COMM lost way to many times

[philso]

Originally Posted by AllenGregoryIV

One of the other things that hasn't been metioned is the use of ferrules with the weildmuller connectors. We used them on nearly all of our competition robot connections and have had zero issues.

Here is the white paper from Weildmuller recommending their use with stranded wires.

We import ours from China using Aliexpress but you can also get them from US electronics vendors, ebay, or Amazon.

Make sure you are using the right crimp tool and gauge ferrules. Also get 8mm long ones.

We got our ferrules from www.ferrulesdirect.com and our crimper from amazon. We found that installing ferrules on our 16AWG wires (i.e. from the compressor) made them too large to work with the Weidmuller connectors.

The cable that came in our KOP had a right-angle plug on it. Once it was plugged in, wrapping a long cable-tie around the radio and the plug worked well to keep it plugged in very securely. We were able to pass the cable-tie under the radio, between it's "feet", so it could not slip off.

[waialua359]

Originally Posted by JDNovak

I believe you are on to something. Static problems can be so elusive because changes in humidity and surface contact materials are so difficult to track.I can relate one situation where static buildup definitely was the culprit and why it may relate to your problems.

We have used dozens of UsDigital MA3 absolute encoders for steering feedback over years so we know their characteristics pretty well. We have also used aluminum AM Performance wheels that contact the carpet even with tread attached. One off season we decided to replace the performance wheels with AM Plaction wheels for a demonstration robot. Within a couple of hours of run time, we had two encoder failures. The outputs would start to jitter and eventually get too bad to use. After realizing the wheels were Polycarbonate and probably isolating the chassis to allow static buildup, I attached a chain to drag the ground and never had another problem.

Many people are running pneumatic wheels this year without problems but most tires have additives that are conductive. Looking at the beach tires, they don't appear to have additives that other tires have. Is it possible that they are insulating enough to allow a charge to build on your robot? We have also had more static problems when running belts. They tend to create static like a Van De Graaff generator.

I'm not sure if a chain is the best idea this year with the defenses but something to reduce the potential between the carpet and the frame might avoid the big discharge when you touch the metal defense. If you can find a static gun some of this can be more scientifically proven.

For those that might say that the control system is immune to static, after working with plastic extrusion for 30+ years I have seen enough electronic equipment affected and destroyed from static buildup to believe that nothing is immune. Some designs are more tolerant but at some point a large enough discharge will cause damage. Also, many devices will continue to operate after being zapped but are degraded and prone to erratic performance and failure until replaced.

Mr. Novak,
thank you for your expert opinion.
I think we definitely need to look into this.
We lost COM way too many times at only our Hawaii regional.
It never happened after a hard bump or hard defense where we would suspect cables being loose. We are quite certain it is not our connections.
Hawaii has VERY high humidity.
The thing that caught my attention is getting shocked by static electricity when touching the robot this past weekend.

[philso]

Originally Posted by waialua359

Hawaii has VERY high humidity.
The thing that caught my attention is getting shocked by static electricity when touching the robot this past weekend.

High humidity and static electricity are usually mutually exclusive. One method used in industry to control static electricity is to humidify the air. I have not had static problems at work since moving from Eastern Canada to Houston but then our equipment does not have so many rotating parts. It is possible that the pneumatic tires used by many teams this year may generate much more static charge than other types of wheels used more commonly in previous years.

Try wiping some fabric softener sheets all over the tires. They leave a mildly conductive film that will help dissipate the static charge that accumulates on the tires. You would want to wipe all the way to the shafts or metal hubs that are in contact with the shafts. Hopefully, this will dissipate and equalize the charge over the whole robot so there cannot be an electro-static discharge from one part of the robot to another. It may be necessary to re-apply the fabric softener periodically since it may rub off. Alternatively, there are industrial anti-static coatings and sprays such as Staticide, but they are probably harder to get than the fabric softener. I also just found these instructions on how to make your own anti-static spray.

Perhaps CTRE can add some clamping devices such as TVS' across the inputs and outputs of the PDP and VRM. They would add only a dollar, or less, to the part costs though they would need some board space. The wiring in our robots is short enough that a TVS in the PDP would probably limit the voltages in the whole circuit running to the radio quite effectively.

