[Diy_efi] Feedback - keep it coming :-)

Steven P. Donegan steve
Mon Feb 26 18:26:26 UTC 2007


I am open to all feedback - and will adjust the design as seems
appropriate. The Super cap was strictly for baby UPS capability -
although it will indeed deal with low frequency garbage quite well. I
will indeed adjust my current design, which in most parts mirrors
Megasquirt already (which works), based on the group's inputs. If I
didn't want feedback I wouldn't be doing this in a public forum - so no
worries :-) The bridge was simpler, single component answer to the more
complex belt and suspenders approach the megasquirt folks took (and that
is not in any way intended to be a negative comment about megasquirt -
they are the reason I am doing this project - ie my inspiration). And
given that the system runs on 5V and 3.3V the battery would have to be
at least 2 dead cells down before the voltage drop across the diodes
would really make any difference. At that point there is not enough to
fire injectors or coils properly so it's kind of moot... I may also go
to more than 2 layers - at least on the full ECU design. The PIC based
design given it's low frequency operation and full shielded case design
probably would not gain anything by more than 2 layers... And I already
have my circuit design software routing based on shortest traces and
fewest number of vias as design goals...

And - BTW - I am a hobby type - my daily job is designing systems and
software, not boards - so in the PCB area I am an amateur compared to
many other folks who do that type of work for a living. So by all means
keep the feedback coming!!!



On Tue, 2007-02-27 at 07:09 +1300, Bevan Weiss wrote:
> Hi Steven,
> I think having a super cap is a good idea on the microcontroller
supply 
> lines (to act as a temporary power source) however am not sure why 
> you're using a bridge (bridge rectifier I assume).  If you're after 
> reverse polarity protection then you should really just use a single 
> series diode. This ensures that you have the reverse protection 
> available, but also that the system won't operate with the polarity 
> swapped, a second plus for this configuration is the reduced voltage 
> drop due to only one rather than two series diodes.
> 
> I would still add at least one 1000uF capacitor, as well as several 
> smaller capacitors.  The shotgun approach is the best (cover a range
of 
> capacitor values), unless you're willing to perform some more 
> complicated frequency analysis of your capacitor spread.  The
frequency 
> analysis approach is generally used when it's important to minimize
BOM 
> cost through using the minimum number of capacitors.  Shotgun
approach 
> often results in additional capacitors being used than are really 
> required, but this isn't normally a problem.
> 
> Is there any particular reason you're going for a super cap on the
RTC 
> rather than using a more standard coin cell battery?  I haven't
looked 
> into temperature ratings for super caps lately, but last I looked
they 
> weren't great.  Something like -5degC to 45degC..  They also have
quite 
> high ESR making them bad for ripple reduction as Will has already
mentioned.
> 
> I applaud you for taking on this design in such a public manner.
Sorry 
> if it seems like we're telling you how to suck eggs, just want to
make 
> things easier for you :)
> 
> 
> Bevan
> 
> 





More information about the Diy_efi mailing list