2015 Between-Match Delays

[ayeckley]

Originally Posted by Conor Ryan

Field and connectivity issues are rare, but do occur.

While this was generally true at Pittsburgh it was not the case at Buckeye. I'd estimate (based on my personal observation) that on Thursday less than 25% of Practice Matches required FTA/FTAA involvement including going out to the Alliance Stations. That rate shot up to over 80% for Qualification Matches on Friday, and back below 25% on Saturday.

Anyone know which field is at Windsor?. I'm tempted to say the field at Buckeye was #11 but there a >50% chance that I'm mistaken.

[Jared Russell]

mDNS can make establishing the connection to the robot take for-freaking-ever. If anything, the large number of game pieces and long robot setup times help to mask just how long the delay can be.

[MrTechCenter]

We ran into a similar issue at Central Valley in Week 2. We were behind schedule quite a bit, so we had teams unplug their driver stations immediately after the match ended and the next match would plug their stations in as quickly as possible. The driver stations for the next match would connect to the FMS, but the next match wouldn't be pre-started yet, so they wouldn't see communications with their robot yet. Most of the time, we had to have teams restart the driver station application. But even then, the delay wasn't huge.

I think the biggest delays come during eliminations, because alliances will face back-to-back situations and this requires a field timeout. Tack on a timeout or two and there's a slightly longer today. Still, most regionals are ending on time or earlier than expected.

[MikeE]

Originally Posted by Jared Russell

mDNS can make establishing the connection to the robot take for-freaking-ever. If anything, the large number of game pieces and long robot setup times help to mask just how long the delay can be.

Once a good rhythm is established the various sources of delay tend to align, i.e. they all mask each other and it appears smooth and painless on the surface.
At a couple of my events we timed the inter-match delay and it was interesting to see how much perception and reality were uncorrelated.

[epylko]

Originally Posted by ayeckley

...
Anyone know which field is at Windsor?. I'm tempted to say the field at Buckeye was #11 but there a >50% chance that I'm mistaken.

I'm at Windsor. How do you tell what field is here?

[ayeckley]

Originally Posted by epylko

I'm at Windsor. How do you tell what field is here?

The road cases should have the number printed on them. I don't recall where they were stored last year; I think in the north-west corner of the gym near the mentor lounge.

[epylko]

Originally Posted by ayeckley

The road cases should have the number printed on them. I don't recall where they were stored last year; I think in the north-west corner of the gym near the mentor lounge.

I'll check tomorrow. One of the volunteers said that the US kits stay in the US and the Canada kits stay in Canada. Either way, I'll try and verify tomorrow.

[GeeTwo]

Usually at Bayou, by the time the score had been settled, the robots cleared and replaced, and the field reset (which usually went pretty briskly considering that there are over 200 game pieces), the match was able to start within another minute or so; Chris Copleand often started announcing matchups before the last team was off the field to help achieve this.
About one match in four or five seemed to have extra delays. Of the ones where I could see or later heard the reason, two were improper robot placement (hanging over the landfill or auto zone) and the rest (about five or six) were connectivity issues (one robot, the rest driver station). I suspect most of the others were connectivity, as well; I could occasionally see one team working hard behind the polycarbonate while the others waited.

[Greg McKaskle]

The "mDNS" issues are something that is still being investigated. The better name for it might be "slow DS-robot connection on field", because there are a few glitches here and there that show similar symptoms, but different reasons for the delay/breakage. Also, the necessary DS restarts are caused by a few error conditions that require the TCP comms between FMS and DS to restart. They weren't discovered during beta and show up more often on some laptops than on others.

Good news is that some fixes are being rolled out with field configuration. Others would require new DS or library releases, but there is considerable overhead with a mid-season release. So I guess the summary is that your perception is correct, and it is something that is being worked on more than it may appear.

Greg McKaskle

[EricH]

Originally Posted by epylko

I'll check tomorrow. One of the volunteers said that the US kits stay in the US and the Canada kits stay in Canada. Either way, I'll try and verify tomorrow.

