Society of Robots - Robot Forum

General Misc => Misc => Topic started by: Rcreator on October 20, 2011, 06:53:39 AM

Title: comparison between axon II and chipkits
Post by: Rcreator on October 20, 2011, 06:53:39 AM
hello, you compared axon with arduino, axon was much better for big projects but then few days ago ive seen something called chipkit max32 released by microchip using 32bit microcontroller 80mhz clock more flash memory more ram many io etc and surprisingly it was 49.50 while axon 2 is 108 dollars and which as far as i know uses 8bit microcontroller on 16mhz clock what is the difference between can someone enlighten me why is axon 2 over the double of the price and make me feel comfortable to choose it over the chipkit max32?
Title: Re: comparison between axon II and chipkits
Post by: joe61 on October 20, 2011, 07:39:54 AM
Apples and Oranges basically. Chipkit uses a 32 bit PIC, which is another world. Whether Microchip will support it long terms is yet to be seen, and the programming tools are limited compared with Atmel processors.

The difference between Axon and Chipkit is much like the difference between Axon and Arduino. Axon has the ability to easily interface many servos, RC gear, et al built in, while you have to add stuff to Arduino/ChipKit to do the same things. By the time you've added the other stuff the price difference isn't so great.

Chipkit is an interesting beast though, being 32bit. If it could be programmed using Linux I might take a look at it, but I'm not going to run Windows just to try it.

Just my $.02 worth

Joe
Title: Re: comparison between axon II and chipkits
Post by: Rcreator on October 20, 2011, 12:30:52 PM
hi joe,
As far as i recall chipkit is using same programming tool as arduino and controlling a servo to any angle was 1 line(one function call) at arduino using its servo library after connecting any pwm port to signal cable and giving power from an external powersource.

I dont understand what you mean by easy interface am i missing something here what has to be added?
Title: Re: comparison between axon II and chipkits
Post by: joe61 on October 20, 2011, 01:22:13 PM
Well, you have to setup the external power source, for one thing. With the Axxon you just plug the servo into the board. Since most servos come with a three-pin connector it's a bit problematic to bring power and signal together from two different sources.

I was thinking more along the lines of the number of peripherals available on the Axxon vs Arduino:

58 I/O Total
16 ADC
25+ Servos
I2C, SPI
3 UART + USB
Up to 8 external interrupts
15 PWM Channels
64KB Flash, 4KB EEPROM, 8KB SRAM
6 Timers (four 16-bit, two 8-bit)
pre-programmed with a bootloader - no programmer required
numerical LED display
built in 3.3V, 5V, and unregulated power buses
external memory support (port A)

I'm not a sales guy or anything, but I think that comparing them based solely on the amount of money require for the initial purchase misses some things.

A common bit of advice is to decide what you want to do and then pick a setup that will do that. When you decide what you want to do, also keep in mind what you're likely to want to do in the future.

Joe
Title: Re: comparison between axon II and chipkits
Post by: joe61 on October 20, 2011, 01:48:24 PM
I was just looking at the Microchip descriptions of the chipkit to be sure I'm not saying anything dumb. It looks like the chipkit is a 3.3v system, which (If that's right) you might want to think about before going with it.

Joe
Title: Re: comparison between axon II and chipkits
Post by: Rcreator on October 20, 2011, 03:05:13 PM
what is the disadvantage of 3.3v would it make any difference when i use external power source for motors?

Max 32 features:

 Max Speed MHz    80mhz
 Program Memory Size (KB)    512kb
 RAM (KB)    128kb
 Temperature Range (C)    -40 to 105
 Operating Voltage Range (V)    2.3 to 3.6
 DMA Channels    8
 SPITM    4
 I2CTM Compatible    5
 A/D channels    16
 Max A/D Sample Rate    1000
 Input Capture    5
 Output Compare/Std. PWM    5
 16-bit Digital Timers    5
 Parallel Port    PMP16
 Comparators    2
 Internal Oscillator    8 MHz, 32 kHz
 RTCC    Yes
 I/O Pins    85
 Pin Count    100

it doesnt say about interrupts or UART but in the descriptions it says Programmable vector interrupt controller and Serial Communication Modules allow flexible UART/SPI/I2C™ configuration does these mean the same thing i dont know
Title: Re: comparison between axon II and chipkits
Post by: joe61 on October 20, 2011, 04:05:53 PM
what is the disadvantage of 3.3v would it make any difference when i use external power source for motors?
I honestly don't know, I've never tried to run a servo at anything but 5-6V. One of the EE types here can probably say for sure.

