Register | Login
Attackpoint - performance and training tools for orienteering athletes

Discussion: Reliability of Emit timing

in: Orienteering; General

Apr 28, 2010 1:08 PM # 
Austris:
Hi,

Recently my Emit chip stopped working after just 2 years of infrequent use. OK, I know these things happen, that is not the topic.

I gave my chip for repair. Got it back. If first race/training I was surprised how slow I am. In next race I could not believe my result. Then we did a small test: I ran for ~5 min, result was shown ~10 min.

What it means: you could modify your Emit chip to make your result better. I guess, in big races running time is not taken by Emit but by a separate timing system. Well, that is not the case in local events.

Does anyone else have heard about such odd behavior? Has anyone tested original Emit chips if the timing is accurate?

Austris
Advertisement  
Apr 28, 2010 2:38 PM # 
bgr:
While orienteering in Norway, Sweden and Finland at the same time(the three corners) I had a similar experience. I finished the race only to be extremely disappointed with my result. A little later another guy mentioned his emit wasn't working and he had gotten an incorrect result. Later, while looking at my results I figured out they were all exactly doubled. I mentioned it to the other guy and he checked his results and it was the same for him, doubled. We had both gotten our emits wet by jumping in some water and going for a swim(during the race). It was right after that they stopped functioning properly.

The next day we raced again. Mine worked, his didn't.

I emailed emit to find out how that happened and didn't get a response.
Apr 28, 2010 3:17 PM # 
andrewd:
use SI?
Apr 28, 2010 3:53 PM # 
Austris:
As a participant you are forced to use the system organizer has chosen. So "use SI" doesn't help at all.

I'm just trying to find out if this problem is known to others.

Austris
Apr 28, 2010 4:06 PM # 
Jagge:
I can remember 5 or 6 cases of emit cards running totally wrong speed. I believe all emit cards drift 1-2 second during one hour race. That's why those should never be used for timing. Doing timing like that that is not allowed in Finland. And not elsewhere either i guess, that's just for training use and splits. I wonder where you were racing and were those actual races or just trainings.

I don't think SI is much of a option these days. for example
http://www.o-worldcup.spb.ru/component/attachments...
~3 eur punch cards. not bad.
Apr 28, 2010 4:17 PM # 
Jagge:
In latest WOC two medals were given to athletes who did not punch correctly. One of them punched "too fast", the other punched with dead punch unit and got no any mark at all. But they were qualified. Jury saw they should not be punished for crappy punching system, SI.
Apr 28, 2010 5:29 PM # 
Tundra/Desert:
Jagge, even in the land of SFR—SI is very much competitive with SFR. There is a lot that people don't like about SFR. What goes for SFR is the superior results analysis (e.g. you get min/km for each leg, leader min/km, % behind leader each leg, etc. at download), but that is just software and people have written software for SI to have the same or better capabilities. What goes against SFR is the nonintuitive "contact" punching (no worse than for Emit though). I don't know exactly what SFR prices their stations at, but SI will dump-price their fingersticks (packaged in soft plastic, not hard shells) at nearly the cost of the RFID chip when it's forced to compete with SFR directly. (I didn't say this.) I also don't know much about the reliability of SFR stations.
Apr 28, 2010 9:22 PM # 
jonm:
IOF rules require a single clock for finish timing (or they did when I was last involved in a WRE event). For EMIT this effectively means using a seperate finish timing system and the EMIT card for "punch checking".
So this situation is avoided for the very big events but obviously not smaller ones.
Apr 28, 2010 10:38 PM # 
graeme:
Does anyone understand why SI fingersticks are so expensive? They've gone up by about a factor of 4 in ten years. Is it just monopoly pricing?
Apr 29, 2010 12:32 AM # 
robplow:
SFR looks interesting and more competition might bring prices down. But if they want to expand they will need IOF accreditation. So far only SI and Emit have that.

There are also a couple of companies in China (Learnjoy is based in Beijing and is affilliated with Orienteering Association of China - there is a link on the OAC website but it doesnt seem to work: the other is ChinaHealth from Shenzhen near Hong Kong) - that make punching/timing systems (similar to SI). I once helped someone from one of these companies contact IOF about getting accreditation but as far as I know - despite several attempts - he never even got a reply.

Does anyone know what sort of "deals" IOF has with Emit and SI - eg is there any exclusivity arrangement?
Apr 29, 2010 3:04 AM # 
GuyO:
Maybe the IOF thinks that SFR -- like a LOT of what comes out of China -- is an unreliable knockoff.
Apr 29, 2010 4:30 AM # 
AZ:
IOF rules require a single clock for finish timing (or they did when I was last involved in a WRE event

I'm not sure this is a correct interpretation of the current (2010) rules. In my reading a punching finish (i.e. using SI or EMIT) is allowed by the IOF at WMOC and WRE events, but not at WOC, JWOC, and World Cup events. Here are the rules I think apply:

23.5 The finishing time shall be measured when the competitor’s chest crosses the finish line, or when the competitor punches at the finish line,...

EDITED TO CLARIFY / FIX HTML ISSUE:

23.7 < Applies to WOC, WCup, JWOC > In interval start races, an electronic finish timing system shall be used, allowing the competitor to finish by running across the finish line.
Apr 29, 2010 4:42 AM # 
jjcote:
I read 23.7 to mean that your time is when you cross the line, not when you punch, meaning that the time from the gizmo you're carrying doesn't cut it.
Apr 29, 2010 4:49 AM # 
Tundra/Desert:
SFR is somewhat reliable (I haven't encountered a failure in several hundred punchings), and it's certainly not a knockoff. The original engineering is in St. Petersburg and the software is quite original, head and shoulders above Kraemer's creations. I don't know if they are interested in an IOF certification. They'd have to get more of a market share in Russia to start with, their stuff appears to be rarely used outside of the Russian Northwest.
Apr 29, 2010 5:22 AM # 
AZ:
Hm, some crazy HTML stuff obscured my post. Apologies for the careless proof reading of my post.

