Actions
Forum Thread Details
  • Replies 17 replies
  • Subscribers 2567 subscribers
  • Views 1583 views
  • Users 0 members are here
Related

Freescale RD3803MMA7660FC review issues

gandrewstone

Hi 7660 reviewers,

 

I've been holding off on my review because I haven't had time to put the device in my own circuit.  Lack of code really is the issue isn't it?  I mean, frankly all that software they gave us and the USB->serial->i2c board is just a cute toy.  Or more realistically they probably made all that for marketing reasons and figured that it could be pawned off as a dev kit.


It would be pretty easy for them to give us a nice API that calls into an unimplemented I2C "underface".  What I mean is that they wouldn't provide the I2c implementation; all the uPs have those anyway... they just provide calls into a "prototypical" I2c library (read, write, etc) and you would translate their calls into your uP's I2C library.


I've been holding off on my formal review on my blog until I can get some code written.  But its interesting I kind of feel like they are getting free labor here... I mean its worth more then $35 bucks of my time to write that code.  But it bugs me that Freescale has clearly already written that code, but for some reason only released an incomplete header file (for example, it does not mention that the I2c address is 0x4c).  Ultimately Freescale is really losing out; the quality of our reviews would be much better and would come out much sooner if they had provided that code.

 

Here are my initial observations for you other reviewers, summarized.

 

0. No code: see rant above :-)


1. There seems to be no way to program the uP on the USB->Serial board.  So I can't change the code to try out my application.


2. The USB protocol is documented, so I could write my program to run on the laptop and use that protocol.  But that would be a lot of throwaway code, and not really test issues like whether my uP can handle the data stream.


3. Putting the chip on its own .1" daughterboard is great!  I can prototype with it.


4. You can buy that little daughter board separately for $35, vs $130 for the whole kit.  Since that's really the only valuable part for real engineering eval, that's what I would buy.


5. I have a handy 5v uP, other uPs have 1.8v.  Wouldn't it have been a lot more useful if their helper board did logic level conversion instead of USB to serial?


In sum, seems like a nice little chip, with a poorly conceived "development" toolkit.  I guess the magic of open source will eventually fix that, esp. if I get some time this weekend! :-)


If you have managed to upload code to the uP or solve any of these other issues, send me a message!

 

Cheers!

Andrew

Parents

  • Now the details

     

    Here's the link where you can find all the necessary resources for building the JB16 BDM debugger (like in my case) http://forums.freescale.com/t5/OSBDM-JM60-Open-Source-BDM-for/USBDM-Version-2-JB16-Version/td-p/49363

     

    Thanks to pgo, the Freescale forum user who spent so much of its time having this neat thing working and being open source).

     

    As a personal advice please check the Freescale website for available samples. JB16 is still in sampling program so you can have a few chips of this kind. I personally got more than one and used some other in a small USB I/O expander project of my own. Quite a handy tool.


    I want to try to build the OSBDM version of the debugger since this kind of debugger supports some other families of Freescale uCs (HCS08, RS08, Coldfire V1 - V4, and the 56F series of DSCs). I think the idea of having a unified debugger is great!

    The OSBDM is based on the MC9S08JM60 uC, which is in sampling program as well. So, if you want a simple version of the debugger, go with the JB16 version. On the other hand the OSBDM is more versatile but the final decision on which one to start is yours.

    Being open-source, you have access to the firmware source code as well as the host DLLs. And there's support for Linux as well.

     

    Keep in touch

     

    Bogdan

Reply

  • Now the details

     

    Here's the link where you can find all the necessary resources for building the JB16 BDM debugger (like in my case) http://forums.freescale.com/t5/OSBDM-JM60-Open-Source-BDM-for/USBDM-Version-2-JB16-Version/td-p/49363

     

    Thanks to pgo, the Freescale forum user who spent so much of its time having this neat thing working and being open source).

     

    As a personal advice please check the Freescale website for available samples. JB16 is still in sampling program so you can have a few chips of this kind. I personally got more than one and used some other in a small USB I/O expander project of my own. Quite a handy tool.


    I want to try to build the OSBDM version of the debugger since this kind of debugger supports some other families of Freescale uCs (HCS08, RS08, Coldfire V1 - V4, and the 56F series of DSCs). I think the idea of having a unified debugger is great!

    The OSBDM is based on the MC9S08JM60 uC, which is in sampling program as well. So, if you want a simple version of the debugger, go with the JB16 version. On the other hand the OSBDM is more versatile but the final decision on which one to start is yours.

    Being open-source, you have access to the firmware source code as well as the host DLLs. And there's support for Linux as well.

     

    Keep in touch

     

    Bogdan

Children
  • in reply to bbolo1

    Hello,

     

    just wanted to point out that the OSBDM in its present form is "unified" only in hardware, but the firmware has to be reprogrammed every time you want to work on a different type of target.

     

    The USBDM by pgo has a single firmware that works for HCS08, HCS12, Coldfire V1 (CF1), but the OSBDM has to be reprogrammed with different firmware for each of these target families. And naturally you need some other BDM pod to do the reprogramming ...