My guess is that it would need a booster of some kind. You'd have to take it into consideration if you want to get most (?) other shields, which generally are 5v and would need level shifters.

Quote
Max 32 features:

I haven't used one, so I don't know how well any of it works, and should probably drop out of that discussion. While looking around though I noticed this (http://ruggedcircuits.com/blog/?p=202) which brings up a couple issues of compatibility with existing shields. Depends on what shields you want to get I guess.

Quote
it doesnt say about interrupts or UART but in the descriptions it says Programmable vector interrupt controller and Serial Communication Modules allow flexible UART/SPI/I2C™ configuration does these mean the same thing i dont know

I'd guess it's similar to the Arduino llibraries. Dunno. If you want to do something that the libraries haven't provided for though, you have to learn the PIC. No doubt that's why they're making this attractive :-)

I'm kind of playing devil's advocate at this point. I bought one of the picKits a while back, and never could come up with anything to do with it. Maybe I'll see something in this discussion ...

Joe
Title: Re: comparison between axon II and chipkits
Post by: Soeren on October 21, 2011, 10:28:13 PM
Hi,

what is the disadvantage of 3.3v would it make any difference when i use external power source for motors?
No disadvantage.
It's very easy to interface to 5V boards if needs be and 3.3V is how the future is going, while 5V stuff is being phased out faster than some people realize (regrettable though, would be nice if they could just coexist).


Max 32 features:
[...]
Don't forget it's USB OTG and USB host facility and the hardware RTCC (Real-Time Clock and Calendar with Alarms) to name a few more.
It's Arduino (downwards) compatible, It's not an Arduino!


it doesnt say about interrupts or UART but in the descriptions it says Programmable vector interrupt controller and Serial Communication Modules allow flexible UART/SPI/I2C™ configuration does these mean the same thing i dont know
Yes it goes all the way to 20Mbps!!

There's no sense in comparing a 32bit controller board to the 8bit Axon II.
The PIC's 32 bit controller board is brimfull of goodies, faster and better equipped, which is no wonder for a 32bit controller and the reason that they can make it at a more realistic price is because they have the production facilities in place to churn them out (and the production price of the chip itself is negligible between 8 and 32 bits, it's the development that costs) - All Microchips boards are comparatively cheap considering what you get and usually (always?) carry the largest version of the controller in its "family"/series.

Next time I order from Microchip, I'll probably grab one or two, just for the heck of it - they're extremely cheap considering they're 32bit boards.
Don't worry too much about the "missing" power planes ~$60 will buy you loads of pin-headers and stripboard

Microchip supports their products long term, as long as it doesn't mean keeping dinosaurs in chains ;)
And don't ever buy the bull about Microchips tools being inferior to those from Atmel - If that were true, PIC's wouldn't be the most used controller among people who earns their living with microcontrollers.
That said, while the controller is from Microchip, the board is from Digilent and further...
"chipKIT uses a modified version of the original Arduino™ IDE for compatibility with existing code examples, tutorials and resources."
The operative word here is "modified", as it has to be to include all the extras that an Arduino doesn't have.


You can download the datasheet for the controller here (http://www.microchip.com/wwwproducts/Devices.aspx?dDocName=en545660) if you want to study it more.
Title: Re: comparison between axon II and chipkits
Post by: WaterPig Master on October 22, 2011, 01:51:37 AM
I feel a bit silly even posting because I've never used either of them, or an arduino, but as far as I can tell:

Axon (especially with webbotlib) is perfect for putting together advanced robots very quickly. It is more likely to appeal to people who are less interested in low-level programming and EE work.

The chipkit is like an arduino on steroids. It's a fairly highly powered embedded computer with mucho mucho peripherals, but you'll have to do the EE work or buy shields in order to interface it with lower level circuitry.

