The Judge timer

[Cory]

One more comment: This time variation due to track setup inconsistency is why I am an opponent of the "cumulative time" scoring method.

Validating the consistency of the track setup across a set of, say, 20 to 40 heats is probably both difficult and impractical. I know of no one who has tried.

[Mike Parrish]

Are you 100% certain that the track was not bothered in any way during your timing runs?

I don't believe that the track was bothered while we were running the cars.

Two more questions:

1. Which is the more controlled system, your track setup, or the circuitry and logic inside The Judge?

2. Can you reproduce the problem?

1. I believe the Judge would be the more controlled system.

2. No, we have not been able to reproduce it yet.

I am also wondering about my ram deficient computer communicating between Hyper Terminal and the software we use.

[gpraceman]

I am also wondering about my ram deficient computer communicating between Hyper Terminal and the software we use.

That shouldn't make a difference. The timer sends the times data as ASCII text, which you see on the HyperTerminal screen. If there is a timing accuracy problem, then it lies with the timer or the inputs to the timer (start switch and finish line sensors).

These are the only problems that I have seen with The JUDGE:
+ If the manufacturer's operating procedure is not followed you will get all times being 0.0001 seconds or all over 9 seconds. A slight open-close (start switch bounce) will give the 0.0001 seconds and resetting the start gate at the wrong time will give the times over 9 seconds. Also, if you are using race management software and the software tells the timer to reset when it has already started its automatic reset sequence, the timer will get confused. Depending on which race managment software you use, this may be seen as no times or all times over 9 seconds.
+ Letting the batteries run low. This will cause the timer to be hard to reset. I recommend using an AC adapter insted of batteries.

To me a blip in the times, that is otherwise unexplainable, is more likely due to how the start gate is operated. If the start gate is not of the spring-load open design, then there can be a good deal of variation, since a strictly human open start gate is prone to a human's inconsistency. The timer will start timing when the switch is open enough to break/make the circuit, even if the start pegs have not rotated away fast enough to let the cars be unrestricted in starting to roll.

[Mike Parrish]

To me a blip in the times, that is otherwise unexplainable, is more likely due to how the start gate is operated. If the start gate is not of the spring-load open design, then there can be a good deal of variation, since a strictly human open start gate is prone to a human's inconsistency. The timer will start timing when the switch is open enough to break/make the circuit, even if the start pegs have not rotated away fast enough to let the cars be unrestricted in starting to roll.

I agree, this would be the case if the times got slower, but, this set of cars got faster. We have not been able to duplicate this yet.

[ExtremePWD]

It was a surprise to see the faster times with this set of cars. In about 6 runs, our fairly equal cars ran in a range from 3.15-3.19, with one run in the middle of a set of 6 where all three of the cars ran 3.08/3.07.(?)

Sorry for the late post - I just joined recently but was interested in the problem submitted for analysis. The 0.06 second improvement converts to roughly 9 inches if you use more of an average velocity. However, if the fault occured .06 seconds into the start of the race then the cars are at a very low velocity and the fault happened within the first 1/2". (just guessing, I did not take the time to calculate. If someone is not comfortable with my guess I can calculate it more precisely.) I am not familiar with the Judge but two possibilities come to mind. The first is relative to the RAM deficiency comment. If the Judge's on-board electronics do all of the switch polling then the computer is just an interpreter of the Judge output and thus would not impact the times. If the Judge is a real-time data pass through device and it requires the computer to actually poll the system switch states then the computer may have not been consistently polling if it was overloaded with unrelated processing. However, my leading thought is the possibility of incidental contact with the starter switch. This could be by gate rebound (which is consistent with the travel of the cars within 0.06 seconds), a person bumping the gate into the switch or a person brushing the switch with a part of their body. It may appear to the timer that the gate was momentarily closed then reopened signalling the start of a new race. This theory can be supported by intentionally flashing the gate switch closed momentarily in the middle of a race and checking the impact on the timer output. It would at least help us understand how the Judge reacts to a midrace switch momentary activation. I would also watch the starting gate dynamics if it is spring loaded to see if the gate can rebound back close to the switch. A manually operated gate could have just had an operator hesitation at the wrong time.

[ExtremePWD]

I felt bad about not doing the calculations and I did not want to be the provider of misleading information, so here goes:

X=(0.5AT*T)Cos(TANG)

A=32.2 ft/sec/sec (the acceleration due to gravity in the vertical direction)
T=0.06 seconds (the time that the constant acceleration is acting on the car)
TANG=track angle (assumed to be 45 degrees)

X=0.5*32.2*0.06*0.06 * Cos 45 (feet)

X=0.041 feet, 0.492 inches

The car would travel approximately 0.492 inches in 0.06 seconds if all sources of friction are ignored. In reality there are frictional forces resisting the motion of the car and the actual distance travelled would be less. I guess I was a little high on my estimate.

