105 replies to this topic

[RCT2123]

Shame to hear about your power supply. Nice equipment though. Keep the pics coming!

[ThisIsGabe]

http://i.imgur.com/6sNCFZU.png
Update time once again!
Top left - these are my VFD capacitors. Typically you use these to try and increase the output voltage at moderate to high frequencies. They're 4700uF caps, but they will work at up to 100 volts. Additionally that voltage is enough to knock you into the hospital, so this goes onto the list of things to look out for if someone tries to replicate this.
Top right - Motor controllers. I wanted to try building an H-bridge, more because I was bored last friday night than anything else. In the final version I'll be using these TA7291SG(O,J) Toshiba Semiconductor and Storage | TA7291SGOJ-ND | DigiKey , which have all the components of a traditional H-bridge but with voltage and temperature sensing (which would have been handy, I forgot a pull up resistor on one of the transistors and just about burned the markings into my thumb when I grabbed it to see if it was loose) Still need to make/design control circuits for the solenoids, but that's not really a huge rush. Chinese new year has shut down DirtyPCBs' factory, so they'll be down for the count till the 27th.
Bottom - hopefully what shall be the final revision of the actual car's connection to the track. It's come a long way! This one not only has 6 bearings to act as the wheels, but two smaller ones on each holder to allow the wheels to rotate side to side as well as up and down. I have to wait a little to 3d print then machine the parts, just because it's extremely cold in the 3d printer lab, and this causes all of the 3d printed parts to lift off the build platform and warp.

I've been programming for the PLC! Redid nearly all of the I/O and track sensor decisions to keep them manageable, and so far I have E-stop, gates and block routines done. I did this while mooching snacks at a superbowl party xD This weekend I want to try testing out the PLC and interface converter to see if I can make everything play nice. I finally have all the right cables (fingers crossed) so we should be fine. Only problem I can see is my serial cable and the USB to serial card I have to use in order to connect it to my laptop. Only one way to find out!

I also made a few phone calls and emails, and there might be a way to fix the power supply without tearing it out, apparently with my model of supply shorting the outputs to a secondary connector on the outside will cause the controllers to all reset. Huge props to Unipower Corp. for taking the time to help me with a problem on one of their legacy products that they aren't getting any money out of.

More to come!

[ThisIsGabe]

Superfast update everyone!
https://www.dropbox....rogram.ino?dl=0 <- click on that link and you can see the current version of my VFD code. You'll have to open it in the arduino programming library or with notepad++ (which everyone should get because it's amazing Notepad++ Home ). It's got pin assignments, ramblings and some ASCII art (because it's fun and keeps me organized)
https://www.dropbox....ogram.docx?dl=0 <- this is my PLC program "layout design thingy" basically a planning document for the PLC code. It goes with this I/O spreadsheet here -> https://www.dropbox....LC IO.xlsx?dl=0

Tried talking to the PLC today, it didn't respond. I think it's my serial cable which has the Tx and Rx swapped. I got it for free from my school's tech lab, so I'll just make another regular cable sometime next week.


More to come! Check the dropbox links if you want (quite literally) an instantaneous update. (if I forget to post here for another three weeks again) xD

Edited by ThisIsGabe, 15 February 2015 - 04:09 PM.

[ThisIsGabe]

Much like the last update, this one is fast, but I can measure how fast it is! *cymbal crash*
http://i.imgur.com/KzrQilE.jpg
This is a velocity sensor using two of the photointerrupters, 16 of these pairs will be used on the launch. The speed for these should be high enough, but all I had to test the speed was a piece of paper to swing through them, so I wasn't able to get up into multiple meters/second
This sensor array took much longer than I want to admit, because of the need for multiple input triggers I spent a while trying to get the sensors to only trip once. They had the nasty habit of constantly triggering the arduino's input and telling me the velocity had gone off to infinity. Took about 50 lines of code to smash all the bugs out!

More to come!

[ThisIsGabe]

Midnight Caffeine Fueled Update time!
http://i.imgur.com/CcuX0KB.png Climb the ladder with me!
On the right, miscellaneous documents to help me keep the I/O matched up. On the left, the PLC code for the Gates, Estop reset and one manual mode command. The brackets ] [ are "readers" which allow the signal to continue if the state is closed, ]/[ is if open. ( ) will power on the output specified, ( L ) latches the output or bit high, and (U ) latches the output or bit off.
To read the ladder program, you start on the left and move down, and at each rung, you move left to right. If all the conditions are met, then the outputs are turned on (or off), and if one of the conditions is not met, the rung is ignored.