Rule 23.7 does indeed mean exactly what JJ says, but what got lost in the HTML is that this rules applies only to WOC, JWOC, and WCup events - so for "normal" WRE events (and for WMOC) it is okay to time using Emit or SI

(I have corrected my previous posting)
Apr 29, 2010 6:56 AM # 
Jagge:
In Finland using Emit cards' internal clock for timing is allowed only in training events, not for any races, not even for small ones. I hope/believe that's how it is in Norway too. If it's not prohibited by IOF rules for WRE, that just tells how SI oriented IOF rules and thinking is.

Looks like I wrote nonsense last night. Those SI punching incidents did not happen in (ski-o) woc, it was World Cup and EOC in Ski-Orienteering. Too fast punching in relay, punching with dead unit without getting any signal in men's long. Both competitors were qualified. I wonder is this the way these things will be handled from now on.
Apr 29, 2010 8:05 AM # 
jonm:
@Jagge SI has the clock in the units. This is a better architectural way of ensuring consistent timing, whatever you believe on the other merits of SI vs EMIT.

@Graeme Markup and IP :)
Apr 29, 2010 8:26 AM # 
robplow:
Couldn't Emit produce special start and finish units that write the time to the card? Then the split times would be based on the runner's card's internal clock - but the overall time would be based on the clocks in the punch units - same as SI

GuyO - reliablity and breach of copyright are certainly issues I would expect to be considered in the certification process - but there didn't seem to be any process.
Apr 29, 2010 9:18 AM # 
Jagge:
jonm, I don't see it that way. SI way is much much worse architecture. Someone needs to set times to all those units, precious volunteer work wasted and leaves space for human errors (forgetting to set time right or programming unit wrong).

It's pretty good as it is. You have clock at start and competitors start without any start punch. At finish line you have online control. So start time is based on start list & clock at start (regular start clock, not necessarily emit one) and finish time is based on computer's clock at the time of online finish punch. Splits are from card's clock, possible card's drifting goes to split from start to #1. If one likes to get splits right, just put online control before start too, compare real time between online splits against card's split and stretch splits accordingly if times differ.

All you need is to make sure those two clocks are running same time and same speed, start clock and finish clock (laptop). No need to set times to any punch units or do any punch unit programming. Just think how much volunteer hours is wasted for programming SI units and setting times right. Emit organizers just take units from garage, put them to forest, pick them up later and put them back to garage. No need for programming or time setting - less volunteer work, no human errors.

robplow's idea of writer units wouldn't be bad for trainings, but I am not sure is there any real need for it.
Apr 29, 2010 9:52 AM # 
AZ:
Hey Jagge, there is an SI Time Master available to organizers that save almost all of that volunteer time. When I pick up the SI units from the garage I spend approximately two to three minutes setting the time on all of the units. And I never program the units, I just use 'training mode'. So I'm not sure this is the best way to argue for the superiority of the Emit system architecture (though I know there are many other arguments to support the idea ;-)

For those of us not so familiar with Emit = can you tell us more about this online finish control. It has a cable to a computer? How far away can the computer be?

Also how does the organizer know which Emit unit to put at each control site? Are they identified in some way, or does it perhaps not matter?
Apr 29, 2010 9:55 AM # 
jonm:
Thats all implementation and nothing I disagree with. My point is simply you cant garantee EMIT card clocks are the same and you have no way of checking them. Architecturally therefore one clock is a better way that 100s. Your description of using EMIT as a punching (not timing) system by externalising the timing is exactly my point .. your back to 1 clock.
Apr 29, 2010 10:11 AM # 
robplow:
Jagge now that you explain how to use an online control for the finish - that achieves what I was thinking anyway.

My big problem with SI is that if a punch unit fails to work, the runner is penalised - either through disqualifiation or loss of time punching a back up unit (or even worse using a pin punch on the map).

In theory Emit avoids this by having an automatic manual back up (yes I know that the back up paper with pin pricks is not very reliable - especially in wet conditions - but the concept is right).

I just dont see how it is fair that whether or not the punch unit works correctly is the runner's responsibility (as it is in the rules for SI).

I seem to remember you writing about this at length in a previous post some time ago Jagge.
Apr 29, 2010 10:21 AM # 
Jagge:
Well, you need to have all punch units at one place to do the timing in with the Time master in few minutes. If they are already in forest, it takes a whole lot longer because you need to get to each control. So you will have to do the timing several days before the event, so clock's may drift and be off at the day of the race. And if you put them to forest at the very last moment, you can't do proper test runs days before the race and also you'll have to use those valuable last hours for playing with punch units with hurry - insted of just making sure everything is ok. That leaves room for human errors.

Online finish is just regular online control. I don't know how long the cable to online control can be. Longest I have seen is about kilometer. If used as finish unit it is usually fow meters away from the computer.

Emit punch units has code. Some use same emit codes as control codes, some not. At forst it was not recommended to use same codes, because one might cheat by having own punch units at car and just punch there. But these days codes seems to usually be tha same. Anyway, you there is mapping emit code -> control code, so basicly it doesnt matter. Usually we put punch units to forest several days before the event, and do lots of test runs for all courses. Good training and that also makes sure we know what sort of units we actually have out there. And all of them actually work. So during last day course setter doesnt have to worry much about punch unit, he/she can instead for example make sure maps an control descriptions are ok. I am pretty sure this possibility to make most of the work days in advance that gives time for ensuring all is right is the biggest reasons why we haven't seen much disasters here. We have limited amount of volunteers available and it sure doesnt sound good at all if the most competent person(s) (course setter in worst case) would mainly playing with punch units during last day.
Apr 29, 2010 10:42 AM # 
AZ:
With the SI system, I use the Time Master on all controls immediately before I put those controls into the forest. I do this a few days before the event and don't worry if they drift. The only controls that I really care about are the Start & Finish controls - which can be properly synchronized very quickly immediately before the event. So I may get a small drift which will result in splits being slightly wrong, but all competitor times will be accurate. So I think SI & Emit are quite similar in this respect of volunteer time.

Though it is not always the case that things go well. On the Barebones.ca website there is a list of "errors we have made" and quite a number have to do with SI units, including a rather bizarre one at a skiO event where for some reason we did not use the TimeMaster to synchronize all controls and had some new ones on European time zones and some stadium ones on Day Light Savings while the other old ones weren't. So in this case we were lazy with our procedures and as a result had some awful problems (nothing that our top timing volunteer couldn't fix by wasting a few of his precious hours after the event)
Apr 29, 2010 11:10 AM # 
IanW:
There have been several instances of EMIT clocks running at half-speed - due to the binary nature of whatever clocks they are using - and hence recording double the elapsed time.

