pic: 3CIM Ball Shifter

[cad321]

loading photo ...

I love the 3-CIM shifter, but the ratio is terrible. 17 ft/s is a great high gear for 3 CIMs, and 5 ft/s is a great low gear for 3 CIMs, but speeds of 9-12 fps low or high gear don't benefit much from the extra CIM. In order to see a noticeable increase in power/acceleration with the extra CIM, your high gear needs to be pretty high (I'd say 17 ft/s is a good starting point), and your low gear needs to be pretty low (4-5 ft/s is what I'd personally do with 3 CIMs).

tl;dr - the extra CIM isn't going to help unless your high gear is really high, and your low gear is really low.

[KrazyCarl92]

Originally Posted by Andrew Lawrence

I love the 3-CIM shifter, but the ratio is terrible. 17 ft/s is a great high gear for 3 CIMs, and 5 ft/s is a great low gear for 3 CIMs, but speeds of 9-12 fps low or high gear don't benefit much from the extra CIM. In order to see a noticeable increase in power/acceleration with the extra CIM, your high gear needs to be pretty high (I'd say 17 ft/s is a good starting point), and your low gear needs to be pretty low (4-5 ft/s is what I'd personally do with 3 CIMs).

tl;dr - the extra CIM isn't going to help unless your high gear is really high, and your low gear is really low.

You could go after a traction limited at 40 Amps per motor Low Gear in the range of 7-9 fps with 3 CIMs, whereas this range is around 4.5-6.5 fps with 2 CIMs per gearbox. However, then you're running the risk of popping the Main Breaker with a 240 Amp traction limit with just your drive train. I would tend to go with a 160 Amp traction limit in low gear on my drive train even with 6 CIMs, just to be sure I won't pop my main breaker in any reasonable sustained pushing match. At least until I would get more experience working with 6 CIMs in drive gearboxes. Because of this preference, the advantage of 6 CIMs in a shifting drive train would be better acceleration at higher speeds, not a faster low gear (even though reasonable arguments could be made for a faster low gear).

I do agree though, I already much dislike the gearing differential between low gear and high gear of the VP ball shifter for 2 CIM gearboxes, and it seems even less attractive for 3 CIM gearboxes.

[apalrd]

Originally Posted by Andrew Lawrence

I love the 3-CIM shifter, but the ratio is terrible. 17 ft/s is a great high gear for 3 CIMs, and 5 ft/s is a great low gear for 3 CIMs, but speeds of 9-12 fps low or high gear don't benefit much from the extra CIM. In order to see a noticeable increase in power/acceleration with the extra CIM, your high gear needs to be pretty high (I'd say 17 ft/s is a good starting point), and your low gear needs to be pretty low (4-5 ft/s is what I'd personally do with 3 CIMs).

tl;dr - the extra CIM isn't going to help unless your high gear is really high, and your low gear is really low.

I couldn't disagree more. I think 5 ft/s is a terrible low gear speed.

Think about what your functional objectives are for the drivetrain in general. Set some hard numerical goals. Make sure you include all reasonable aspects, including weight, power input (motor selection or options), and dynamic performance goals. Top speed is actually a terrible metric of drivetrain performance. Acceleration and dynamic performance are huge, especially when driving fast, but actual top speed should not be a functional objective. Most objectives in FRC are time to x, where x is game dependent. We like to use sprint distances, the time it takes to travel x feet from a standstill, where x is dependent on the game. Sometimes there are several x's, depending on the game. 15'-20' long distance analysis is usually good for mostly open field games, while 10' is usually good for short distance pick and place maneuvers, and you can prioritize based on your game strategy.

THEN, start turning those functional objectives into machine attributes, with numbers and analysis to back them up. Do some math, using one of the many spreadsheets available on chief delphi if needed, to see what the minimum hardware possible to achieve your functional objectives is. Can you meet all of your dynamic goals with a single speed gearbox for your motor input? Good. The fun stops here then. You should then use hard test data, usually by building it and testing it fully weighted, to see if your correct.

As you get better at this, and design, build, and test more drivetrains, you will get much better at setting the functional objectives correctly to start. I also suggest you benchmark other robots and teams, to see what kind of performance they can get with what power input they have, to set your functional objectives.

IF you decide you can't meet your functional objectives with a single speed gearbox, then you can evaluate a multispeed gearbox. Since it's easiest to start with COTS, you can evaluate what ratio spreads you can get from the commercial suppliers. AM sells 4:1 and 2.56:1 spread, while Vex sells 2.2:1 spread, but it's possible to make something different by changing the mating gears. With the dog shifter, you also have to be careful your gears don't crash into the dog.

