Timing Systems

[Dean Los Angeles]

About making the system work over the distances involved at Bonneville, salt, weather, and other conditions notwithstanding -- it's doable, especially with digital signals, and those being sent repeatedly until the receiving station has at least one complete set of data from each.  At a reasonable repetition rate of a few dozen bursts per second -- there wouldn't be any delay in getting the data to the computer.  The computer does the math and reads out the times/speeds, just like now.  Only the method of getting the tripped light signal from the light to the computer changes.

It all sounds so easy until you start looking for real parts. The light system is an on/off switch. This has to be processed by something. That means a digital processor and a transmitter/receiver. Then you need a computer and all new software to turn that into the good old timing slip. Putting a transmitter/receiver into the loop adds latency and affects accuracy. If the timer is at the light then the cost goes up even more. And calibration is a nightmare. Sure it's doable, but the cost is very high.

[Seldom Seen Slim]

Dean, you've missed some of my points.  There's already a computer in the system, for example.  The same one, in the same location, is all that's needed.  It would need to do another function -- instead of simply comparing the time between two contact closures it would have to read a message that contains the information about when those two contact closures took place -- information that would be contained in a pair of data messages from the remote transmitters.  The remote transmitter would only have to transmit a message telling what time the event happened.  The receiver at the remote location would be listening to a continuous clock signal so it would have freshly-updated times to add to the message saying that the light had been tripped.

And there's absolutely no "latency" to consider.  It doesn't matter if the data is transmitted from the lights in microseconds or tenths of a second -- because the data from the light will contain a clock signal telling when the event happened.  In other words, if you ask someone "What time is it when event #1 happened?" -- and he replies "It happened at 10:00:02:.00032549 O'clock" -- it doesn't matter when he tells you that information -- the event's time of occurrence still is given to you in very precise terms -- and the latency of the report doesn't matter.

Yes, there's a need for a receiver and a transmitter at each timing light.  I acknowledge that.  They would not affect accuracy since they'd be transmitting data including the time of the event.  See above paragraph.  The timer would not be at the light -- it'd be the computer back at the timing stand, just like now.

Sure, some new software would have to be created, and new hardware, too.  But - the cost should be on the order of what's spent for today's hardware (new wire), "firmware" (fixing and cleaning the wires after use), and software (the labor involved in installing and removing the wires).

Next?

[Dean Los Angeles]

the data from the light will contain a clock signal telling when the event happened.

The timer would not be at the light -- it'd be the computer back at the timing stand, just like now.

This would involve a clock signal at every light that would have to be calibrated to a master clock. And no, the timer would not be at the computer. The time comes from the clock at the light and the computer just does the math. The clock time from mile 3 would have to be compared to the clock time from mile 4.

This is not only a technically difficult and expensive project, but you would have to have calibration that is traceable to the National Institute of Standards and Technology.

[Seldom Seen Slim]

I'm in agreement -- each light would have a colck in it -- that's continuously receiving clock data from the master clock.  Each clock would therefore be tied to the master, and there's be (in effect) only one clock with multiple remote units.  But only the master would determine what time it is.

And no, there would not be a timer (which, to me, implies a device to compare times, or to record a "start" time and a "stop" time) at each light -- just a contact transfer that would indicate when the beam was broken, and a data generator that would read the time (from the internal clock that's sync'ed to the master clock) when the beam breaks and encode it into a data packet containing the information that A: the beam was broken; and B: the beam was broken at such-and-such time.  The data packet would be transmitted to the timing stand where it would be inputted to the timing computer.  That 'puter would calculate the time between lights using the times from the succeeding clocks, and once the time between lights is known, using the information to determine the speed between the lights.  So - you're right, in that the time between the clock time from mile three (light) would have to be compared to the time from the mile four (light).  That's an arithmetic problem -- subtract time "b" from time "a" to get result "c".  Many of today's computers can do such a trick.

And why would the calibration need to be traceable to NIST?  If the current survey is so -- the proposed system's physical measurements would be the same, so therefore also traceable.  If you imply that the time signal must be tied to NIST -- clocks that are traceable to NBS (such as those on radio SSV, WWVB, etc) are cheap these days -- and might be able to supply the master time signal instead of having it originate at the timing stand.  Error correction on those isn't difficult -- so any variations due to atmospherics are relatively easy to correct.