But as has been said before, the electronic system was never designed to be a timing system, it was introduced to record and check correct punching. The split times and other timings are an added benefit. And both systems have their problems - I don't think you can put 100% trust in either to work consistently perfectly.

Unlike in Finland, clubs in the UK regularly use both systems as the sole timing device. Fortunately this is rare at the larger events because online and radio controls are often used, but not always. I could see the clock drift being a problem at sprint races for instance...
Apr 29, 2010 11:23 AM # 
Jagge:
You can do errors, no matter what system you use.

For some reason people seem to program their units and run around with them at last minute. And keep failing to set time right. Like in tiomila (of was it 25manna) couple of years ago they had several finish units runnig different time, so results were quite different than the order they were seen punching at finish line. I think they afterwards fixed some that for first teams, but I am not sure did they ever fix it for the rest. If they did it took so many days I never bothered to check - we were disqualified for too fast punching at last control or something, as usual. Maybe all organizers just don't know the right way to do it, or they have older equipment. Or they need to borrow units for larger races and they have to program them to make them work together with their own equipment. And even if you if you need to set time right for just couple of units you can do human error and just forget one of them.

robplow, Emit back up paper survives just fine if you put some tape on it. But Emit back up concept goes all wrong (at least) in one place. You can get epunch without getting back up punch. It needs more carefull and slow punching to get the pin mark and because you get epunch without being so carefull people just get used to be fast & sloppy. Its common people get all epuches but only half of the pin mark. So when something goes wrong, back up paper is often useless because in 50% of the times pin mark is not there. I believe back up mark should be the one you get easier & faster and the real one only if you are not too sloppy. Like this we would know competitor one punched at right control. And disqualify only if too many real punch are missing (runner is too sloppy puncher to be qualified) - even if back up punches are just fine. You need to put limit somewhere. But it's not possible because it is designed the way it is.
Apr 29, 2010 12:57 PM # 
Austris:
Many people see it from the organizer's point of view. While valid view, I think the main focus should be on a participant. How easy and fair it is to the participant.

Of course we can dream of the ideal system, but what we have is Emit and SI. Having used both of them quite a lot, I would say SI is more user friendly.

Although, I once had a bitter experience in relay comp "mispunching" at the last control (organizer's fault because we were winning team and the unit had gone to sleep), I waited for 2 seconds to be safe, but couldn't confirm the punch because of extreme noise in finish area.

In case mispunching with SI, there is often a record in the checkpoint. The problem is with rules, which don't allow to take that info into account.

I think Emit has a major showstopper - user unit can die any moment during the race. OK, there is backup, but crappy backup.

For SI timing, I think it is very fine even if units are not in sync, because as long as the finish units has the same time for all competitors, it is fair to everyone.
Apr 29, 2010 1:12 PM # 
simmo:
Maybe in Finland it is ok, but there are many countries where controls put out more than 12 hours ahead (the previous afternoon) would be stolen or interfered with by non-orienteers (eg hoons, hunters, trail-bike riders, dope planters, etc).

Usually you couldn't put out controls on roads (for easy courses) early - sometimes you wouldn't put these out until just before the event. Often you would also have a hard course drinks control on a road, that too wouldn't be placed until the day of the event. Sprint events - you wouldn't want to put the controls out in a town until your control minders are there just before the event.

As AZ says, the only times that matter are start & finish so does anyone care if an SI station in the forest is out by a few minutes? And if the station doesn't work, everybody has to manual punch so its the same for everybody. If runners don't wait for the beep and light (and manual punch if it doesn't), they deserve to be mp. If the station does beep but for some reason the runner's download shows mp, a protest can be made and determined when the station is retrieved and interrogated.

Its impossibly expensive for countries like Australia where only SI is used to change to Emit. Maybe we could change to SFR as it is much less expensive, but we couldn't do it until it has IOF approval.
Apr 29, 2010 1:42 PM # 
Cristina:
You can get epunch without getting back up punch.

Indeed. My experience with this has soured my view of Emit greatly. I think Austris is right to think about the user experience.
Apr 29, 2010 3:09 PM # 
Jagge:
And if the station doesn't work, everybody has to manual punch so its the same for everybody.

No. Some competitors may visit the control before it dies. And if there is several units, some are lucky enough to punch with the ok one. Very unfair in relays, mass starts, chase starts, if there is several units unit at last or second last and one of them is dead. The unlucky one who uses the dead one has no chance in sprint finish. That has happenned in 25manna, same happen two months ago in ski O world cup, men's long mass start. Two units, the other one was dead. Not fair at all.

If runners don't wait for the beep and light (and manual punch if it doesn't), they deserve to be mp.

In that ski o world cup they did not think that way. They qualified two competitor (medalist) who did not wait for light/beep and never got any punch mark to their SI sticks. And that the way it should be, it's stupid if competitors has to wait for signals / computers to communicate during a running race.
Apr 29, 2010 3:31 PM # 
jjcote:
interfered with by non-orienteers (eg hoons, hunters, trail-bike riders, dope planters, etc).

In the USA, we don't have any trouble with people hooning our egs.
Apr 29, 2010 3:36 PM # 
expresso:
Our club has 1-2 SI stations stolen per year. Along with the cost of stand & bag, this gets expensive (~$160USD each). Also, we are often sending stations in for battery replacement, not cheap either. And then we also flash BIOS updates into them as part of winter maintenance. Although we'll likely never get to experient with EMIT here in the US, I expect these costs and hassles are less for Finnish clubs.
As a competitor, I really don't like the clunky EMIT chip.
Apr 29, 2010 3:50 PM # 
jmnipen:
In my orienteering career ive used both emit and Si frequently. SI as a kid, and now Emit unless going internationally. During the time ive had emit i havent had problems with it ever, though i replaced my emit with a newer one because the last one is 8 years old, as they should be replaced every 5 years.

Before, i prefered SI, but after using emit for a couple of years, Ive realised that SI just sucks bollox. Its slow, there are huge complications when organizing with SI, the punches are heavy, they are easy to loose, and they even make noise when you punch!! (how am i supposed to use stealth, when that thing can be heard 40 meters away?). Not only that, getting to the control, it takes almost just as long time as punching with SI, as with traditional punching system...

