Observations during testing.

[mstompro]

I ran some tests this weekend in preparation for my race next weekend.  I set up four antennas with FX7500 reader.  All-in-all I had pretty good results in terms of read rates, but did notice a couple things that make me wonder.  I noticed that several times when I ran through one of my tags would read several seconds before the other tag (they were programmed with different bib numbers).  That got me wondering if two runners were very near each other is it possible that it could have them out of order.  I then ran across the line with another person who was a few feet behind and occasionally the finish order would be incorrect.  Is this pretty common or does it have something to do with my setup?  

Another thing I noticed...I conducted numerous tests with tags on my shoes and tags safety pinned to my shorts on my hips.  Without fail, the antennas would ALWAYS pickup the hip tags first when worn together.  If I had hip tags and ran close behind a runner with shoe tags, lots of times the antennas would pick me up first.  

Another question.  Is it better to place the antennas directly across from each other or slightly offset.  I thought heard someone say that if they are directly across from each other they could interfere with each other?

[rtspt.usa]

i have found with only side antennas that sometimes a placement could be off a little most likely due to the position and angle of the tag(s) in relation to the position of the foot with in the runner's stride.

i would definitely off-set your antennas so they are not shooting directly at each other.

[Scovel Racing]

I've timed two cross country events so far (about 7 races in all), Order mishaps do happen every once in a while, the most important thing at the meets for me is to make sure the order of the top 10 is correct. Part of the job of the timer is to make these corrections, but sometimes its just to hard to see the bib numbers in time, so for a backup I stream a live GoPro feed to my computer and then recorded the stream using FRAPS, this way if I have a question, I just write down the apprx time and check the video before printing results.

[Todd Gould]

First:
The Microsoft .Net framework being used for TIME values does not provide .001 resolution.  This is why (if you have noticed) that the last 3 digits do not seem random.  They seem to repeat.  So do not think that we have 1/1000 second resolution.

Second:
Due to the above, the software will sometimes be confused on what order 2 chips finished in if they are Very Very close together.  This happens about 2 or 3 times in a bike race with 30+ people crossing the line in 8 seconds.  Hence why I have seen it much more than anyone else.  This is where the GoPro is invaluable for getting order correct.

How to recognize these errors:  You need to look down the Finishers list on the timing screen.  You will see two competitors with the same exact time down to the 1/1000 of a second.  The time displayed may or may not be the time recorded. Yes.  I meant that last sentence. (See below)

What I do:  For every one of these occurrences, I watch the GoPro video and note the finishing order of these mass finishes  (It is only 8 seconds of video....).  Then I match to the results in the timing screen and manually adjust the finishing time of the competitor that should have finished first when there is a tie. You have to do this for every tie you find as the order reported in the timing screen may or may not be the order reported in the results report. (remember there is no tie in order only in time) So even if you have a tie and the order is correct, you should manually fix it because when you print the results the order could be different.  I have also noticed that sometimes when I adjust one time of one of the 2 that show the same time, and hit enter to make it final, the time that was showing for the one I did not change will change on it's own.  (Almost like it is a display problem on the timing screen).  I spoke to Brian about this over a year ago and he pointed to the .Net time service as being the problem.

I time with 8 antennas on a single FX9500.  4 overhead reading the helmet chip, and 4 on the side reading the fork tag programmed to the same number.  When We have these mass finishes that are doing close to 30 MPH, I thank the lord for the Agee Race Timing software as there is no way this could be hand scored.  We rarely miss a tag with our setup.  If we do, we just add them in from the video.  I still get nervous every time I see the big pack flying into the finish.  But my nerves calm down once I look at the video and see that we did not miss one single tag.  Great software, redundant tags, good tag placement and good antenna setup are the keys.  Pre-race tag placement checks and corrections at the start line help too!

 

[mstompro]

How do you guys use the GoPro?  I presume you wait until of the race to review it or do you have configured to play back directly wired to your computer some how?  Where do you have it positioned in relation to the finish line? 

 

Lets say I notice that the RFID system mistakenly lists two runners out of order.  I'm presuming the best way to rectify that is to just change the time of one of the runners?

[rtspt.usa]

the gopro can be streamed directly to your computer.. what i am not sure is if it can be viewed while recording??

you would also want it (or any other camera) aimed in such a way that you are getting a time clock in part of the picture (if possible).

i cheat and use a full frame time synced video camera called an IdentiLynx from FinishLynx (the photo finish folks) you can review while recording and is time synced. so i just have both Agee and Lynx running on the same machine and can check things very easily while the race is going on.

[SportMadeSimple]

I have done several High School XC races and the finishing order is more important than the time.  There are several instances in each race where athletes are out of order and I've used a DVR Security Surveillance system with 4 cameras to record the finish line.  I had promised a video, but haven't made one yet (sorry).  I've been thinking a lot about this issue and I think that one of the possible resolutions could be to prioritize the antenna.  What I mean is that if your antenna were numbered 1 through 4 as they get further from the finish line then the software could adjust based on this.  So a runner that is picked up by antenna 4 (and not the previous 3) could be given an adjusted faster time.

Another possible solution would be to have an interface addition where close reads could prompt the timer to resolve the finishing order. ... I'm not confident I know the reason why it happens (I assume it is just the variance in the chips read range?) but I feel like I need to find a better method to resolve these issues during the race.

I'm interested in the Identilynx idea ... What would be involved in implementing that system? 