For me, the massive appeal of the Axon and webbotlib is crushed a bit by it's flaky Mac OS X support, whereas I believe the chipkit has a native IDE - and I have a nice PIC programming workflow set up already. But if Axon was ever to offer solid OS X support, I would almost certainly choose it over the chipkit for advanced robotics projects

Cheers,
Barnaby
Title: Re: comparison between axon II and chipkits
Post by: joe61 on October 22, 2011, 04:09:47 PM
The operative word here is "modified", as it has to be to include all the extras that an Arduino doesn't have.
What are you thinking about here? My understanding is that they ported the Arduino IDE so that you could do the same stuff on the chipKit. I haven't seen where they allow more. In fact there are some things it doesn't do that the Arduino does, like pin change interrupts on any pin, et al.

I'm not trying to start a pissing contest here, I don't care what chip someone wants to use, I'm just kind of interested in this. The PIC is potentially better for what the OP wants to do, as is an Axxon. Depends on what that is. He started off by talking about servos, and the Axxon supports servos better than Arduino (or ChipKit as far as I can tell). I didn't get that point across very well, granted.

Joe
Title: Re: comparison between axon II and chipkits
Post by: Soeren on October 22, 2011, 05:24:17 PM
Hi,

The operative word here is "modified", as it has to be to include all the extras that an Arduino doesn't have.
What are you thinking about here?
Is this a trick question?  ;D
I don't see how I can express it any clearer?
The Arduino doesn't include eg. USB OTG, high speed comms etc. etc.


My understanding is that they ported the Arduino IDE so that you could do the same stuff on the chipKit. I haven't seen where they allow more.
You didn't see it, so it doesn't exist??  ;)


In fact there are some things it doesn't do that the Arduino does, like pin change interrupts on any pin, et al.
Well, the pin change interrupt on the Arduino is purely a software library and can thus worjk on both.
If you wanna compare "real" interrupts, mindyou, that the 32bit PICs interrupt controller runs circles around the ATmega in a way that you cannot even compare the two - it's like David and Goliath less the sling shot.


I'm not trying to start a pissing contest here, I don't care what chip someone wants to use,
You could have fooled me, when trying to call an 8bit basic board superior to a 32bit multi peripheral board.


I'm just kind of interested in this. The PIC is potentially better for what the OP wants to do, as is an Axxon.
In what way do you think that the Axon II outperforms the ChipKit PIC32 board??


Depends on what that is. He started off by talking about servos, and the Axxon supports servos better than Arduino (or ChipKit as far as I can tell). I didn't get that point across very well, granted.
You must be thinking on the pin headers here, as they both the Axon II and the Arduino has got the same motor.
And anyone claiming that a $108.- 8bit board is better than a $49.- 32bit board due to a couple of rows of cheap header pins is substantially biased (to put it in a nice way).

Anyone looking at it with a clear and unbiased mind cannot be in any doubt whatsoever.

But there's more folks... No steak knives though ;D but the chipKIT Uno32 with the same impressive 32bit controller goes for a mere $26.95.
In other words, you can get 4 x 32bit boards for the price of a single 8bit board!
And don't forget that there's a real company behind the ChipKit boards, that's a value in itself.

Loyalty is good, but blindly giving advice to newbies, against ones better knowledge, or without even checking the facts is doing bad.
Title: Re: comparison between axon II and chipkits
Post by: joe61 on October 22, 2011, 08:24:10 PM
The operative word here is "modified", as it has to be to include all the extras that an Arduino doesn't have.
What are you thinking about here?
Is this a trick question?  ;D
I don't see how I can express it any clearer?
The Arduino doesn't include eg. USB OTG, high speed comms etc. etc.
No, I meant what does the modified Arduino IDE make available that the regular one doesn't? Of course there's more stuff on a 32 bit processor vs an 8 bit one.

Quote
My understanding is that they ported the Arduino IDE so that you could do the same stuff on the chipKit. I haven't seen where they allow more.
You didn't see it, so it doesn't exist??  ;)
I don't think I said, that - let me check ... no, I didn't say that.

