They’re done!
After my Digikey and Avnet shipments arrived, I could finally finish out the 8 remaining RageBridge2 boards in the panel. Using my handy dandy hot-swap test rig, I’ve determined that they at least power on right now without exploding, which is a good start.
Before I upload running code, I’m planning on adding a mix option to the firmware which can be activated by soldering a little jumper wire (a future board rev will add a header that can be used to select it). This way, the controller can be used with non-mixing radios (like 2 channel pistol grips) to let one channel control steering (differential) and the other speed (magnitude).
Update 16 October! The mixing firmware has been implemented.
The two RageBridge2 boards I assembled last time have been tested in real-life driving already. First, I installed it in Null Hypothesis in place of the RageBridge v1:
Then I gave board serial #2 to Adam and had him install it in his pre-Dragon*Con speedballed robot Blitz.
And then, we faced off at the Somerville mini Maker Faire:
Basically what it reduces to is two somewhat gimpy bots driven by two out-of-practice drivers. The motor that smoked in Blitz was on the front lifting tines, and it was being driven by a Victor 883.
But despite my best efforts to test these boards out in a strenuous condition, Blitz and NH are kind of… the same. They both run 4 overvolted drill motors (and keep breaking them… clearly these things have gone totally to shit) coupled to 4 inch wheels, and while NH has a higher top speed due to my more overvolting, the RageBridge boards limit the motor current such that the motors kind of behave the same anyway.
I need more data points before declaring the RageBridge Certified Legit™. That is why I will now officially announce the….
ragebridge breeding program
These 8 boards will be released “into the wild” as much as possible such that I am not in control of their use and installation save for a short user manual page. I’m therefore going to put the call out for participants.
You may qualify for the RageBridge Breeding Program if…
- You have a load, such as a robot, vehicle, grain thresher, etc. which runs between 18 and 36 volts, has 2 DC motors, and would like about 40-50A, but can live with 60 amps at most. This is because the RageBridge has intelligent constant-current limiting set to 60A. So if you don’t mind a little… anemic acceleration on your twin Magmotor mountainboard, that will work just fine. At this time, the 2 output channels cannot be combined into a single high-current channel. The 36 and 24v boards are currently not the same in order to test out different components, but the production version will run the whole range of voltages.
- It exists right now and you can swap RageBridge into it. Sorry, projects which will be finished Real Soon Now cannot qualify.
- It uses an R/C radio as an interface. RB currently cannot take analog inputs (e.g. throttle pedals) or TTL serial commands. “Vehicle firmware” which does take these inputs is one of the things I intend to work on as soon as possible, and serial commands is also on the list of things to add. The firmware is easily flashable in the field for when that happens; you will need an FTDI cable. After all, I do write the firmware for this controller in Arduino…
- You are willing to send $50 USD to help me offset the cost of boards and parts. If it works great for you, then awesome – 36V dual motor controller for $50, and I’m not gonna ask for it back! But…
- You understand that you are not receiving a finished product and there is no warranty on functionality or reliability. Do not use this to hoist people up elevator shafts or to run your heart-lung machine, for instance. Would you even use an Arduino to run a critical medical device like that?! Each RB board will be tested before it ships out and I can only take “returns” for DOA.
- You are willing to send me video, pictures, and/or test data showing the board working in your environment.
If you agree with the conditions, send me an email (charlesg@ this 3-lettered university name .edu) with the subject “RAAAGE” (that’s 3 A’s) and the following information:
- Your load details – what it is, and what voltage it runs, so I know which board type to send you.
- How soon you intend to install and test the RB board
- How you intend to mount the board in your load (via its mounting screw, heat shrunk, above a fan, on a boat with a goat, etc.)
Because there are only 8 boards available, I clearly cannot take everyone. I’m seeking “diversity of loads” here – vehicles, robots, elaborate dentistry tools, etc. I will set a deadline for “applications” of MIDNIGHT on Wednesday, October 17th (that means as soon as the clock turns to 12:00 and it becomes the 17th). Eastern Standard Time. I’ll notify both accepted and rejected parties then. They’ll ship during the week.
