Why not integrate Jaguars with Diodes?

[Ether]

Originally Posted by techhelpbb

make an additional module or a different PCB for the reverse polarity protected version.

I really don't see the harm in that idea do you?

It would require more thought to ascertain whether it makes business sense. There are NRE and fixed costs involved.

[techhelpbb]

Just wanted to add this...when I get a chance I'll test my other idea...not that what I am about to post here is anything magical as it's basically similar to what's on the robot already...but something to consider as it would seem faster and designed for precisely the sort of protection we are talking about.

http://circuitprotection.ru/upload/p...04-Hbridge.pdf

Costs about $5.50 in small quantity:
http://www.mouser.com/ProductDetail/...9dFfZLyw%3d%3d

[jp1247]

Something kind of interesting. Today, while flashing tan jags, I had connected the first black jag backwards, turned it on, and the breaker immediately popped back and forth rapidly. I smelled burning electronics. I turned it off, and switched the cables. It still works, and all is well, perhaps I just lucked out?

[Ether]

Originally Posted by jp1247

I had connected the first black jag backwards ... perhaps I just lucked out?

The question you should ask yourself is, what caused you to connect it backwards, and what could/should be done in the future to prevent that from happening again.

[Matt Krass]

Originally Posted by jp1247

Something kind of interesting. Today, while flashing tan jags, I had connected the first black jag backwards, turned it on, and the breaker immediately popped back and forth rapidly. I smelled burning electronics. I turned it off, and switched the cables. It still works, and all is well, perhaps I just lucked out?

I suppose if it was really quick, the body diodes might not get destroyed. Though I would have expected the reverse voltage to travel down through the regulators and take out the control logic too, that's what seems to be happening on my end of things...

Matt

[Kevin Sevcik]

Originally Posted by techhelpbb

Just wanted to add this...when I get a chance I'll test my other idea...not that what I am about to post here is anything magical as it's basically similar to what's on the robot already...but something to consider as it would seem faster and designed for precisely the sort of protection we are talking about.

http://circuitprotection.ru/upload/p...04-Hbridge.pdf

Costs about $5.50 in small quantity:
http://www.mouser.com/ProductDetail/...9dFfZLyw%3d%3d

Your mouser link is to a 36V, 30A PTC. If you'd bothered to look at the data sheet, you'd notice that that particular device has a trip current of 50A. That's potentially a recipe for annoyance with the way we use Jags. Not to mention that the time to trip at 100A is a blistering (literally) 4.5 seconds. I regret to inform you that the 40A breaker you already have to have in circuit is going to trip faster than that PTC, in 0.8-1.8 seconds.

[sand500]

My team broke a jag because we switched the polarity.

[Ether]

Originally Posted by sand500

My team broke a jag because we switched the polarity.

The question your team should ask itself is, what caused you to switch the polarity¹, and what could/should be done in the future to prevent that from happening again².


¹ in industry this is sometimes called "root cause analysis"

² in industry this is sometimes called "lessons learned" or "continuous improvement" or "corporate learning"

[techhelpbb]

Originally Posted by Kevin Sevcik

Your mouser link is to a 36V, 30A PTC. If you'd bothered to look at the data sheet, you'd notice that that particular device has a trip current of 50A. That's potentially a recipe for annoyance with the way we use Jags. Not to mention that the time to trip at 100A is a blistering (literally) 4.5 seconds. I regret to inform you that the 40A breaker you already have to have in circuit is going to trip faster than that PTC, in 0.8-1.8 seconds.

This is the datasheet for the Snap Action breaker in the 2012 KOP.
http://www.snapaction.net/pdf/MX5%20Spec%20Sheet.pdf

150 % overload = 3.9 - 47 seconds
175 % overload = 2.2 - 9.2 seconds
200 % overload = 1.5 - 3.9 seconds
250 % overload = 0.8 - 1.8 seconds
300 % overload = 0.5 - 1.1 seconds
400 % overload = 0.3 - 0.6 seconds
500 % overload = 0.2 - 0.3 seconds

Correct?

[Joe Ross]

Originally Posted by AlecMataloni

You could always buy one of these instead: http://www.vexrobotics.com/victor-speed-controller.html.

You'll blow up a victor just as fast as a jaguar if you hook it up backwards.

[AlecMataloni]

Really? I'm sorry to say that I didn't know that. Thank you for informing me. I've only seen people discussing the issue with Jaguars, and I incorrectly assumed that the Victors didn't have that same issue.