Quote
In fact there are some things it doesn't do that the Arduino does, like pin change interrupts on any pin, et al.
Well, the pin change interrupt on the Arduino is purely a software library and can thus worjk on both.
Umm, no it's not. The Atmel chip has the ability to fire pin change interrupts on any pin.

Quote
I'm not trying to start a pissing contest here, I don't care what chip someone wants to use,
You could have fooled me, when trying to call an 8bit basic board superior to a 32bit multi peripheral board.
Nope, didn't say that either.

[...]

Quote
Anyone looking at it with a clear and unbiased mind cannot be in any doubt whatsoever.
Wow. Well, thanks for that clear and unbiased analysis.

Title: Re: comparison between axon II and chipkits
Post by: Admin on October 24, 2011, 08:06:23 AM
No flaming allowed everyone ;)

I agree with joe61's Axon vs Arduino comparison comments. And as for cost, sometimes it's worth spending more if it can save you hours of wiring and frustration - but that depends on how much you value your time. :P

This may or may not matter, depending on your project, but you should also consider the compactness and size of the boards. Or if you're running on battery, the energy efficiency. Sometimes you have a ton of serial devices that need to be connected, so you need as many UARTs as possible. If you only had a few servos on your bot, running a few jumper wires isn't a big deal. But if you were building a bot with 20 or 30 servos like a hexapod or biped, you'd be much better off with built-in power buses. Again, as joe61 said, it depends a lot on your project for what is important. Even more, if you plan to build more robots in the future, consider that your parts will be scrapped and reused - can your mcu choice handle more complex bots in the future?

As for 3.3V, the Axon Mote uses 3.3V. You'll have some issues with certain older electronics that only works with 5V, but the industry is moving to the 3.3V standard and most stuff coming out today favors 3.3V over 5V.

ps - as Soeren said, I'm very biased in my comments ::)
Title: Re: comparison between axon II and chipkits
Post by: Rcreator on October 25, 2011, 03:14:40 PM
alright thanks everyone
Title: Re: comparison between axon II and chipkits
Post by: Webbot on October 25, 2011, 03:38:52 PM
For me, the massive appeal of the Axon and webbotlib is crushed a bit by it's flaky Mac OS X support, whereas I believe the chipkit has a native IDE - and I have a nice PIC programming workflow set up already. But if Axon was ever to offer solid OS X support, I would almost certainly choose it over the chipkit for advanced robotics projects
The current 'flakey Mac OS X support' is really because there is NO ide - so folk use AVR Studio which is Windows only - or whatever other environment they choose. Mac etc will still work but its up to you to find all the missing tools (editors/compilers etc) -  ie the support is not 'flakey' - but rather it is 'not a complete environment'.

Am currently working on 'WebbotLib Studio' a java based IDE - which should therefore work on Windoze, Mac, 'nix etc.
Of course it will work with the Axon but also all the other boards currently supported.
Early days yet.
Title: Re: comparison between axon II and chipkits
Post by: joe61 on October 25, 2011, 03:42:36 PM
There's an Eclipse plugin for AVR programming as well. I've never tried it though so I don't know what it's like.

Joe
Title: Re: comparison between axon II and chipkits
Post by: WaterPig Master on October 25, 2011, 04:05:35 PM
Quote
ie the support is not 'flakey

You're absolutely sight — but it does irritate me that admin claims 'mac OS X compatability' for the Axon when actually, said 'compatability' is minimal! And 'documentation' is a few forum posts by people who've had a go at it.

Webbotlib studio sounds great! I'll be sticking to my cheapskate PIC setup at the moment, but I'll keep an eye out for this.

Cheers,
Barnaby
Title: Re: comparison between axon II and chipkits
Post by: Webbot on October 25, 2011, 04:13:15 PM
There's an Eclipse plugin for AVR programming as well. I've never tried it though so I don't know what it's like.
I use it - it's not great and has loads of bugs. Try having lots of separate Java projects open as well and then save a single 'C' file from your AVR project - then sit back and watch the sun set whilst your computer rebuilds every single project from scratch. Aargh. Look at all the project toolchain options - webbotlib supports about 10 processors so if you want debug and release versions from the same set of source code then its a nightmare to configure. Change one C file and wait whilst it rebuilds all 20 versions. I don't use it anymore. Eclipse is fine for 'professionals' but its a big ugly download to keep up-to-date and properly configured for newbs.

