BLHeli - Double brushless for mCP X - with governor coming

[coredump]

Thanks for the hex file. I use all 20 pins of the parallel programming plus a few Vcc, AVcc and Gnd.

It is just the normal parallel programming mode. Nothing fancy.
I can post some picture later tonight.

[coredump]

Here are some pictures.
1) taking the MCU off
2) put it on the programming board
3) AVR dragon HVPP to reset the chip
4) put it back on walkera esc

Update:
I connect the motor and battery. It has sound tone.
So I haven't totally fry the MCU yet.

[4712]

Wow, this is a bunch of work to save the tiny mcu, but one time done... You are equipped for all others...
I hoped there would be a "small" simple way just to clear the chip...

As I do not own the dragon, i decided to replace the mcu with a ATMEGA168.

Back to the questions:

I have another discovery. Why the walkera fuse picture show in
the BLHeli 3.0 picture has CKDIV8 set? Wouldn't that will slow down
the system clock by 8 times?

CKDIV8 sets the startup frequnezy to 2Mhz, which is in the safe operation area for all allowed Vcc voltage values.
In the Macro "Initialize_MCU" the devider is set to 2 after reset, which sets 8Mhz and is still in the safe operation area. For 16Mhz the Vcc value should be above 3.8V, if my simplified calculation is right (refer to the diagramm in the datasheet).
Using 16Mhz would automatically mean the destruction the EEProm content, when used.

I am pretty sure the original fuse bit on the
walkera ESC did not have the CKDIV8 bit set.

Which does not mean, that the devider is not set in the code...
Or it would be the simple reason why Walkera did not integrate a TX programming mode, which would require the EEProm usage. And may be one more reason to set the RSTDISBL fuse the other lock bits.
Anyway it is not recommended by Atmel to leave the safe area.

One more: Be sure to set the brownout level for 2.7V!

I could sen You the bootloader, but the description is in German (code is commented in English ).
http://www.mikrocontroller.net/topic/95839

[coredump]

I am not aware of any other small and simple way. There are other circuits to reset with HVPP mode,but the number of connection pin is very similar. If I need to unlock another
walkera ESC, I just need to take it off, put it on the programmer and put it back on.

There are those programmer adapter can save me the soldering to the programmer
board. It cost $30, seems a little expensive. If I have a lot of chips to unlock that
might worthwhile. Right now using the hot air on the board which is not too bad.

Thanks for the tip regarding the DIV8. One way to do it is use 8M to do the
initial programming read/write EEPROM. Then switch to 16M mode on actual
flying.

OK, I change the fuse to use brownout at 2.7v.

Is the serial boot loader compatible to GPL? BTW, the pull request means
the pull request on github. I recall I saw one of your pull request before.

[4712]

Originally Posted by coredump

One way to do it is use 8M to do the
initial programming read/write EEPROM. Then switch to 16M mode on actual
flying.

Mmmhhh... without any EEprom usage: unpredictable results...
With EEprom usage: definitely No way... that would destroy the EEprom content, the only question is: when.
There might be a way to reduce the chance of corruption, but if the MCU jumps suddenly somewhere, You're lost anyway.

Is the serial boot loader compatible to GPL? BTW, the pull request means
the pull request on github. I recall I saw one of your pull request before.

It is not my code, Hagen does not provide any information under what license the code should be...

[coredump]

Ah, thanks for the tip on the overclock. Haven't take a close look on that part of the data sheet yet.

[cestb]

Sorry wrong post

[fac13]

The hex file HC5A1SA48V.tail.hex on github includes addresses above 4k. See lines 258 and onwards.

I think the target ESC uses an ATMEGA48 variant, and they have just 4k of flash.

When I try to flash that file, using avrdude, it fails saying line 258 is out of range.

When I try to do my own compilation of the asm file, taken from the same github zip, Atmel Studio fails with a message saying "end of .cseg at 0x109c is beyond end of memory at 0xfff".

I've clearly missed something, or misunderstood something, because it looks on the face of it as if both the asm and the hex file are unsuitable for the processor on a HC5A1SA48V, but I know that can't be right.

What have I got wrong?

[coredump]

I compile those file and put it to github.
I haven't test that particular Esc. What Mcu is on that chip? Edit: atmega48 found it. I recall all of them had 8k code space. This one is special.

Let me see what I can do to reduce the code size.


Sent from my iPhone using Tapatalk

[coredump]

Btw I have one like that use atmega8.
So 8k code space.



Sent from my iPhone using Tapatalk

[4712]

Atmega 48 has 4K only. So try:

.equ TX_PGM = 0   ; Set to 0 to disable tx programming (reduces code size)

.
Github is a PITA for Windows users .
You always have to reformat the text files; at the end of line there is missing CR or LF (don't remember).

