Anderson PowerPole for CAN bus?

[otherguy]

Originally Posted by mbshark

We are currently trying to decide between APP 15 amp connectors and the pwm connectors for the CAN connections on the Talon SRX's. I am wondering if the APP 15 amp connectors would be frc legal as the maximum gauge wire is 20 AWG and the Talon SRX's CAN wires are 22 AWG. I would hate to have the team rewire the whole thing during a competition because of a research mistake.

The connector is rated at 15amp. You're not actually going ot be putting 15 amps through it. There's no rules preventing you from using connectors rated for higher currents than required in your application.

My guess is that the 15A connector was suggested because it's the smallest gauge connector available for the power poles. You're likely to have trouble crimping a larger connector on such small gauge wires, and if you succeed without breaking the wire, you're likely to have failures down the line.

[AllenGregoryIV]

Very interesting thread. We're going to try out the APPs on our prototype robot since that's what we have right now but if we run into any problems I like the idea of combining these two solutions.

Originally Posted by Peter Johnson

We're using Hansen Hobbies' 2-pin 0.1" LP (latching polarized) connectors (http://www.hansenhobbies.com/product...nlpconnectors/) for our CAN bus. They use different pin inserts than their other 0.1" connectors, but you can use the same crimper for both. They're a tiny bit longer when mated than a standard PWM but they're much more robust thanks to the latching feature and the polarization keeps you from accidentally swapping CANH and CANL. Basically we put a male and and a female connector on each Talon; makes daisy chaining a cinch.

Originally Posted by otherguy

During beta testing we decided to crimp the CAN talons with a three pin header with the center pin left empty. We did this on both ssets of CAN cables on the controller.

Reasons for doing it this way:

• we already had the crimpers and connectors for PWM cables.
• female headers prevent exposed contacts that could short to the chassis
• using a 3 pin connector with the center pin empty allows the controllers to be plugged in to the roboRIO maintaining support for PWM capability.
• it makes it easy to daisychain controllers together, just use a 3pin male header (see below)

The pin headers for connecting two female cables: http://www.hansenhobbies.com/product...pt1in_srfma40/

Using the 3 pin versions or the Hansen Hobbies locking polarized connectors would allow you to still use PWM if you wanted to but also prevent you from flipping polarity on accident. The 3 pin housings aren't much more expensive than the 2 pin. You can daisy chain with out much hassle or the need for special connectors. From one of the non-talon nodes you would just need a few wire pairs with just the male or female ends of the connectors attached.

[Al Skierkiewicz]

Allen,
We like the Victors and will likely use them for this season. There are a variety of other connector solutions available through outlets like Digikey (who is a big FIRST supporter in Minnesota) and others. As Don has pointed out above, solutions need to maintain buss impedance in some fashion (it is a transmission line interface after all.)

Mb,
There is relatively little current flow in this bus so it does not come under the wire gauge rules. Please remember that the those rules relate current protection to wire size. A dead short on the CAN buss will not burn up the wire.

I would use a locking connector of some type that is easy to terminate and difficult (read as 'impossible') to connect backwards. The RJ-11/45 used on the Jaguar is a pretty reliable device, crimpers are available everywhere for cheap and the connectors are easy to get. As always, I tend to discourage solid wire on moving devices but there is plenty of stranded wire options for these connectors. Spend the money for the right connectors and crimper if you want reliable operation.

[IndySam]

So is the simple solution of wire nuts not the way to go?

[Al Skierkiewicz]

I would not use wire nuts.
If you have access to DB-9 hardware, that is not a bad way to go either. Easy to terminate, crimp or solder, with standard hardware everywhere, including IDC (ribbon cable) termination.

[philso]

Originally Posted by Al Skierkiewicz

M,
There isn't a rule that specifies connectors for the CAN bus but I have to ask "why"? This is a critical part of the robot and should be treated a such. For that reason I highly recommend as few connections as possible. If you are going to use the APP connectors, be sure to use the 15 amp terminals and if your wire is very small gauge, I recommend you strip twice the length needed and fold over the exposed wire before inserting. This will give you a more reliable crimp. Don't forget the need to terminate the bus.

The CAN bus connections are critical just like many others such as the power to the RoboRio. Since the CAN bus wires are integral with the Talon SRX, teams who use a crimp to join two of them together will have to cut the wires when re-using the motor controllers, making the wires shorter each time. The wires can be soldered and then un-soldered but soldering has it's own problems. This is where teams have to learn to do a good job and learn to enforce an adequate level of Quality Control (treat them like one would treat the braking system in ones car).

The CAN Bus wires are much smaller than the wires normally used with the Anderson connectors. They may have to be folded over more than two times for the crimp to hold the wire in properly. If not done right, the wire may pop out with a slight pull later. I am also concerned with the relatively high un-mating force of the Andersons. Many people pull on the wires to un-mate connectors even though that is bad practice. With the 14-10 AWG wire normally used, the wire and crimp have a decent chance of staying together. With the small CAN wires, I can foresee people breaking the wires.

