At long Island teams were looseing their jaguars and even during the build season they were blowing out a lot. They are much better then the Victors for speed control, but they seem to be less reliable.
In testing, we had several failures. We sent them back to the manufacturer for replacement and failure analysis. Other posts on C/D have commented on them as well. We decided to use the Victors and wait until the Jaguars have been upgraded to be safe. 
I think part of it could be the fact that they are new.We have used the Victors for many years (as long as I have been in FIRST) and teams know how to use them. We know what they can take and what they can’t. Give teams 5 years with the Jaguars we will probably not be blowing them that quick by then.
We did not use the Jaguars, I think during the off season I may play with them a little, the extra size worries me slightly but I think they have a lot more power than the Victors. Anyone have any ideas on how to analytically test this? I would be willing to help/do this testing if someone would help me write up requirements and design tests. (Not sure what people want to know about them)
All we did was prevent the code from sending the two extremes of the pwm signal within N duty cycles of each other. More specifically, we try to ensure that only a maximum of 12V potential is seen across the motor outputs on the Jaguar. This keeps the motor from going full forward to full reverse in a split second, which has led to some teams seeing a 24V difference (from -12V to 12V) for a split second across the motor output terminals. This trips a fault circuit on the Jaguar, which takes 3 seconds to reset. I suspect that it also severely shortens the life of the Jaguar. Further discussion on this topic by some very highly reputable sources is here.
So I’d say you should blame software, not the hardware :rolleyes:.
chris, most of the jaguars that were burning out at long island i hink was because teams mis wiring them. although i know team 870s jaguar was hooked up right and burt out.
Assuming the reliability issues can be solved, the Jaguar is an improvement over the Victor. With twice the update frequency, more PWM output “steps”, and improved linearity, you get better results out of feedback loops when the Jaguar is doing the speed controlling.
Unfortunately, all of our evidence is anecdotal rather than quantitative. However, I will point out a few things you can count on.
Victors were a proven product. As such, when something went wrong with a victor, it was very rare to see a team come here and complain about it.
In the same vein, the newness of the Jaguars tempts a lot of people to jump on Chief Delphi the second they have a problem and post about it. Chief Delphi is a very non-representative sample of the actual durability of the Jaguars.
I’m not sure how willing IFI or Micro-Luminary would be to share comparable failure rates, but that’s the only way you’re going to get a real answer to that question.
Regarding the power question, we hooked a Jaguar and a Victor up to the same motor, which was connected to an encoder. The Victor produced a 7% higher rate on the encoder, but the speed curves looked horrible compared to the Jag.
I’m not sure exactly how a change in free speed rate correlates to an actual power change - we didn’t check the amperage output of both under load because we don’t have a multimeter that runs that high.
Before we get back into a full out war on Victor vs. Jaguar it is important to note a few things that are thrown about.
- Both controllers at full throttle put out a constant voltage which is the battery less losses in the wiring, connectors and series resistance of the FETs in the controllers. The Jaguar FETs are a little lower but the Jaguar must switch to zero periodically to charge up the cap in the charge pump circuitry.
- Linear? This varies with motor type and between motors of the same type. All works itself out with practice or some form of feedback.
- Anecdotal or not, the Jaguars are seeing (at a higher rate) failures that were not seen with the Victors. They are working on it and in my opinion they are trying hard to give you a product that is reliable.
- The output frequency of the Jaguar is 15kHz while the Victor is 150Hz. I suspect that this difference with the KOP motors is the big difference teams are seeing with output speeds.
- The Jaguar has a current sense resistor in the negative side of the H bridge. Although low resistance (equivalent to 6" of #10 wire), it does effect the output voltage a little. This is the same resistor that the Jaguar uses for over current sense. Once reached, it takes a minimum of three seconds after the fault has been removed before the protection gives control back to the output.
- Failures have occurred on all types of motors under all types of operation and Luminary is trying to get through all of the data and their own testing to determine where the problem is actually occurring. In spite of the years of electrical advice (mine and others), teams still manage to get conductive debris inside the controllers regardless of their manufacture. A few wayward strands of aluminum or a pile of filings will cause a FET to self destruct in a visible and noxious way. The result is usually a controller that is unrepairable.
