Chief Delphi - Battery connected backwards to robot

Quote:

Originally Posted by DonRotolo (Post 1569376)

I think most of he vulnerable items would fry at nearly the speed of light; realistically too quickly for a reaction to help.

Yes, any devices that will die will do so within microseconds.

Quote:

Originally Posted by Al Skierkiewicz (Post 1569465)

A reverse diode across the input to the PDP would be huge and the short reverse voltage is likely to still produce damage in downstream electronics. Such a diode would need to withstand 600+amps and might allow a volt or more in reverse. The main breaker may still take a second or two to trip at that current. Considering the normal human reaction to the trip, this must survive at least two events before team members will go looking for the actual problem. Inspecting I have been lucky to find two batteries in team's pits that had been wired backwards but never used and one that was being wired backwards while I was in the pit.

Diodes large enough to clamp the battery for long enough to trip the breaker will be about the size of the main breaker we are currently using now, even if schottky diodes are used (lower voltage drop and power dissipation). It may be possible to add a semiconductor fuse in series with the existing breaker. These fuses blow very quickly, allowing the use of a smaller diode. Unfortunately, semiconductor fuses are rather expensive (over $100 each).

Quote:

Originally Posted by juchong (Post 1571096)

While reverse polarity protection is usually overlooked since it's assumed that the user should never reverse the input polarity, it's good design practice to put some safeguards in place given that the components teams are required to use are expensive. Many teams are working with the single module given to them at kickoff and would rather spend their budget on travel costs rather than blowing up electronics.

A cost effective, very easy to implement solution would be to use a reversed P-Channel MOSFET in series between the battery input terminal and the PDP distribution node. Afrotechmods has a great video on the subject here. A transistor such as the SUM110P06-07L-E3 would work for the FRC PDP.

If you needed more current capacity (240A pulsed for the P-Type vs 440A pulsed for the N-Type) you could apply the same circuit, substituting a P-Type for an N-Type such as the STB120NF10T4, on the negative (ground) side of the circuit and achieve the same result. Both of these transistors have an RdsON = ~10mOhm meaning that the maximum power dissipated in the transistor under full load will be 4W @ 20A continuous.

Either one of these configurations would adequately protect not only the PDP, but everything else connected to it.

I am not sure the current rules allow such devices. Even if they were to be integrated into the PDP and FIRST approves it, they would still take up a lot of space.

The voltage drop of the P-channel MOSFET linked would be in the range of 0.75V at 100A and about 3.2V at 240A. The N-channel MOSFET would have similar voltage drop. There are already many threads about the dangers of voltage drops leading to brown-outs and radio and RoboRio resets.

The voltage drop across the MOSFET will also lead to power loss of about 69W at 100A and 397W at 240A. The small thermal mass of these devices means that what robot builders would consider a short peak (say 10's of milliseconds to half a second) would have to be treated as a continuous current and the heatsinks required to keep the temperatures in check would be quite large. Please note that the Safe Area of Operation charts are for case temperatures of 25C indicating the data is only good for a single pulse or applications with long cooling times between the pulses (not FRC application).

The data on the datasheet is all for a device temperature of 25C. Both the voltage drop across the MOSFET and the power dissipation will increase as the device temperature increases making the situation worse.

It is possible to use larger devices or multiple devices in parallel to reduce the voltage drop but they will still take up a lot of space. Perhaps in another 5-10 years, we might have devices with voltage drop low enough to make a series protection device practical.

Overall, it is probably best for teams to learn to check their work many times before applying power. It is not a bad thing that team members learn that some things in life are just very unforgiving, like drinking and driving, treating firearms without sufficient care and connecting electrical equipment up backwards.