I have two malfunctioning units. Both will emit smoke when driving the motors forward.
Opening them up reveals that one or more FETs (I am presuming they are FETs) show signs of heat damage - the plastic around them is melted, the grounding tab is discolored. Also, the central cap in one unit is bulging slightly.
A multimeter test indicates low resistance between V- and M+ on both units. I think this is consistent with the observed behavior (i.e. smoke) when driving the motors forward.
My plan is to simply replace the damaged FETs and the central cap in one of the units.
Has anyone done this before? Does anyone have any suggestions/comments?
If you want to repair it for the experience, go ahead. (I’ve never tried it, but those big FETs might be hard to solder to a board with a lot of little surface mount components nearby.)
If you just want a new Jaguar, follow this procedure to set the RMA process in motion. TI will generally replace them for free if the failure is consistent with a number of common failure modes—however, I just saw a notice saying “June 11, 2010 is the last day RMAs will be accepted for Jaguars received in the 2010 kit of parts.” If they don’t keep taking RMAs for the “grey” Jaguars (i.e. to keep the customers happy), they may end up giving a lot of business away to IFI (which has historically had good customer support for failed units).
One of our electrical mentors fixed a couple fried Jags for us a couple years ago. I’ve no idea how he went about doing it, but they are in fact functional, we just can’t use them in comp seeing as how they’re now modified control elements. Not sure how he went about doing it though, I wasn’t even on electrical at the time
Erik,
To CSI the one Jag with the bulged cap. That occurs when the power input polarity is reversed. Electrolytic capacitors hate having their polarity reversed. As the Jag is not protected for this, there is likely a lot more wrong with it. Call it a paperweight and have done with it. For the other one, closer examination is needed. When Fets short or have metal flakes help in their demise, there is likely other damage that is difficult to see. Things like missing circuit traces that have vaporized or dead FET gate drivers. Some of these issues back up into the power supply chips.
I’ve never done this successfully, due to U6 and U7 being fried. However, I have fixed a Victor by pulling some MOSFETs from fried Jaguars. (When the power is reversed, it tends to be the small electronics and the capacitor that are fried. When you apply power to the wrong side, it’s just the MOSFETs that blow up.)
Go ahead and try. However, make sure to use a limited-current power supply (a SMALL 12v battery) and a fuse when you try it. You will probably know immediately; if it does not turn on, or if you get magic smoke, it’s gone.
It takes some experience. Luminary Micro uses ROHS solder, which does not re-flow well. I had to mix it with lead-based solder to even take it off.
Adding a bit of flux (I use water based, but any flux should work) helps with reflow quite a bit. I don’t do any SMT work without plenty of flux, and I typically use ROHS solder. It’s a good idea to clean the flux off when you are done (I’ve used isopropyl alcohol, and a hair-dryer to dry it afterwords to great effect)
If the input power polarity has been reversed, the unit is scrap. But, we have repaired 2 of the Jags by changing FETs. I do not consider this to be a modification as it is not modified only repaired. Am I wrong?
Also, the ones that we repaired did not " smoke" they simply stopped working.
The control system is designed to allow wireless control of the ROBOTS. The Classmate
PC, FirstTouch I/O module, cRIO-FRC, speed controllers, relay modules, wireless bridge,
batteries, and battery charger shall not be tampered with, modified, or adjusted in any way
(tampering includes drilling, cutting, machining, gluing, rewiring, disassembling, etc.),
Since tampering includes disassembling, any repair to a jaguar would be illegal under last year’s rules.
[quote=<R60>]The control system is designed to allow wireless control of the ROBOTS. The Classmate
PC, FirstTouch I/O module, cRIO-FRC, speed controllers, relay modules, wireless bridge,
batteries, and battery charger shall not be tampered with, modified, or adjusted in any way
(tampering includes drilling, cutting, machining, gluing, rewiring, disassembling, etc.), [/quote]
I hope “tampering includes disassembling” was not intended to prohibit taking something apart if it was going to be reassembled in original working order before use. Note that teams are instructed to disassemble the cRIO in order to install the gaskets.
The fancy battery charger cabinets using the guts – disassembled, but unmodified – of the KOP Schumacher chargers are a bit more problematic. I’d like to see some productive debate about the wording of that particular rule this year.
Well, lets look at its intent:
Does it give an unfair advantage to your team?
If repairing a Jaguar that you accidentally fried gives you an unfair advantage, then we have a LOT to discuss.
Similarly, with the battery chargers. All you have to do is measure their current to see if they charge at greater than 6A. If they don’t, then they’re fine.
You can assume the GDC disagrees with me. They haven’t stated the intent of the rule, and so it probably differs vastly from mine.
For example, I would say that modifying the resistors in RC servos to make them variable-speed motors - as long as they’re limited to 54 oz-in of torque - is perfectly fine. (Then again, modifying the gearbox would fix that pretty easily. Maybe the rule should be based on power, not torque. )
Let’s take a look at unfair advantages.
There’s many advantages one team can have over another - mentorship, money, tools, school board support, student experience, student motivation. But those are considered fair (except perhaps the money). So the crux of the issue is on what is considered unfair.
My initial definition of fair would be if everyone could do it if they chose. In the case of programming autonomous and implementing sensors, I’ve been told that’s not possible. However, that’s integrated into the game itself, and widely assumed to be a fair advantage.
Is unfair determined by how common the knowledge is? Yes and no. The rules are designed to allow some innovation. Without it, we’d just be building from a kit. But there are some things (for example, custom driver station software, or 3-phase motor controllers, or low-pressure inflatable manipulators, or engineering an impact-absorption system ) that are considered out of reach. These still qualify as innovation, but are considered unfair advantage. Some of them also fall under safety hazards, but for the same reason - they don’t believe we have the experience or mentoring to do it properly.
