WildStang 2015 Beta Testing blog

[SteveGarward]

So, one day, we got this email...

2015 Beta Testing - You've been selected!

The sound of students heads exploding could be heard from my house 35 miles away.

We're excited to be one of the beta testing teams for the new 2015 control system. We've set up a blog to document some of our experiences with the new system. As we post new information, we'll update this thread to let you know.

The blog is here: http://wildstang.org/blog/

First post is up now: First impressions (http://wildstang.org/blog/?p=108)

[SteveGarward]

Our apologies. And earlier issue with images not showing has now been fixed.

[billbo911]

Our apologies. And earlier issue with images not showing has now been fixed.

I'm still not seeing any images on either link.

I am using Chrome on Win 8.
I also tried IE under Win 8 and have the same issue.

[SteveGarward]

Ok, this time for sure! Pages my be cached, so if they're still broken, try a shift+refresh.

[SteveGarward]

Hi everyone

We've been doing some more work, setting up a benchtop testing board. We're working on a bigger update, but here's a little more info as we progress..

http://wildstang.org/blog/?p=176

[rich2202]

Thanks for the info. I look forward to updates.

A suggestion for future blogs would be:

Side by side comparison of the new/prior power distribution boards. You have already mentioned stuff like no LED to show circuit is working.

Similar type comparison when you test the pneumatic setup.

Similar type comparison to the old digital and analog boards. Same number of connections? Do they work the same? etc.

Thanks!

[SteveGarward]

Thanks for the feedback.

We honestly thought there had been a number of 'before and after' comparisons by other teams, so wanted to focus on what we had found interesting or important to note about the new system. But, we'll keep it in mind for future posts.

[JesseK]

Steve - this looks great so far! Thanks for taking this approach (pointing out what you find interesting). I've seen two BETA teams up close now and had no idea there was a cover on the PDB for the battery cables. It's one of those 'hidden in plain sight' things that's pretty nifty. There are also important nuances you're explaining which aren't always brought across in the questions asked when seeing a board in person. For example, the 5V 2A from the VRM may not be an option for custom circuits if it continues to be dedicated to the wireless.

[Ether]

I've seen two BETA teams up close now and had no idea there was a cover on the PDB for the battery cables.

See 0:28 in this (http://youtu.be/ol7BdxSHwkM?list=PLG_KOHBuXHNdUeoDkGSFQhrOJ7KvZbkB Q) YouTube video posted Aug 27th.

[JesseK]

See 0:28 in this (http://youtu.be/ol7BdxSHwkM?list=PLG_KOHBuXHNdUeoDkGSFQhrOJ7KvZbkB Q) YouTube video posted Aug 27th.




I can't read their voices while sitting here watching my wife's guitly-pleasure TV with her ;)

I'll check their videos out sometime, thanks for the link.

[SteveGarward]

Hi everyone

New post (http://wildstang.org/blog/?p=204) up today about our plans (and changes to those plans) for an MXP expansion board.

Any feedback welcome.

[SteveGarward]

It's been a while with no updates, so how about two in one day?

We've been working on a part to help with cable retention and are ready to release some details. We're still iterating over designs, but here's some information so far (http://wildstang.org/blog/?p=225).

[SteveGarward]

For example, the 5V 2A from the VRM may not be an option for custom circuits if it continues to be dedicated to the wireless.

Thanks for the feedback. For the above, the VRM has two outputs for each of the voltage/current combinations. So, two each of:
5V @ 500mA
5V @ 2A
12V @ 500mA
12V @ 2A

If you want to use the same circuit powering the wireless is up to you.

[Joe Ross]

Thanks for the feedback. For the above, the VRM has two outputs for each of the voltage/current combinations. So, two each of:
5V @ 500mA
5V @ 2A
12V @ 500mA
12V @ 2A

If you want to use the same circuit powering the wireless is up to you.

The outputs are not separate. For example, there is one 5V @ 2A supply with 2 connectors. For that reason, I speculate that nothing else will be allowed to be connected to the 5v 2A supply besides the radio, but I have no insight into the rules.

See the "Briding Ouputs of VRM for more Current possible" thread on the beta forums.