Tomorrow I think I'm just going to binge program and try (emphasis on try) to bang out the complete program by the end of the day. Break out the coffee! ;)
More to come!

[ThisIsGabe]

Midnight caffeine fueled update two!
https://www.dropbox....rogram V1.0.xps The first version of my ladder logic program!
There are some bugs left, some more sensing that I want to add, and some loops that need to be refined (using latches and unlatches isn't great practice), but at least this gives you an idea of what the final version will look like!
Going to try and iron out the rest of the bugs tomorrow
More to come!

[ThisIsGabe]

Well, I finished the PLC program, but I hit a bit of a snag. Originally I had a cheap chinese USB to Serial adapter, but for some reason I was unable to communicate with the PLC. After taking the thing apart, I found that it was very cheaply constructed, using no FTDI or similar chip to adapt the two, and only the Rx and Tx lines are connected. Unfortunately, the PLC's communication requires full serial port capability, so I bought a new driver that does have all of the serial pins working. Serial port - Wikipedia, the free encyclopedia (DB-9 is the pinout we want)
The adapter should get here on Tuesday, so I'll try and get it working
It finally got warm enough in the 3d printer lab for prints to come out properly, so I'm going to try and get in car prints, which will let me try and get it ready for the final machined version.

More to come!

[RCT2123]

I missed like 4 updates so this was really fun to read through. Things are getting really cool now. Can't imagine how much you are learning.

[ThisIsGabe]

I missed like 4 updates so this was really fun to read through. Things are getting really cool now. Can't imagine how much you are learning.

Thanks! I'm learning a ton from all this, especially with working on PLC programming.
I hit another peak and sag last night around 1 am - while I am able to speak with the interface converter, apparently the AIC is not the final piece of the puzzle to talk to the PLC. >:I
What I need to set up is this: USB -> RS232 -> RS232 Cable -> 1761 NET AIC -> Pheonix Cable -> 1747 NET AIC -> Wonky ethernet/flattened ethernet cable -> SLC 5/01 PLC
If it seems convoluted, that's because it is, and the confusing way the communication standard for DH485 is set up has been giving me massive headaches. However, it's still cheaper than buying an 80 dollar USB -> SLC 5/01 adapter! Also this way is more likely to be compatible with my raspberry pi.
So we'll find out on Tuesday when the new parts arrive if this fancy setup will work properly.

I also was finally able to plug in my mega power supply...and it didn't work. Booooooo
So the darned thing is in pieces being inspected. I've been narrowing down faulty components since tuesday. So far I'm sure that the output side of the switching transformers is working properly, and that three of the power diodes are potentially faulty. I also know that the controller's 15 and 5 volt rails are good, which probably means the controller didn't get blown up (hopefully). The front panel LED also works, and the fan runs very well (it's got metal blades and is about 3 times the weight of a PC fan!) This involved extremely dangerous poking around 300 volt capacitors while still plugged into a 240 volt wall outlet. Don't try this at home! Once I get the controller working again I'll post some photos of the recovery process.
I also was able to get a reasonably good print of the magnet holders, and they work nearly perfectly! Just need to print out the rest of the car to test the sizes, then I can move on to machining the car parts out!

More to come

[ThisIsGabe]

*lightning rumble*
I....LIIIIIIIVE!
Exams suck. I've been slammed with school, and one of the busiest 2 months of my life. But now here we are~
First order of business - I've finished (literally 20 minutes ago) the VFD design. http://i.imgur.com/IcHKIJc.png LOOK HOW FANCY THAT SILKSCREEN AND COPPER LOGO IS
30 photo interrupter inputs (with indicator LEDs), an H bridge and solenoid driver (to allow for a drive tire on the launch), resistor voltage divider for the three phase motors, and the bus voltage, communication to my SLC PLC (more on that in a sec), and 5 rollback relays. BUT.
I'm not going to be sending that PCB off to the fab house juuuust yet~ I've got another project - building a custom ATX 1kW power supply - that could use some board space, and I still have about 30 square inches left on the board. I hope to finish the design on that in the next week or two, then I can bang them out to the fab house. Besides, I dont really have a good soldering iron with me so I can't do much with the PCB even if I got it right now
Second thing - my PLC program. I've been getting some help from Brian Ondrey on the programming for this. He knows SO much about PLC programming, and if he reads this someday THANK YOU AGAIN! So my program has been steadily improving along with my knowledge of RSLogix. Not quite good enough to share here, but getting there!
Thirdly - I'm in GLORIOUS, SUNNY TULSA! Emphasis on SUNNY. I'm interning at a...power company...that lots of people know about. Nice place! And it's interesting to experience a desk job after 2 years of theme park and cafeteria.
Fourth....Ebay...you shall be the death of me. I now am in possession of about 5 kilowatts of power supplies. 1.5 kW of which work xD SO I may be swapping the 50 volts for 5 volts.. peak 200 amps. I'm still working on Ol' Switchy, but it never hurts to have a backup.
More to come!