Although the emit is big, you can easily fit the emit to the punch, so its quick, and the best part of it is that since there is a backup, even if it skrews up, you can still can get away from a DSQ guaranteed, whereas SI you cannot.

The best part is that when using the emit, its easy to organize training events with it, and there are expensive objects out in the woods, so the damage isnt great if its stolen.
May 1, 2010 3:21 AM # 
simmo:
@ Jagge. It's also unfair if later starters have heavy rain, or if a flag has fallen over and can't be seen immediately. Later starters in many events have 'elephant tracks' to follow - is that fair? When manual punching was used, runners who didn't check they'd punched correctly in the box would be MP, why should it be different for electronic systems? Organisers need to ensure all stations are working, have a good battery charge, and test them immediately before the race. I've made several thousand SI punches and not one fail yet.

I agree that having to do a manual punch where others may not have had to during a sprint race or a chasing/mass start race may be a problem if there is a 1 in 10000 chance a unit may fail. In sprint, organisers need to be extra careful that all the stations are working, and they can do this because they are all close, and each control could have a minder (especially at a big race) who can test their station throughout the race and change it if necessary. They don't have mass or chasing start at WOC (except relay), precisely because there is a lot of luck involved. If there is a small amount of bad luck with a station fail would that have as much influence on the race as following a better orienteer, or a forking system that isn't 100% fair?
May 1, 2010 8:55 AM # 
Jagge:
Back to topic, emit timing reliability. Like I wrote, there is two clocks, one at start (regular start clock) and one at finish (laptop's clock). Times are based on those and to get times write you need to set times right. Quite often that's not the case. organizers simply forget to set times right. And runners will get different time with their own watch. Fortunately it is same for everybody, and it is easy to fix afterwards and system allows fixing it even during the race (add/subtract x seconds form those competitors results who started from start B ). So basicly forgetting to set times right will not cause any damage. The danger here is start clock is so much off some competitors may not get to start in time. Note also there can be several puch units at finish line and it will not cause any trouble - finish time comes from computers clock, so you can have several units or change units on the fly without having to set times right or anything.

If start punch is not utse, the biggest danger is letting some competitors start too early. It may not get noticed if clear puncing is not made at some accurate minute before start. This can be erased by having start puch, but usually there is no start puch. For example last weekend that happenned in Norway, they noticed afterwards from gps tracking last runners started one minute too early and the winner changed several hours after the event. Start is the place yuo need to carefull if emit timing without start puch is used. I guess that's the case with any punchng system if start punch is used.

So there was start and finish, then we have punching at forest. Punch units and cards are silent, so competitor will not know is car or punch unit dead or not. That's the best part pf emit timing. Electronics will not affect your time anyhow. You don't need to wait for any signals, you don't need to re-punch or pin punch if punch unit or your card has died. So if all punch units are ok or dead, it doesnt matter - the course will take exactly the same amout of minutes and seconds and you wil be punching at winish line at the very same second. As it should, it's a race and timing should not have any affect on competitors time. This is why emit timing is not only reliable, it also accurate. If you think SI's waiting, repunching and pin punching you can't say reuslt times are accurate at all, competitors times are affected so munch.

So afterwards we can considier what we do if there was dead unit in forest. for eample if punch unit is dead and also broken in a way it's not easy to get pin mark, we can simply not disqualify anyone who should have punched there during the time frame dead unit was there. And if/when you replace the broken unit you dont need to think about timing - there is no clock inside the unit. You can just have pile of reserve units - basicly taken straight from your garage - and replace dead one if needed.

One more place to go wrong: you need to know who's running with which card. If you don't have any record of that, you will just have lots of times and result times for emit card numbers, but you can give prizes or disqualify anyone - you dont know who were the runners behind card numbers. That happenned once in tiomila. They used rental cards only, did not register cards and got just lots of splits and result without knowing who actually ran with them and which team. I don't think they ever got any proper results. But I guess you need to know that with SI too.

I think if you think that all trough you'll notice there is not so many places you can go wrong as organizer. And as competitor you dont need to think about electronics, try to see flashes in bright sunlight or try hearing beeps when people are cheering / close loudspeakers. Or be afraid someone hears the beep when you punch in relay/mass start. The bulky card and akward punching is the price we pay to get for all that (big card becaus clock and electronics are there, akward punching because of back up AND it also forces competitor to put card close enough and gives electronics enough time to do it's thing). When you get used to that, you usually see the benefits and are happy to carry that card and puch like that to get results right and to have less pain when it is your turn to organize a event. And what it comes to punching speed, with emit it's up to you how fast you can punch. With SI it's up to electronics, not you. There is clear difference and emit is clearly better in this aspect too.

So what it comes to topic, Emit timing is the most reliable we have. Actually the only somehow reliable timing/punching system there is. It's as even more reliable than good old pin punchig, pin punches used to broke too and some had to use less optimally placed punces or even try punching with broken pin punch.

--

Simmo: What it comes to elephant trails, that just fair if the last best starting slots are earned by running well in qualification. Some even see following fair for same reason (I don't). Forking is never fair, but we can't do much about it if we has to use forking. We can't do much with weather either, but puching system is not like weather, it's not beyond us, not given by gods. we can use fair one. It is lame to say punching sustem can cause several seconds unfairnesses just because there can be rain or elephants can be on some runners way. Of course some things may affect more, but that's bad excuse for adding more unfairness on purpose, if we really don't need to. Or consider a punching system is not bad because there is bigger unfainess possibilities.

Also, as you wrote with SI you need to be specially carefull with sprint - I guess in mass starts and relays last controls too - to make sure there is no dead units. it's not that easy as unit can die any moment. Thats why it happens, like couple months ago in ski O European champs / world cup. Twice. I am sure they were carufull enough, but it just happens. Note anso, being extra carefull here and there as you said needs voluneer work, it need experienced volunteers and it also leaves space for human errors. Yet more room for human errors and need for volunteers, all that just becase the punching system is designed in a way it is. You wrote there is no they don't have mass or chasing start at WOC (except relay). True, but I don't think the reason is luck. And we have those in world cup and other races. I see SI is fine for hobby o, but it really should not be used in any real competitions, just as emit's internal clocks should not be used for race timing.

