Re: Denso Window Motors, Warning - Don't use with Jaguars !
 

The following is just an observational opinion.

1) We know the Victors work well with the window motors and have for years.
2) The window motors work well with the Spike relays.
3) The Jag is a new(ish) animal to us and we don't have quite enough experience or information to make a solid determination about them, yet.

The simple, and I really mean simple, solution is to use either a Victor or Spike to control them. (I know, duhhh!!)

Now, if I'm not mistaken, using the Jag on CAN, don't we have a lot more data available to us coming back from the Jag? (Sorry, I haven't even touched CAN with them yet.) If so, could that information give us a lot more insight into what is at the root of the observed problem?

I know it is competition season right now, but I bet the CD community can figure this one out quickly. Ra, Ra!! (The sound of rustling pom-poms in the background.)

Re: Denso Window Motors, Warning - Don't use with Jaguars !
 

NI has Greg McKaskle (impressively knowledgeable and helpful and available) to help with cRIO and LabVIEW issues.

Doesn't TI (or Luminary Micro?) have an applications engineer "assigned" to FIRST to assist with issues like this?


~

Re: Denso Window Motors, Warning - Don't use with Jaguars !
 

Close. The PID loop is evaluated at a rate of 1 kHz.

Re: Denso Window Motors, Warning - Don't use with Jaguars !
 

What is your source for this information please.

Thank you.


~

Re: Denso Window Motors, Warning - Don't use with Jaguars !
 

http://www.youtube.com/watch?v=tf_QyFlbA8w


~

Re: Denso Window Motors, Warning - Don't use with Jaguars !
 

Please allow me to clarify a few things... I spent some more time in the shop this evening, and I would like to correct a few things I said, and emphasize a few others.

1. The Jag's do NOT produce a +/-12 square wave with an idle command. They behave just like a Victor (albeit at the significantly higher switching rate). EricVanWyk corroborated the fact that I had incorrectly reported this earlier.

3. The Jaguar does NOT seem to exhibit any sort of instability. Also an erroneous notion on my part. It is well-behaved across the full range of control with the Denso.

4. There IS still a problem with the Jag driving a Denso motor (and not with a Victor).

5. The problem is NOT due to overheating and / or the PTC. It IS very easy to trip the PTC resistor, but the failure we see is quite easily separated from the PTC overheating, particularly when observing the current.

I have instrumented both a Jag and a Victor and monitored both the voltage and current waveforms with a scope. I will attach some plots below for reference. The waveforms do look differrent. The Victor current waveform is a series of impulses, while the Jag (because of its higher frequency) is much smoother and essentially constant.


The failure mechanism:

The Denso motor fails to start reliably even when 100% drive is applied (and the PTC is not tripped). The voltage waveform is correct, and the current is high. It just acts like it's stalled even though there is no mechannical load. We can easily get this to happen with just the motor (nothing connected to the output shaft).

The motor does not start even if you try to "help" it move in the commanded direction.

When the motor is "stuck" like this, you can coerce it to start by turning it in the OPPOSITE direction. A slight nudge in the opposite direction makes it suddenly free up and turn in the commanded direction. It is as though there is some sort of backlash in the worm gear, and shifting this backlash causes a release of tension, freeing the motor to turn in the opposite direction.

We can NOT get the motor to fail in this way with a Victor powering it. There was some mild evidence of sticking, but it always freed up with a command authority much less than maximum (maybe 25%) wheras the Jag could not start even with maximum command authority, happily drawing > 10 amps (until the PTC tripped after many seconds).

With the Victor, the motor growls as it begins to start. With the Jag, it's a high pitched squeal. Could it be that the growling vibrates the worm gear enabling it to move the backlash just enough to start when the high-frequency of the Jag cannot?

Scope traces of Voltage and Current... Jag and Victor.

Scope Traces -- Trace1 = Voltage, 10V/div. Trace #2 = Current. 100 mV/Amp

Plot #1 = Jaguar at ~50% negative, Current -5A
Plot #2 = Jaguar at ~95% positive, Current +5A

Plot #3 = Victor at ~10% positive, Current pulses to 7A
Plot #4 = Victor at ~20% positive, Current pulses to 8A. Note distortion of voltage waveform (reverse emf?)
Plot #5 = Victor at ~99% positive, Current now almost constant 1-2A

Bottom line. I still believe there is an issue with the Denso motors powered by a Jag. The issue may be largely mechanical, but it's the combination of the two that seems to be problematic. We will be replacing the Jag with a Victor. Also note, as previously discussed, this motor does heat up quite quickly, and the PTC does trip easily. I don't think it is good for heavy continuous use whether you have a Victor or a Jag.

-Tom

Re: Denso Window Motors, Warning - Don't use with Jaguars !
 

He's a TI employee working on the Jaguar.

Re: Denso Window Motors, Warning - Don't use with Jaguars !
 

That sure fits the facts (the new facts :-).

The pulsing current produced by the Vic knocks it loose, whereas the smooth current produced by the Jag does not.

Sounds to me like the motors are defective; not commutating properly. Could be a marginal design or a bad lot.


~

Re: Denso Window Motors, Warning - Don't use with Jaguars !
 

See, now that's what I'm talking about!

Re: Denso Window Motors, Warning - Don't use with Jaguars !
 

Tom,
The DC shift between pulses is normal for a motor that is moving with the bridge turned off (open). You are reading generated open circuit voltage of the motor moving. The low throttle value on the Victor exhibits similar, but lower voltage level because the motor isn't turning as fast. It appears that Jaguar is shorting or otherwise providing a low impedance in between pulses. Is the jaguar in "brake" mode while the Victor is in "coast"? I know that the Victor will switch to brake (both low sides of the H bridge turned on.) when the throttle is zero +/- the deadband. According to the manual, the Jaguar only turns off one side of the bridge with PWM, while leaving the opposite side FETs turned on. That may be what is making the Jaguar waveform look cleaner but still doesn't explain why the two are so different when not supplying current to the motor. Although the small zero volt pulse in the Victor waveform may be the second leg of the H bridge still turned on for a short time.

Re: Denso Window Motors, Warning - Don't use with Jaguars !
 

Al, is this true regardless of the Vic's brake/coast jumper setting ?


~

Re: Denso Window Motors, Warning - Don't use with Jaguars !
 

No, I am sorry. This is only in brake mode. In coast, all FETs are turned off in the Victor.

Re: Denso Window Motors, Warning - Don't use with Jaguars !
 

I was wondering if anybody has resolved this problem further. Our team used 4 Denso motors for our 4 wheel independent steering. We kind of suspected we had this problem from the beginning. We had some programming and mechanical issues that masked it. Now that we have solve the programming and mechanical issues it is very apparent that we are plagued by this problem. It is very intermittent and random. The steering motors have very low load on them and do not get warm unless we run for extended time. (10 - 12 minutes). On Monday night we plan to replace the Jags with Victors. Is this our only option? Do the victors solve the problem? This is puzzling problem. Unlike the other post we can not get the motors to run again unless we power down and reboot. Affects us for the duration of the match. Any other ideas or comments?

Re: Denso Window Motors, Warning - Don't use with Jaguars !
 

In software, try setting the output to 0 and then back again. Does this fix it?

Re: Denso Window Motors, Warning - Don't use with Jaguars !
 

For what it's worth, the anecdotal evidence I've heard with Jags and Window motors... Problems arise when set to "brake", but go away when set to "coast". I skimmed the thread to look for something like this and couldn't find it, and I have no idea why.