Thursday, February 28, 2008

Signal Processing, Power Bus and Power Board

The signal processing board, as you can see above in it's not-so-finished state, is in the process of being designed using Eagle Layout Editor... The yellow lines you can see up there are all the connections that need to be made between each of the little red resistor and op-amp chips and green headers...It's called a 'rat's nest' for a reason. Anyways, the board is currently in a much-more-finished state as you can see below:

The job of the signal processing board is to take the raw output signals of the sensors, temperature, humidity, incline, as well as the motor controller output (in current and voltage for velocity and torque) and process them such that the computer can make out the position and status of the robot. Additionally, the computer's D/A output is run through the board to filter it before it goes back to the motor controllers.

Eagle is sweet. It is about 40000000 x 10 ^ inf times cooler than Cadence...Though the tool panel on the side is a bit weird and scrolling is entirely counter-intuitive and the library indexing system sucks, it is incredible to be able to drag the parts around the board and have it recalculate the shortest paths for common signals for you...totally sweet.


If all goes well, we should be able to order the board by Friday. Once the board has been ordered, time will ramp up again on the power board (which changes the battery voltage and current into smaller voltages for use by the computer, radio, gps, radar, sensors, sig processing board, etc.) The power board will be a CNCed two-sided copper-board, machined entirely hopefully over the course of a half hour...woohoo!...using a design in SolidWorks...wow! Thanks Pete Fontain! It's all designed and ready to go into SolidWorks...which, by the way, is the best program ever...it's a happier, friendlier world than Pro/E. I highly recommend the experience...I once wrote a paper about the CAD experience and the creation of the universe in a class on Jewish Mysticism...yay liberal arts...

Anyways, the power bus is also designed...We're going to use screw terminal blocks to secure the battery series packs in place. We were a bit worried about adding in a heat sink for the power diodes, but we had the sweet idea of using the copper bus block itself as the heat sink, since the case of the diode works as the terminal just like the pin out. It should simplify the design since the terminal blocks won't have to be machined out anymore...It will be pretty simple really. Fuse holders still need to be ordered, however...Once the funding switch gets turned back on that order will go out ASAP.

So that's about it with electrical stuff. Electronic stuff is going better I'm told...Yay for debugging the GPS...

Wednesday, February 27, 2008

Quicky

We have Google Analytics on this blog and we've seen certain areas of the country where people seem to be very interested in the project but we have no idea who they are. Please let us know if you're one of the people that checks this blog frequently, particularly if you live in or near Cohoes, NY; or Billerica, MA. If you are interested in sponsoring the project in any way, please contact me: elas1@dartmouth.edu

Every step of the way we've encountered more obstacles but tomorrow should be the day that we get wheels on the bot and have something that rolls, even if it doesn't drive.

Drew made a breakthrough with the GPS tonight in determining why it doesn't work. He figured out that we needed a chip to convert the signal from the GPS into the standard RS232 that we need for the processor. We still have to get it communication properly but we'll be working on that in the next couple days.

Gus is laying out the signal conditioning board tonight and we'll be sending that out for manufacture by Friday.

Our electronics enclosures may be backordered for a while, so we'll have to make temporary boxes while we test our system.

Friday, February 22, 2008

Step by step

We got the chassis back from welding today. Inky did an awesome job despite his claims that he might be hungover and that his eyesight isn't what it used to be. Overall we were very impressed.

We're shooting for full mechanical integration by tuesday of next week, and we'll keep you updated as we sprint towards the deadline.




Drew testing the chassis pivot (and looking like a tool)


Thursday, February 21, 2008

You're sending signals

This past week was chock full of successes on the electronics side of the project. We were finally able to get the motor controllers to work properly after many hours of playing with the loop gain and current limit adjustments and making sure we were operating in the proper mode. We can now precisely enforce the motor velocity with a control signal from our processor board.

We were also able to get the processor to communicate with a laptop using the radio modem and a serial connection. We then got a Java GUI rigged up that was able to send characters to the serial port and communicate with the processor. Bottom line, we can drive the robot by remote control over 10km away....once we have a robot.

Mike's been having a tremendously difficult time getting the GPS to work. The company tech support doesn't even know what kind of connectors are used to connect the board to anything else, so he had to rig up a custom adapter. We can't see any signals coming through the serial port, so right now we're scrambling to get that to work, otherwise we're dead in the water. Hopefully Novatel will return our calls at some point....

Picture me Rollin

It's been a little while since the last post, we've been insanely busy trying to get Yeti's wheels on the ground. On the mechanical end, we've been held up for a week on the welding due to the repeated snowfall. Our welder has been working overtime keeping the snow-clearing equipment for Dartmouth running, so we've been nervously biting our nails as the deadline slid. The bulk of this got done today, so we should have the chassis and hub components assembled tomorrow.

We've ordered enclosures from Carlon to house the electronics on the robot. These will be arriving on Monday, so we should have most of the bot assembled by the middle of next week. This has been an interesting lesson on how to motivate subcontractors when they have no stake in the final product and you're not paying them.

Saturday, February 9, 2008

More Pictures



Tuesday, February 5, 2008

Fabrication Pictures

Chassis Components


Chassis Components and Weld fixturing


Fixturing from the side


Pivot - Bushing and pivot surface


Motor Mount plates - these get welded into the ends of the motor housing tubes.


Hubs - Half completed hubs - These are being machined by Kevin Baron for us.





Drive Axels - these connect the hubs to each gearbox shaft


Drive Axel with custom Keyseat cutting broach. We've having fun working with both metric and english units. At least our robot shouldn't crash into mars if we make an error.

Sunday, February 3, 2008

GSSI site visit

On Friday morning the team had the opportunity to head down to the Geophysical Survey Systems Inc. headquarters in Salem, NH. Gus, Drew, Kevin, and Eric delivered a presentation on the GPR Robot project to several of the engineers and marketing people down there. The presentation consisted of a background on the purpose of the robot, the technical details of its design, and what we've accomplished so far.

After the presentation, we talked with the engineers for a while on the details of integrating their radar system with our robot.
The GSSI engineers were very concerned about both radar interference from the radio and the drive motors, and the radar reflections from the robot itself. It became clear that we'll have to do extensive testing to determine the magnitude of interference and undesired radar reflections, so it's a good thing that GSSI has generously offered to loan us a radar system for testing. We'll have to test the robot under different conditions to assess the interference.

Ken suggested taking the robot up to Tuckerman's Ravine on Mt. Washington in the spring so we can test the radar on snow over 20' deep. He also suggested 'air launching' the robot to isolate the radar antenna and robot from all other surrounding matter to get a clean picture of the reflections off the robot. Needless to say we're a little bit concerned about how to safely get our robot 50' in the air, but we'll look into various suspension schemes once we get to that point.


After the meeting, Ken took us on a tour of GSSI where we saw the manufacturing and test facilities. We got a hands on demo on how to use GPR and were all quite surprised at how easy it is to use. The pictures below were all taken with a b-side camera phone so the quality isn't stellar.