[ThisIsGabe]

Well, I got some good/bad news...
Bad news - after spending almost half my weekend, I have determined that my 5/01 CPU is in fact haunted. This was determined after allllllllllllllllllllllllll the methods of PC to PLC communication were tried in every possible combination, taking my nice logic analyzer to the communication cables, and the use of enough expletives to make a sailor blush. I was recommended to try kicking it, but I think it'll be faster to just have it exorcised.
Good news, I'm improving the PLC I'm using to an SLC 5/03! This CPU (while $100 *heavy sobbing*) will allow me to not only communicate DIRECTLY with my computer using its built in serial port (DH485 WAS INVENTED BY SATAN, I SWEAR), but I'll also have an improved instruction set, so I can do fancier logic.
SO
While I wait for it to come in, I'm getting started on my "panelview" clone. What's a panelview? On nearly every roller coaster, there's basically a screen that allows the ride operator to interact with the ride. The level of interaction is really dependent on the manufacturer and the specific ride. They look kinda like this http://www.aotewell....0-Terminals.jpg . I can tell you on Maverick, mostly the panelview was just used as a display, and we never really had to mess around with it because Intamin loves using dozens of pushbuttons, while on Gatekeeper I believe you use the panelview to do things like restraint unlocks (B&M coasters usually don't have too many buttons).
Problem is...a panelview is really expensive :c so I'm going to be using a raspberry pi touchscreen, running (probably) pygame to interface with the PLC. Because a panelview is designed to communicate directly with a PLC, while my raspberry pi isn't able to by default, I plan on using serial monitors to listen in on the PLC and PC's communication so I can spoof a PC connection with the raspberry pi.
More to come!

[ThisIsGabe]

Good news everybody!~
PLC programming is just about done!
I've made the code that runs, it's got about 8 errors that it can throw in addition to a generic ESTOP, an automatic stop (which allows the trains to reach a holding position rather than stopping instantly), and a whole new I/O setup.
The code compiled and RAN on my new 5/03 CPU. I want to rewrite my error scanning routine, but for now the code is in an operable state. Hooray! @n@
~~~~~~~~
"Now you're playing with power"
I bought one more power supply...last one I promise
Not 1kW..not 2.5kW...but 6KW!
It's a massive 42V, 137A power supply designed for servers. I probably won't use all of the power it's capable of producing, but it'll give me even more wiggle room.
In other news, its big and heavy and I felt bad making the poor UPS man carry it up three flights of stairs. I might buy a second one to do an extreme teardown because just looking through the grills shows off lots of reeeeeally cool PCBs and they're only 50 bucks
~~~~~~~~~~~~~~~~
I'm also working on the CAD for the cars themselves. While the magnet holders I do have done, I need to figure out a good way to connect the wheel assembly to the magnets, because I feel like they need at least 2 degrees of motion, or I could just go for a fixed setup and loose some of the runner wheels. It's something that I'll just have to work on and model to figure out which way works best and which will allow me to meet my deadline.

I've given myself a deadline of October 31 to finish this whole project. Partly because if I don't finish it by the end of 2015 that makes two years of hammering away at this, and partly because I want to have something REALLY cool to flounce around at the IAAPA expo in November in the hopes of scoring an awesome co-op for 2016.

So for now, here's my todo list
~Write error sweeper code for PLC
~Panelview for PLC
~Finish roller coaster car
~Get PCB for controller made
~Make LSMs and car

More to come!

[ThisIsGabe]