[SteveGarward]

Yes, sorry. I should have been more clear and said connectors.

Thanks Joe for clarifying.

(Late nights and long weeks make Steve something something...)

[JesseK]

As an EE in college, we lived for the blinking LEDs. If they didn't shine or blink correctly, our circuits were wrong. Even now, as a systems programmer (think Java), when I put something on an embedded system (think Vuzix glasses with a custom circuit) I still 'live and die' by the blinking of LEDs.

It is good to see we are passing on this obsession to the next generation :D

[SteveGarward]

It is good to see we are passing on this obsession to the next generation :D

On our team? You have no idea... :ahh:

[Alan Anderson]

If you want to use the same circuit powering the wireless is up to you.

A 2015 robot rule might make that decision for you.

Based on the available power from the VRM, and on the expectation of having the same wireless bridge next year as we used this year, I wouldn't count on being permitted to use the VRM 5 volt output for anything else.

[SteveGarward]

I agree, and rules trump anything I ever say.

For powering custom circuits (which have mostly been LEDs lately), we have been using a separate regulator, fed by 12V from the PD board.

[The Lucas]

I like your Cable Magic post. I was hoping one of the beta teams would print something like that, since the DIO connections no longer have a lock.
I have a suggestion to possibly improve the DIO cabling further by using Latching Polarized (LP) Connector Housings like these 1x3 Female Housings. (http://www.hansenhobbies.com/products/connectors/pt1inlpconnectors/pt1in_lp_1x3/) At first glance, it seems that with a few cutouts for the latch and the notch, Cable Magic could be modified to accept both LP and standard PWM connectors. The question really is if the resulting part would be rigid enough with less plastic, and withstand repeated latching.

I have used the LP connectors in prior years for intermediate sensor and solenoid connections. They have done a good job of preventing accidental disconnects and reverse connections (we had a really bad year with electrical tape covered PWM intermediate connections so we decided to upgrade). I really like the prospect of individually latching DIO channels on the RoboRIO, not sure if it is feasible until I get a test unit. Any insights you and the other beta teams could provide would be much appreciated.

[SteveGarward]

The software subteam has just published a blog post (http://wildstang.org/blog/?p=348) on their initial work with the Java platform on the roboRIO.

Three in a week? Somebody stop us!

[SteveGarward]

A quick update - we made a mistake and posted an incorrect board design for one of the MXP expansion boards. The 'original' board design was an older version - please see the update on our blog (http://wildstang.org/blog/?p=364) if you were looking at using this board design.

[JesseK]

Can you ssh to the RoboRio?

[Greg McKaskle]

Yes. ssh, sftp, webdav and the web services are the primary tools for viewing and modifying roboRIO settings. There will still be a formatting/imaging tool for the basics.

Greg McKaskle

[JesseK]

Cool, thanks Greg!

[billbo911]

A quick update - we made a mistake and posted an incorrect board design for one of the MXP expansion boards. The 'original' board design was an older version - please see the update on our blog (http://wildstang.org/blog/?p=364) if you were looking at using this board design.

What is the value of "R1"? I assume it is for current limiting of the +5vdc.
What is the value of the filter cap "C1"?

Wait, never mind. It looks like that is just a power indicator LED with current limiting resistor.

[SteveGarward]

Wait, never mind. It looks like that is just a power indicator LED with current limiting resistor.

Correct - this is just an LED using the 5v DC from the MXP to know it is connected. The resistor R1 is just a current limiting resistor for the LED. Value should be around 150R for an average 2v @ 20mA LED.

We changed from surface mount to through hole at the last minute for this set of boards (since there was only 1 SMT part left..), and didn't notice we had missed the value on the schematic.

[billbo911]

Can I make a suggestion for the next Rev. of your MXP Expansion board?

As it is designed right now, it doesn't appear to have any method of securing it to the RoboRio other than the MXP connector it's self.

How simple, and cost effective, would it be to lengthen the board so that it overlap the two nutserts in the top RoboRio? Then just have holes in the board line up with the nutserts. That way you would have a securing method on the other end of the board from the connector.

[Mark McLeod]

Did you mean this version (http://wildstang.org/blog/?p=204)?