[John Shreffler]

Hi. John here. I make The Judge. I have been reading over the posts and find them interesting. A few comments on points raised:

A start gate that wants to open, but held back by a latch, is far superior in keeping results consistant. I supply a magnetic switch which is the most reliable start signal I can find. It operates at about 1 inch separation, but with latched opertation, the time between the cars suddenly seeing a clear road and switch operation is minimal, and always the same.

The Judge does all the elapsed time calculation and sends the finished result back to your computer. You can see the raw data in Hyperterminal. Your computer is not involved in this in any way.

Low batteries used to be a problem, but the current model Judge uses a new microprocessor that functions down to less than 3 volts, still talking to your computer even when the batteries are so low that the red LEDs no longer light.

[GoneFission]

In my Pack's derby this past January, both of my boys (Webelos II and Wolf) had a single run (out of 6) that was ~0.050 higher than their usual lap times. I was monitoring the computer and sitting next to the starter. So far as I can tell, the variance was valid. I couldn't see the entire race behavior for those heats, but nothing stood out from those races other than the resulting finish order.

Typically we see a variance of half a second or more from a starter error. There was no track shift that could explain the change.

[Cory]

Typically we see a variance of half a second or more from a starter error. There was no track shift that could explain the change.

I'm not trying to be smart-alecky, but I want to be clear on what you mean by your final sentence.

Do you mean there was no track shift period?

Or do you mean there may have been a track shift but not one capable of causing this big of a change?

Thanks.

[GoneFission]

I meant that no one bumped the track. Not even a little. There was in this instance, no track shift occurred that would explain a time variance. It was there (in different lanes for each car) and then was gone.

I think it was a placement issue - one race didn't get optimal alignment on the track starter gate. I see this when rolling the cars with the boys on a table. You put the car down and think it's straight but it isn't. As soon as the car goes you see that you had the wheels were pointed straight, but once the axle moved, the wheels shifted. It's a small amount but it's enough, I think.

Next year when my pack sets up their track, I will spend some time testing car alignment variations and see what the time variance is.

[Cory]

Next year when my pack sets up their track, I will spend some time testing car alignment variations and see what the time variance is.

Fifty milliseconds is well outside the range I saw when I did the "plop-down staging" test years ago. My average difference was about 10 milliseconds and max difference was about 20 milliseconds.

Of course, that was on a shortish track, ~24 feet. On a longer track or on an aluminum track the penalty could be more severe, esp. on aluminum.

Were there any other commonalities of the anomolous runs of your two sons, e.g., same lane?

[GoneFission]

This was on a 40 foot Piantedosi 6 lane track. The Cubmaster and I spend a couple of hours on setup to make sure that the section interfaces and guide strips transition as smoothly as possible before bolting securely.

The runs I spoke of were on different lanes and 45 minutes or so apart due to their ranks. The only thing buffeting the track is the start gate operator. We used a manual pull-down, I think for the last time.

Our last Cubmaster filled in for one rank and botched the first start. I saw >1 second more than usual times on the computer and asked him if the start was good. He shrugged, so I asked if we needed to rerun the heat. He told me that it would probably be best. The restart times were back on par.

Gotta us get a mechanical release!

We don't watch for individual car statistics until a rank is finished because of the time it would take to look at each one. We DO watch for group heat times to be consistent. If the groups are consistent, then car variation is considered luck of the draw. The gate operator never works a rank they have a son competing in.

[Cory]

Gotta us get a mechanical release!

Amen that. It might be interesting to compare this year's consistency statistics with next year's when you have the mechanical release.

Also, will be interested in seeing the results of testing staging variations on a longer track.

Thanks again!

[FRANKLIN WHALEY]

We use the JUDGE system. If you're using the new magnetic contacts on the starting gate the speed at which the gate opens makes a big difference. We went to a spring loaded open gate and the times are more consistant. But they went up from 2.41 to 2.48 on the same car. All you have to do to get a better time is manually open the gate slowly. It gives the cars a chance to start rolling while the gate is still opening. As a matter of fact you can just take the spring off and let it drop by gravity and still get the faster times. It would most likely be closer to the times of systems using sensors in front of the starting pins. With the spring loaded gate we have never had a difference in time that wasn't just the car hitting the rail more or just plain having a bad run. We test cars and more importantly the track for consistancy with the Judge and have never had any malfuctions. Don't know if the starting gate could be your problem, but I thought I would throw this out there. Maybe it will help someone.

[Mike Parrish]

While setting up our track for our first outlaw race of the season at church on Saturday morning, we tested the theroy of slow vs. fast gate starts (ours WAS a manual pull down until Saturday afternoon).

Run 1: Fast start, time 3.18
Run 2: Slow start, time 2.98
Run 3: Medium start, time 3.04
(all runs same car in same lane)

After a quick trip the the hardware store to buy a deadbolt, 45 min to change the spring location on the starting gate rod and install the deadbolt to release the gate, we now have a starting system in which the spring tension pulls the pins down, out of the way of the cars.

Now the same car in the same lane ran several runs that were within thousandths of a second of each other.

Mystery solved.