WebbotLib Studio will also have Project Designer built into it. So 'drag' a new motor controller onto your project - this will regenerate any code, compile and easily upload to the mcu, and run. All from the same IDE.
The code editors have syntax highlighting and will be linked into the existing help system. Code assist should therefore be possible ie just type in 'motor.' and it will list all the functions (methods) you can do to the motor.

But the Eclipse plug-in is available today - blast it.
Title: Re: comparison between axon II and chipkits
Post by: joe61 on October 25, 2011, 04:20:15 PM
Well, I wasn't necessarily suggesting it. I've used Eclipse for C/C++, PHP, and other stuff, and I don't really like it much either (although it's pretty good with PHP and XML stuff). I'm a dinosaur though. I think emacs is an IDE.

Joe
Title: Re: comparison between axon II and chipkits
Post by: Webbot on October 25, 2011, 04:43:47 PM
— but it does irritate me that admin claims 'mac OS X compatability' for the Axon when actually, said 'compatability' is minimal! And 'documentation' is a few forum posts by people who've had a go at it.

Its not my role to defend admin and I'm sure he will have his own view. But here is my 2 cents worth: admin has created some hardware products (the Axon range) which are microcontroller boards. They don't use windows, mac, 'nix or ANY other operating system. Same for all (well 95%) of all microcontroller boards on the market. So they aren't compatible with ANYTHING.

Here's what I think you mean by 'compatible'-
To program any board from any manufacturer you need to have a development environment: some kinda text editor to write your code, some kinda compiler to turn it into micro controller code, and some kinda programmer/loader to burn it onto the board. Admin doesn't provide this development environment himself coz others have already written the various elements and coz he's a hardware guy. So when talking about 'support' for any OS it just comes down to finding all the relevant 3rd party bits that are out there on the web for each of the stages I've mentioned. Lets not under-estimate the number of combinations - ie 'how many operating systems' x 'how may text editors' x 'how many compilers' x 'how may hardware programmers and boot loaders' = a very large problem. Just check out this forum for the number of posts just related to the AVRISP MKII programmer - none of it is 'admins hardware' problem. Support can become a nightmare. Try asking ATMEL, who produce all the chips, to recommend a development environment that works on all computer operating systems - they wont be able to give you an answer either.

Admin owns the hardware/board - what goes on at the PC/Mac/Unix end is up to you - but the forum tries to bring together peoples experiences of the tools that are out there.

Then there are other folk like me. Just another forum member. My 'expertise' is in writing libraries for the boards (webbotlib) and also in trying to create computer based apps (Project Designer etc) to make life easier for the newbie to get going. The code i create is in java and so should run on Win/Mac and Unix. But I'm just a hobbyist doing this for fun and don't have access to 'one of everything'. So can I test every version of windows xp/win7, mac, unix, different programmers etc := no. But collectively we can and so life gets easier - until Mr Gates et al create new versions of Windows that no longer work etc etc. It will always be ongoing.

But I guess what you are trying to get at is having a Mac environment that lets you easily write projects, compile and upload them without hunting around for all the required bits. Thats where I am going with WebbotLib Studio.
Title: Re: comparison between axon II and chipkits
Post by: Webbot on October 25, 2011, 04:45:39 PM
Well, I wasn't necessarily suggesting it. I've used Eclipse for C/C++, PHP, and other stuff, and I don't really like it much either (although it's pretty good with PHP and XML stuff). I'm a dinosaur though. I think emacs is an IDE
Joe - I've used Eclipse every working day for around the last 10 years. I think its great. But the AVR plug-in from a 3rd party is the thing that is full of bugs and drives me mad.
Title: Re: comparison between axon II and chipkits
Post by: joe61 on October 25, 2011, 04:57:06 PM
Here's what I think you mean by 'compatible'-
To program any board from any manufacturer you need to have a development environment: some kinda text editor to write your code, some kinda compiler to turn it into micro controller code, and some kinda programmer/loader to burn it onto the board. Admin doesn't provide this development environment himself coz others have already written the various elements and coz he's a hardware guy. So when talking about 'support' for any OS it just comes down to finding all the relevant 3rd party bits that are out there on the web for each of the stages
It might also be said that avrdude was originally written for the Mac, and there are lots of text editors around. A little effort in learning how makefiles work, and you have avr support on a Mac.

