Looking for a bit of guidance

Chris H wrote:
> In message <MIudndHpR67NtubUnZ2dnUVZ8vidnZ2d@lyse.net>, David Brown 
> <david.brown@hesbynett.removethisbit.no> writes
>> However, you should note that there is good reason for there being 
>> limited information about the mpc5xxx, and limited response to your 
>> questions and posts.  These devices are not particularly widespread - 
>> there are few developers that use them.
> 
> I would not agree there... In some industries it is very widely used.
> 

Yes, but these are mostly areas where there are a small number of 
developers, even though the quantities of devices are large.  They are 
not nearly as popular as, for example, their predecessor the 68332, nor 
related devices such as the MCF523x (which have a ColdFire core, and a 
similar TPU).

>> Those that *do* use them are often in companies that have longer term 
>> projects with larger development groups, and most of their questions 
>> and information are held within these groups - the automotive industry 
>> is a prime example, and a major user of these devices.  The other 
>> chips you mentioned are much more common among smaller developers and 
>> hobby developers, so there tends to be a lot more sharing of 
>> information and questions.  This is why your are more likely to find 
>> the information you need from "professional" sources - training 
>> courses, and information and examples coming with expensive commercial 
>> development tools and evaluation boards, rather than "amateur" or 
>> "low-cost" sources such as web pages, hobby boards, and so on.
> 
> I agree.  Most people using PPC  are in large(r) companies have direct 
> contacts in Freescale and ST.
> 
> 
>> Depending on how far the project has come so far, you should be trying 
>> to do a cost-benefit analysis as to whether to continue using this 
>> chip, or moving to a different architecture.  The mpc5xxx are 
>> expensive devices, and expensive for development.
> 
> Quite so. I would recommend a cost benefit analysis.   Take your time 
> rushing in now could be a fatal and expensive  mistake.
> 
>>  If you need the features (such as the TPU units), they are excellent 
>> value for money, but if not, you are better off with an alternative.
> 
> There is only one TPU compiler available AFAIK (se stock them :- )  and 
> the TPU is a specialist area on it's own.
> 

I understand that development for the TPU is much easier these days with 
the eTPU and ByteCraft's compiler.  But the great majority of TPU users 
are just that - users.  For example, we have used TPU's in the 68332, 
MPC561 and MCF5234 for functions such as synchronised PWMs, encoder 
inputs, and UART channels without doing any TPU development as such. 
It's even easier now with Freescale's online tools for packaging the 
required functions in a ROM image.  *Developing* TPU functions looks 
like a lot of fun, but I haven't found a good excuse for doing so yet.

Chr15 wrote:

> I hear everything your saying.  But, sadly, I don't have the option of
> just simply, 'handing' the chip back, 

I don't think anybody said anything about it being _simple_ to reject 
the project in its currently defined shape.  Nevertheless, we have 
strong agreement that it has to be done.

So somebody has made a technically wrong decision.  That leaves you with 
two main options: point out this fact now, at the risk of making 
somebody look stupid to their peers (and maybe angry at you), or point 
out this fact later as an explanation why your company sank so much 
money into a doomed project.  That one would make you look stupid.  It's 
put up or shut up time.

The next step of escalation may have to be to switch into "defensive 
engineering mode" on this.  I.e. stop worrying about the project as 
such, and start worrying about how you're going to prove it wasn't your 
fault.

You'll have to excuse me if I don't get back to everyone's concerns, but I
think I've done a good job in summing up :)

>That was my first thought too.   A PPC is a hell of a system to learn if

>all you know are PICs and H8's.  Who ever gave you this project is 
>either an idiot or trying to get you fired.

I don't disagree with any of you in this regard, the more I read through
the manual the more complicated it becomes, not to say I don't want to
learn about it; the whole reasoning behind me posting here actually :)

Im well aware of the scarceness of this processor due to this mostly
professional use, and I've tried to explain a few times how over board this
processor is going for what they want to do.  I don't think its either
idiocy or someone wanting to cause me grief, but the individual that chose
this processor lack of understand of embedded tasks (its for a College
research group, not comerical/industrial use so cost isn't a factor).