[techhelpbb]

So I'll post this in the form of a question:

Other than the consumption of say 1.75W of battery power for each Jaguar, and the space it consumes...is there any reason that someone couldn't put a a single pole relay normally open contact in series with the H-bridge and then put a diode in series with that relay's coil so that if the power is backwards the coil simply will not energize and connect the H-bridge at all? Would this not completely elimate the risks to the body diodes? Couldn't you perhaps reduce the energy required to power that coil using the control electronics if you wanted to do it a more complicated way (considering you'll only close that relay contact when the polarity is correct and your controller here can easily pump out PWM?)

If so what rating would you all think is fair for those contacts?

How about 70A like these:
Mouser, 70A, 14VDC relays

[Kevin Sevcik]

That would provisionally work, yes. The issues I'd want thoroughly addressed before I'd use it:
1. How reliable is it under FRC service conditions? We're talking pretty exciting bumps and crashes, and available voltage dipping into the 9V range under heavy current draws. And you have that diode dropping .7V of your available voltage for energizing your coil. It'd be unfortunate if the coil dropped out under those conditions when an unmodified Jag would hum merrily along.
2. I don't have the Jag schematics up at the moment, so I can't say that one of these will work. You may need two to completely disconnect the H-Bridge to actually protect everything important.
3. You do realize you're proposing to slap a 1" cube into the Jaguar's innards? The Horribly Oversized and Oppressively Bulky Jaguar? If this turned out to be the solution, I think I'd take a leave of absence from CD for a year to save myself the howls of outrage.

[techhelpbb]

Originally Posted by Kevin Sevcik

That would provisionally work, yes. The issues I'd want thoroughly addressed before I'd use it:
1. How reliable is it under FRC service conditions? We're talking pretty exciting bumps and crashes, and available voltage dipping into the 9V range under heavy current draws. And you have that diode dropping .7V of your available voltage for energizing your coil. It'd be unfortunate if the coil dropped out under those conditions when an unmodified Jag would hum merrily along.
2. I don't have the Jag schematics up at the moment, so I can't say that one of these will work. You may need two to completely disconnect the H-Bridge to actually protect everything important.
3. You do realize you're proposing to slap a 1" cube into the Jaguar's innards? The Horribly Oversized and Oppressively Bulky Jaguar? If this turned out to be the solution, I think I'd take a leave of absence from CD for a year to save myself the howls of outrage.

1. No doubt...testing is absolutely critical. Some of the Panasonic 40A relays seem to have drop out voltages well into the 7VDC and below range. Not sure that's really a good idea as a contact rating, but there are plenty of choices.

2. At some point either on the positive supply side, or negative supply side of the H-bridge you should only need to disconnect one contact to break the circuit. Course there is a current monitor resistance in that circuit, but breaking the circuit should protect that as well.

3. As I pointed out above, the idea in my mind was to consider this as a purchasing option. Frankly, with reservations, I agree with other mentors here that damaging something by powering up wrong can be a good lesson. However, I could easily see how it could be a painful lesson for poorly funded teams (even a show stopper). I don't think we should all pay for mistakes we might not make (in power, in size, in weight). Then again if someone has concerns like this perhaps there is value in considering the possibilities if they are willing to pay for it.

Additionally, I'm not sure you can escape the whole extra cubic inch issue. Even if you use a MOSFET(s) in the existing available space in the Jaguar the space involved seems to approach that size. Perhaps it could be built as a module outside the existing Jaguar and attached by someone that won't get it backwards? In that manner the Jaguar itself is the same...but sadly you'll not be able to power up the Jaguar at all like that (I had entertained the idea of having the Jaguar being able to even communicate on the CAN bus that the power is reversed).

Perhaps a connector could be made available on the outside of the Jaguar. Designed so that what plugs in there can't be plugged in with the wrong polarity. At that connector you could either plug in a 'jumper' which is merely a piece of sufficent gauge wire, or you could plug in the reverse voltage protection module which itself is designed to mate that connector with guarenteed polarity. The idea being that a team that has this concern could buy the modules if they want them, have them attached to the Jaguars, and then remove them later and put the 'jumpers' in their place (after they are sure they haven't made a serious error).

[Al Skierkiewicz]

Originally Posted by jp1247

Something kind of interesting. Today, while flashing tan jags, I had connected the first black jag backwards, turned it on, and the breaker immediately popped back and forth rapidly. I smelled burning electronics. I turned it off, and switched the cables. It still works, and all is well, perhaps I just lucked out?

JP,
I would consider that Jag dead. When you stress components to the point where they smell like they were burning, they probably were. It may be working now but it will come back and haunt you when you need it the most.

Relays of the type needed for our service will likely be susceptible to mechanical shock due to the need for large contacts. There is no substitute for checking and rechecking before applying power. I highly recommend more than two people check all wiring and someone should never check their own work.