[FTC]: Samantha? Like it or hate it?

[jamie_1930]

I knew from the start that I wasn't going to be fond of the new samantha module and fcs setup, especially after today's events at the Clarkson Regional, however this is not an attack on those that run the regional it is a new system and things like this can happen and I hope we can come to the root of why here.

My initial thoughts were with the comparison to FRC I feel like while using new technology to pilot the robots does open up more possibilities it will also allow for more problems. Right now I feel like many of the problems of the new samantha module come from interference, as well as the fact that it is brand new this year. For instance it seems like now every new gadget has wifi capabilities and if every cell phone and ipod that's in the arena is putting out a wifi signal then there will be a compounding rate of interference with the wifi used to communicate with the field. I liked bluetooth because it seemed like whenever a problem occured it was my fault not someone else and that meant I could correct the problem the next time we went out for a match. Also while many things do use bluetooth the limited range means someone on the other side of the arena won't be interferring with your connection and there is a smaller number of such devices being used.

If anyone sees something wrong with what I said above please correct me and if anyone that was at Clarkson today has any more insight on the problems that occured I would love to here what happened. Please feel free to post about any problems you see with the samantha and possible fixes to these. We're engineers and we're here to fix problems, please don't post here trying to attack or blame anyone else that isn't going to solve anything, plus while some field techs may make mistakes sometimes we would too if we had their job.

[PhilBot]

Wellll...

You said a lot, but didn't really say anything specifc.

Exactly what were the problems that you saw at the regional?

Maybe someone has seen them before....
Maybe you can prevent them from occuring again.

Just staying there were problems, but that's normal .... wasn't particularly helpfull or engineering-like. More like most news reports these days...

Try this approach next time........

-----------------------------------------------
We saw several problems at the xxx regional.
a) xxx happened whenever yyy occured
b) some people had zzz happen to them...

The problems seemed to be Samantha based.
I say this for the following reasons:
a) reason 1
b) reason 2
c) some other mitigating factor....
Has anyone seen simmilar occurances?
Could there be another reason for the observed events?
---------------------------------------------------

That's more the way an Engineer or Scientist is meant to think.

Phil.

[EthanMiller]

I was also at the regional - It's hard to be specific about the problem because there wasn't any communication about them - an announcement about technical difficulties, followed by "We'll be back in half an hour", then two hours later, Alliance selection. Each team had only 2 rounds of play, except the first place team (which had 3 matches) and the last place team (which had 1).

It's worth noting that the communication breakdown wasn't only between the event staff and the teams - I spoke to many event people trying to figure out what was going on, and noone knew anything.

This also screwed up alliance selection - how do you pick alliances when you haven't seen everyone play? Additionally, they redid stats halfway through, so two teams who hadn't seen the updated lists had to pick without time to think about who they would pick.

Overall, though, other than the communication, Clarkson regional did do an excellent job of finishing the event, given the technical problems. Worth noting, there was a lot of testing of the equipment the night before, with no problems.

It's reminiscent of FRC with all the wireless issues with the new system.

In terms of the actual issues, sometimes communication would simply stop between the robots and the FMS, sometimes there were delays of up to five seconds, sometimes the Samantha modules wouldn't connect. Without communication, though, there wasn't a lot of seeing the what caused it.

[PSHRobotics]

The actual hardware of the Samantha's seems to be flawed - or at least the configuration system. Two of my team's sister teams had their Samantha's stop functioning after the inspection judge "downloaded" the new configuration onto them.
Also, at the St. Louis regional, there were many problems with disconnections of the Samantha. The problems were persistent throughout the competition and the referees decided to allow some robots to disconnect because "there was no considerable change to the outcome of the match". There was only one match that was replayed. Interestingly, that match had a relatively minor disconnect (the Samantha reconnected after a few seconds, while in other matches the Samantha would stay disconnected the entire time or would continuously disconnect and reconnect to the point where the robot was unusable) and was also a blowout (one alliance had a massively larger amount of points than the other).
The problems continued into the Finals, where 2 out of 3 matches had disconnected during both autonomous (stopping the autonomous program executions) and in teleop.

Bluetooth had some major flaws, but it still seemed much more consistent (at least once it finally connected to the system) than the Wifi with the samantha did.

[Andrew Schuetze]

Any possibility that the issues were related to the following experience and solution posted in this thread?

[emmell]

I'm sorry to hear about the problems in New York and St. Louis. I was the FTA for the Kell Robotics FTC Qualifier in Georgia and we saw none of those problems. This was the first qualifier for our state and so we had to flash each of the Samanthas with the configuration for the FTC_FIELD and FTC_PIT routers. Out of the 11 teams, we only had 2 problems with Samanthas in getting the flash to work properly (the FTC_FIELD worked, FTC_PIT didn't). So those two teams had to be tethered in the pits to practice. But no one had any problems communicating with the FTC_FIELD router.