This is good -- you're helping me fine-tune the idea.  Now all I've got to do is find someone with more digital packet expertise to do the actual design of the system.  The theory is well on the way to full completion.

[4-barrel Mike]

"Theoretical timing systems produce theoretical results."     

Mike

[jb2]

Okay,
Somewhere Jack is falling out of his chair laughing.

Jim

[Glen]

Whats wrong with the wired system we now use. It's accurate to 9 decimal points gives instant results and has all data you need on the print outs. The wireless systems have been looked into and they have tons of design information for doing it. The cost is very expensive, would require all new inventory and storage of same.

Laying the wire is a days work, the same for pick up. It is checked as they rewind it and wiped clean. The course is surveyed and marked and SCTA has a back up state of the art survey system. The timers are certified every year and surpass all standards.

So why do we need to change what works. We have not missed a timed run in several years on the salt. FIA and FIM approved as well.

[fredvance]

If its not broke dont fix it.

[McRat]

That's my problem.

"If it ain't broke, fix it until it is" 

[Peter Jack]

I resemble that remark!

[F104A]

Geeez Guys, that was exhausting! I think I'll get a cold beer and go out to the shop and think up something else...........Ed

[Malcolm UK]

If I understood the original question it was particular to tim9ing for an FIA supersonic speed record attempt at a loxation that would not be the salt flats.

The FIA tightly prescribe the methods of timing that they allow and I beleive it stresses that it is a case of 'timing'.  They want time duration for the measured distance with the speed calculated from the time.  Any speed recording device does not appear to meet that requirement.

As the mile or kilo has to be measured to an accuracy of 1/10,000 of its length it is clear that the FIA thoughts are on starting an approved and certified for accuracy timing device at the start of the distance and turning it off at the end of the distance.

Automatic timing to an accuracy of 1/1000th of a second and the recording has to be produced "directly by the passage of the vehicle without human intervention".

GPS speed devices are not in compliance with that rule.

Just bite the bullet and pay Dave Petrali as Thrust SSC and Breedlove did.

Malcolm UK

[NAE_DAG]

Being that I have a bit of experience in defense electronics, I thought Nah.. no way.  It turns out the "trick" is to use two GPS units each capable of tracking 14 satellites, one being stationary reference, and the other mounted and in "kinematic" mode.  The DoD, Russians, or whomever can dither the signal all they want, there can be tropospheric distortion, including phase changes, and whatever- the two units see the exact same signal and subtract the error as they compare the data streams. Other fancy stuff I don't understand yet is performed on the signals- statistics being one... and viola! 20Hz, 10mm horiz accuracy from crappy signals.  The DoD limits the use of these units to under 1000 knots, and 30,000ft altitude.  God I hope we don't approach either. These units are used commonly for automation of road graders and other similar construction equipment. See Topcon.com
-Steve

[Dean Los Angeles]

We can thank the soldiers in the Gulf war for the accuracy of today's GPS. The military degraded the civilian GPS signal through selective availability so that the enemy would not have the ability to use GPS for weapon navigation. During the Gulf war there was a shortage of military GPS, so a lot of soldiers bought their own. Selective availability was turned off so the soldiers could use the better accuracy. In 2000 is was turned off permanently.

It's 60,000 ft, not 30,000 ft. If your GPS is above 60,000 ft and going 1,000 mph it will stop functioning. Kinda like if it was stuffed into a re-entry vehicle with a warhead.

Topcon and others have figured out ways to achieve stunning accuracy, but at really prohibitive prices. But if you want to grade that building site to phenomenal flatness, the technology is there. I just can't see it being used for LSR timing.

[tomsmith]

There is always a better way to do something.  The question is why do you need to do it.  If the present system works, why spend the money?  What do you gain except for the satisfaction and fun of doing it?  I confess that I ignore my own advice, especially regarding my computer upgrades. 

Incidentally, in the early sixties the FAI required optical tracking for aircraft world records.  The FAI didn't trust radar tracking (which was much more accurate).  The accuracy of optical tracking was positioning within a foot or two from about 25 miles away.  I suspect that people would get a chuckle nowadays from optical tracking, just like they do for my 1950 Harley Hummer.