I havent raced hundreds of races with SI, just tens of races. So far about 3-5% of units ahve been dead. On average in every second race there has been at least one dead unit.
May 1, 2010 10:34 AM # 
Tooms:
Succinct.
May 1, 2010 2:19 PM # 
expresso:
@Jagge
Oh come on !!
3-5% of the units were dead ??
You were making a compelling case for EMIT but lost my confidence with this "statistic" and how you can't seem to find any of the good in SI. Your argument is unfairly skewed, IMO.
May 1, 2010 3:23 PM # 
Jagge:
You are right my statistics are skewed and not fair. I think I have run wrong races. Not much "regular" middle size SI races, just tiny races in spain/estonia etc - less important local races where organizers just take it easy for purpose. And big mass races like 25 manna 10mila, where units get exhausded for such a big amount of runners during short period in cold weather. And most of the dead unit cases happenend long time ago, I rememer one park sprint in Estonia where almost all of the units were dead because of some sort of cell phone ringing by while carrying units problem. It would take lots of races to fix my % and since I don't race much with SI these days my statistic are still badly skewed. Like I wrote, I have run only tens of races with SI. I guess your statistics are skewed too, to the other direction for opposite reasons. I guess the truth is somwhere between and much closer to your statistics.

And I simply don't know enough about SI to know much good things about it. My SI stick looks good, nice red color with golden tip. And they did not charge us anything when we had to send all of our club's new sticks back to SportIdent to be checked after three of them failing to work properly at 25manna. That was nice too.
May 1, 2010 7:58 PM # 
Jagge:
Off topic, but since simmo asked here is few positive comments about SI. It may sound strange, but I believe SI may have potential to be as good or even better than Emit at least in some perspective. It just would need to be used a bit differently and change rules a bit. About like this:

- Turn off light and sound signal. Punching would become a bit like punching with Emit - competitor would not know is unit dead or alive. All he can do is punch and go on. No need for pin punch back up or secondary unit, all can use the same unit, works it or not. This would make it easier for organizers. If unit dies no ones run will be affected so there is not that big hurry either to replace the unit.

- since there is no signal, competitor will have to learn to keep stick long enough in the hole to get the punch. We might say one should keep it there half second. If units are as fast as SI claim, it should be no problem.

- some competitors push limits and punch faster, and some will punch too fast. Here SI would beat Emit hands down. SI has back up in punch unit and back up mark gets always done first and the actual mark after that (unlike case Emit, actual mark first, back up later if ever). So we could handle these too fast punchers with rules without having to disqualify for being sloppy just once. So, if a punch is missing competitor could ask back up inspection. Inspection fee would be 10 euros per missing punch. Maximum 3 punches, if there is more it would be DQ for punching too fast on purpose. If back up punch is found, competitor would be qualified. If not disqualified. If dead unit is found at that control and it appears to be died before the competitor is supposed to have been there, the competitor is qualified and inspection fee returned.

Like this no one would be able get advantage by punching too fast (inspection fee would make them learn), no one would be disqualified for being sloppy just once, dead units would do no harm, no beebs so you can punch stealth, slightly less work for organizers (no need for back up pin punch etc) except the inspection team would be needed. But inspection fees should make it worthwhile. The drawback is possible delay of final results, but that would not be too bad if there would be inspection group ready and waiting, possibly already in forest laptop in back bag and cell phone in hand. If equipment works as well as simmo says, races would go pretty smoothly like this. If its' sprint, controls are not that far away after all. There would still be all SI timing and programming annoyances but the dead unit re-punch delay thing would be history, and SI might be almost equal or better than Emit from competitors perspective.

But I guess that all is all too far out for SI folks who are so fond of those beeps and flashes.
May 1, 2010 8:52 PM # 
Cristina:
Regarding punch speed of SI units - with the newer, faster SI sticks (SI-6 and beyond?), I find that I can not punch faster than the unit can work. In other words, by the time I have inserted the SI stick all the way, the unit has already beeped and flashed. So, with the newer system punch time (as long you actually punch and not make a half-hearted attempt) is not really an issue.
May 2, 2010 3:55 AM # 
simmo:
Jagge - right with you there. Maybe SFR might be more willing to eliminate their station signals before SI?
May 2, 2010 6:02 AM # 
Jagge:
I think you can turn light and sound off from SI units if you like. I think it's up to us, SI users.

SFR would be more intresting if they would have designed it the other way around, cheap & lightweight batteryless $3 control unit and expensive & possible bulky punch card. Don't know about back up, but stealing or breaking $3 control unit would not be that big deal. It might be possible to sell some sets to current emit / SI areas just to be used for urban events.
May 2, 2010 6:13 AM # 
simmo:
I might buy a set to use in schools, 15 stations and 30 cards would be around 840EU. SI school set of 14 stations, 20 cards and split printer (no software) is 3500EU!
May 2, 2010 6:14 AM # 
AZ:
"those beeps and flashes"

In an informal survey of under-8 year olds, SI was found to be 100% preferred to Emit. And with the under-4 year olds it was even more so!

Do they have kids courses with Emit ;-)
May 2, 2010 6:48 AM # 
Jagge:
"Do they have kids courses with Emit"

Back up spike is too dangerous for kids. But if you take it away, it's safe. But no back up, kids would hate that for sure. I also suggest using emit instead of SI for kids under 1 year old. They might try eating the tiny SI stick and get suffocated. Bulky emit is far more safe - at least if battery inside it does not start leaking.
May 2, 2010 9:14 AM # 
gruver:
You're all arguing over something that may be superceded soon. We will carry a black-box GPS unit for controls and software will determine control visits (GPS track passing within x metres of each position). There will probably still be a flag at the controls and we might as well put a pin punch there for the backup, but we save heaps of euros on the control stations. Timing by clocks at the start and finish, this is easy for everyone to understand.
May 2, 2010 9:46 AM # 
Jagge:
gruver,I think improving gps accuarcy to that level will take so long we will see lie detector punching first. We'll have just flags and ask at finish did you or did you not not visit and touch all right control flags in right order and did you check codes too.
May 3, 2010 12:06 AM # 
TheInvisibleLog:
I think someone is working on the GPS/no control option for street orienteering.

