pic: Battery connector.....

[Thad House]

Originally Posted by MrBasse

I'm suprised that nobody has said anything about the fact that a robot that can do that repeatedly is designed wrong. If you stay within the limits of your supplied hardware you wouldn't have this issue. Nowhere in the rules does it say you have to draw the amperage that you are. Would it make for a slightly les competitive machine? Maybe. Would it reduce the risk of melting connectors and destroying components? Definitely.

It just seems to me the a lower amperage rated connector is something that would result in an engineering challenge. I'm pretty sure that is what most of us are here for.

This seems comparable to someone complaining that their drill press doesn't work very well as a mill...

I would actually have to disagree with your first statement. It has been repeatedly said that the breakers are in the system to protect the wiring and connections. If the breakers are not doing their job, and wiring is melting before the breakers do break, something in the rules have to change, because thats when things start getting dangerous. Either a new connector that can actually handle how much current the breaker will allow, or a smaller breaker, which is absolutely not the answer. So its a safety issue more then anything.

[MrBasse]

I had a feeling I would be disagreed with on that. I don't have a problem with that a all. I feel that the connector is definitely under rated for its purpose. But what I am saying is that we all know that rating going in. It wouldn't hurt to design for your components.

[Mykey]

Forklift at work...



I tend to agree that the source of the problem is not the connectors. It sounds like you have an issue in your robot.

The only way I can think of their being an uneven current flow is if you have a short through the frame to ground. Also, if one connection starts to fail before the other, then the resistance on one side will increase at a higher rate than the other. This could cause uneven heat production. Seems unlikely though.

[Nirvash]

Just from my experience with electronics and people using undersized connectors, it isn't odd to see one connector fail first or more dramatically.

[apalrd]

Way way way back when the Anderson SB-50 connector was selected and apparently Anderson approved the higher current rating, the kit (this is back when you could only use parts from the kit or Small Parts) contained 2 Bosch drill motors and 2 Fisher-Price motors, and an assortment of small motors. The batteries were the same as they are now, but they used a 60a main fuse instead of the 120a main breaker we now use. Shortly after, the Victor was used, then the IFI control system. Now the rules allow 18 motors of similar power to those 4 drill and FP motors, robots have nearly tripled in speed, and the games are significantly more complex.

Now, we design to the 120 amp breaker, and put a lot of effort into keeping that breaker from tripping. Sometimes we design a bit too close, and it safely trips, and we take steps to prevent it from tripping again. We know what it will take, thanks to data sheets and trip curves. We already do a lot of design based on the main breaker and 40a branch breaker trip curves. We know when they will trip, we can run simulations of accel events, travel times, and distances to optimize our gear ratio to our strategy. We know how far we want to push the main breaker and we certainly know that when we exceed it, we have to scale back a bit but the failure is safe (even if it could cost us a match). In addition, in our pursuit of perfection, we design to push the robot as hard as possible during a competition. Maybe we push too hard. Then we go and drive the competition-designed robot for hours to practice, or at fast pace off-season events, and the issues show up.

We don't have a number or 'melt curve' for the 50 amp plastic connector (and I can not design my robot to 50a instead of 120a and stay competitive) so we can't design around it. Sure, we could empirically melt a whole bunch of connectors with a several hundred amp current limited lab power supply and huge load (the only reasonably safe way to do it) to get data on connector melting vs time vs load, or we could spend $5 more and buy a properly sized connector so the engineered electrical limiting device (the breaker) trips first and safely instead of the undersized connector.

The whole point of the main breaker is to safely cut power when we exceed the electrical limits imposed by the FIRST electrical system. We shouldn't rely on a plastic connector to fail (possibly shorting and causing several hundred amps from the battery to start a fire) as an energy limit. The connector is far too small for our application (even if it was approved for the current draw way back when) and the company even makes a connector the right size for us.



It's not the robot, it's several robots over several years of practice and off-season events, designed at relatively common FRC speeds with quite efficient drivetrains. We run them (on the practice bot) for hours at a time, running roughly 50% duty cycle, but we can't upgrade the connector on the practice bot without upgrading all of the batteries (which we also need to take to competition, and plug into the comp bot).

[MrBasse]

And again, I'm not saying that you aren't doing anything that a lot of FRC teams are doing everyday. I think it is safe to say that everyone pushes the limits of the Anderson connector. However, I think what we can learn is that you have pushed it too far a few times. Either by design or by use. I'm not saying you should design for 50A, as that would take the amazing machines you make down quite a few levels. I agree that FIRST should upgrade the connector to a higher rated connector. However, I also know that they haven't. I also haven't seen hundreds of other teams reporting melted connectors. So perhaps it is something in the way you have designed/built/run your machines.

