I’ve been a mentor for about a year now, and I’m still confused as to why the chassis of the robot is not allowed to be electrically “bonded”. I certainly don’t know everything about safety or robotics, but I just got a docking station for a robot certified against the UL1740 spec (http://ulstandardsinfonet.ul.com/scopes/scopes.asp?fn=1740.html), in which the chassis is required to be bonded, essentially a single point connection to the power inlet’s ground, capable of (in our case) handling 20A fault current from any piece of “exposed metal” if I recall the spec properly. This is in place in the UL1740 spec to aloow a loose wire to properly clear its fuse when it makes contact with the chassis. There’s MILSPECs and NASA documentation for this sort of thing too… I have read a few posts about 2007 battery incidents, but I fail to see the logic of how those support isolation of the chassis.
Like I said, I certainly don’t know everything, so if someone could explain this, or point me to an oficial FIRST post on this (am I the only one who has issues navigating their site?) I would certainly appreciate it.
Thanx in advance!
I believe what you are looking at with the UL1740 is “grounding”. Usually as a whole a metal machine in house or business is supplied by a power source that is only protected by its protection device, a breaker or fuse box and if the ungrounded conductor touches a grounded or grounding conductor or the metal frame without a load this will cause the protection device to break.
With robotics we have breakers on every positive wire that leaves the fuse block and the other end is connected to our speed controller then to a motor. Electric motors are designed where the case of the motor is isolated from the conductors and by isolating the robot we also don’t have any risk of someone touching the frame and the controller or the frame and becoming the path to a load and having a heart condition or a weird anomoly where the brushes of the motor have a failure and touch the motor casing.
Why is a AA size battery covered in paper or plastic? To keep the sides isolated from simultaneous contact.
So why do we isolate? Just because.
Without the actual text of the requirement, I can’t comment on it. However, I would bet that Mike AA nailed it.
In FIRST, we isolate the chassis to protect us from shorting the battery through the chassis. If it is normally attached, then we only need a single fault (positive wire hits it) to ruin our day. If it is normally isolated, we need two faults (positive and return wire).
It appears I’m now expanding on the above points because I write too much. Nevertheless…
Actually, that rule explains the reasoning a rather lot more than most others. I think it’s primarily in place to prevent teams from foolishly trying to save weight by using the robot frame as a current return. Your summary of that standard sounds very much like it applies to robots connected to wall outlets, as opposed to our DC battery powered robots.
In the case of equipment connected to wall outlets, it certainly makes sense to electrically ground the chassis and prove there’s an adequate current path to ground from any exposed metal. If the chassis weren’t grounded, or some part didn’t have an adequate path to ground, a live wire touching that part wouldn’t pop any fuses. That would leave that part at some significant potential relative to ground, and anyone touching it could be shocked. But this is only possible because the power supplied to the machine is quite literally referenced to the ground you’re standing on, and people are relatively low resistance sacks of salt water.
In our battery powered robots, this isn’t the case. The battery is floating relative to ground, so you could touch, grab, caress, or lick the positive terminal (or anything connected to it) and not feel a thing. Provided that you AREN’T touching anything connected to the negative terminal. Since you can only be shocked by touching the positive and negative terminals at the same time, your robot only becomes dangerous if there’s both a loose hot wire touching metal and a loose return wire touching metal. If we started connecting our frames to the negative battery terminal, we’d actually be increasing the likelihood of getting shocked because we’d be providing much more convenient return paths if a hot wire were to come loose. At which point we’d then have to either ignore the increased risk or go through a similar qualifying procedure to show that all parts of the robot have a good path back to the negative terminal, etc. So I think the reasoning it that it’s easier for teams to keep everything isolated from the frame, and it’s easier for inspectors to determine everything’s isolated from the frame. As opposed to making sure everything on the robot has an adequate path back to the negative terminal.
Good points all, and yes, the UL1740 spec was written more with AC powered robots in mind (automated welders and the like) but does still apply explicitly to “mobile robots”. I know at least on the robot that we have docking to our 1740 dock the negative terminal of the battery is connected to chassis for two reasons, EMI compliance as well as fault return path (for popping fuses). From my own oppinion, I’d rather have a loose wire hit the chassis, pop a breaker or fuse and be a dead wire before it hits an IC on one of my pc boards… This is, of course, on a completely enclosed robot, where the only shock a human can really get is an ESD one (a reasonably nasty one at that). The FIRST robots, being open framed certainly do present a different situation, and you’re starting to sell me on the isolation…
It would be neat to hear an expert from UL chime in on this…
Thanx for the responses!
Jonathon,
There are a variety of issues that your specification is intended to cover that does not apply to the problems encountered in FRC robots. The high current capability of the power supply in either fixed or mobile robot services can allow a significant voltage to be developed should the robot frame not be bonded to a good ground and power supply common. This is based in part on the known failure of fuse components and the developed hazards due to the failure of these protection devices. Entering into these specifications are such documented variables as common skin resistance, breakdown of typical footwear, hazardous voltage levels, etc. all designed to prevent injury to anyone standing near or on a piece of equipment or in casual contact with it.
The deliberate prevention of using the robot frame as a power supply common or current carrying conductor is in direct response to the experience of robot interaction in an FRC event. As the game changes from year to year, there are often severe robot contacts that include arms and other appendages becoming entangled in the opposing robots. It would be possible therefore for an arm to contact the power supply distribution terminal, pre fused or protected, and pass current through both robots when and if the frames should touch. (i.e. the path from positive terminal through arm/appendage to robot frames and back to the negative terminal) By isolating power from the frame, a continuous circuit cannot be made simply by robot to robot contact. Not only does this practice prevent damage to control components, it minimizes fires and battery damage on the field. Remember that two robots entangled as described above, would continue to complete a maximum current circuit even if the robots were disabled.
This brings up an interesting issue with the new control system as the NI CrIO case is connected to power supply common to be in compliance with specifications you have described but puts FRC robots at some risk not previously encountered. I am sure that FIRST engineers are working on this problem as we speak/write.
On the Power Distribution module (aka PD), there is a 1.1A (hold), 2.2A (trip) PTC (Positive Temperature Coefficient) “breaker” in the return path for the 24V supply for cRIO. That device will limit current “through” the cRIO chassis to only a few Amps.
How will we inspect for chassis isolation since every team that mounts a cRIO to a metal chassis will have an intentional low impedance connection to ground? We haven’t figured out the details yet but we’ll probably need to unplug the cRIO from its power supply when performing the standard chassis isolation test.