PLC program has been 100% completed! All the subroutines and code works properly, the PLC survived a 48 hour run test without a fault, and now all that I need to finish is the panelview!
~~~~~
Panelview:
After spending about 2 hours reverse engineering the serial communication between my PC and PLC, I happened to find a long datanovel about the DF1 protocol that you can use to communicate with SLC5/03-05s it's a very nice protocol, and if you've ever wanted some light reading, check it out. It's a fairly basic call and respond protocol without encryption, and with two different optional types of checksums. Have a look!
So I spent about 2 days reading this book and figuring out how to write a driver for the raspberry pi, I happened across a great tool called Advanced HMI. It's a totally free program that uses visual basic, and a minimal amount of coding. I've got the beginnings of my GUI done, but I'm still working on figuring out how to properly send commands to a SLC. Currently I am able to read data values though, so that's good.
~~~~~
Supercell VFD:
Drive transistors are cool and all, but I think my VFD is just too simple!
I want something that'll be long and hard to design!
I want a super complex PCB that looks cool!
I want the ability to make a three phase full inverter!
This should be a nice long and complex project that takes years!
What's that? TI already designed something that would work for this application?
Well I guess you could take the EASY way out
So I guess I'll have to remake my main PCB again to include this driver....oh well~
~~~~~
Car and LSMs:
I'm working steadily on getting these done, and I've decided on a wheel layout which will be much closer to regular K'nex cars, with 6 on the front and six on the back, and the groups will rotate together.
For the magnets, I'm split on whether coils or loops more like "fullsize" LSMs will be better. This will probably be something that I test once I get the cars finished that way I can test them in real life
Fullsize LSM

More to come!

[JayketheKid]

This project is truly awesome. I can not wait to see more!

[ThisIsGabe]

After 12 hours, I've finished the majority of the circuit board layout. I want to make some traces bigger, add some more vias and a few logos, but its mostly done.
It's been a very interesting weekend xD

[ThisIsGabe]

SUPERMASSIVE UPDATE!!1!1

Okay, PCB has been finished
http://i.imgur.com/FysB3hE.jpg
This has the soldermask on the inverter legs removed so I can solder huge wires directly to the PCB to make sure it can handle the current. I just need to wait on an email from the PCB maker to see if they'll let me have that extra breakout segment or not, then I'm going to place the order for it (if I don't decide to stick more stuff on the silkscreen cause it looks pretty). I'm also going to be making an attempt to sample most of the TI components, because I'm a poor college student and Texas Instruments is a nice company who gives free components to poor college students who promise that they'll mention how nice of a company TI is.

http://i.imgur.com/iRzOhS7.png version "lets just make it work" of the wheel assembly. It's simple, its quick, it means I won't have to try and 3d print critical components, so it's generally a good idea.

https://sites.google.../knexlsmlaunch/ <- I'm making a website! Well...I'm having google make me a website! Someday I'll host my own site, but for now...this'll be a nice way to present the project (once I finish the project and the site)

I actually almost posted this before I remembered, but I finished the control panel viewer!
http://i.imgur.com/dohXiXd.jpg
It took weeks of coding to allow me to add this highly complex and difficult interface between a computer and a PLC! *cough* I found this amazing software called AdvancedHMI which did it for free and all I had to do was spend hours fighting with it to make it and the PLC shake hands. But once they did it was smoooooth sailing. Well, jittery sailing because you program AdvancedHMI with Visual Studio, and that program is a memory hog, but you get the idea!

More to come!

[ThisIsGabe]

Circuit board arrived yesterday! It turned out amazingly well, now I'm waiting on the rest of the components to arrive from Digikey (which I just placed the order for). Surface mount parts and solderpaste should be arriving in the next week. Look at how pretty the PCB is! http://i.imgur.com/NCOb4dZ.jpg

http://i.imgur.com/7rB9DTK.jpg As I was looking for relays for the LSMs, it became apparent that the majority of digikey relays weren't up to par. The only relays that would have been able to work with the project required a backorder in NOVEMBER. So Ebay came to the rescue again and now I have the rollback relays! http://i.imgur.com/7rB9DTK.jpg Plus they cost about half as much as digikey's relays AND came with rackmounts.

Texas instruments are officially awesome, especially Jim Denison from their customer service department. They sent me the MOSFETS, drive controller and temperature sensor free of charge and SUPER fast. Like next day in my hands fast. I will definitely be using TI in more of my projects, their datasheets are great and they've got a nice staff.

Hopefully next week I'll be assembling the PCB and reflowing it (I'm probably going to use a skillet or lab hot plate). I also got a 3d print of the roller coaster parts and now in the next week or so I'll be making the car out of my aluminum stock

https://sites.google...xlsmlaunch/home Still a work in progress, but I'm adding more stuff every day and keeping it nice and up to date

We're in the home stretch!
More to come!

[ThisIsGabe]

Power supply cables came in! Big, beefy and best of all, the power supply passed the smoke test! Generates the proper 42 volts across the output!
Getting the arduino code done is a chore, but it's kind of a fun thought process. Still waiting on PCB components!

[ThisIsGabe]

Got the drive board 100% assembled - all that remains is to finish the programming for the arduino!
I was able to test out the LEDs, and fortunately all of them were installed in the correct directions!
Hopefully tomorrow/thursday I'll be able to mill out the aluminum car parts.