The ratio spread is the most important attribute about a shifting gearbox. Modifying the final drive ratio (final chain or gear reduction) is super easy given the vast assortment of COTS 20dp spur gears, but the shifted gears are much harder to change. Initially, do your analysis with a COTS shift spread and modify the final drive ratios to get your desired performance in either or both gears (depending on what your desired performance is).

You will find that some of your performance metrics are easier to meet in low gear, and some are easier to meet in high gear, and some aren't that great in either gear. This is where, if possible, you would ideally like to optimize the shift ratio spread to put your two gears where they best meet your targets. This is also where you learn the importance of setting your functional objectives properly, and simulating properly. All of this comes with experience, which I don't expect you to have yet, but you can start from the vast knowledge of the Chief Delphi community.


My ideal vehicle speeds (useful for comparison, not terribly useful for analysis) of a 6-CIM 2 speed drivetrain are around 8.5 and 15 fps for low and high respectively (90% eff speeds), based on several years of simulation, test data, model improvement, and more simulation. But I don't have enough data for 6-CIM drivetrains, I need to build, instrument, and test one to see how I like it, and improve the model accordingly to design the next one.

Is your functional objective to be traction-limited at 40a/motor (240a for the entire gearbox, note) in low gear? Are you actually going to push a solid wall and hope the main breaker (120a) doesn't trip?

@krazycarl92, IMHO the simulation and data I have suggests a lower ratio spread is better, and that the VP ball shifter has too much spread. It's all about what you really want, and really need, and how you want to use your gears.

Originally Posted by apalrd

I couldn't disagree more. I think 5 ft/s is a terrible low gear speed.

The point wasn't that for any regular old shifter you need a super high gear and a super low gear. It was that by adding another CIM motor to each side, you don't see any major benefit until upwards towards 16-17+ ft/s, and any speed below 5-6 ft/s. I got this information from JVN's acceleration calculator. You can see by testing multiple ratios that between the speeds of 8 ft/s and 12 ft/s there is almost no noticeable difference in acceleration at all between 4 and 6 CIM drives (even between 13-15 ft/s the change isn't large enough to be too noticeable, though it is mathematically larger than the change found in the 8-12 ft/s range). This "magic window" holds some of the most favored speeds in FRC unfortunately, which is why I suggested a really low gear and a really high gear to fully take advantage of the extra CIMs. If the shifter was a 4-CIM, then 10 ft/s is a wonderful low speed since you're getting everything you can out of your motors.

[apalrd]

Why do you want a really low low gear? What objectives does it achieve?

Why do you want a really high high gear? What objectives does it achieve?

I don't really care if you optimize the benefit of adding the extra motors, the goal should be to optimize the system as a whole.

There are a lot of variables that you aren't likely seeing. Not limited to electrical system resistance, battery current capacity, motor heating/thermal performance, and the acceleration/current profiles which are all important differences between 4 and 6 motor drives. I can't say for certain how important they are, since I don't have data proving most of it one way or another, but I have lots of subjective and theoretical suggestions.

I am well aware that 6 motors is a marginal performance gain theoretically. The limiting factor moves from the motors to other things somewhere between 4 and 6 CIMs, which is why you don't see much improvement.

[donkehote]

I think the 6 cim drive systems should incorporate amp clamp style amp meters feeding info back to the crio. Used to trim the victor output to keep the amp draw on the main breaker below a threshold (dont know what would be safe but you get the idea) therefore allowing non traction limited drive. when in a push fight, you may get close to the amp limit, and doing this kind of monitoring of the overall amp draw could save you popping the main breaker, and going inop for the match.

I know those amp clamps are quite cheap, like less than $60, however, using induced voltage on a small gauge wire rapped around the battery lead might provide a similar kind of overload sensor.

[bardd]

Did you make sure the plates will hold up? It looks like the top CIM's mount will bend back the first time the robot get hit from the side. I think you should run a stress analysis before continuing your work. Other than that I really like your design! Looks very streamline. Do you intend to build it or are you designing just for fun?

[KrazyCarl92]

Originally Posted by apalrd

My ideal vehicle speeds (useful for comparison, not terribly useful for analysis) of a 6-CIM 2 speed drivetrain are around 8.5 and 15 fps for low and high respectively (90% eff speeds), based on several years of simulation, test data, model improvement, and more simulation. But I don't have enough data for 6-CIM drivetrains, I need to build, instrument, and test one to see how I like it, and improve the model accordingly to design the next one.

Is your functional objective to be traction-limited at 40a/motor (240a for the entire gearbox, note) in low gear? Are you actually going to push a solid wall and hope the main breaker (120a) doesn't trip?