[waialua359]

Originally Posted by philso

High humidity and static electricity are usually mutually exclusive. One method used in industry to control static electricity is to humidify the air. I have not had static problems at work since moving from Eastern Canada to Houston but then our equipment does not have so many rotating parts. It is possible that the pneumatic tires used by many teams this year may generate much more static charge than other types of wheels used more commonly in previous years.

Try wiping some fabric softener sheets all over the tires. They leave a mildly conductive film that will help dissipate the static charge that accumulates on the tires. You would want to wipe all the way to the shafts or metal hubs that are in contact with the shafts. Hopefully, this will dissipate and equalize the charge over the whole robot so there cannot be an electro-static discharge from one part of the robot to another. It may be necessary to re-apply the fabric softener periodically since it may rub off. Alternatively, there are industrial anti-static coatings and sprays such as Staticide, but they are probably harder to get than the fabric softener. I also just found these instructions on how to make your own anti-static spray.

Perhaps CTRE can add some clamping devices such as TVS' across the inputs and outputs of the PDP and VRM. They would add only a dollar, or less, to the part costs though they would need some board space. The wiring in our robots is short enough that a TVS in the PDP would probably limit the voltages in the whole circuit running to the radio quite effectively.

Thanks for the insight and links. We will definitely look into this for Championships.

[spacepenguine]

This is a great discussion on static electricity. Supposedly the robot frame is isolated from the electronics, but that doesn't mean it can't cause problems.

I only saw a small mention of VRM wiring, but suspect some teams may be hurting themselves by pairing up the radio and LEDs/lights on the 12v/2a bus of the VRM. There is a rule and checkbox on the inspection form to prevent this, but I'm not sure it was consistently applied across districts and regionals. We're all volunteers, after all.

From non-FIRST experience I would hot glue or permanently fix in anything with a barrel connector. They like to work their way out with vibration. We've seen this a lot with CMU's RoboBuggy.

[KrazyCarl92]

Team 5811 experienced issues with both radio and RoboRIO reboots at the Greater Pittsburgh Regional in Week 2. In about a third to half of our matches, we would be dead on the field for some period of time. We made an action plan for our Week 5 event at QCR to mitigate these issues, assuming that power connections to the concerned components was the root cause.

1. We had been using ferrule crimps on our RoboRIO power connection. After being eliminated in Pittsburgh, I had the opportunity to thoroughly check Team 3138's beautiful wiring where I noticed that they were using ferrules on every single connection EXCEPT the RoboRIO power connection. They stated that the clasp that screws in from the side only gets a single point of contact with a ferrule connector (due to the way the ferrule connector does not seat nicely within the RoboRIO), but that bare stranded wire actually provides a superior connection in this instance. Easy enough: we would trim our ferrule crimp off those wires, trim, then strip to establish a better RoboRIO power connection.

2. We noticed at Pittsburgh that the Radio barrel connector for power was at best a "janky" connection, causing us to lose radio power every so often during normal play. To mitigate this failure, we decided to try Power Over Ethernet (PoE) as supported by the Radio. We prepared a wire with the appropriate pinouts labeled prior to the event. Unfortunately, the wiring diagram from Team 3015 that we were using was vague, and used both solid and dashed lines to indicate which pins should be powered on the Ethernet cable. We decided to try the solid lines only first for fear of blowing out the radio (12V power to pin 4, Ground to pin 7 from the VRM 2A circuit). Voila! it worked in the pits and the radio was turning on! Unfortunately, we went to the field for the very first practice match and the robot would not connect. We then decided to use the dashed lines in the wiring diagram as well to test out how that would work (12V power to pins 4 & 5, Ground to pins 7 & 8 from the VRM 2A circuit). This did not lead to better results, and upon testing it turned out that the pins on the Ethernet connector and the wire itself for pin 5 did not have continuity (not sure why, faulty cable I suppose). At this point, we didn't have extra Ethernet cables to spare and didn't want to waste other teams' cables testing, so we decided to take another approach...