There's a tutorial on how to set that up on ladyada's web site: http://www.ladyada.net/make/usbtinyisp/avrdude.html (http://www.ladyada.net/make/usbtinyisp/avrdude.html)

Quote
But I guess what you are trying to get at is having a Mac environment that lets you easily write projects, compile and upload them without hunting around for all the required bits. Thats where I am going with WebbotLib Studio.
My brain just caught up with this. I haven't tried your stuff yet because I so rarely boot into Windows. I might have to learn an IDE now...

Thanks

Joe
Title: Re: comparison between axon II and chipkits
Post by: Soeren on October 25, 2011, 05:02:35 PM
Hi,


I agree with joe61's Axon vs Arduino comparison comments.
Yes, you've said so a few times before and when comparing with the original Arduino, I somewhat agree on a couple of your arguments :)
However, this was more a comparison between the Axon II and the chipKIT Max32, so it would be much more interesting to hear your comments on that and even on the Axon II vs. chipKIT Uno32 (where you can use your "4 times" rule, as it's priced 4 times lower than the Axon II ;)).


And as for cost, sometimes it's worth spending more if it can save you hours of wiring and frustration - but that depends on how much you value your time. :P
Banging a "shield" with header pins on top of another board doesn't really take much time (or money)


This may or may not matter, depending on your project, but you should also consider the compactness and size of the boards.
I thought you were all for the modular approach (which btw. the Arduino concept is a good example of).
It's hard to modularize anything, if all the stuff have to go to a single small PCB and while the small size is good if you're making tiny robots, it won't matter for a larger 'bot.


Or if you're running on battery, the energy efficiency.
I haven't been able to find the power drain for the Axon II, but it would surprise me if it wasn't higher on a 32bit 80HHz controller compared to a 8bit 8MHz controller.
However, the exact drain of the controller is of very little concern in a robot running sensors and actuators
(The 32bit PIC goes up to 90mA at 80MHz and 6mA at 4MHz when none of the "running power" saving modes are used).


Sometimes you have a ton of serial devices that need to be connected, so you need as many UARTs as possible.
Perhaps the 6 UART modules, the 4 SPI modules, the 5 IIC modules, the CANbus controller, the Ethernet controller or the IrDA with hardware encoder/decoder of the PIC32 will be of assistance  ;D


If you only had a few servos on your bot, running a few jumper wires isn't a big deal. But if you were building a bot with 20 or 30 servos like a hexapod or biped, you'd be much better off with built-in power buses.
That would need to be a substantial powerbus. 30 servos with a drain of up to at least 1A a pop and...
Quote from: axon2_datasheet
The Axon II has been tested to handle at least 6A on each power bus, but can in theory safely handle up to ~13A.
So, the Axon II won't be able to handle 30 regular size servos, according to your own datasheet.
A true modular approach will lessen the demand of any power bus, as they won't have to handle the total current of all involved consumers.


Again, as joe61 said, it depends a lot on your project for what is important. Even more, if you plan to build more robots in the future, consider that your parts will be scrapped and reused - can your mcu choice handle more complex bots in the future?
Yes, I believe a 32bit controller will be able to handle more complex tasks than an 8bit controller. Do you really disagree?


ps - as Soeren said, I'm very biased in my comments ::)
You have a "right" to, of course, but there's a big difference between being biased towards ones own "child" and neglecting the obvious advantage of getting 4 times the bit-width controller board with 8 times the Flash memory, 16 times the RAM at less than half the price, or, settling for the chipKIT Uno32 with slightly less impressive data, at a 4 times lower price.
Title: Re: comparison between axon II and chipkits
Post by: Webbot on October 25, 2011, 07:12:52 PM
Soeren - I'm not going to answer the questions you've raised, nor those of anyone else, but would encourage a different view point....