We did have 2 issues during our qualification matches. One team got hit hard and Samantha dropped communications. The other, the NXT locked up and wouldn't respond to anything (we saw that happen with Bluetooth as well). Samantha stayed connected to the FCS during the lockup, so I don't know what happened there.

One thing that was great was that we didn't have any problems getting the teams back up and running for matches 2 through the end when they reported back to the field. The teams put the robots on the field and turned them on and the FCS found them right away and we were ready to start the next match. As a tournament organizer, I find that to be the best leap in progress with working with these robots with a centralized FCS. So, don't be too hard on Samantha yet.

So, a couple of tips that the teams and organizers should know:
1) Teams: Do not bring your routers to the competition. Keep the wifi to a minimum to keep it open for the competition.
2) Organizers: Make sure you set up the FTC_FIELD on one channel of the WiFi spectrum and FTC_PIT on a different one. Having them on the same channel will just confuse things.
3) Teams: Make sure your Samantha is not buried in the bowels of your robot. It's best to mount it vertically and with the red button on the top (the actual antenna is on top of the unit).
4) Teams: Get a small 1 foot USB extension cord and plug that into the Samantha unit and zip tie that securely to Samantha and the bot. You can then plug your USB cable from the NXT into the extension cord. Then, when the organizers need to flash Samantha, you can just disconnect the USB cable from the extension cord and plug the flash drive into it. Then flash Samantha and reconnect the cable. This will also relieve any stress that directly affects the Samantha unit. Also, get a small USB A-B cord (1'-1.5' feet). Secure these cables to the robot with zip ties.
5) Teams and Organizers: If you do have two routers (FTC_FIELD and FTC_PIT), make sure you wait 30-40 seconds during the flashing process to get both network's information loaded onto the Samantha. The Red, White, Blue lights should light up in succession twice (once for the FTC_FIELD and the second time for FTC_PIT). Then connect to the FTC_PIT router while in the pits to test out the bot. When you come to the field for the first time, you may have to press the red button on Samantha to sync. Then you'll be set for the day.
6) Teams: Make sure you compile and download the Samostat program to your brick. You can find it two folders under the Program Files\Samantha Field Control System folder. One for LABView and the other for ROBOTC. Pick the appropriate one for your system and install it on the brick. FTAs will use that program to see your configuration status for Samantha and make sure you can talk to the field and pit routers respectively.

I hope this will help you. Good luck to all the teams and all event organizers out there.

[jamie_1930]

I just realized something about what may have happened at Clarkson. When you went through the final phase of inspection and did the testing match, the FTA working the pit loaded the samantha wifi configuration. This configuration must have been the FTC_FIELD configuration, because it's the same configuration we used on the field. emmel said that at georgia the had a FTC_FIELD router and FTC_PIT router, meaning that if you were testing in the pit you would have to load the pit configuration and if you went onto the field you would load the field configuration. Now at Clarkson we had an area to test your robot with samantha in the pit, but never had to change configuration if we were on the field and then went to the pit and vice versa. Meaning that we had two seperate routers running the same configuration, while the matches were carefully looked over it was almost like it was free rein in the pit. If you wanted to run a test match you just need a field tech to come over and start the match. I feel like every time I came out of a match I saw people on that pit field. So what I'm trying to say is that at Clarkson we had two routers running the same configuration, and sometimes running matches at the same time, this seems exactly the kind of situation that could lead to the problems we saw.

[ssi]

Actually, the Samantha holds TWO configurations at one time. It can know how to connect both to an FTC_FIELD and an FTC_PIT at the same time. That is why they did not have to load a separate configuration for each area.

[emmell]

Originally Posted by ssi

Actually, the Samantha holds TWO configurations at one time. It can know how to connect both to an FTC_FIELD and an FTC_PIT at the same time. That is why they did not have to load a separate configuration for each area.

Exactly right. David beat me to replying.

[wilsonmw04]

Just coming back form the Richmond Regional, I have to say that Samantha worked much better than the Bluetooth solution from last year. Matches all ran smoothly throughout the day. There were 4 communication issues my team encountered today. Twice, our NXT froze in the middle of matches after sustaining a shock during normal game play. Samantha stayed connected on both occasions. Second, one of our alliance partners lost communications with the field. The cause of the disconnection was a poor power connection. The last was a miscommunication with the field staff about which controller set my team was using during an elimination round. None of these were do to the Samantha module.

[Andrew Schuetze]

the later posts in this thread seem to validate that the Samantha unit is working but we do need to learn what causes large scale crashes to improve on the best practices for tournament organizers and team behaviors. I think the post by emmell is a must read at this point for everyone either competing in or running FTC events.