Preliminary specifications:
- 18 to 36v nominal operation
- Absolute maximum voltage for “24v” boards is 37v, absolute maximum voltage for “36v” boards is 57v.
- 40A continuous at room temperature with no additional convection or insulation.
- Constant-current region above 60A (your load will see a maximum of 60A, plus or minus a bit of ripple).
- R/C servo pulsewidth range of 1900us to 1100us with 1500us center. +/- 20us deadband.
- 5V power available for receivers
- 4.5″ x 2″ outer footprint
- 1.7″ x 1.7″ hole pattern for mounting, sized for #4 screws.
the ragebridge fundraising program
I intend to send out one more board revision before releasing the design to manufacturing. This new revision will wrap up some of the hardware loose ends like component placement, manufacturability, trace routing optimizations, and like, making the RESET button actually reachable with an average finger.
Boards and parts (especially parts for 10…) are still really expensive, especially in low test quantities where I can’t take advantage of price breaks.
If you like the RageBridge and its new addition of a constant current output mode to the small- and medium sized robot/vehicle controller world, and want to support its development, then please let me direct you to my donations page. There is no suggested donation, and contributing does not constitute a pre-order for any product!
The whole mission of the RageBridge project is to try and make a compact, reliable dual channel high-powered motor driver using 21st century designs and parts, with no “fanciness” like dual encoder inputs, built-in PID controls, 8 different ways of taking commands, etc. in one unit. It is a rebellion against the new trend in robot controllers to integrate all of these functions and charge more money for it while not really being that good at, you know, driving motors. Basically, I want to give the venerable Vantec series and Victor 88x series of controllers a this-decade equivalent. Raw, simple power for applications which demand it, and nothing more.
the ragebridge kickstarter
Support us on Kickstarter! The link to RageBridge’s campaign is…
…
Not yet. I fully intend to pitch this sucker up on Kickstarter when I have a demonstrably working device. No promises and renderings here – I’m fully intending to use KS to gauge the initial market and to tell me how many times to hit the button. So if you’re just totally into a public fundraising campaign or find paypal links repulsive, hang tight!
dear friend , too bad i donot live in states , i wish i could apply to an alpha tester !!
but can i ask you a couple of questions that confuse me when i look into your design :
1-why use 15v supply for ir21844 (isn’t 12v is enough )
2-why use dual direction current sensor at motor , wouldn’t a uni dir sensor between V+ and upper Mosfet do the job ?
3-why use 10k resistor from gate to source
4-why the use of resistor 1k from sensor output to micro input , doesn’t this change the Vout of sensor
5-wouldn’t TVS diodes on upper Mos or RC snubber across motor necessary to protect against spikes in case !
wow that was a lot ! :) thanks for your patience :)
keep up the hard work , you might succeeded in building a pro H bridge after all.
Hey Bob,
There were actually quite a few international requests. One board is currently going to Austrailia and another to the Netherlands. You should totally have asked.
To answer your questions,
1. I was interested in having as low on-state resistance as possible, and to get the FETs out of the “miller” region quickly. Hence, a high gate drive voltage and reasonably fast turn-on times (100-200ns).
2. One current sensor on the V+ rail does not provide enough information. Because this is a reversible controller, the most direct measurement path is to place a current sensor inline with the motor output. Either way, a bidirectional sensor is needed because of this controller’s implicit regeneration capability.
3. It is a pulldown resistor to shut off the FETs in case the gate drivers fail. Just a safety precaution.
4. That was initially to isolate the microcontroller from a potential source of noise, the current sensors, since they were very close to the high-dV/dt switching of the FETs. It turned out to be not needed, so they were eliminated in this latest version. They will influence the reading, but only very slightly – microcontroller input pins generally have very high input impedance and capacitance on the order of picofarads.
5. On a “single quadrant” controller, yes. Not only are these controllers 4-quadrant (the motor is always being driven somehow – there is no freewheeling period, conventionally), but the FETs themselves contain reasonably competent diodes – Large power FETs have “intrinsic” diodes when biased backwards. They serve the purpose of flyback diodes in a more simple motor driver.