What I like about the Axon series, especially the Axon II, is that its like a micro controller and a bread board all wound into one. The built in power bus means I can easily plug in servos, sensors, or hookup SparkFun like boards very easily with just some bought patch cables. I haven't needed to use a soldering iron for a long time. The Max32 board, like Arduino, doesn't lend itself easily to that.

Yes we could start by
Quote
Banging a 'shield' with header pins on top of another board doesn't really take much time (or money)
but most of us don't have the time or inclination when it just gets to getting something to work. After all (as the electronics guru) I'm sure you breadboard designs before committing to etching PCBs etc.

The advantage is that by continually working with one board is that you understand it better rather than using something new for every project. Having said that then once I've got stuff working I will often design a custom board for the finished design (maybe a smaller processor,  specific i2c connections etc etc).

Comparing the facilities of one chip against another (number of IOs, UARTs) etc is fine - but if you always want the best then you will be adopting a new one every week as the technology moves on. IMHO the Axons have more than enough  power for most things that most people are doing. A faster chip will just mean that its spending more time doing nothing.

Ah yes - but what about price - the Max32 is cheaper and faster? But where do you get support? I believe that this site has an abundance of support for AVRs - created in many ways due to admins hardware offerings. You can ask questions here about the Axon series and get a fast turn around whether it be a hardware or software question. Ask about the Max32 and I would imagine you would get a fraction of the response. That has to be worth something and Admin understandably may want to recoup his time/commitment to date by pricing his hardware accordingly.

You pay for what you get.
Title: Re: comparison between axon II and chipkits
Post by: Admin on October 25, 2011, 10:15:02 PM
Quote
— but it does irritate me that admin claims 'mac OS X compatability' for the Axon when actually, said 'compatability' is minimal! And 'documentation' is a few forum posts by people who've had a go at it.
This is definitely a fair argument. Webbot's response is also pretty good. I plan to rewrite/improve a bit of my documentation for the Axon series soon, so I'll incorporate better instructions for Linux/Mac. I don't use either (I have Linux installed and use it on rare occasion, I just like Windows more . . . cough).

Quote
this was more a comparison between the Axon II and the chipKIT Max32, so it would be much more interesting to hear your comments on that and even on the Axon II vs. chipKIT Uno32 (where you can use your "4 times" rule, as it's priced 4 times lower than the Axon II ).
disclaimer: I haven't used the chipKIT
If it is 100% Arduino compatible that is good, as it already has a decent sized community. Obviously it's processor specs are bigger:
80 Mhz 32-bit MIPS
512K Flash, 128K SRAM
compared to the Axon:
16 MHz 8-bit MIPS
64K Flash, 8K RAM

But in reality, you don't need more than ~15k Flash, 1k RAM, and 1MHz for a decent line follower robot. I've never programmed a robot that needed more than 2K of RAM (except for a mapping robot in which RAM/EEPROM must be fully utilized). 16 MHz will give you 230kbps baud for your UART, more than enough. 80MHz is still too slow for computer vision or any other fancy computer algorithms, but it will help if you're doing inverse kinematics (if you don't know what that is, you don't need it lol). As for flash, 64K is a LOT. I've never used more than ~45k in my life, although one owner of the Axon told me he exceeded 64K so had to fiddle with his program to keep it under (he was doing some crazy IK stuff with it).

The Arduino Mega has the *exact same* I/O as the Axon, while the chipKit PIC32MX795F512 appears to be imitating the exact same I/O. But it also has a real time clock (awesome!), CAN (anyone even care about that?), IrDA coder (useless for robots; I made one myself for ~$5), and ethernet (not that great for most robots as it requires a wire to be attached).

If the chipKit imitates the Arduino power supply (I can't find the schematic), then you need two batteries to use it - one for the mcu and another 6V for servos. The Axon just needs one 6V battery. The chipKit also lacks power headers, forcing you to either make for yourself or buy the $20-$30 shields.

If I had to choose between the Arduino and the chipKit, I'd trash the Arduino in an instant. It's frankly garbage in my opinion.  ::) The chipKit however is a much harder comparison to make . . . again, it depends on what your project requires.