@krazycarl92, IMHO the simulation and data I have suggests a lower ratio spread is better, and that the VP ball shifter has too much spread. It's all about what you really want, and really need, and how you want to use your gears.

I fully agree that whatever drive train and gearing is chosen, it should be chosen as a solution to a set of functional objectives. Ideally these objectives are based on proper strategy, physical analysis, and experience. (This sounds an awful lot like engineering)

The two main functional objectives I would aim to achieve with most games and a 6 CIM gearbox are (over simplification):
-Traction Limited at 160 Amps Total for the drive train (best met in Low Gear)
-Time to travel a particular distance, based on strategy (best met in High Gear)

Usually the time to travel reasonable distances in FRC is optimized by gearing which corresponds to a top speed somewhere between 13-18 fps, accounting for voltage drop and inefficiency. This doesn't change terribly much between 4 and 6 CIMs. Our functional objectives and corresponding gearing seems to agree here.

Now looking at our other objective, we have to do some math to figure out what gearing we want. I arrive at the 160 Amp total current draw at the traction limit because 40 amps per motor has been used as a low gear standard by many teams for many years on 4 CIM drive trains largely without issue. Then last year, we hear about many 6 CIM drives tripping the main breaker. To be sure that our robot won't trip the main breaker, it would be nice to design it such that it won't, which we can achieve with proper gearing. These claims are also supported by the main breaker spec sheet. A properly functioning breaker will be at risk for tripping after 10 seconds at 240 Amps, whereas it will go for over half of the match at 160 Amps. This is also neglecting any other sources of current draw on the robot.
To do that math, we will need some assumptions:
-145 pounds for battery + bumpers + robot
-6 CIMs in drive train
-4 inch wheels
-CoF=1.2
-Efficiency = 80%

Our gearbox will need to be designed to overcome a tractive force of 174 pounds. If I want this traction limit to occur at 160 amps of total current draw, that will require that the drive train achieve this traction limit at 26.7 amps per motor. Based on motor curves, this corresponds to 63.2 oz. in. of torque output from each motor. This is equal to 3.95 in. lbs. Which means that the drive train as a whole will input 23.7 in. lbs. before taking efficiency into account.

Our desired output torque is 348 in. lbs. This leads to a gearing of 16.5:1, using the square root of 0.8 to account for efficiency (making the best guess assumption that efficiency losses apply evenly to speed and torque).

This gearing corresponds to a top speed of 4.8 fps using a speed derating of 85%. (5.6 fps theoretical)

This shows that a low gear speed of 5 fps for a 6 CIM drive train is perfectly reasonable given this particular design objective.

The gearing spreads for a gearbox to complete both functional objectives I mention would be between 2.6:1 at the lower speed end and 3.6:1 at the higher speed end.

Commenting on another person's experience because they speak of drive train specs in terms of top speed is undue. By that logic, teams like Simbotics and The Cheesy Poofs who have published on their website their robots' top speeds in each gear would really be in need of experience. This actually shows a wealth of experience because it takes tremendous experience to be able to look at a design, hear a number in fps, and then immediately know what this means for the performance of the drive system. Based on a number in fps or gear ratios, I can look at a teams robot design and understand what the implications are for their drive performance. If something seems off or unreasonable about gearing decisions, I will ask a team about their thought process. At this point, I usually learn something new or realize the team doesn't know what they're doing. It is the industry standard to speak in terms of top speed; we wouldn't answer the question "What's the top speed of your robot?" with "We travel 10 feet from a dead stop in 0.95 seconds."

[apalrd]

I agree that top speed is useful for comparison and benchmarking. OP also said it was his first gearbox design, so my experience comments are justified. We don't design around top speed, but we talk about it a lot when we compare gear ratios quickly. But for analysis, we switch back to our accel times, accel curves, and stuff.


The question really is if one of your functional objectives is traction limited push current.

We no longer use that as our metric for low gear. We design low gear to meet a lot of the short game dynamics objectives, and we're getting better at quantifying current. We spend a lot of time in high gear high current scenarios, so any way we can reduce those (high gear launches and zero-point turns especially) by improving low gear performance (and actually using it for short game driving) is a huge improvement to us in overall electrical draw.

[Chris is me]

While this discussion has quickly gone on a tangent, I do have a question: Other than lowering current draw for pushing situations, what's the point of a low gear? Unless you're traveling quite fast, your acceleration won't be notably better. For short distances it seems a low gear would be helpful, but because the distances traveled are so short, you're not saving that much time anyway.