This project is more of a one off project, with hardware already built up
around the processor; without successfully running code on it(wish I could
insert a facepalm smiley here), not a practice I encourage or agree with. 


>Do you have any connection to Freescale or ST?

>The PPC parts are wonderful parts and very powerful I have used them, 
>The part seems a bad fit for the job in hand.

The group did have a connection to ST, but it turned out not as strong as
I'd had hoped (The MPC56xx looks to be new-ish and different enough than
the 55xx to make it unique, thus not a lot of support on their end) as for
Freescale I did manage to create a tie with the local distributor in order
to obtain a [donated] evaluation main board which is a huge step in the
right direction (considering this group only had access to the mini board
prior to me being here).  

I fully agree with every ones opinions in that this part is way, way,
overkill for the task asking to be done.  Even hardware issues(pad spacing
in relation to the clock frequencies being ran causing EMC issues on user
built boards, for example) have been quite extensive.  I think this part
was chosen because of its common use in automotive applications (this small
group focuses on Hybrid and Power electronic research).

>>I hear everything your saying.  But, sadly, I don't have the option of
>>just simply, 'handing' the chip back,

>Why?

As I should have stated before, hardware has already been built up around
the processor, causing the research group to be kind of locked into this
choice.  And me being the 'low man' on the totem poll so to speak (as well
as coming in after the choosing of this processor), I figured I would try
to 'roll' with the punches in order to try and figure out what needed to be
done.  

I am well aware of the task ahead (and its possible impossibility at my
level of experience), but hopefully I've explained my self a little bit
better than in my original post :)  I think I'll start talking to my ST/FS
contacts more.

The goal of the device is to implement a PI based phase shift control of a
PWM wave, according to current/voltage/temperature AD readings, to charge a
high voltage battery off the power line, as well as have some CAN
communication with other devices.  

On Jan 25, 5:39=A0am, Chris H <ch...@phaedsys.org> wrote:
> In message
> <ebf7d25b-42ce-4478-8be3-ae8c5d990...@w1g2000prk.googlegroups.com>,
> steve <bungalow_st...@yahoo.com> writes
>
> >On Jan 23, 6:57 am, "Chr15" <csema...@gmail.com> wrote:
> >>re.
>
> >We use the 5xxx processors everywhere, =A0I would suggest buying an eval
> >board, most of your questions (and ones you haven't thought of yet)
> >will be answered quickly just getting the eval board up an running
> >(which should be very quick) and looking at the sample code provided.
> >The freescale forums are where you need to be
> >http://forums.freescale.com/freescale/board?board.id=3DMCUCOMM
>
> Do you have any connection to Freescale or ST?
>

no.

> The PPC parts are wonderful parts and very powerful I have used them,
> The part seems a bad fit for the job in hand.

I disagree,  I think everyone here is confusing the 560xP series with
it big brothers, the 560xP is a cheap small (as small as 64 pin LQFP)
bare bones device with a z0 power core, no MMU, no TPU, no floating
point,  no cache, max 60Mhz,  with just a handful of peripherals, it's
not a big 500 pin ball grid array like it big brothers.