- At this point it seems that the greatest Jaguar failure is a controller that only produces output in one direction. If yours failed in another manner, Luminary needs to know.
I like them, they are pretty cheap, double update frequency, and work pretty well… The only issue that I have is the quality… They arn’t the best quality wise, but they do work pretty good… The on off fan is pretty sweet, and yeah, they work great… I would happily use them from now on…
If better documentation is provided to all teams then we should see less wiring faults and better coding tactics. Thankfully we had an NI Tech AND Luminary Tech running around assisting with both wiring troublehsooting and coding help which greatly reduced the amount of Jag’s given out at the event… (as a matter of fact I DIDN"T GIVE OUT ANY Jag’s at NJ b/c of them). If any were given out it was by Luminary b/c they were doing on site RMA for “real” failures. It’s sad how not every venue had the same kind of service =.
I only had to authorize one Jaguar from Spare Parts at SBPLI, but teams had spares of their own to draw from, so it’s harder to track all the failures. The one I did authorize was for Bay Shore, 271, and they burned out a bunch on their practice 'bot before the regional, so they didn’t have any spares of their own.
Out of curiosity (and forgive the stupid question) what exactly does this mean from a real world standpoint? I mean more is usually better but there comes a point where you just don’t need it. (Not saying that we don’t need the extra output frequency, just a general statement)
And I too would like to see more documentation as well. I was severely disappointed when I saw how little there was.
That still is good stats =)
Your just a ball of good news today =)
If anyone is curious a bunch = 5 in 1 week, they all came from the same order (Our Spares) so we are assuming it was just a bad batch, because on our real bot we’ve had only 1 failure out of 9.
Edit: Other then that I love the jags, they’ve given us to have alot finer control over our bot ( mainly our head )
Andrew,
There was quite a bit more documentation on the Jaguars including schematics available on the Luminary website.
As to the output frequency vs response. The brush/commutator length converts into a repetitive frequency calculation dependent on speed for each motor. All KOP motors are not the same brush/commutator design. The FP is significantly different than the CIM. The difference in output frequency of the controller determines how effectively the controller can make current changes in the winding it is connected to, through the brush assy. The higher frequency may interact differently on a particular motor due to winding inductance or the number of voltage transitions it would make while connected to one particular winding. The mass of the armature also comes into play when comparing motors as does the mechanical load on said motor. While the Jaguar may be more “linear” on one type of motor than the Victor, it might not be on another. Take a close look at the motor curves and then try to interpolate different loads into the response.
To my knowledge, neither 1618 nor 2815 burned out any Jaguars this season. (1618 used four on their robot, 2815 used four Jaguars on their manipulator and two Victors on the drivetrain.)
They’re still a little big for my liking, but I’m intrigued by their potential.
what we have noticed, jaguars burn up before the motors do (in a situation of overstressing motors), which is pretty handy honestly rather change jag than motor.
To add two cents worth, we switched from Victors to Jaguars
on our drive in order to enable reliable operation of current sensed
traction control. The 120 hz chop rate of the victors was too low
to be smoothed out adequately with an RC fiter we put on the
current sensor. We did note that we lost a little top end speed
with this change, but the resulting traction control was worth
the trade.
The 40 amp current limit on the Jaguars was pretty aggressive
and we did hit that limit during testing with a gearbox that was
binding a bit. We did not have any Jaguars fail.
Eugene
We use two Jags. One for each side of the drive-train. Everywhere else we are using victors. We have had no failures of either the Jags or the Victors on either our comp bot or practice bot. I will say this, we mount all of our speed controllers vertically above our drive motors. This ensures the minimal amount of particulate to fall into the enclosure of the Jag. One key difference between the Jag and the Victor, the Jag does not have a conformal coating, the Victor does. I believe that this is a contributing factor to the number of failures seen.
We also mounted our Jags vertically, and have yet to lose a single one on either our main machine or our practice bot. I think that since particulate seems to be one of the single biggest factors in the failure of the Jags thus far, vertical mounting is (at the moment) the way to go.