[stem12bfalk]

Samantha was easier to use than Bluetooth. We had very few problems with Samantha and the problems that we did have were easy to fix. It took us a while to figure out how to set up and configure the Samantha but that was because Samantha was new and none of us knew how to use it. We followed the directions that were on the FIRST website and played around with it and we got it to work. Most of the problems we had involved the computer not connecting to the router or not using the correct templates for autonomous and teleop. The only other problem we had more than once was when the Samostat said that the IP address of the Samantha was 169.254.___. __. That problem was easy to fix all we had to do was turn the robot and the brick off and then turn them back on. If that didn’t work we unplugged and then plugged back in the router or reconfigured the Samantha. Our Samantha died at the Maryland competition for unknown reasons but my guess would be that it had something to do with being in a cold car for too long or something like that. We had problems with FCS not connecting to Samantha when the computer did. Trying to rescan for the Samantha didn’t work so we had to exit out of the FCS. We had to turn on the Samantha before we looked for it using FCS or it wouldn’t find it even if we rescanned. Better instructions for setting up the Samantha would have been helpful. The only complaints I have about Samantha are that we could only get one which meant that if it broke we were screwed and that some of the tech people at the Maryland state and Pennsylvania York qualifier competitions knew less than I did about Samantha and FCS.

[aklego]

I'm FTA'ing at our state tournament and just encountered an odd issue that is extremely successful at preventing Samantha connections. A team.gif of the wrong size on the NXT will totally foul up the works. It makes the FCS post a variety of errors.

Credit where credit is due. The kids figured it out.

[aklego]

Whew. I just finished FTA'ing our state championship. I'm sold on the concept of samantha but not the current hardware. I FTA'ed at a regional several weeks ago and there were huge differences. At the regional, NONE of the teams had used samantha before. (All were from rural Alaska and their schools were 100% Mac based, hence no FCS.) As a result, all of their modules were brand new. We plugged them in and, excepting the low battery issues, they worked perfectly.

Yesterday was quite different. (We had some interference from an existing network that was jumping channels but that was quickly resolved and the affected match was replayed. And, contrary to rumors, there was not a jammer in the building.) The biggest issue was the samantha robustness. Or, more correctly, its lack thereof. Known powerhouses that had been using their modules for months were the ones having problems. Their modules had intermittent faults. Most frustrating, unless the module was really toasted, the problems would only occur during a match. The culprit, confirmed by the onsite FIRST rep, is weak solder joints. The "Samantha Best Practices Guide" is only a bandaid, and does nothing for teams that may have inadvertently damaged their module before it was published. We also had lots of locked up NXT's and the previously mentioned guide would indicate that samantha hardware faults were again the cause. Burned out LED's were also not rare. Bluntly, samantha was beta tested on 1,645 unsuspecting teams this year. The test results are in and the samantha hardware did not pass.

FIRST would do well to publicly admit the issues sans candy coating and explain how it is going to be rectified next year. It is safe to say that teams are tired of control issues.

As an aside, one thing that would be helpful would be to provide significant realtime information to teams. The small battery icons next to team numbers on the timer display are a great start but they are hard to see, especially in the heat of battle. I would not have been adverse to having a rep from each team looking over my shoulder at the FCS screen--It would have made my job easier for teams to see the disconnects as they happen.

Having said all that, from the reports of some events, our tournament was an unqualified success. We had 40 teams, each had 5 matches and we finished qualifiers a mere 15 minutes behind schedule.

[PhilBot]

I diagnosed a related issue last night.

An intermittant Network connection issues was traced back to an intermittant power connection at the Sami module.

Turned out NOT to be the module, but the plug-in cable/connector combo.

On closer inspection the red wire seemed stressed at the entry point to the white plastic connector, and slight wire jiggling could cause the Sami to reset.

The Sami module was tested on a spare power cable (where the heck did that come from ??????) and was proven to be robust. So we replaced the cable. All was good.

We disected the "stressed" part of the suspect wire, and found NO broken conductors. We can only assume that a prior incident had caused the actual metal barrel of the power plug to become deformed (spread) enough that wire movement caused intermittant connections with the mating part inside the Sami module. VERY hard to diagnose, but causes problems whenver the robot took a hard bump.

Once again, this is a VERY rough and tumble environment and teams MUST be educated on how to tie down cables and protect connectors. I'm still amazed to see how many NXT's and Batteries fall out during a match.

As an aside... I can't imagine LEGO ever assumed that the NXT USB cable connector would see the wear and tear that FTC is now forcing it to endure. Multiple plug/unplug cycles during code testing, then it becomes a CRITICAL link for the actual competition. This is probably a $0.50 part.

Seriously !!!

Phil.