[Ohio Race Day]

Regarding setting up antennas across from each other.  According to what I gleaned from an earlier post (see the link below), the tag confusion happens when you have two same strength readers with antenna pointed across from each other.  When you only have one reader then there is no tag confusion. See BA's answer in another thread.

https://groups.google.com/d/msg/agee-race-timing/LCgU3ka_DKE/yEVBMNqc3YAJ

[rtspt.usa]

i still wonder if antennas -whether on the same reader or different reader pointing at each other -- and that close to each other-- would not cause some interference...especially on the same frequency...

[rtspt.usa]

an Identilynx camera with an auto capture plug in (only records when there is movement) would probably run about 2,500 from Lynx System Developers.

i think a good alternative is your dvr system which if it's like my home security system can record and play back at the same time..

another, and cheaper altogether would be to have 2 cameras set up and recording simultaneously, then one can be taken off line and reviewed then put back one line, etc..
For XC the respective rule books are pretty clear (although the hs rule is really outdated).. at least for ncaa and usatf all xc finishes within a tenth need to be reviewed and verified anyway so chip time is never taken in as is.

The ultimate XC set up is my Ipico system with FinishLynx FAT camera and Identilynx --all synced together.. so when the chip times come in you verify each person very easily --everything is on the same screen and seamlessly integrated.

here is a video of me using all of this at a meet.  image on the left is the FAT side camera. image on right is Identilynx full frame. Box that pops up is the chip times and order which all are synced together.. operator selects the correct person according to image(s). It is usually in the correct order especially with 2 chips but there are times when it needs to be reversed.

https://www.youtube.com/watch?v=hRC0_Cd-uoM

[Scovel Racing]

@mstompro
GoPro Wireless Setup
1) Download VLC Player (allows you to stream the GoPro feed to your computer)
2) Purchase Fraps (allows you to record videos even when they are minimized)
3) Turn on your GoPro "App Wifi" then connect to the GoPro "hotspot"
4) In VLC Click Media->Open Network Stream->copy/past http://10.5.5.9:8080/live/amba.m3u8
5) stretch the VLC Screen so that its maximized, Fraps will record based on the screen size
6) Start recording using the Fraps Record hotkey (make sure to change it to a hotkey that isn't related to Agee, to prevent mistakes)
7) Minimize screen until you need to stop recording (it will continue to record at the maximized screen size)

Need to review the video right away but want to continue recording?
Stop and restart the recording, this will instantly finalize the video file so you can view it and start a new recording.

unfortunately the gopro stream has about a 2-5 second delay, so it doesnt help at all if your watching in real time

[Brian Agee]

" All-in-all I had pretty good results in terms of read rates, but did notice a couple things that make me wonder.  I noticed that several times when I ran through one of my tags would read several seconds before the other tag (they were programmed with different bib numbers)."

You mentioned wearing two or more tags that were programmed to different numbers. If you slid two tags down the same shoe lace (even if they are programmed to the same bib number) then this will cause poor read rates. If you placed a tag on each shoe then it should have picked you up every time.

Also, I recommend that you place your antennas about 6 inches past the finish line. So if the center of your mat is the finish line then you'll want your first antenna to basically be at the end of the mat (like how you have the antennas at the end of yoru patio in the 3rd picture).

"That got me wondering if two runners were very near each other is it possible that it could have them out of order."

Yes, in really close finishes it's possible that it could pick up the tags in the wrong order. RFID works extremely fast and will get them in the right order most of the time, but it's possible that the tag that came across first just happen to be at a bad angle for that split second that it should have been seen first. If you see that it got someone in the wrong order, just edit their time and bump it up a fraction of a second so that it bumps that finisher up one place.

"Another thing I noticed...I conducted numerous tests with tags on my shoes and tags safety pinned to my shorts on my hips.  Without fail, the antennas would ALWAYS pickup the hip tags first when worn together."

Now that is interesting! I'll have to test this myself and see what happens. Thanks for sharing!

"Is it better to place the antennas directly across from each other or slightly offset.  I thought heard someone say that if they are directly across from each other they could interfere with each other?"

Antennas connected to the same reader can be placed directly across from each other without any problems because the reader only 'listens' for tags through one antenna at a time. To the human eye it appears that all of the antennas are always one, but in reality the reader cycles through the antennas one at a time (extremely fast of course) to avoid tag confusion.

If you have a second reader on your finish line, then of course those two readers may interfere with each other, but I really don't think you would ever notice any problems. In the Marathon yesterday at one point we had our backup system's antennas working right alongside the main system's antennas and both systems were picking up the tags at the same time without any problems.

[Brian Agee]

Yeah, for those that are curious, here is what I originally sent Todd about the 1/1000 second precision over a year ago:

Yeah, late last year someone else pointed this out to me because they found that he very last digit was always a 3,7, or 0. After digging into this for a while I finally found that it's something stupid that Microsoft does. You can read about it by looking at the "DateTime" section on this link: http://msdn.microsoft.com/en-us/library/ms172424.aspx

Basically Microsoft says that SQL Sever (not really .Net) is where the problem is. Specifically it says "Date and time data from January 1, 1753, to December 31, 9999, with an accuracy of one three-hundredth second, or 3.33 milliseconds. Values are rounded to increments of .000, .003, or .007 milliseconds."

I have a solution in mind for this that I'll eventually add in, but for most of my customers that level of precision is just fine, in fact I've had a handful of request just to remove that last digit.