Quote
I thought you were all for the modular approach (which btw. the Arduino concept is a good example of).
It's hard to modularize anything, if all the stuff have to go to a single small PCB and while the small size is good if you're making tiny robots, it won't matter for a larger 'bot.
With the Arduino, you have the option of buying expensive shields, or wiring up a power bus and finally attaching your gizmo to it. With the Axon, you just attach your gizmo to it without the extra cost of a shield. A GPS just needs 3 wires attached to the Axon and you're done, but with the Arduino you need to spend 50% more on a GPS shield. The Arduino can't have more than one shield at a time, another drawback.

Quote
I haven't been able to find the power drain for the Axon II, but it would surprise me if it wasn't higher on a 32bit 80HHz controller compared to a 8bit 8MHz controller.
The waste is from the choice in voltage regulator. The Axon uses an LDO that works as low as 5.3V, whereas the Arduino regulator requires a 7.2V battery (they use an el-cheapo regulator to keep costs down). The efficiency is much higher for the Axon. I don't know what regulator the chipKit uses so can't comment . . . but I'm betting it's expecting a ~4V+ supply in with a 3.3V out at 90mA just for the processor.

Quote
So, the Axon II won't be able to handle 30 regular size servos, according to your own datasheet.
That's if you use only one battery (which isn't likely, considering you need 30A). You can distribute the batteries across the power bus, so if you use two batteries spaced out, no trace would see more than ~8A.

trace: 8A->battery<-8A-8A->battery<-8A

Quote
Yes, I believe a 32bit controller will be able to handle more complex tasks than an 8bit controller. Do you really disagree?
Unless you need some serious precision math or large memory for mapping/photos, there isn't much a 32bit can do that an 8 bit can do with respect to a homemade robot. If the goal was to make a computer, I'd say none of these boards are the appropriate choice . . .

Quote
at a 4 times lower price.
A better comparison is with the chipKit Mega board, which has an equal amount of I/O and is $50 before factoring in making/buying a powerbus and additional battery.

Besides, Project Designer (which btw is the most awesome robot programming tool ever) doesn't support the chipKit :P ::)
Title: Re: comparison between axon II and chipkits
Post by: WaterPig Master on October 26, 2011, 02:03:05 PM
Quote
This is definitely a fair argument. Webbot's response is also pretty good.

Updated documentation sounds good. I'm not complaining about lack of documentation/support for Mac users (when it comes to electronics/robotics, we're used to it  ::)), merely the fact that you're implying that the axon is equally well supported on all the platforms.

I suspect that your axon mote will become the more prolific product, as it's more unique. But IMHO the thing that really sets the products apart is the software used to develop them. Take that away and theyre all pretty much of a muchness, albeit in different shapes.

Quote
But I guess what you are trying to get at is having a Mac environment that lets you easily write projects, compile and upload them without hunting around for all the required bits. Thats where I am going with WebbotLib Studio.

That is indeed what I'm after and sounds marvellous! Looking forward to its release, might finally buy an axon when it is released! This is what I mean by software support defining usefulness.

Cheers,
Barnaby
Title: Re: comparison between axon II and chipkits
Post by: Cristi_Neagu on October 27, 2011, 08:12:15 PM
It would be great to see WebbotLib support for the max32. I'm currently designing in my spare time a educational robotic platform (industrial robot, that is), that would support a 6 DOF robot, at least 2 other driven axes and vision via Blackfin. Since we are talking about inverse kinematics, at least 6 quadrature encoders, and at most 16ms processing time per step, I don't really know about the Axon... I mean... it can do it, in theory, but how well? On the other hand, the max 32 with 80MHz at 32 bits and 83 IO... that'd be super. But since I want the final product to be easy to use, the work environment provided by WebbotLib is a must.

Also, Admin... Don't you have some plans for something more powerful than the Axon?
Title: Re: comparison between axon II and chipkits
Post by: Webbot on October 28, 2011, 11:31:06 AM
It would be great to see WebbotLib support for the max32
I agree - but adding the PIC architecture, or ARM for that matter, to the existing AVR wouldn't be a simple task.
With the future WebbotLib Studio also providing toolchain support then it would also mean finding open source compilers, programmer/loader code etc.
I'd love to take over the world - but just not enough time during my coffee break  ;D