Searching \ for '[PIC]: Component reduction...' in subject line. ()
Make payments with PayPal - it's fast, free and secure! Help us get a faster server
FAQ page: www.piclist.com/techref/microchip/devices.htm?key=pic
Search entire site for: 'Component reduction...'.

Exact match. Not showing close matches.
PICList Thread
'[PIC]: Component reduction...'
2002\06\14@115831 by Pic Dude

flavicon
face
Been trying hard to squeeze my PIC boost gauge (from some months
ago) into a round gauge enclosure, and that means elimination of
as many external parts as possible.  So I'm looking at every
component and thinking of why I really need it.  Converted the
circuit from 10 direct-driven LED's to 12 MUX'ed LED's (3x4 config)
and then this occurred to me last night...

http://www.avn-tech.com/stuff/mux_no_transistors.jpg

Not shown in the pic is that all LED cathodes are to the left.
Also, the dashed line represents the division between the main
board and the display board (yes, I need 2 round PCB's).

I could possibly eliminate the driver transistors since I have
extra ports to generate enough current.  The reason of tying the
lines together is that I can reduce the number of wires going
between the main board and the display board.

I think I'm not violating any PIC rules here:
- I can keep the tied-together lines at the same logic-level by
       keeping them on the same port.
- By multiplexing, only one set of 4 is on at any point in time,
     so I'm within the port's total max current.
- Also within the chip's max current.

BTW, I need the 16F872 since I need an A/D.  Nothing smaller w/flash.

I guess what I'm looking for is the experts' blessing on this.  Is
there something wrong here that I am missing?  Or shall I start
my PCB layout....?

Thanks,
-Neil.

--
http://www.piclist.com#nomail Going offline? Don't AutoReply us!
email spam_OUTlistservTakeThisOuTspammitvma.mit.edu with SET PICList DIGEST in the body


2002\06\14@121528 by Tan Chun Chiek

flavicon
face
Do note that power source/sunk by PORTA and PORTB is limited to 200mA,
shouldn't cause much of a problem if you are driving the LED within specs.
Since the LEDs are muxed, it's only switched on 1/12 of the time, you might
like to put a buffer and drive them harder to achieve the same brightness as
driving them directly.

Regards,
Tan CC

{Original Message removed}

2002\06\14@122450 by Pic Dude

flavicon
face
Actually, the LED's are on 1/3 of the time, not 1/12.

So far, I've found that muxing (with driver transistors) at
20mA is bright enough.  And actually too bright at night.
In this case, it should be the same, but I thought if it
got to be a problem, I could always drop to some 10mA LED's
etc.

Cheers
-Neil.



{Original Message removed}

2002\06\14@140641 by Doug Butler

picon face
From a BSEE with 19 years experience with uP stuff... Looks good, go for it!

Doug Butler
Sherpa Engineering


> {Original Message removed}

2002\06\14@171452 by Bob Ammerman
picon face
Looks li9ke it should work just fine.

Bob Ammerman
RAm Systems


----- Original Message -----
From: "Pic Dude" <.....picdudeKILLspamspam@spam@AVN-TECH.COM>
To: <PICLISTspamKILLspamMITVMA.MIT.EDU>
Sent: Friday, June 14, 2002 11:57 AM
Subject: [PIC]: Component reduction...


{Quote hidden}

--
http://www.piclist.com#nomail Going offline? Don't AutoReply us!
email EraseMElistservspam_OUTspamTakeThisOuTmitvma.mit.edu with SET PICList DIGEST in the body


2002\06\14@171656 by Bob Ammerman

picon face
> Actually, the LED's are on 1/3 of the time, not 1/12.
>
> So far, I've found that muxing (with driver transistors) at
> 20mA is bright enough.  And actually too bright at night.
> In this case, it should be the same, but I thought if it
> got to be a problem, I could always drop to some 10mA LED's
> etc.
>
> Cheers
> -Neil.
>

Neil,

If you really want to get cute you could sample the instrument cluster lamp
voltage (via a voltage divider) with another analog input and use it to
perform duty cycle brightness modulation of your LEDs.

Bob Ammerman
RAm Systems

--
http://www.piclist.com#nomail Going offline? Don't AutoReply us!
email listservspamspam_OUTmitvma.mit.edu with SET PICList DIGEST in the body


2002\06\14@173601 by Pic Dude

flavicon
face
I actually do have that in place as well, and unfortunately
I couldn't eliminate that transistor.  The original version
of this boost gauge used a DIP switch for dim/bright, and I
coded the park-light signal to that since it was simple.

Here's the real fun part -- my Audi has a lot of other red
LED's (which is what the boost gauge was built to match)
and somehow it detects dark, dims the LED's gradually, and
slowly raises the backlighting on the analog instruments.
Although I could use a separate photocell for this, it would
be really sweet if I could find that signal and use it for
the boost gauge as well.

Cheers,
-Neil.



