A real puzzle - Electromechanical fault
 

We started build season with a plan to use mecanum wheels on our robot. We built up a basic robot and tested it on the bridge. We found that the mecanums didn't give us enough traction to climb the bridge. However, the robot worked fine at that point. We could drive, turn and glide.

We decided to switch to conventional wheels. We made the programming change from holonomic to arcade drive. Now our robot will drive forward and backward, but whenever we try to turn, we see no motion, hear a slight decrease in the speed of our cooling fan, and the RSL goes into fast blink mode, indicating a fault. As best we can figure, our power system is becoming overburdened and the voltage to our router is getting too low, so it resets.

The battery has a fresh charge.

If we disconnect either side of our drive train (two motors at a time) electrically, we still get the same behavior, so I don't believe we have one bad Jaguar or CIM.

If we lift the wheels off the floor and try to turn, all the wheels turn in the proper directions, no faults.

We were driving our wheels directly off of CIMple boxes, and I was concerned that the gear ratio was too low for conventional wheels, so we switched back to the mecanums and continued to have the fault.

At this point I'm out of theories. Anyone have any to offer, or diagnostics to try?

Looks like another difficult Thursday at the regional...

Re: A real puzzle - Electromechanical fault
 

Important questions:
How many wheels and what configuration? Long robot or wide robot?
What size wheels?
Did you chain all your conventional wheels together on one side?

Fast blink on the RSL means you're losing comms, which mean you're dropping your battery voltage extremely low, which means you're drawing one heck of a lot of current. If things work fine with the wheels off the ground, you don't have a short and this is a mechanical issue with your drivetrain. Which, yes, means a pretty hectic Thursday. All the more reason to pinpoint the problem and go in with the right solution.

Re: A real puzzle - Electromechanical fault
 

With a good battery and the correct wiring, the system should protect itself from router brownouts. You do have the DAP-1522 powered through the 12V-to-5V converter plugged in to the dedicated boost-regulated 12V connector on the Power Distribution Board, right? Losing communication when trying to turn suggests that you might have it connected to one of the unregulated red Wago terminals instead, or perhaps directly to the 5V camera supply without the converter.


You didn't mention anything about rewiring your motors when you made the wheel switch, but if you did, take a very careful look at where all the wires are going. If something got swapped, you could end up seeing really weird problems with two different Jaguars' outputs connected to the same motor.

Re: A real puzzle - Electromechanical fault
 

If you're driving directly off a cimplebox, then I would say your gear ratio is too low. Especially when you turn.

Re: A real puzzle - Electromechanical fault
 

@Kevin: 4 wheels, each directly connected to their own CIMpleBox with a single CIM driving. Long robot. 6 inch wheels. No, the conventional wheels were just stuck onto the hubs off the CIMpleBoxes.

@Alan: Yes, we're using the 12V to 5V converter off a branch circuit. No rewiring done.

@Coffeeism: That's actually what I'm hoping, because I know what to do about that. However, I'm puzzled about why it was working in holonomic mode originally.

Thanks everyone for the suggestions. If you got more, I'll take 'em!

Re: A real puzzle - Electromechanical fault
 

4 wheel, long wheelbase, not enough gear reduction, and it doesn't move, but draws a lot of current....that sounds about right. You have a mecanum drivetrain, but no mecanum wheels. You probably should not expect it to be able to turn, sorry.

As for why it doesn't work when you put it back, I don't know.

Re: A real puzzle - Electromechanical fault
 

What's a "branch circuit" ?

The 12V should be coming from the 12V regulated supply on the PDB.

Re: A real puzzle - Electromechanical fault
 

Also see R42B. The current draw incurred by turning your robot causes a voltage dip which reboots your radio. The regulated supply highlighted above is designed so that voltage dips of the magnitude you're seeing (I'm assuming that your bridge isn't connected to this power supply) won't power cycle the DAP.

Did you switch out your battery at any point? If you didn't, that would probably explain why the issue recurred after you returned to mecanum.

Ether, branch circuits seem to be implicitly defined in the manual as anything powered by a fused connection.

Re: A real puzzle - Electromechanical fault
 

Order Omni wheels and put them on either the front two wheels or the back. This will allow you to turn.

Long wheel bases with 4 traction wheels at the corners will not turn.

Re: A real puzzle - Electromechanical fault
 

Could it be that originally your robot was under a lighter load without the rest of your components?

If you can afford the weight you can swap out your cimple boxes for toughboxes. My team is running 8" mecanum wheels of 4 toughboxes with no issues.

Re: A real puzzle - Electromechanical fault
 

Try a different battery. Your battery may show correct voltage but may not be able to supply current under load. This is a classic symptom with car batteries.

Dr. Bob

Chairman's Award is not about building the robot. Every team builds a robot.

Re: A real puzzle - Electromechanical fault
 

How low can we pull the battery before the PDB boosted/regulated 12V output droops too low to keep the 12V-to-5V alive to keep the DAP-1522 connected?

We are a real current hog this year and I worry about this.

Joe J.

Re: A real puzzle - Electromechanical fault
 

The 12V-5V converter is good to about 7V input.

Re: A real puzzle - Electromechanical fault
 

The two PDP protected 12v & 24v outputs are good down to about 4.5v, maintaining their 12v/24v.
The 5v PDP output get cut off at around 5.5v.
Beware sharp sudden spiky dips at low voltages.

See Power Distribution Board.pdf, pages 6 & 7 Specifications for minimum voltages for the 24v protected supply, 12v protected supply and 5v outputs.

P.S.
Correction: the regular 12v outputs do not get cut off at low voltage. They follow the battery on down.
But, something cuts controlling PWM/relay outputs at less than 6v, so the effect is the same.

Re: A real puzzle - Electromechanical fault
 

Thanks. I should have looked for the manual myself.

In practice, do they have caps on the inputs to survive small spikes that dip below these values or does noise make the radio go dark?

Also, have people had any success monitoring battery voltage and then modifying motor outputs to prevent such brown outs? It seems like it would be reasonably straight forward: when battery voltage dips below X find the largest users of current (probably wheel motors) and scale their output back until the battery voltage recovers. With a reasonably fast loop time, I suppose that we can keep from browning out in the first place (you'll lose power on the motors but you stay connected -- seems like a bargain worth making).

Joe J.