Hey Steve looks like a total win! I bags the first set that you sell Ha ha! They look destined for Wombot!
Miles
_________________
Miles Blow - Julie Pitts
www.mulesfilm.com.au
www.wombokforest.com.au

-Pickasso- Vivid Sportsman champion 2015

Temperature limiting code is implemented.

Still need to play with the values. But my testing with a heatgun (despite some melting) appears to work.

Steve
_________________
Steven Martin
Twisted Constructions
http://www.botbitz.com

Just a quick note on my thoughts as to how to implement max slew rate.

1. Average 5 RC pulses. (Bit of a hack but should have the same effect) However will slow all responses of the ESC.

2. Ramp using the main loop delays (AVR libc delays are terrible so you probably wont have any idea how fast its actually ramping) Still not sure how to integrate this with braking delays and such.

3. Put it in timer overflow interrupt which is probably best option. But similarly to 2 I am not sure how to do it. And as main loops handles states might get a bit messy.

Also Tz85a tutorial video is on my private youtube. I need to cut a few things but hopefully should be up later tonight.

Can anyone suggest a desktop recorder for Ubuntu to record AVR studio programming?
_________________
Steven Martin
Twisted Constructions
http://www.botbitz.com

quote:

---

Originally posted by marto:
Can anyone suggest a desktop recorder for Ubuntu to record AVR studio programming?

---

Thanks aaron will get to it tonight. Trying out youtubes online video editior atm.
_________________
Steven Martin
Twisted Constructions
http://www.botbitz.com

Basic coding technique for Slew Rate limiting.

State machine for output section with update rate limited by master timer/interupt tick

Rx received pulse width -> Target PWM%
Time since last output update elapsed ? No - Exit without update
New PWM output = difference between Current PWM and Target PWM with change for this step (delta) limited to slew rate
Ouput new PWM rate, update current PWM to equal new
end of state machine loop

Thats the basic principle. I can go into more detail if needed.

Good Work Steve !
_________________
Great minds discuss ideas. Average minds discuss events. Small minds discuss people

Thanks Brett. That should fit well into my current setup I will use interrupt to set flag @ 1khz and then trigger based on that.

The other thing I am thinking about is calibrating the RC input. At the moment I have just calibrated it manually to the HK transmitter but that is a bit tedious.
Could do auto center on startup but once again that causes issues. I don't really want to introduce wiring a switch either.


I am pretty keen to get all thee details implemented and get feather ESC code semi stable so we can do some alpha then beta testing.
_________________
Steven Martin
Twisted Constructions
http://www.botbitz.com

Max slew rate implemented.

Still untested.
_________________
Steven Martin
Twisted Constructions
http://www.botbitz.com

Exponential rates now in code.

Still untested.
_________________
Steven Martin
Twisted Constructions
http://www.botbitz.com

It has been 10 weeks since I first looked at this and I think we have come quite a long way, towards getting these controllers working and tested.

The current state of the code is pretty well sorted for a the Atmel controller line. I think most of the base features are in there and we are at the stage where only testing will show whether they are really useful or not.

To that end there is now 6 Beetle ESCs "In the wild" and after a bit more testing we may see some feather ones filtering through as well.

Some of the key things which need to be tested:
- temp limiting values.
- the ability of the feather controllers to drive drills
- compatibility of the RC values
- current capability compared to other ESCs

The code should be reasonably generic for all ESCs. The current version should work on the Plush30, Tz85a and Tz12a with the available header files.

For future development I think feather ESCs may be a little ambitious. I think they should work but @ $40 + programming I am unsure whether it offers much of an advantage for your average builder. The plush 30s are probably a little over rated for the beeltes. (Aaron has a video of them driving 100w scooters) So cheaper alternatives may be more suitable.
I also hope to get the 2in1 code up and running if I can solve the problems I am currently having.

Still looks like it next 2 months should be even more interesting as I start to put to them to use.
_________________
Steven Martin
Twisted Constructions
http://www.botbitz.com

Two Turnigy K-force 100A ESCs arrived today. When I opened one up it turned out to be almost identical to the 120A boat controller I already bought . The only differences are a separate BEC PCB and a fairly useless water cooler on the boat ESC, while the 100A controller has a finned cooler and a better plastic case for $13 extra - go figure!



This probably means the 100A rating is realistic with a larger heatsink so if its hackable, the boat ESC is my fave model.

Looks similar to Tz85a. Where is the MCU? Got picture of the other side? I am thinking that a lot of these have a generic layout for the fets, then a separate controller board.
_________________
Steven Martin
Twisted Constructions
http://www.botbitz.com

Here is the FET side:



The MCU is on the underside of the logic board and totally inaccessible; there are 4 pads near the red power cable that almost have to be the programming port so all is not lost . There are 7x3 fets on this side and 1 x 3 on the reverse side - I can't imagine why they would put one FET per half bridge away from the heatsink!

In the interest of Beetle ESCs what current rating would be required? And how many cells do people want to run? There is only 4 cell ones @ HK but I think the 4S is limited by the BEC. I would be willing to try some 6S batts on them but I am not sure if anyone is running a beetle at anywhere near that.

Steve
_________________
Steven Martin
Twisted Constructions
http://www.botbitz.com

If there 100 amp controllers wouldnt we be looking more towards featherweight controllers and not beetle.
_________________
Andrew Welch, Team Unconventional Robotics