Robot doesn't turn in high gear

[Chuck Glick]

I've seen their setup first hand. They have four 8"(?) custom traction wheels in back and 2 AM omni's in front. After watching them at Philly, their low gear was relatively fast, very similar to our speed, but we used the AM Single Speed Gearboxes on a 1 to 1 output to wheel ratio. So I would suggest that you increase the size of your sprockets on the wheels, decrease the size from the gearbox, or even both. That hopefully should help with the turning problem.

-Chuck

[newton418]

Team 418 used a 6 wheel drive, 4 small CIM-DeWalt transmission drive train. We lowered the center wheels 3/32", which is enough to sink the center wheels in the carpet, but still have all 6 wheels contact the ground. In 2nd gear we realized we had some difficulty turning (the RC was resetting, I believe), and after talking with the electronic guys we think we figured out why. When designing the drive train, I figured we could run each small CIM at 40 amps (since they are on a 40 amp breaker); however, I was informed that the power distribution block is rated for 85 amps (I'm not exactly sure, but around this number). This means we weren't getting the current, and thus the torque, we had planned on. Perhaps you guys are facing a similar problem.

[Al Skierkiewicz]

Originally Posted by newton418

When designing the drive train, I figured we could run each small CIM at 40 amps (since they are on a 40 amp breaker); however, I was informed that the power distribution block is rated for 85 amps (I'm not exactly sure, but around this number). This means we weren't getting the current, and thus the torque, we had planned on. Perhaps you guys are facing a similar problem.

This is another one of the myths of electrical design. The Rockwell blocks are designed for continuous current of 85 amps per block. The 40 amp breaker will function with up to 600% of it's current trip rating for several seconds. Neither of these devices limit the current to that of their ratings. But there are limiting factors in the wiring of the robot. Things to consider are the series resistance of the wire, terminals, Victors, and distribution. A simple rule of thumb I use is the "wire/foot" or WF. It is simple to equate series resistance with the resistance of a foot of #10 wire which is .001 ohms. Using this as a rule of thumb, then the battery has 11 WF internally, the #6 feed wiring has 0.5 WF per foot of wire (you must include both the postivie and negative wire) while the #12 has almost 2 WF per foot, the Victors have 4 WF of series resistance and you can have between 1 and 3 WF for every crimp if not done with proper tooling, not tight on the block or screw terminal or not fully crimped. So, how does this affect your design? At 100 amps of current (roughly the stall current of either Chalups motors or the old FP motors) each wire foot will drop 0.1 volt. Just take a look at the circuitry that feeds your motor, add up the resistance using this method and multiply by 0.1 to find the voltage drop to your motor.
Ex. 4 ft #6 wire, breaker panel, one victor, 4ft of #12 from fuse panel to Victor to motor, 20 connections= (4x0.5)+1+4+8+20=35x0.1=3.5 volts drop in all losses at stall. This translates to a loss of about 1/3 of the available current or about a loss of almost 100 oz.in. of torque and a drastic change in the efficiency.

[newton418]

Originally Posted by Al Skierkiewicz

This is another one of the myths of electrical design. The Rockwell blocks are designed for continuous current of 85 amps per block. The 40 amp breaker will function with up to 600% of it's current trip rating for several seconds. Neither of these devices limit the current to that of their ratings.

I was aware that the 40 amp breakers could be pushed past their limit for small amounts of time, but decided to design the drive train such that we wouldn't need to constantly push the breakers, since we could get sufficient performance at 40 amps. I am much less familiar with the Rockwell blocks, since they are new and I do not work on the electronics. I figured that since they are rated for 85 amps they could be pushed past this limit, but I have no idea by how much or for how long. Could you further explain what the Rockwell blocks can actually handle?

As for the resistance in the circuitry, we kept the wire lengths to a minimum, and will definitely be more conscious of it in the future; however, I do not understand how this would fully explain our problem. At 40 amps, a small CIM outputs ~.8 N*m (please correct me if my calculations are wrong), and IF we could run them significantly beyond this, say, near stall current, then even if you take the ~2.2 N*m stall current then the wire resistance would have to be somewhere around 76 WF. Though I am not the most familiar with our electronics system, I am confident it does not have this much resistance. Now, I did make several assumptions when calculating the torque needed to turn in 2nd gear, so in actuality .8 N*m of torque might not have been enough; but, assuming that the difference is negligible, how do you explain the problems we faced?

Also, I have just talked to our electronics/code mentor and he suggested that we bypass the Rockwell block to see if it is (at least in part) the problem. He also suggested we monitor the voltage on both sides of the Rockwell block using a USB DAQ, so when we get those results I will make sure to follow up on this.

[MrForbes]

Originally Posted by newton418

Could you further explain what the Rockwell blocks can actually handle?

They can probably handle a couple hundred amps for a very short time...

what happens with electrical parts when a lot of current goes thru them, is that they get hot....every part has resistance, and resistance turns electrical power into heat energy. When the metal parts get hot enough, they melt thru and can no longer conduct electricity.

The point of having circuit breakers and fuses, is to put a "weak link" in the circuit that will fail first, and prevent the other wiring and components from overheating.

The power distribution system for the robots is a well designed system, and will do it's job if you follow the instructions carefully so you build it as designed (and make sure to tighten all the wire connections firmly)

[Al Skierkiewicz]

Originally Posted by newton418

Could you further explain what the Rockwell blocks can actually handle?

