is much closer to the curve.
Thanks for sticking with me on this one Ether! :)
Thanks for sticking with me on this one Ether! :)
and thank YOU for following up and not quitting.
vBulletin® v3.6.4, Copyright ©2000-2017, Jelsoft Enterprises Ltd.