I appreciate the response and I see where you are coming from, but perhaps that should be where you go back and modify your approach a little rather than complain about the supplies you have been specified. Last year we ran our robot for hours and never melted a connector, the year before the same, the year before the same. Thousands of other teams have done the same (though maybe they are all being oddly quiet about this). It's not an attack at you, just a sign that something you are doing isn't quite jiving with the required components. The downside is that this problem creates a very unsafe condition.

[apalrd]

I'm not complaining about the part as specified, I'm complaining that we are stuck using a part that was specified 15 years ago with 1/5th the power in the kit and 1/2 the main fusing capacity.

I don't think it's an issue of FIRST actually wanting us to use the connector for engineering reasons, I think it's a case where FIRST is using the same connector they have been using because they have been using it forever and have no motivation to change. I am trying to give FIRST a good motivation to change to an appropriate connector.

Since we went up to #4 main battery cables, we haven't melted any connectors. Using #4 was our engineering change to both melted battery connectors and main breaker trips, and it generally worked.

[EricH]

Originally Posted by MrBasse

Last year we ran our robot for hours and never melted a connector, the year before the same, the year before the same. Thousands of other teams have done the same (though maybe they are all being oddly quiet about this). It's not an attack at you, just a sign that something you are doing isn't quite jiving with the required components. The downside is that this problem creates a very unsafe condition.

If I may ask: How many motors are you guys running at a time? 4 CIMs? 6? A couple of window motors? A compressor? And how long at a time? 2.5 minutes? An hour? 6 hours?

See, here's the thing. Sure, hundreds of teams may have run their robot for hours over the season last year. But there are teams who will literally stop driving around only to change or charge the battery in any given multi-hour session. These teams may have 6 CIMS, a pair of MiniCIMs and a pack of other motors of the 550 and 775 varieties, plus the compressor, active at the same time, for much if not all of that time. That's one large amount of current potentially being pulled. Now, I haven't heard of any failures locally, but that doesn't mean that it's not happening--or not at the edge of happening. (The other possible explanation is that the connectors are wearing out due to fatigue/heating of the area around them, then not seating properly and adding a spark to the mix--then they melt, then they burn.)

And I'm pretty sure the Bees tend to run on the high end of both total run time, and time in a session, as well as lots of motors, and they tend to run enclosed bots, which will get some heat build up inside. If you're running on the low end of all of those with an open bot, you probably won't see anything similar.

[sanddrag]

I'm curious about the condition of the contact itself, prior to melting. Was it significantly worn or scratched from touching wires directly to it to test motors, or clipping alligator clips to it?

[MrBasse]

Originally Posted by EricH

If I may ask: How many motors are you guys running at a time? 4 CIMs? 6? A couple of window motors? A compressor? And how long at a time? 2.5 minutes? An hour? 6 hours?

See, here's the thing. Sure, hundreds of teams may have run their robot for hours over the season last year. But there are teams who will literally stop driving around only to change or charge the battery in any given multi-hour session. These teams may have 6 CIMS, a pair of MiniCIMs and a pack of other motors of the 550 and 775 varieties, plus the compressor, active at the same time, for much if not all of that time. That's one large amount of current potentially being pulled. Now, I haven't heard of any failures locally, but that doesn't mean that it's not happening--or not at the edge of happening. (The other possible explanation is that the connectors are wearing out due to fatigue/heating of the area around them, then not seating properly and adding a spark to the mix--then they melt, then they burn.)

And I'm pretty sure the Bees tend to run on the high end of both total run time, and time in a session, as well as lots of motors, and they tend to run enclosed bots, which will get some heat build up inside. If you're running on the low end of all of those with an open bot, you probably won't see anything similar.

With a four CIM drive and a compressor and a 775 powered shooter we would run for two to three hours of practice regularly. Just as a fun question... Would you run your KOP compressor for an hour straight and expect it to maintain peak function?

[Dale(294engr]]

This is a contact resistance problem, (one black, one red)

caused by improper connector seating (i.e. not fully inserted / mated)

exacerbated by crud, dirt, scratches on the contact surfaces +alignment

worsened with use (insertion cycles), years of use / misuse

(so these units more likely to be relagated for use only with practice robots)

analysis:

Two considerations:

P=I^R 1st order heating /melting localized at conductors contact point

2nd order Vdrop=IR = degraded robot performance: less I due to less V

A good estimate for average competition Robot current draw is ~50A

50^2= 2500 * RohmsPerContact =.001 ohm (1 milliohm arbirtary example)

P=2.5W =no problem with heating, no time limit

now include intermittent multi motor high current start / reverse periods .1-.3 sec of currents up to ~400A ^2 = 160,000 *.001 = 160W which is only a problem if it exists too long i.e. allowing build up heat at connector contact(s) junction.

at ~1sec expect some melting!!