[fac13]

[QUOTE=4712;4000070]Atmega 48 has 4K only. So try:

.equ TX_PGM = 0   ; Set to 0 to disable tx programming (reduces code size)

.

Thank you 4712 and coredump! That change reduces the compiled size to 3636 bytes, and that flashes without problem of course.

I haven't tried connecting up and using the esc yet but I'm a lot closer than I was earlier today, I wish I'd asked first but assumed I'd done something wrong.

Thanks again for the fast replies

[coredump]

Originally Posted by 4712

Atmega 48 has 4K only. So try:

.equ TX_PGM = 0   ; Set to 0 to disable tx programming (reduces code size)

.
Github is a PITA for Windows users .
You always have to reformat the text files; at the end of line there is missing CR or LF (don't remember).

Install windows msys git and choose check out convert and check in UNIX (default)

Git is great.

[SadSack]

Hello
Been lurking here for long time now.
Well this is amazing work sskaug, Thank you
I will try Silabs version at some point but got enough atmel hardware to have a go with.


Coredump :
Like your adapter i made one up yesterday to recover 328 and quite a few others TQFP and few MLFs quite a few more to go when needed.



There is Eagle files if your interested HERE

not the best layout but works well enough.

Thanks, Phil

[jsg001]

I've been working on getting the BLHeli code to run on the HK 074000030 1S/7A esc. It seem to be compatible with the HC5A1SA48V code; that is in terms of port locations and battery voltage cuttoff. However, the motor does not seem to spool up completely. I think it may not be leaving startup mode. That is a guess, I'm still working to understand the code. I see posts in this thread about such an issue, but it's not clear to me as to what to look at first to solve this. Any help would be appreciated.

Secondly, I think i may have shorted something on the main control board while working on it. The main motor control signal still looks good, but the servos are now responding erratically. That is even with all the other modifications removed. Maybe something in the gyro area gone bad (guess). I defluxed and washed the board. A visual inspection under the microscope does reveal any problems. Any ideas on this board?

Thanks,
john

[Supaplums]

Hi all. Been out of the game for a month or so, somewhat because of frustration with this little guy. Anyway, still having the same troubles. Here is my layout: BLHeli 12 amp flashed main with spin 16,300 main motor, 3A BLheli flashed tail with HP03, V1 board. My problem is that the main motor does not seem to want to spool up. When you put the throttle up it just pulses short little bursts to the main without being able to completely spool. The tail seems to be working fine. Took Dylan's advice and changed the programming of the start up acceleration to a bit higher. It didn't work so I put it all the way up, still nothing. Also figured I'd try raising the start up velocity, still no good. Anybody have any idea why my motor doesn't seem to have enough power to spool up? It has spooled up before by the way, just not consistently. Seems to need a push. Any assistance in this matter would be greatly appreciated. Thanks!

[Spunjah413]

What version are you using? I still like to flash and use V2.4 on the Main, as i have never had any problems using it on any motor (Ad i have basically used every one mentioned in this forum). I would try erasing the chip manually threw the menu's, then flashing an older version and see what happens.

Another problem could be a bad ESC. I had one i was building that would only pulse the 915 and never start. Eventually it failed and melted on me during programming. Ive had 3 instances already of either bad ESC's or Faulty flashes that needed a Re-flash.

Last thing i would check is Gear Mesh, and Motor mount screws. Make Sure there is play between the pinion and gear. Also make sure the mounting screws are not hitting any windings of the motor.

[Supaplums]

Thanks for the reply! Pretty sure nothing's hitting the windings cause it has spooled in the past. Also notice it will spool up with only minor problems when I take the blades off. Some kind of power problem I guess. I think my version is 3. Unfortunately I don't have the ability to flash. Had it done when I bought it. I would think its a battery problem but I fly my other BL with them all the time.

[jsg001]

Here are my solutions to the above problems I posted about regarding the HK 07000030 1S/7A ESC....

(1) There is a startup transient on the power line that made the cell count routine incorrectly identify identify the number of cells; hence, the majority of time on startup it was in a power limiting mode for my 1S setup. I verified this with a quick code change then as a permanent fix I added a 1.0uF cap on the resistor divider network. Works perfect now with original code.

(2) The main controller board issue I was having was user error. I accidentally bound to the wrong helicopter on the transmitter with a different swash plate type.

I felt the governor at 1.5 was overdamped in my setup with the HP05S. I bumped the proportional term up to 3.0. Will test some more and make a few test flights, but on the bench it looks good.

[jsg001]

This looks useful...

HK part number 332000007 Atmel Atmega socket firmware flashing tool

http://www.hobbyking.com/hobbyking/s...ng%5FTool.html

I don't like flying with all of the SPI programming wires hanging off the ESC; concerned of shorting something with a broken wire in a crash.

This is a the first I've seen something like this. Are there other such devices? Anyone using something similar?