‘The secure’

http://wildstang.org/blog/wp-content/uploads/2014/10/MXP-mounting-276x300.png (http://wildstang.org/blog/wp-content/uploads/2014/10/MXP-mounting.png)
This board layout is larger to span from the MXP connector to the two screw holes for mounting at the top. This allows it to be firmly secured to the roboRIO.
Download the board here (http://wildstang.org/blog/files/expboard-passive-mounting_holes.brd).

[SteveGarward]

Thanks Mark, beat me to it.

The only downside to this board is the additional cost of empty space. Probably worth it for security.

[billbo911]

Did you meant this version (http://wildstang.org/blog/?p=204)?

Yep, that'll do it! How I missed that, I am not sure. Thanks.

Thanks Mark, beat me to it.

The only downside to this board is the additional cost of empty space. Probably worth it for security.

That extra space can be utilized for many things, including security. You may find a need to add additional custom circuitry, this space would allow for that. The security alone will be worth every extra penny spent!

[SteveGarward]

That extra space can be utilized for many things, including security. You may find a need to add additional custom circuitry, this space would allow for that. The security alone will be worth every extra penny spent!

After yesterday's update on the forums here (http://forums.usfirst.org/showthread.php?23689-Active-or-passive-MXP-board), be careful what custom circuitry you add. Looks like most things would classify it as an active device.

[Greg McKaskle]

Out of curiosity, just how much does security cost? Or do you have price estimates for the different versions?

Greg McKaskle

[SteveGarward]

The board that spans the roboRIO to reach the screw holes comes to about 3" x 3". We were going to use oshpark.com again as we have had good results from them in the past. They charge $5/sq in for 3 copies, which comes to about $15 per board. Add header pins, and it's not much more than that, depending on how you buy.

The board could be made narrower - we were sticking with the footprint of our other board.

My only issue is that for a passive board with only breakout pins, there is a lot of empty space.

The mini version is a total of around 6sq in, so about $30 for 3. So I guess security costs about $5 per board. :)

[Greg McKaskle]

Good to know. Thanks.

Greg McKaskle

[Michael Blake]

Hello... thanks for publishing such an informative blog!

3481 is one of the Texas Beta Teams. Have you guys seen this new REV board, it looks like it does everything you guys are talking about and will be available soon through AndyMark?

https://www.facebook.com/REVRobotics

--Michael Blake

[SteveGarward]

Here's an update on our original MXP board design (http://wildstang.org/blog/?p=376), and what we've decided to do with it. Short version, we not aiming to get it approved by FIRST at this point. Too many other fun projects and things going on for us to be able to work out inventory and vendors etc.

[billbo911]

I noticed while reading this (http://www.ni.com/pdf/manuals/376099a.pdf) document that the RoboRio connects to a Host computer via USB. (Pg. 5)

Does this mean that the RoboRio is programmed via USB and not ethernet?
Can we still "Run" code before building it and the deploying it? (I certainly hope so!)
Can we still "Run" and deploy code wirelessly? (Again, I hope so! I've seen one too many laptops go flying.)

(Sorry, I haven't seen this mentioned yet or I just haven't looked hard enough.)

[Joe Ross]

I noticed while reading this (http://www.ni.com/pdf/manuals/376099a.pdf) document that the RoboRio connects to a Host computer via USB. (Pg. 5)

The USB is an additional network interface, in addition to ethernet and wireless. You can program over any of them. I think your questions are LabVIEW specific. The LabVIEW process is exactly like before with the exception that reboots are no longer necessary when running code as startup.

[Mark McLeod]

I noticed while reading this (http://www.ni.com/pdf/manuals/376099a.pdf) document that the RoboRio connects to a Host computer via USB. (Pg. 5)

Does this mean that the RoboRio is programmed via USB and not ethernet?
Can we still "Run" code before building it and the deploying it? (I certainly hope so!)
Can we still "Run" and deploy code wirelessly? (Again, I hope so! I've seen one too many laptops go flying.)

(Sorry, I haven't seen this mentioned yet or I just haven't looked hard enough.)

Program via USB or Ethernet or wirelessly.
The USB is recommended for imaging the roboRIO, but once it's imaged, either connection method will work.

