How *should* timers respond to ties?

DIY timing systems
Post Reply
steve547
Apprentice
Apprentice
Posts: 7
Joined: Sat Feb 14, 2009 5:52 am
Location: Oregon, USA

How *should* timers respond to ties?

Post by steve547 »

I'm in the middle of scratch-building my own timer (yeah, I know I'm insane, but it's a hobby) and at this moment (with the big race on Tuesday... no pressure or anything...) I'm working on the part of the firmware which reports back the race results to the computer.

Ordinarily I'd just round off the timer values to 3, 4 or 5 decimal places and call it good, but also show the actual place markers in the order the timer saw the cars, like:

A=3.5324" B=3.5324# C=3.2423! D=3.8742$

But I saw in another thread here that there was an opinion that it would be better to have the timers just completely acknowledge the tie. (And I should mention I'm using GPRM 8.0.1322.12 in case that's significant as far as what GPRM's going to expect to see in the reply.) So if that's the case, what would be preferable?

How about repeating the place character in case of a tie, and skipping down to pick up the rest? For example, if there was a tie for 2nd place, the place markers would be !,",",$,%,&, or a three-way tie for first would be !,!,!,$,%,&.

In other words, the above tie would come out as:

A=3.5324" B=3.5324" C=3.2423! D=3.8742$

Would that work? Also, I gather from what I've read that some timers report a standard set of lanes even if they're not active or implemented. (A 4-lane timer (like mine) would report 6 or 8 lanes in the responses with 0.000 for the extra lanes.) Is that necessary?

Thanks in advance for this info!
User avatar
gpraceman
Site Admin
Site Admin
Posts: 4926
Joined: Fri Jun 20, 2003 12:46 am
Location: Highlands Ranch, CO
Contact:

Re: How *should* timers respond to ties?

Post by gpraceman »

steve547 wrote:But I saw in another thread here that there was an opinion that it would be better to have the timers just completely acknowledge the tie. (And I should mention I'm using GPRM 8.0.1322.12 in case that's significant as far as what GPRM's going to expect to see in the reply.) So if that's the case, what would be preferable?
If setting up GPRM to use the customer serial timer interface, the place characters will be totally ignored. GPRM will base everything off of the times being reported.
steve547 wrote:Also, I gather from what I've read that some timers report a standard set of lanes even if they're not active or implemented. (A 4-lane timer (like mine) would report 6 or 8 lanes in the responses with 0.000 for the extra lanes.) Is that necessary?
With the custom serial timer interface, it would be better to only report times for the lanes that the timer physically supports.
Randy Lisano
Romans 5:8

Awana Grand Prix and Pinewood Derby racing - Where a child, an adult and a small block of wood combine for a lot of fun and memories.
steve547
Apprentice
Apprentice
Posts: 7
Joined: Sat Feb 14, 2009 5:52 am
Location: Oregon, USA

Re: How *should* timers respond to ties?

Post by steve547 »

@gpraceman, that's perfect! Thanks.
User avatar
SlartyBartFast
Master Pine Head
Master Pine Head
Posts: 272
Joined: Mon Feb 12, 2007 11:30 am
Location: Montreal, Quebec

Re: How *should* timers respond to ties?

Post by SlartyBartFast »

One thing to note:

- Don't let the timer report times to a greater accuracy than the electronics allow.

Probably not an issue if using a fast processor. But me careful with code and signal timing.

Just off the top of my head, so I may be wrong on this.
steve547
Apprentice
Apprentice
Posts: 7
Joined: Sat Feb 14, 2009 5:52 am
Location: Oregon, USA

Re: How *should* timers respond to ties?

Post by steve547 »

@SlartyBartFast: yes, that's a very good point. I think it's easy to get carried away with reporting "accuracy" and not taking into account the point where you're no longer measuring the cars as much as the sensor reaction times (etc.).

In my project, the microcontroller (a PIC in this case) is tracking to the microsecond (0.000001 sec) but I know I lose a few microseconds in the process of responding to each lane crossing. I'm allowing a configuration option to report precision but typically I'm reporting out to the millisecond (0.001 sec) or maybe 0.0001 sec, but I don't think I'd push it more than that.
User avatar
SlartyBartFast
Master Pine Head
Master Pine Head
Posts: 272
Joined: Mon Feb 12, 2007 11:30 am
Location: Montreal, Quebec

Re: How *should* timers respond to ties?

Post by SlartyBartFast »

I keep meaning to start the various projects I have in mind and never seem to get past the research stage.
At first I was going to port the derbytimer code to an AVR and actually have a kit to build an AVR programmer. But now I've found a development board that can be programmed and powered directly from USB.
http://www.sparkfun.com/commerce/produc ... ts_id=8971
I figure that with a 80MHz controller and 78 I/O pins I could come up with all sorts of projects. The basic Bitwhacker functionality is interesting too.
Best of all, Bruce's derbytimer software should work directly with little if any rework.
Post Reply