Back to SI. Anyone aware of the nature of the problem with the withdrawn 565 firmware? Our club members had three failures to record punches on SI-6 sticks at the recent Australian Easter carnival. In at least two cases there was no likelihood of over-fast punching. We used the fact that the events with the mispunches were using 565 firmware to argue for interrogation of the control boxes and had the three runners reinstated as a result. Without the evidence of the firmware withdrawal on the web site, each would have been an mp result.
May 3, 2010 11:20 AM # 
gruver:
http://freeog.sourceforge.net

Independently of these guys, I've been doing some experiments. Taking a GPS on a course and comparing the track with the control coordinates to see what sort of tolerance is needed to cater for the errors - and how this varies with the terrain. Example - on a steep fault-scarp area (400m elevation range) with forest cover, a tolerance of 23m would have given me a "pass". Used a conventional handheld (SIRF III chip). Believe we could in principle use it for certain competitions now - eg MTBO.
May 3, 2010 1:00 PM # 
ndobbs:
I thought EMIT did flashes! Marcus P used an organisers' brick in a NOC sprint a few years back. It died halfway around, so the emit units stopped flashing, so he started punching "properly". Result: a big fat DSQ.

I'm not sure emit is silent either. One frequently hears punching sounds like "stupid@%#*@emit bricks#@#%!".
May 3, 2010 4:48 PM # 
Jagge:
It makes big difference do you just make visit to the top of a 5m cliff right above control or do you actually climb down, punch and climb back. GPS accuracy should be less than 50 cm, also altitude, in all rock holes and urban canyons you can imagine for control sites, all the time and instantly - not after minutes measuring and averaging. And unit should be cheap and lightweight. GPS is not even close that, and most likey will not be for a long time if ever.

But it might work somehow together with pin punching. Organizers would not check punches, maybe just winners when class is ready or if gps track auto inspection alerts there may be something. They would just archive cards, publish gps tracks and let competitors study each others routes afterwards. if someone finds something strange, he could request punch card to be examined afterwards. Possibly organizers might do some sort of fast transillumination scanning to cards too and publish those images too. I'd think it's unlikely anyone would cheat - maybe just some misspunching without noticing it. But that might not be good solution for big races like oringen or any championchip races or multi day races where changing results afterwards is out of the question. And not sure would it be that good solution for any type of event compared to current ways to do it.
May 3, 2010 9:48 PM # 
gruver:
I'm surprised at your pessimism about the rate of change of technology Jagge. I think that there will be many commercial motivations to reduce the price and improve the accuracy, including in the "urban canyons" and indoors where it doesn't work at all now. But in a small club in a small orienteering country far from the orienteering "centre" I am content to observe, to experiment, and not spend too much money on SI stations:-))

Actually, we still use "question and answer" controls for some events, and you know what? We don't check the answers. The knowledge that they CAN be checked is a good "lie detector". So your suggestion of just archiving the cards is a good one.
May 3, 2010 10:59 PM # 
O-ing:
Most times GPS is good for 23m, as you say, in good conditions maybe 10 to 15m. That's nowhere near good enough, as Jagge says.

There are currently GPS systems that can deliver cm accuracy - "Real Time Kinetic": but these systems are extremely costly and require an accurate base station to be set up and triangulate with the hand held unit. It just isn't the right technological solution to the problem (much like trying to use a GPS for navigation).

Jagge: Thanks for the above - it made my day. Going to such extraordinary lengths to explain the issues and solutions is wonderful. Keep at it.
May 3, 2010 11:00 PM # 
Cristina:
I think there's something pretty fundamental about runners having to physically touch or pass over a very exact location. As Jagge says, GPS is not nearly accurate enough to ensure that people are actually visiting the exact feature, and some features have unsymmetric approachability - you wouldn't want people to be able to 'visit' a control while on the wrong side of an impassable fence in a sprint, for example.

I don't much care what kind of technology is used for the 'punching', but I think it should require that the runner actually touch (or run on top of) a location. Emit and SI both work for that.
May 4, 2010 9:39 AM # 
mikee:
Reliability of SI:
http://www.h-o-t-t.com/hott/Blog/Entries/2010/5/4_...
???
May 4, 2010 9:40 AM # 
gruver:
Thank you for your feedback, which has shifted my thinking a little. But I can still see the prize in the distance - saving thousands of dollars of hardware out there in the forest.

Yes its hardly possible that a GPS will ever be accurate enough to determine whether a person was at a position within an arms reach, and competitors navigating with a map need something to tell them they are "there". So we will still have a flag. Now any electronic punching method (SI, EMIT) needs to have a manual backup before it is acceptable. As Jagge suggested this could be an old-fashioned pin punch.

Now why did we progress beyond cards? Primarily the difficulty of checking them. So we wouldn't. Jagge has given us some methods where the threat of card inspection could ensure compliance. Secondly, the split data. Seen in this light, the GPS data only has to be accurate enough for useful splits. The result doesn't depend on them, their function is competitor feedback and entertainment. We would actually get MORE feedback and entertainment being able to see the whole route.
May 4, 2010 12:18 PM # 
TheInvisibleLog:
Maybe the solution you seek is simple RFID tags as controls.
May 4, 2010 12:36 PM # 
Jagge:
I have been thinking of using something like this to record SI beeps or oher punching sounds, should work pretty well if you can carry it close to SI stick. You should get accurate time when SI sound happenned, together with gps track it should proof somehow you were at last pretty close to a beeping SI unit at the right (gps accuracy?) control. You could also use it as stand alone puncing system - just buy different kind of horns, bells, triangles and other musical instruments and let competitor play them at each control. Just download mp3 file, seach for those sounds, check sound is right and pick split time.
May 4, 2010 1:12 PM # 
iriharding:
maybe we should expand our thinking about what could be done.

Some clubs have done radio controls in a very few locations. What if the control stations not only recorded the time on the e-punch stick, but also were part of a communications network (cell phone technology, wifi or other) that communicated the punch and runner info to HQ. HQ could then display a live leader board with splits.