Yes, we can still "run" debug mode before downloading the code permanently.

Yes, we can do all that wireless, via Ethernet, or via USB.

[SteveGarward]

We have deployed code via USB, Ethernet and over WiFi. So far, all is well.

[Greg McKaskle]

Just to pile on. The roboRIO identifies itself to USB such that an interface driver is loaded on Windows. The driver will bridge ethernet over USB, allowing all of the other programming tools, from cmd line to Eclipse or LV IDEs to work transparently using UDP, TCP, ICMP, HTTP, or anything that works over ethernet.

The USB interface will always be on the 172.22.11.xx subnet, just in case you were wondering.

Greg McKaskle

[SteveGarward]

Greg: Knowing that it's using mDNS, is there a defined/standard set of IPs that it tries to assign to the common parts (eg, roboRIO, driverstation)? I guess really only the roboRIO is common now..

This relates to... We were looking at having an offboard unit connected via ethernet for logging direct to SD card, among other things. If the offboard unit is listening for logging over UDP, we need to know its IP address. Ultimately, we need to know how likely it is to get a new address. We can always set it in the bridge as a defined IP, but don't want to cause any conflicts.

If that makes no sense, let me know. :) (It's been a long day..)

[billbo911]

Thanks everyone for the replies!
Wow, it sounds like a lot of the connection and deployment issues may be addressed and resolved. Awesome!

... I think your questions are LabVIEW specific. The LabVIEW process is exactly like before with the exception that reboots are no longer necessary when running code as startup.

Joe,
You are absolutely correct. I hadn't considered, or realized, I was looking at this from just one point of view. We have always used LV up until this year. We are now moving to Java with the help of a new mentor who does programming for Oracle.

[Joe Ross]

You are absolutely correct. I hadn't considered, or realized, I was looking at this from just one point of view. We have always used LV up until this year. We are now moving to Java with the help of a new mentor who does programming for Oracle.

C++ and Java don't have the distinction run and build/run as startup. This is true for the cRIO and the roboRIO. However, a build does not take very long and occurs as part of the deploy process. New for the roboRIO is that a reboot is not necessary for a C++ or Java deploy.

[Greg McKaskle]

mDNS works independent of DHCP or static IPs. It uses a device or service name to locate the IPs of the devices you are looking for.

The DHCP servers, to this point do not do anything special to allocate IPs. That is, of course possible, and you are also able to statically assign IPs that are in the subnet that the DHCP server will never hand out.

I would not recommend taking a guess at what the DHCP server handed out based on order of request.

By the way, the recommendations for this are still in work, so sorry if I'm being vague.

Greg McKaskle

[SteveGarward]

Beta testing update: We've done nothing!

Well ok, we've done a lot, but not much beta testing for a while now. Some of the reasons can be found in our latest blog post (http://wildstang.org/blog/?p=466).

[SteveGarward]

A while ago, we posted about our 'Cable Magic' to help hold the PWM/IO cables in place with the new RoboRIO.

Today, we have released the CAD files (http://wildstang.org/blog/?p=472) for this. If you make it and try it out, please let us know how you go.

[adciv]

Could you post a JPEG/PNG picture on your site? Not all of us have the cad programs installed...

[SteveGarward]

Sorry, there were photos in the original post (http://wildstang.org/blog/?p=225) about it, so I didn't put any up in this one.

Here is one image though:
http://wildstang.org/blog/wp-content/uploads/2014/10/cable_magic_v3-300x205.png

[adciv]

Thank you. I'm interested in this, we'll have to see if we can print one ourselves.

[SteveGarward]

For those of you at home still playing along, we have released a video overview of the 2015 Control System components and some of our findings over here (https://www.youtube.com/watch?v=XQRCMeyuPs0&list=UUqm4SpmkRy7NgaAC-z1Twww).

Please remember, this is based on the beta system, and things may change in the final system released to teams.

[SteveGarward]

Way back when beta testing was still a thing, we created a couple of boards for the MXP. In December/January, we finally had time and space to build up one of each, and this week we managed to fire them up for a quick test.

For completeness, we have posted some pictures of the final boards on our blog here (http://wildstang.org/blog/?p=558).