Realistically.. Anderson contact resistance greater than .001 ohm happens
then at .002 ohm double the above.. .003 triple, .004 quad etc.

2nd order performance problem is connector contact IR drop,
i.e. Voltage drop (V-IR)
or sag as it may be referred to

400A *.001 = .4v loss * 2 for both sides) = .8v droop or 11.8v available to all robot systems

(brief battery Vsag is ~= .1v/50A so 400A is 4v additional loss
.. 12.6 -4.8 = 7.8v available to motors (and all electronics)
(simplified, disregards that motors will draw less current as Vbatt sags)


(12.6-.8v = 11.8v for new battery, fully charged, discharging at 50A so must subtract an additional .1v for internal battery Resistance sag at 50A delivery, Batt Voltage now is 11.7v Best Case!)

.. critically, motors provide less torque as less peak current is available as more voltage is dropped across contact resistance(s)
(plus all other connections & wire, to & from all motors)

.. before the melt and fall apart.. melt progresses quickly past a certain point, the contact resistance quickly rises, hastening the melting process due to I^2R heating leading to a forced disconnect due to melting apart.

The single surface contact area used by Anderson has always concerned me.
Seems a bifruc pin into a barrel would guarantee more surface area at the cost of a slightly larger physical connector (or perhaps two, one single for each side - just need to assure logistics of improper polarity mating.. opposite sex for + & -and a way to prevent plugging Battery into itself unfused..

Perhaps ALL batteries should be fused at the battery side.

I have witnessed upset robots battery terminals short against aluminum frame, overheat and smoke. Batteries have lead internal "fuse" safety link connections between the 6 ea 2v cells.. so depends on the condition in competition of battery charge state /condition plus resistance of terminal to frame "short" to determine when & if link melts (fuses), opening the circuit, or just smokes until the stored charge is consumed... in heat!

Good luck in competition

.. and may all your connector contact reisistances << .001 ohm until after the championships...

[nuttle]

I'd also very carefully check the condition of the contacts. Could you have a charger cable that is scratched up and damaging the battery connectors? This could be the type of thing you'd only see upon close inspection. Also, anyone happen to know how many mate/unmate cycles are specified before contact resistance becomes too high? For example, plating wears away, metal oxidizes, and you now have much more resistance? Or, maybe the spring loses enough force that the contact is no longer very good?

Also, can't count the number of times I've seen someone use the leads as a carry handle, but that's another topic.

[Al Skierkiewicz]

Guys,
Hold on so we get some real facts listed. The SB 50 was not introduced when we went to the IFI controllers. It became mandated much later. WildStang used a similar connector for many years prior to the start of the SB50. The battery is capable of 600+ amps when fully charged and that drops off with discharge. The main breaker is also capable of 600+ amps for several seconds without trip. The SB50 is rated for 50 amps continuous, that is every hour for every day, for ever. The rating is verified by UL testing based on temperature rise over long term use, that is more than a day at continuous current. The test includes a safety margin meaning a fixed amount of excess current will not raise the temperature beyond the safe limits set by the UL test. The UL test data is on the Anderson website if you search for it. As I remember the safety margin is like 100%. That is 100 amps continuous for more than a day without exceeding the set temperature.
Andrew, there are many reasons for connector failure, that is why I would like to look at the connector if possible. I have seen some failures, I would like to add to my data with your connector.
Please remember that the 120 amps forktruck connector may be terminating a 24, 36 or even higher voltage. Please do not be lulled into believing your are talking apples and apples when just looking at the connector.

[EricH]

Originally Posted by MrBasse

Would you run your KOP compressor for an hour straight and expect it to maintain peak function?

Not necessarily--it'd probably do a couple of thermal shutdowns, for one thing; for another, it isn't rated for that. If my robot were going through air that fast, though, I'd have it shut down for an hour so I could hunt down the leak in the system.

That said, if I were using a lot of air, I'd reasonably expect a pretty good amount of runtime on the compressor in that hour.

[Nirvash]

Originally Posted by MrBasse

Would you run your KOP compressor for an hour straight and expect it to maintain peak function?

If by KOP you mean the VIAir, with proper cooling? Maybe. The older Thomas? With cooling, yes.