Originally Posted by dellagd

We're planning on using some crimp connectors to just crimp the wires together once we have all of the Talons mounted. Minimizes the probability of failure and we're not too concerned with having to cut the wires when were done, saying as we'll only be doing it a few times a year in a worst case scenario.

We just started using them at work and got them from Home Depot or Lowes until our regular wholesale suppliers got them in stock for us.

Originally Posted by DonRotolo

CAN really doesn't like connectors, but the Talons seem to require some sort of solution. Powerpoles are overkill and provide a relatively large impedance 'bump'; I';d recommend 0.100 headers like those used for PWM connections.

Whatever solution you choose, make SURE that CAN Polarity is absolutely correct, or you'll be unhappy with robot performance.

Some fairly expensive equipment or some very specialized 3-D field-solver software would be needed to determine, objectively, the effect of a connector on the CAN Bus impedance. In general, a connector with contacts that have dimensions and physical configuration most similar to the rest of the transmission line (the wires) will probably introduce the least disturbance to the impedance and the least corruption of the signal. Thus, the locking connectors from Hansen, or something like them, are probably the best solution. An added bonus is that the "Latching Male Housings" from Hansen seem to have a shroud so that when male pins are inserted into them, they cannot short to anything unless one pokes something into the end.

http://www.hansenhobbies.com/product.../pt1in_lp_1x3/

[philso]

Originally Posted by Al Skierkiewicz

I would not use wire nuts.
If you have access to DB-9 hardware, that is not a bad way to go either. Easy to terminate, crimp or solder, with standard hardware everywhere, including IDC (ribbon cable) termination.

The CAN Bus wires that come with the Talon SRX's are probably larger than the wires in ribbon cable and might not work with IDC DB-type connectors. There are IDC connectors for larger wires but it would take some research to find some that are readily available in the small quantities that FRC teams would want to buy (i.e. no need to order in 1000's).

It just occurred to me that we may be using some locking IDC connectors with 18AWG wires at work. I will have to have a look around when I get to work today.

[JDNovak]

We have had some interesting debates about this while trying to define the best practice for controller locations. It probably warrants it's own thread but I'm too lazy to start one. We currently have defined 8 subsystems here we might use closed loop control with Talon SRXs. The side debate is whether to distribute them close to the motor / feedback location with a longer and more vulnerable CAN bus or place them in a group with a short CAN bus and longer motor and feedback leads. I leaned toward the former but lost the debate and the 8 Talons are in a small group which my be best as a CAN wire problem takes down the whole bus where a feedback sensor wire affects one device. As a side note we have been distributing the CAN bus around our robots every year since 2010 with minimal problems after debug.

To the point of the thread, since the SRXs are close to each other, we pulled out the terminal blocks from past KITS of PARTS (I think they were donated by AutomationDirect) and used them to daisy chain through. It works well for the layout.

As for the feedback sensor wiring, I had already ordered the latching plugs from Hansen Hobbies that Peter mentioned above. I have been using them for months on my 3D printer without problems. I'm passing the hot end temperature thermistor through one that is constantly being jogged at high speed and it has caused no problem. I wouldn't be afraid to use them on the CAN bus but I use grey wire nuts on small wires all the time too.

[notmattlythgoe]

Has anybody looked around at JST connectors?

[ozrien]

CAN connector coming soon, see attached image.

[AllenGregoryIV]

Originally Posted by ozrien

CAN connector coming soon, see attached image.

any idea on pricing?

[GeeTwo]

For some authoritative documentation on connectors for CAN, see section 4.6 of:

http://www.ti.com/lit/an/slla270/slla270.pdf

Thanks to Ether for digging this link up and posting on another thread!

[ozrien]

Originally Posted by AllenGregoryIV

any idea on pricing?

http://www.vexrobotics.com/217-4429.html

[wireties]

Originally Posted by IndySam

So is the simple solution of wire nuts not the way to go?

The best connectors have a profile similar to the diameter of the wire. The reflections occur at impedance mismatches and the impedance of the connector is largely a function of the geometry. In an ideal world the wire and all the connectors would have the same impedance at 1Mhz and the terminations at each end would be 2X that impedance, nominally 50-60 ohms for CAN. Note that this is impedance, not purely resistance.

We have used PWM connectors and Molex connectors (for 20AWG wire) with zero issues running 9 Talons. Powerpole pins are overkill for the current and are larger than 20AWG wire - I would shy away from using them.

Good luck!

[Ken Streeter]

Originally Posted by ozrien

CAN connector coming soon, see attached image.

Looks like a good solution - I like it!

Personally, I'd love to see that block be part of the TalonSRX if a hardware rev gets performed in future years.

For that matter, I really wish the 10-pin "encoder connecter" were also weidmuller connectors -- the 5x2 .05" header connector is less than optimal for interfacing to directly.