With OR epunch software you can already have a live "finish" line display of results (on the web too) ; this network idea is taking it one step further (and increasing the spectator interest).
May 4, 2010 2:56 PM # 
AZ:
Regarding Sandy's "rant" about SI - I agree with her that there is something wrong in the SI system - simply too many mis-punches from people who clearly are not "punching too fast". Having a history of real-time software development my spidey senses are tingling and I suspect some kind of software bug, perhaps in an interrupt handler. But this is pure speculation. Does anyone know of any systematic testing that has been done on the SI system? Perhaps all our frustrations could be channelled to debugging this system that isn't going to go away ;-)
May 4, 2010 3:37 PM # 
Uncle JiM:
SI admits to a problem with there system when it comes to using SI 6 cards with the new firmware, 565. We had 3 athletes MP, using SI 6 cards at our recent Easter event. We got them all reinstated once we checked the SI box. All athletes reported that they saw the unit flash, that is why we asked the organisers to check the SI box.

Anybody know if Sandy has a SI 6 card, and was Tio using the new firmware
May 4, 2010 4:25 PM # 
Steffen:
I still remember from punching cards times that many people told the were one hundred percent at the control and punched even there were no sign of a punch on the card. So to be honest I didn´t trust anybody who told me that he was 100 percent and the control and punched.
May 4, 2010 4:42 PM # 
Jagge:
I also remember how one guy were 100% sure he punched and no sign of the punch. But after further investigations we noticed there was a slightly strange punch mark, we did not have such a pin punch in forest at all. So he had punched on wrong box, precisely on previous punch and most of the pins hit the same holes as the previous punch. If I can remember right the new punch did only there was just one new pin mark (or two). At first it was disqualification for sure, but then he got result. The we noticed it would be possible to punch also exactly on one or two previous punches in a way there would be no sign of the punch at all. You would have to punch to wrong box, of course, so you can disqualify for sure, but still he might have been be telling the truth when he claimed he punched.
May 4, 2010 5:01 PM # 
jjcote:
Many times when checking cards, I have had other people tell me that they could not see any sign of a punch, or could not understand the punch patterns. 90+% of the time, I look at the card for 5-10 seconds and find all of the punches. Sometimes it's more complicated to figure out. When there really is a missing punch, it's usually from a control with drinking water. Once in a great while we can't find the punch, send for the person, and when he looks at the map, he realizes that he didn't go to the control.
May 4, 2010 7:33 PM # 
Steffen:
When I said no sign I mean this. I´ve controlled punched cards for long time enough and a punch over one other is figured out in seconds. But the normal thing when you come to finish is that you are sure to have all controls. But like always there are people who could agnize they and some who could not. And Jagge what I what to now. How many pays you EMIT you that you make the this advertisement here. I mean the topic of this thread are strange and serious problems with EMIT cards not with SI.
May 4, 2010 7:53 PM # 
AZ:
Steffen - in the old days I would allow my self to be convinced with such arguments (that the problem is with the people, not with the hardware). However, even though I'm sure there are cases in which people are at fault, I started to change my thinking when two people who punched immediately before me at a control were disqualified for bad punching at that very control. So I am convinced that these two made legitimate punches and that the hardware failed. I am also convinced that a good percentage of people who claim to have made a good punch are not mistaken - that they did indeed do everything right. On the other hand, I see no reason to expect that the SI system is bug-free. I am a software developer by profession and know of absolutely no software that does not have bugs. So I am inclined to think that there is a very big bug in the software that is almost impossible to duplicate and therefore extraordinarily difficult to find and correct. That makes more sense to me than trying to maintain, against growing evidence, that the people are punching badly. I think we should try to help find the bug. Perhaps a $1,000 prize for the first person to be able to duplicate the problem??
May 4, 2010 8:16 PM # 
Tundra/Desert:
Adrian, I do hardware and firmware to feed the map-sport habit. $1k is at least an order of magnitude off, given how elusive the issue is.
May 4, 2010 8:31 PM # 
mprg:
I know this thread's original topic is Emit timing, but I would like to make a related comment about my experiences of clock drift on SportIdent units.

The units I used for this test are all BSF8-DB, the new small ones with the LED screen on the back. They all have a manufacturing date of approx April 2008 and they all have firmware version 5.65 (now withdrawn, we will have to downgrade them all). I programmed the units on February 12th and I made sure that the laptop I used was synchronised to Internet time just before I started.

I woke 20 of the units up this evening, May 4th, 81 days later, and compared the time displayed on the units' LED screens with my laptop which was still on recently updated internet time. My comparison method was crude, I simply looked at the LED screen and then the computer's clock and noted the difference as best I could.

I was very surprised to see the following:
- only one of the 20 units was still on the same time as my laptop
- eighteen of them had drifted forwards by between 1 and 15 seconds
- one of them had drifted back by 2 seconds
- the average clock drift was +6 seconds and the mean was +7
- five of the units had drifted forward by 10 or more seconds
- nearly all of the drifts were different.

Taking the worst case, +15 seconds in 81 days, equates to about 0.2 of a second per day. I know this does not seem like a lot, but if you programmed your finish units a week before an event and one finish unit drifted forwards and another back, you could have a greater than 2 second discrepancy in finish times. What I learned from this little test is ALWAYS program the units the NIGHT BEFORE the event, and look at the LED just before they are placed. Also I think this means that SI is unsuitable for permanent, or semi-permanent, courses.

I am not a hardware specialist, but it does seem to me that this amount of clock drift is high. Since the original point of this thread concerns timing accuracy in Emit, maybe getting reliable low-power clocks is harder than I thought.

has anyone else has experienced similar Sportident clock drifts?

Marcus Geoghegan
Ajax OC, Dublin
May 4, 2010 8:38 PM # 
Eriol:
I think the biggest mystery with SI is that Swedish people never experience that horrific bug that makes you mispunch even though you punched, saw the flash and heard the beep. Or is it just the national character that makes them too humble to ever complain about an obviously flawed system?
May 4, 2010 9:48 PM # 
AZ:
In defense of SI for permanent courses:
1. it helps to have an SI Time Master, that can used to set an SI control's time simply by touching the two devices together
2. really there are only two super-important controls to ensure have the correct time: the start & finish. If the others drift it just means splits are off. Start & Finish should be synced once a week or so.
3. once every few weeks a volunteer can go around the course with the SI master and sync the times on all the units.