Well, I’ll stop my rant for now. Perhaps this should be another thread.
OK, Al goes out on limb…
Disassembling robot electrical components is a violation of the rules. Your inspector will ask you this question during the inspection process. R60 is very clear in what it includes, battery chargers included. There is a good thought process behind this. For robot components like the motors and servos, modifying them can/will give teams an advantage due to a change in power output or power curves. The same logic can be applied to modifying speed controllers with low resistance FETs, spikes with higher current relay contacts, etc. The modifications as listed and the exceptions as listed all are designed to minimize advantage and insure safety as designed. In the case of the battery chargers, should one of these devices be allowed to be opened by anyone that is not authorized, serious injury can result. First simply cannot make a blanket policy that covers all those possibilities and keep everyone safe.
Marshall, the servos cannot be modified simply by adding a resistor. They are feedback controlled positional devices with an active motor driver and feedback mechanism internal to the device. Although modification kits are available for specific purposes, we cannot inspect the internal mechanism for each and every robot. Therefore, modification is verboten.
For some of the other reasons, the intent is to prevent you from shooting yourself in the foot. In the case of a speed controller for instance. Maybe you are perfectly able to change out a FET but generally they are not the only devices that have been damaged. You change out the FET and assume everything is OK. When you get onto the field, you find your robot has turned into a flaming inferno, threatening the venue and other robots not to mention any human that happens to be close.
As to the Crio gasket, this is a well thought out procedure to help protect an expensive control component. If followed to the letter, anyone can make the change without fear that disaster will result. In fact if used, you will likely experience a much lower failure rate in competition. This is due to metal flakes your team was not careful enough to prevent falling into the most expensive component on your robot. I can tell you it happens, I have removed buckets of debris over the past two years from teams who thought they did a good job of covering stuff. The Crio is built like a tank but it is not bulletproof!
While some teams mentorship excels in a particular area, all teams do not have access to that level of experience. To allow those teams to use their talents unrestricted is an unfair advantage.
For the rules that don’t seem to fit any of the above, consider those to be simulated real world restrictions. The GDC does want to you to think after all. A good way to do that is to make fake physical limitations, i.e. size, expandability, battery type, weight, motor type and numbers.
If anyone thinks that there are not teams that push the envelope, do things that violate the rules or shoot themselves in the foot, you need to walk around with a UL safety advisor or inspector for a few hours. I have stories that will turn your hair gray.
FIRST provided a pair of Q&A responses in 2010 that pertain to <R60> and the disassembly of electrical parts for cleaning and installation of gaskets. (It would be excellent if FIRST incorporates those clarifications directly into the rules for 2011.)
You could certainly make the argument that repairing the device with a new FET (identical to the original one) satisfies the same test articulated in the Q&A response (“the device is completely restored to its original condition”)…but I wouldn’t recommend it for competition purposes. (I should probably have mentioned that in my original response—I expected you would be able to RMA them and get fresh ones, obviating the need for repairs.)
Firstly, it’s not clear if the intent of the Q&A was to imply that anything that restores a device is acceptable, or anything that restores a device through cleaning and disassembly only is acceptable. (The narrower reading is the safer bet, but the more expansive reading wouldn’t be a ridiculous interpretation either.) Also, because it says “completely” rather than “substantially” or “materially”, you could legitimately question how that’s meant to be interpreted—after all, taken literally, “completely” is a practical impossibility, even though you might never be able to measure the difference. What degree of leniency was intended?
If any of this is unclear, ask these questions again as soon as the 2011 Q&A opens.
I wholly agree that repaired electrical components should not be used in competition. The quality of my solders are not acceptable in a situation where we are already pushing the ampacity of our wires, and where adequate protection is not present for the motors and motor controllers.
In short, a Jaguar is an engineered device. Once I repair it, I’ve introduced a point of failure (at my solder joints).
(However, I am perfectly fine using it for practice and demonstration. I think it’s unlikely that a failed Jaguar will damage other components, assuming you have your robot laid out cleanly.)
I tend to find the restrictions weighted towards electrical and programming, where I see the most potential for innovation.
To give all teams a fair advantage, I should think we put equally strong restrictions on electrical, mechanical, and programming.
That means either lessening restrictions on electrical and programming, or increasing restrictions on mechanical.
Yes you can repair them depending on the severity and extent of the damage. However there are significant flaws with the gray Jags that IMO would make the effort required to repair them in vain. If your issue was with a black Jag I would suggest attempting to repair.
The gray Jags lack ESD protection, contain more FET’s of higher Drain Source impedance and, if I remember correctly, significant board layout differences that effect heat dissipation. The current sense resistor is undersized and adds more series resistance to the bridge adding to the heat issue.
The black Jag contains ESD protection on all of the exposed pins, has a more robust layout, contains fewer FET’s and has a high side current sense circuit that uses a 2mOhm current sense resistor of a higher watt rating. I have yet to ESD a black Jag, I have however ESD’d almost all of our gray Jags.
Even if you are able to repair your gray Jag you will most likely be repeating your efforts in the future.
Since your question was regarding repair, I will leave my interpretation of the legality for a different thread.
But one person’s interpretation of “original working order” may differ from another’s, and both may differ from the manufacturer’s.
I am fairly certain that the disassembly of the cRio was a specific exception granted by FIRST.
My interpretation of the rule in general is along the lines of Al’s comments above - you took it apart, it’s really not OK to use it in competition. YMMV.
)