3. At Greater Pittsburgh we had attempted hot gluing the barrel connector to the radio on Saturday. This seemed to work! We had no further failures of radio connections on the field (all other failures were RIO failures, per the driver station logs). However, when we unbagged the robot at QCR, the hot glue had worn out and the barrel connector was loose again. After we did not achieve success with the PoE solution (lack of resources, not to say the solution itself doesn't work), we decided to go back to the barrel connector for radio power and attempt using a rubber band instead of hot glue to hold in the barrel connector securely. This rubber band wrapped around the back of the barrel connector and the opposite side of the radio to hold the barrel connector in place. I was at first nervous about the extra forces and potential bending loads on the connector, but this seemed to do the trick.

We are happy to report that after successful implementation of solutions 1 & 3, our robot had zero connection failures at QCR! This allowed us to successfully breach the outerworks in every single match, and increase our boulder scoring output.

We've learned a lot through this experience that we will hopefully be able to put to good use to help other teams keep their robots alive during their matches. I would almost even suggest implementing these as inspection standards so teams can only pass inspection with these solutions implemented.

TL;DR: We love ferrule connectors, and they belong on just about every other possible connection except the power on the RoboRIO; bare wire is best there. The barrel connectors that come in the KOP for radio power are ill-suited to the rigors of FRC (especially this year's game); use PoE, hot glue, or a rubber band to mitigate the risk of your radio losing power. It is important that teams be aware of solutions that may help their robots stay alive.

[phargo#1018]

For what it's worth, I did some checking.

The VRM is specified to operate down to 5 Volts. It is also specified to provide between 11.93 and 12.49 Volts, reference CTR's web site.

I thought it might be wise to contact the radio supplier, Open-Mesh. First off, a giant thank-you to Open-Mesh for talking. OM's standard Voltage range is 12 to 24 Volts at 1 Amp. The contact I had at OM indicated that he believes that the radio will operate properly down to 11.8 Volts. He also indicated that the radio should operate properly down to 11 or 11.5 Volts if the output power were reduced.

I will confess, I do not know what setting FRC is using for these devices; I am assuming that they are using the default power. Does anyone know?

I agree with the posts about the barrel connectors, ferrules too!!

If you find your radio is re-setting, you might check the output Voltage on your VRM. Is it less than 12 Volts? If so your device meets the specification requirements, but you might want to use some caution. You should be okay down to 11.8 Volts.

The 11.8 Volts applies at the receiving end after all IR drops are considered.
What wire are you using between the VRM and the radio, 20 AWG, 22 AWG? How long are your wires?

Suppose that you are producing 11.93 Volts at the VRM and you have 3 feet of 22 AWG wire (18 inches out and 18 inches back) and the radio is drawing 1 Amp. 22 AWG wire will account for an 0.050 Volt drop (0.030 @ 20 AWG, 0.077 @ 24 AWG). Any additional resistance due to wire termination at either end will increase this drop.

Bottom line, two suggestions:
1) Know what Voltage your VRM is providing, and
2) Keep the VRM and the radio close to each other!

Hope this helps, I am interested in follow up information!

Thanks for reading!

[philso]

Originally Posted by KrazyCarl92

TL;DR: We love ferrule connectors, and they belong on just about every other possible connection except the power on the RoboRIO; bare wire is best there. The barrel connectors that come in the KOP for radio power are ill-suited to the rigors of FRC (especially this year's game); use PoE, hot glue, or a rubber band to mitigate the risk of your radio losing power. It is important that teams be aware of solutions that may help their robots stay alive.

I would strongly suggest that you leave the wires a bit long if you are not using ferrules for connections where you may want to re-do the connection at some point such as the RoboRio power. The strands of the wires tend to fray and you risk shorting to adjacent wires at the Wiedmuller connectors. The wires are only 3mm apart, center-to-center. The best way to fix wire strands that have frayed is to cut the exposed conductor off and re-strip to the appropriate length. To be able to do this, you will need some extra wire length.

I think it would be easier for people to understand how you use a rubber band to hold the radio power plug on the radio if you post a picture. It is a great idea if am reading you right but it is tricky to envision how the rubber band holds the plug. I am assuming that you have a straight plug.

[philso]

Originally Posted by waialua359

Thanks for the insight and links. We will definitely look into this for Championships.

I hope it helps. I will be cheering for you and your team at Championship. It would be great to see you all back in Texas again.