     

    best regards,

    Robert

  • in reply to Former Member

    Hi Robert,

     

    I have only built the 908JB16 debugger so far. The OSBDM (having the JM60) would be a much better choice due to at least the higher clocking. The idea of reprogramming the OSBDM with another pod anytime you need a target change (different CPU family) seems a little bit odd. A bootloader in the OSBDM would do a much better job. I have found a document on the Freescale website detailing the OSBDM and its internals.

     

    http://cache.freescale.com/files/soft_dev_tools/doc/user_guide/OSBDM-JM60_Users_Guide.pdf?fsrch=1

     

    Have a look onto page 16 on this document. Jumper J3 is used for bootload mode enable. So, I don't see a need for a separate pod to do the OSBDM firmware reprogramming.

    I am planning to build/buy the OSBDM since I am working quite often with a few CPUs from Freescale. I'll let you know about it when I'll have it. Maybe write a review about it ... image Would this be of any use for the other members?

     

    Cheers,

    Bogdan

  • in reply to bbolo1
    Hello Bogdan,

    You're right, I forgot about the bootloader !

    But in the present source code description of the OSBDM one does have to
    reprogramm the firmware for each target family. I downloaded the
    Codewarrior project with the source code, and you have to select the
    target family and recompile for switching between HCS08, CF1 etc.

    I built the JM60 version of pgo's USBDM and it's very fast -- almost seems
    faster than the commercial PEMicro USB-BDM, plus it's nice to have the
    target voltage supply option. (That's another point where pgo's USBDM is
    ahead: it allows you to activate target VCC and select between 3.3V and
    5V in the software dialog, whereas the Freescale OSBDM does this with
    jumpers on the PCB, which is more complicated if one wants to put the
    BDM-pod in a plastic case ...)

    Best regards!
    Robert

    http://www.issole.net/jm60-USBDM.jpg
  • in reply to Former Member

    Hi Robert,

     

    The debugger looks so neat! I was thinking of OSBDM because I might try in the future the 56F series of DSPs. But this is not certain.

    Otherwise, the JM60 version of USBDM is quite enough for my needs. And its performance is even better than the PE's version.

    How much did it cost you to build your version? Looks very good! I might change my mind and go for the USBDM. image

     

    Regards,

     

    Bogdan

  • in reply to bbolo1

    The main expense is to have the pcb made (the mcu being such a fine pitch, one needs a "real" pcb with solder mask).

    The actual parts probably just add to 10 to 15 $.

     

    The other difficulty is that unless you are very good with a fine-tipped soldering iron, one needs a small reflow oven to solder in the mcu (and all the other smd devices go in at the same time), that in turn means one needs a stencil for the solder paste, so the pcb is rather difficult to solder without the stencil and a reflow oven.

  • in reply to Former Member

    You don't need a full oven as there are hand held hot air soldering equipment too.

    Chinese hardware is available for 50€ already and these with solder paste make soldering any SMD stuff a breeze.

    Cheapest reflow ovens are available for around 300€ but they usually are also shipped from China so you will have customs in addition.

    If you have the paste in an injector with a small needle you need steady hands to replace a stencil so this is not required either.

    We have successfully soldered for example LGA14 footprint chips from freescale to our CNC machined pcb's so they do not have any solder mask.

  • in reply to Former Member

    Hello Elias,

     

    I finally built myself a small reflow oven using a toaster oven controlled by a Freescale MCF51AC256ACLKEMCF51AC256ACLKE processor and some triacs to switch the heating elements and a thermocouple to measure the temperature. The total material cost was around 170 euros (depending on the toaster oven), plus some work ...

    At www.pcb-pool.com you can order prototype PCBs for ca 50 euros with a free solder stencil included.

    It was really worth it, because then one switches to using all surface mount parts and the assembly is much easier and more reliable.

     

    Best regards

    Robert

     

    image

  • in reply to Former Member

    Hi Robert,

     

    Toaster oven reflow is also a possibility indeed. But in my opinion an oven is required only if you start to do production batches so for prototyping a handheld hot air is enough. If you want to do multiple boards at the same time an oven is a must (unless you have no time constraints).

    Tricky parts are of course connectors like for microSD cards as they have plastics so you have to adjust the temperature to lower setting usually.'

     

     

    Elias

  • in reply to Former Member

    I successfully use a toaster oven for small numbers of hand made boards, and use hot air for re-work.

    I find hot air for assembling boards to take longer and be more error-prone - especially with fine pitch components, but I guess it depends which tools you are using - everyone's experiences will be different.

     

    PCB-Pool is a good place for getting boards made, they are always excellent quality and for the time being pool orders come with a free SMT stencil, but there's no guarantee how long that offer will last. However they are a bit on the expensive side, and their VAT rate is 21% and you need to pay in euros so my bank charges me an 'oversees' fee and a conversion commission on each payment too. I wish they would open a UK office-even if it was just for billing and payments.