As for the resistance in the circuitry, we kept the wire lengths to a minimum, and will definitely be more conscious of it in the future; however, I do not understand how this would fully explain our problem. At 40 amps, a small CIM outputs ~.8 N*m (please correct me if my calculations are wrong), and IF we could run them significantly beyond this, say, near stall current, then even if you take the ~2.2 N*m stall current then the wire resistance would have to be somewhere around 76 WF. Though I am not the most familiar with our electronics system, I am confident it does not have this much resistance. Now, I did make several assumptions when calculating the torque needed to turn in 2nd gear, so in actuality .8 N*m of torque might not have been enough; but, assuming that the difference is negligible, how do you explain the problems we faced?

Also, I have just talked to our electronics/code mentor and he suggested that we bypass the Rockwell block to see if it is (at least in part) the problem. He also suggested we monitor the voltage on both sides of the Rockwell block using a USB DAQ, so when we get those results I will make sure to follow up on this.

OK,
The Rockwell blocks are simply a piece of copper with a wire clamp at each end embedded in a plastic housing. The rated current is the safe limit when the block has continuous current (measured in days) passing through it when mounted in a closed electrical enclosure. (i.e. no cooling air flowing around it) The safe limit is based on the temperature rise of the block that would cause damage to the wire and block. Since we are only using this block for a few minutes at a time, the heat rise is so low that several hundred amps of current flow will not significantly raise the temperature. However, a high resistance connection (loose, small wire size, improper crimp) will cause significant temperature rise as current flows. I would expect you to find the voltage drop across the Rockwell block to be in the millivolt range.
I did not do an actual calculation but your numbers look like they are in the ballpark. Remember that the actual performance of the motor is skewed down the curves by the series resistance of your system. The motor is tested with a power supply that has no output resistance whereas the battery is .011 ohms. The resistance limits/reduces the available current to the motor and so for a given design point, your speed and torque are reduced. Don't forget to include in your calculations the efficiency of the transmission you are using. As you add up mechanical losses and try to push the motor to gain more speed, the current begins to skyrocket. I think that if you back into the operating point by measuring the wheel speed of your robot in a turn you might be surprised to find that you have moved far to the right on the curve where current and torque is very high, but efficiency has started to drop off. (max power is at 2655RPM but current is at almost 70 amps) Without seeing your robot design and watching it perform it is hard to tell exactly where the problem might be. So a few ideas...check the wheels by hand with the power off and no brake on the Victors. Do both sides move freely and with the same force? Are drive motors on both sides at about the same temperature after a match? Is there a similar amount of wire in each side of your robot? i.e. does one side have a longer length of wire to the block and fuse panel than the other? Do you bring the black wires back to the Rockwell block (as required) or do you instead attach them to the small fuse panel? A picture could tell us a lot.

[Al Skierkiewicz]

Now that I have had some sleep, it is important to also know the diameter of the wheels you are using.

[Jared Russell]

Having seen 1647's drivetrain this weekend, I feel obligated to reply.

Their machine has a 6WD configuration with four custom 8x3" wheels with roughtop tread in the middle and rear, and unpowered omniwheels in the front for support.

I can't comment on the ratios they are using, but I WILL say that I noticed a few things when they were running. One, their right gearbox had a big bend in one of the plates (from what I can't guess). I pointed this out to a team member but I don't think they did anything to fix it.

Two, their machine makes a terrible noise when turning. I think that this is a byproduct of the aforementioned dent and possibly grease/chain tension issues.

But all in all, their middle and rear wheels are VERY close together and I really don't see why they would have issues turning. I don't think that changing the motors will make a huge difference - I'm willing to bet that the problem is in the gearbox, greasing, and chain tension.

That said, I loved 1647's machine, and they were a fierce competitor in the playoffs!

[Al Skierkiewicz]

OK, 4 wheels that are 3" wide is bad enough, but when placed close together, they exhibit less friction than spaced at 25 inches. I wonder if the choice of 8" wheels were included in the original calculations.

[Waynep]

Rusty,
In addition to switching to the small CIMS, another quick fix maybe to recheck your weight distribution. Slide some of your heavier components (battery, compressor) around fore and aft relative to your drive wheels and see what effect this has. Weight distribution has a huge effect on driving characteristics of a machine. Good luck.
-wayne

[Viper37]

This is a classic example of gearing. You wouldn't try to start driving your car in 4th gear would you? The engine or motor's in this case simply cannot displace that much torque from dead stall.

In this instance it doesn't really matter how the weight is distributed. The AndyMark's handle beautifully in first gear, as well as in second. The difference is, in first low speed handling is good, the handling in second gear is at the top of the power band.

[Beta Version]

Originally Posted by Al Skierkiewicz

This is another one of the myths of electrical design.

Thank you. Thank you. Thank you. I can almost *almost* guarantee that your problem is in no way electronic, the cims are getting all the power you need. The problem is in your torque, moment, and traction. Put a long bar/axel through your wheel, weight it with a known weight, then take a fish scale and pull the wheel sideways and see how much force it takes to drag the wheel sideways. Find your k of friction, and calculate how much torque your needing, then see if you need less traction on your wheels. Even if you've lowered your center wheel, your weight needs to be balanced for that to work.