On Jan 26, 11:12=A0am, "Chr15" <csema...@gmail.com> wrote:
> You'll have to excuse me if I don't get back to everyone's concerns, but =
I
> think I've done a good job in summing up :)
>
> >That was my first thought too. =A0 A PPC is a hell of a system to learn =
if
> >all you know are PICs and H8's. =A0Who ever gave you this project is
> >either an idiot or trying to get you fired.
>
> I don't disagree with any of you in this regard, the more I read through
> the manual the more complicated it becomes, not to say I don't want to
> learn about it; the whole reasoning behind me posting here actually :)
>
> Im well aware of the scarceness of this processor due to this mostly
> professional use, and I've tried to explain a few times how over board th=
is
> processor is going for what they want to do. =A0I don't think its either
> idiocy or someone wanting to cause me grief, but the individual that chos=
e
> this processor lack of understand of embedded tasks (its for a College
> research group, not comerical/industrial use so cost isn't a factor).
>
> This project is more of a one off project, with hardware already built up
> around the processor; without successfully running code on it(wish I coul=
d
> insert a facepalm smiley here), not a practice I encourage or agree with.
>
> >Do you have any connection to Freescale or ST?
> >The PPC parts are wonderful parts and very powerful I have used them,
> >The part seems a bad fit for the job in hand.
>
> The group did have a connection to ST, but it turned out not as strong as
> I'd had hoped (The MPC56xx looks to be new-ish and different enough than
> the 55xx to make it unique, thus not a lot of support on their end) as fo=
r
> Freescale I did manage to create a tie with the local distributor in orde=
r
> to obtain a [donated] evaluation main board which is a huge step in the
> right direction (considering this group only had access to the mini board
> prior to me being here). =A0
>
> I fully agree with every ones opinions in that this part is way, way,
> overkill for the task asking to be done. =A0Even hardware issues(pad spac=
ing
> in relation to the clock frequencies being ran causing EMC issues on user
> built boards, for example) have been quite extensive. =A0I think this par=
t
> was chosen because of its common use in automotive applications (this sma=
ll
> group focuses on Hybrid and Power electronic research).

Just a comment, we have to use PPC (or similar) in Hybrid/Power/Auto
applications because we are required to use safety processors
(processors with EEC correction on FLASH and RAM, for instance) which
AVR's/PIC's/H8 etc don't have. If you work in this area you have to
become familar with these type of "industrial" processors (vs the
hobby/commercial processors).

On Jan 26, 9:48 am, steve <bungalow_st...@yahoo.com> wrote:
> On Jan 26, 11:12 am, "Chr15" <csema...@gmail.com> wrote:
>
>
>
> > You'll have to excuse me if I don't get back to everyone's concerns, but I
> > think I've done a good job in summing up :)
>
> > >That was my first thought too.   A PPC is a hell of a system to learn if
> > >all you know are PICs and H8's.  Who ever gave you this project is
> > >either an idiot or trying to get you fired.
>
> > I don't disagree with any of you in this regard, the more I read through
> > the manual the more complicated it becomes, not to say I don't want to
> > learn about it; the whole reasoning behind me posting here actually :)
>
> > Im well aware of the scarceness of this processor due to this mostly
> > professional use, and I've tried to explain a few times how over board this
> > processor is going for what they want to do.  I don't think its either
> > idiocy or someone wanting to cause me grief, but the individual that chose
> > this processor lack of understand of embedded tasks (its for a College
> > research group, not comerical/industrial use so cost isn't a factor).
>
> > This project is more of a one off project, with hardware already built up
> > around the processor; without successfully running code on it(wish I could
> > insert a facepalm smiley here), not a practice I encourage or agree with.
>
> > >Do you have any connection to Freescale or ST?
> > >The PPC parts are wonderful parts and very powerful I have used them,
> > >The part seems a bad fit for the job in hand.
>
> > The group did have a connection to ST, but it turned out not as strong as
> > I'd had hoped (The MPC56xx looks to be new-ish and different enough than
> > the 55xx to make it unique, thus not a lot of support on their end) as for
> > Freescale I did manage to create a tie with the local distributor in order
> > to obtain a [donated] evaluation main board which is a huge step in the
> > right direction (considering this group only had access to the mini board
> > prior to me being here).
>
> > I fully agree with every ones opinions in that this part is way, way,
> > overkill for the task asking to be done.  Even hardware issues(pad spacing
> > in relation to the clock frequencies being ran causing EMC issues on user
> > built boards, for example) have been quite extensive.  I think this part
> > was chosen because of its common use in automotive applications (this small
> > group focuses on Hybrid and Power electronic research).
>
> Just a comment, we have to use PPC (or similar) in Hybrid/Power/Auto
> applications because we are required to use safety processors
> (processors with EEC correction on FLASH and RAM, for instance) which
> AVR's/PIC's/H8 etc don't have. If you work in this area you have to
> become familar with these type of "industrial" processors (vs the
> hobby/commercial processors).

If they are serious about safety, they should have used ROM.  Line
charging of battery is hardly safety related, unless you have thousand
feet of power cord.