I guess what I'm wondering is when someone would want to use a 7-9 FPS low gear with a 14-16 FPS high gear. It seems like to me the only situation where low gear would be more useful than high gear for those speeds would be when you have to push or make really fine adjustments, and both of those are accomplished well with a ~5.5 FPS, 40A traction limited low gear.

I don't have any experience here; I've only ever worked with single speeds, so I'd like to learn.

[AdamHeard]

Originally Posted by Chris is me

While this discussion has quickly gone on a tangent, I do have a question: Other than lowering current draw for pushing situations, what's the point of a low gear? Unless you're traveling quite fast, your acceleration won't be notably better. For short distances it seems a low gear would be helpful, but because the distances traveled are so short, you're not saving that much time anyway.

I guess what I'm wondering is when someone would want to use a 7-9 FPS low gear with a 14-16 FPS high gear. It seems like to me the only situation where low gear would be more useful than high gear for those speeds would be when you have to push or make really fine adjustments, and both of those are accomplished well with a ~5.5 FPS, 40A traction limited low gear.

I don't have any experience here; I've only ever worked with single speeds, so I'd like to learn.

We use our low gear in more of a West Coast Style. It seems like 254/968 (when I was growing up) rarely used their low gear, even for short distances. It was moreso for the few times you were forced to push through defense.

We use it in the same way, and then also use it for thins like balancing on the bridge last year.

For how our drivers operate, I view low gear as insurance that lets you gear as fast as you desire for other objectives, but still have a drive mode where you won't get destroyed by defense (tripping breakers, etc...)

Because of this, we love a real slow low gear.

[Aren_Hill]

Originally Posted by cad321

loading photo ...

Quick bit of advice, it is probably good form (on both the ball shifter and dog gear setups), to put your third reduction after the shifting shaft.

This method can be seen on the Ball Shifter with optional 3rd stage, and the AndyMark Sonic/Super Shifters.

This lowers the overall torque being transferred by the shifting components, which results in less chance of failure and lower required shifting forces.
See this thread for once instance of failure: http://www.chiefdelphi.com/forums/sh...ght=20+carnage

Doing this would also allow you to make your gearbox overall shorter as you can offset the last output shaft in whichever direction you like.

[apalrd]

Our driving theory around low (this was not always the case, but several of us on 33 have converged on this point independently) is that accelerating from a standstil is better in low gear, usually to around 60-80% of the low gear speed, than in high gear. So any time where we never reach peak speed in low, it's faster to drive in low. These are marginal gains in time. BUT, since high gear is in the 'bad side' of the motor power curve for the entire time (with 2:1 or more spread), the current draw from high gear will be WAY higher than in low. For the same or worse dynamic performance. This becomes MUCH worse on a dieing battery, which can frequently be seen near the end of a match.

We've sometimes run auto-upshift software to automate this, it begins the shift around 60% of low gear speed, but the shift takes time to execute under load (not sure exactly what speed it's at by the time the dog disengages).

We also run auton in low gear usually, for precision/control reasons, so a slightly high low gear is good for this. But we could run in high, we just don't, so it's not a huge objective.


I guess this differs from the 'west coast' opinion that 'high gear is where we operate, and low gear is just in case we need to push'. We operate in both gears.

[AdamHeard]

Originally Posted by apalrd

Our driving theory around low (this was not always the case, but several of us on 33 have converged on this point independently) is that accelerating from a standstil is better in low gear, usually to around 60-80% of the low gear speed, than in high gear. So any time where we never reach peak speed in low, it's faster to drive in low. These are marginal gains in time. BUT, since high gear is in the 'bad side' of the motor power curve for the entire time (with 2:1 or more spread), the current draw from high gear will be WAY higher than in low. For the same or worse dynamic performance. This becomes MUCH worse on a dieing battery, which can frequently be seen near the end of a match.

We've sometimes run auto-upshift software to automate this, it begins the shift around 60% of low gear speed, but the shift takes time to execute under load (not sure exactly what speed it's at by the time the dog disengages).

We also run auton in low gear usually, for precision/control reasons, so a slightly high low gear is good for this. But we could run in high, we just don't, so it's not a huge objective.


I guess this differs from the 'west coast' opinion that 'high gear is where we operate, and low gear is just in case we need to push'. We operate in both gears.

I don't disagree with your points at all, I like the level of automation you've built in.

I don't know that anyone has run a decent autoshift out here.

I can certainly appreciate the lesser current draw, a dirty secret of some teams out here is we buy a new set of batteries every year, and if we didn't we'd like show worse performance on the field.

Part of this is we practice A LOT, so we use the last 1-2 seasons of batteries for practice to keep the new seasons worth nice.