Don't rely on the road case number: 11 is the ref case, 19 is spare FMS console, 07 is field supplies IIRC. The small tags might or might not have anything regarding which field.

You want the case COLOR. All the cases in one field match in color--and one or more of them are generally kept right next to the field. Example: Gold field is currently in SVR, after trekking to Inland Empire, CVR, Los Angeles, Sacramento, and Ventura. (Though, I think Gold field is #6, at least it uses radio set #6.)

[plnyyanks]

Originally Posted by Jared

The laptop will say "FMS Connected, correct station", and our robot will show up as connected with the battery voltage and everything, but the field people cannot see that our laptop is connected. We must disable/enable the network adapter, restart the driver station software, and sometimes even restart the robot.

This typically happens when teams are plugged in before the field has completed prestart (and is usually worse on practice day, when we're making substitutions. From being fieldside at two events now, I've found that restarting the driver station or changing the team number almost always solves the problem (when the field monitor shows a ethernet link and nothing else, or everything but roborio). If not, restarting the NI mDNS responder service does the rest. I'm seeing this at Chesapeake, and it's consistent with what I saw at NYC.

I think the faster turnaround time from the refs as they hit their groove (so the field can complete prestart before the next round of teams plugs in) is the biggest factor in these delays not being seen later in the competition.

Originally Posted by ayeckley

Anyone know which field is at Windsor?

I believe the Windsor field was previously at Montreal and North Bay.

Originally Posted by EricH

(Though, I think Gold field is #6, at least it uses radio set #6.)

The radio set and truck number are two different things. Truck #2 (lime green cases) has radio set #4.

[Shrub]

Matches got behind schedule at Northern Lights from 20 minutes to ~40 and back down to <20; (ie a team supposed to be on the field at 12:00 didn't play until anywhere from 12:20 to 12:40), delays were sometimes really long. (Noodle throwing contest anyone?)

North Star, we've stayed about 20 minutes behind. Some matches take as little as resetting the field and the like, some took 10 minutes or more.

I'd say allowing an entire match cycle at 8 minutes as noted before is good timing.

[orangemoore]

At Midwest we have an aggressive 7 minute pace for match times. We were actually on schedule for most of the day and I think at one point we were ahead of schedule.

(I was actually surprised that each team was only given 10 matches. Usually it seems that Midwest pushes the schedule to the max.)

Maybe it helps that there are no rookie teams?

[GaryVoshol]

At the Michigan events I've volunteered at, the time consumers are, in order (from least to most time):

1. Finalizing and announcing scores
2. Field Reset
3. Robots on and off the field
4. Getting robots connected

Yesterday afternoon we stressed that moving robots was more important than moving totes. This improved the timing of getting the robots onto the field, so they could begin getting out of transport config faster.

When things run well, you can keep to a 7-minute cycle, with occasional 6's or 8's. But then you have a big delay where some robot just will not connect, and the FTA's and Orange Hats are working feverishly. Sometimes it takes 2 or 3 dance songs before you're ready to go, and this year there's no way to make up that time.

I didn't count, but in a normal 80-match qualification in FiM, there's probably 2 or 3 times that the green light goes on before the announcer is done introducing the teams. There's maybe 5 or 6 matches where we're waiting for last minute adjustments to the robot after the light is green, and that only for a few seconds. The rest of the time everything is ready to go except the electronics.

[epylko]

Originally Posted by EricH

Don't rely on the road case number: 11 is the ref case, 19 is spare FMS console, 07 is field supplies IIRC. The small tags might or might not have anything regarding which field.

You want the case COLOR. All the cases in one field match in color--and one or more of them are generally kept right next to the field. Example: Gold field is currently in SVR, after trekking to Inland Empire, CVR, Los Angeles, Sacramento, and Ventura. (Though, I think Gold field is #6, at least it uses radio set #6.)

I wasn't able to see any case numbers. Like you said, I was able to see the different numbers on the crate. I _think_ they were red crates, but now I'm not really sure. Doh!