We have a permanently installed course at Canmore that uses this system to keep the units synchronized and it seems to work well enough. In fact, we find most participants aren't as hung up as we are on their precise time - after all there are no ranking points awarded for performance on those courses ;-)

So I would disagree also a bit with Marcus about the need to program all units the night before. I prefer to put up with the drift and put the controls out as early as possible. This reduces stress on organizers and chances for wrongly-placed controls at the cost of some drift in the splits. That is the trade-off that I make. I hate it when we are putting controls out shortly before a race, and much prefer to have vettors just checking (and 'waking up' the SI units) at that point in time.
May 5, 2010 1:36 AM # 
AZ:
I realize the $1,000 is only a token amount compared to the amount of work that will need to be done to investigate this elusive and possibly phantom bug. But perhaps the fame and adoration which will lead to future lucrative contracts would compensate?

I am curious - and might be willing to put up $1,000 for whoever could first provide me a sequence of actions that I would be able to replicate this problem of the SI unit flashing, beeping, and registering a punch but not recording it in a finger stick.

Does such a prize make any sense?
May 5, 2010 2:13 AM # 
jjcote:
In my experience, ordinary quartz crystals without temperature regulation are good for an accuracy of about one minute per month, so that means that you shouldn't be surprised at a two second change after a day. (That's 0.0023% error, right?) (Note that wristwatches do somewhat better because having the watch strapped to your wrist provides some temperature regulation.)
May 5, 2010 7:23 AM # 
Jagge:
Steffen, We used to have pin punches like this:


Punch in the image is "T". As you can see there is symmetric 3x3 matrix of pins. You could punch "I" (three in a row) on "T" in two different ways withour leaving no sign at all. You are quite wizard if you can figure out in seconds if the competitor has actually punched twice or is he just saying so, since there is no visible signs at all. We did millions of punches with these, it is difficult to believe punching two punces on each other like that never happenned. Maybe you used better pin punces, those with more and smaller pins, not in any symmetric matrix.

I have nothing to do with emit, I was't the one who brought SI up on this thread, I have used SI just as a comparison to explain how timing with emit works since thats the topic. I haven't told any of the horror stories we have with SI because those would be off topic. I have just described how both systems work, those are facts and if I have got something wrong anyone can correct me. Anyone can made their own conclusions. If you are not happy with that and think I haven't been constructive, thats your problem. I haven't tried on purpose to misslead discussion from emit's problems to SI.

What it comes to Emit, I have had on strange failure myself. My first card worked perfectly for about ten year (and it's still going strong). But I thought it would be wise to get neew one before battery dies in the middle of some race, so I bough new one with th lcd screen. After one year it broke in the middle of a race. Lcd was still working and time runnig, but it lost connection to outer world after second control. With the button I could see the code nubmer of the latase control (#2), but not the first one. I told it was second punch, se there was one punch befor it in it's memory but there was no way to see the code of it. Unfortunately I had all back up pin marks but the first one :) so I got disqualified. I had all the punches, but there was no way to dig one of them out from the brick. Fortunately I got the replaced replaced for free. There was one year warranty and I had used mine couple of weeks over one year - maybe they did not charge me for all the "advertising" here :)

Anyway, I believe 4-5 years ago they may have did patch of bricks that had far worse failure rate then the older/proper ones. It's dificult to say hoy good are new ones today, time will tell. You don't here too much complaining of these failures in Finland, maybe because most users have older good ones. Also we have lots of training events with emit here, more than actual races (for example here in greater Helsinki area there is several evenst every weekday, 4 open emit training events today, tomorrow five, within ~45 min drive), so failures may happen in those and couse less harm. But if someone from SI country boughts emit bricks and runs with it nothing else than Jukola once a year, that's where the failure will happen of course.

Anyway, emit brick may go crazy like mine did, but I guess more often battery just runs out of juice. That's why we are used to before each season put bricks into freezer, -25C, and then try does it work or not. If it work, battery will most likely survive the coming summer season. I know some clubs do the same with their punch units, but I guess most don't. They just let them die and if it dies during a race they just don't disqualify anyone for not punching at that control, no matter is there pin marks or not. There just will be no splits for that control.

In Jukola its essential to put some tape to make sure back up paper will not fall of and carefully push brick all the way to the bottom at every control. Just as essential as making sure you see the flash with SI. In both cases you should do fine if you do that.
May 5, 2010 7:42 AM # 
IanW:
The new LCD emit cards are more prone to failure than the older cards with no screen. I've experienced two failures myself, one where the card reset itself midway through a relay - not an easy situation to deal with - and have heard of several other occurrences. I wonder if Emit have been too 'clever' with the new cards - there have been several firmware updates to them already - and by their nature they will be less reliable. The only problem I've heard of with the older cards is one of dying batteries - I guess that this could affect the clock running speed?
May 5, 2010 11:22 AM # 
Holger Hott:
It is not (and maybe never has been) allowed to use the EMIT cards for timing at bigger national events in Norway.

One problem with the EMIT display cards is that they record the signal from the online controls, and therefore their memory can be filled up before you reach your last controls. Display cards will not be allowed at the Norwegian Ultra Long Distance Championships this year. All national team runners in Norway got new EMIT cards from EMIT this spring. They did not get display cards, so that maybe tells something about what EMIT thinks of the display cards themselves.

Too bad, I like my display card because I can see right away if I have punched a control, and I can get my splits even when the time recorder stops working (which happens quite often).
May 20, 2010 10:21 AM # 
Jagge:
I have been told there is lots of bad cards among the brand new blue colored Emit cards. They say you can print splits fine on paper, but cards number looks strange and result software refuses to read/accept them and manual work is needed to get splits/punches assigned to the right competitor. Have anyone got similar experiences recently?
May 20, 2010 12:14 PM # 
JLaughlin:
WOC2010 is using EMIT

http://www.woc2010.com/images/otherfiles/pdf/Bulle... page 14
May 20, 2010 2:31 PM # 
jwolff:
When we arranged the Finnish night champs in bad weather last fall, there where about 10 cards that had died during the competition. They were all brand new, i.e. bougth in 2009 although not blue. Cards way older than the recommended 5 years were working fine. Apparently EMIT has a quality problems in recent production.

This discussion thread is closed.