Originally Posted by spacepenguine

Supposedly the robot frame is isolated from the electronics, but that doesn't mean it can't cause problems.

You are correct. The isolation of the electronics from the frame is checked with a hand-held meter set to measure Ohms. The highest voltage that such a test would apply between the electronics and the chassis would be 9V or less since these meters are typically powered with 9V batteries or some number of AA's (I have never seen more than 6). The insulation or air gap needed to prevent 9V from arcing across is pretty small (a fraction of a mm). Electronic components rated to resist static-discharges typically have ratings of at least 1500V, often 2500V. Any half decent static charge would have no trouble arcing over the small air gap (or punching through the thin insulation) needed to satisfy the robot isolation test.

[Al Skierkiewicz]

Originally Posted by KrazyCarl92

1. We had been using ferrule crimps on our RoboRIO power connection. After being eliminated in Pittsburgh, I had the opportunity to thoroughly check Team 3138's beautiful wiring where I noticed that they were using ferrules on every single connection EXCEPT the RoboRIO power connection. They stated that the clasp that screws in from the side only gets a single point of contact with a ferrule connector (due to the way the ferrule connector does not seat nicely within the RoboRIO), but that bare stranded wire actually provides a superior connection in this instance. Easy enough: we would trim our ferrule crimp off those wires, trim, then strip to establish a better RoboRIO power connection.

Carl,
Can you explain this a little better? What clasp/screws are you referring to? It sounds like the brown connector on the RoboRio. If that is the case, I believe that connector does not have the screw contacting the wire directly. Instead it compresses an open contact inside the connector.

As I stated above, please use ferrules that are designed for the termination you are using. The PDP ferrules should be square in cross section to maximize contact and minimize resistance. If you use round (or solid wire) then WAGO states you must derate the contact by at least two wire sizes. A #12 will actually perform as a #16 due to the lack of contact area.

[31Josh61]

Hi,
Team 3161 had the same problem at the waterloo regional. we thought it was about shock and moving the router worked but it think this is becoming a problem. In the finals of waterloo. 1114 lost there comms 2 times.

[IKE]

Originally Posted by Alan Anderson

Why is everyone missing the obvious solution? Use passive Power Over Ethernet (POE) on the port that supports it. The Ethernet connector snaps firmly into place. I've never seen one come loose due to bounces or vibration this year.


...snip....

I have now seen a handful of Ethernet cable failures this season. Several at the laptop (maybe 8 failures of 5 events x 40 robots). About 50% of those were classmates. I have also observed a few "bad" Ethernet cables between the robot and rio (approximately 4-6).

One item that tends to mask the poor connection issue is using a "printer cable" to directly connect into the rio during inspection. While handy, I will be even stronger in my wording with teams that they likely have a cable/connector issue they should look into in those cases.

The majority of failures have been the other "power" related failures noted by many others.

[Alan Anderson]

Originally Posted by IKE

I have now seen a handful of Ethernet cable failures this season.

I saw a few bad Ethernet cables myself. But in no case did the connector ever come out on its own. Unlike the ill-fitting Jaguar CAN connectors, the OM5P-AN's RJ-45 receptacle seems quite reliable. If the cable is good, there shouldn't be a problem.

[IKE]

Originally Posted by Alan Anderson

I saw a few bad Ethernet cables myself. But in no case did the connector ever come out on its own. Unlike the ill-fitting Jaguar CAN connectors, the OM5P-AN's RJ-45 receptacle seems quite reliable. If the cable is good, there shouldn't be a problem.

What was scarier for me was cable staying connected in many cases, but causing intermittent connectivity. Intermittent are often the hardest to diagnose/detect and provide correction for.

[sgeckler]

Originally Posted by Al Skierkiewicz

The ferrules for use with the PDP are not round.

Al, I don't see any mention of ferrules in the 2016 control system wiring documentation I can find. Not the basic and not in the PDP user guide from CTR. Is there something more specific for the PDP that describes the "proper" shape of ferrules you could use? I assume all Wiedmuller would take round? - Sam

Ferrules not mentioned here:
http://www.ctr-electronics.com/PDP%2...7s%20Guide.pdf

Or here:
https://wpilib.screenstepslive.com/s...control-system