{Original Message removed}

2002\06\17@053608 by Alan B. Pearce

face picon face
>Here's the real fun part -- my Audi has a lot
>of other red LED's (which is what the boost
>gauge was built to match) and somehow it detects
>dark, dims the LED's gradually, and slowly raises
>the backlighting on the analog instruments. Although
>I could use a separate photocell for this, it would
>be really sweet if I could find that signal and
>use it for the boost gauge as well.

What they may be doing is using the LEDs as photodiodes during the time they
are off in the multiplexing cycle. If this is the case you may find there is
a specific spot in the multiplex where all led's are off so that on ones do
not confuse the reading. If this is the case then there will not be a
separate photocell and signal available for you to get at, and you may just
have to confuse anyone trying to reverse engineer your product by doing the
same thing :))


hey Jack - why is this analogue input going to the LED array ??? You've made
a mistake on the circuit!!!

--
http://www.piclist.com hint: The list server can filter out subtopics
(like ads or off topics) for you. See http://www.piclist.com/#topics


2002\06\17@114006 by Pic Dude

flavicon
face
> ...using the LEDs as photodiodes...

Now there's a totally new concept to me.  From your description,
it sounds like using the existing LED's, since I don't see any
mention of any special LED's for this purpose.

I believe photodiodes need to be reverse-biased to operate.
Does this mean I can reverse-bias any LED and get it to behave
as a photodiode?

Now you've piqued my curiousity -- how, what, when, where!?!?!?

Cheers,
-Neil.




{Original Message removed}

2002\06\17@114211 by Pic Dude

flavicon
face
Trust me -- I am not about to reverse-engineer the Audi since
they do tooooo many customized things ..... the stereo is
internally connected to the instrument cluster, though I'm
not sure why.  And there is a full diagnostic system built
into the climate control system -- I can push some switches
and have it display vehicle speed, rpm, AC inlet temps, and
a whole slew of other things.

[shudder].


Cheers,
-Neil.



{Original Message removed}

2002\06\17@125823 by Alan B. Pearce

face picon face
>> ...using the LEDs as photodiodes...

>Now there's a totally new concept to me.  From
>your description, it sounds like using the existing
>LED's, since I don't see any mention of any special
>LED's for this purpose.

Ahh I can see a newcomer to the world of electronics, who never scraped the
paint off an OC45 to use it as a photo transistor, or never lifted the lid
on a 4K DRAM to use as a video sensor...

Yes I am talking of using an existing LED as the photosensor. Do remember
that any semiconductor device can be a photosensor, it is not restricted to
a photo diode per se. What does happen is that the characteristics of the
photo diode are set up to get repeatable parameters when used in this mode,
and it may not work as an LED when forward biased because the processing was
not done to enable this. However an LED will work as a photo diode if used
in a photo diode type circuit, and some say that if you use one with the
coloured package then you have a "prefilter" for the appropriate wavelength
of light. You will not get real good sensitivity, but when all you want is
"daylight on/off" to determine how bright the display should be then this
may well be a way to go.


>Now you've piqued my curiousity -- how, what, when, where!?!?!?

I believe this scheme is used in bedside alarm clocks (highly price
sensitive device) to minimise component count, but dimming the display so it
does not keep you awake at night with a brightly lit room, but making the
display bright enough in day time.

--
http://www.piclist.com hint: The list server can filter out subtopics
(like ads or off topics) for you. See http://www.piclist.com/#topics


2002\06\19@133310 by Peter L. Peres

picon face
On Mon, 17 Jun 2002, Pic Dude wrote:

>Trust me -- I am not about to reverse-engineer the Audi since
>they do tooooo many customized things ..... the stereo is
>internally connected to the instrument cluster, though I'm
>not sure why.  And there is a full diagnostic system built
>into the climate control system -- I can push some switches
>and have it display vehicle speed, rpm, AC inlet temps, and
>a whole slew of other things.

This reminds me of some articles about what could happen to you if
electronic appliances would be virused. The Audi has a CAN bus and
occasionally connects to computers which occasionally get updated,
presumably via Internet, right ;-) ?

Peter

--
http://www.piclist.com#nomail Going offline? Don't AutoReply us!
email @spam@listservKILLspamspammitvma.mit.edu with SET PICList DIGEST in the body


2002\06\19@150024 by Pic Dude

flavicon
face
Update...

Remember this thread?  Well, I tried some experiments with
this -- main code does the initialization and ends (endless
loop).  In the meanwhile, a TMR0 interrupt routine muxes
the display -- it dumps a specific hard-coded pattern to
the ports and cycles this as 3 "columns" of 4 bits.

The result is poor -- even though my calculations tell me
all is okay, and I have verified that I have no situation
where the bits tied together are out of sync with each
other, the resulting display is uneven in intensity, and
incorrect -- some LEDs are on that should not be.  Another
odd thing is that it waits about 5-10 secs before it
starts up, after power is applied.

I added a transistor driver to each of the 3 column lines
and it works properly -- nice bright crisp display, correct
output, and starts up instantly.  And yes, the transistors
were added to the point after the 4 pins were merged
together in each group.  I did this to reduce the # of
changes on each test.

The delay in starting up worries me -- I was worried that
it may be due to a conflict on the ports at startup time
BEFORE the ports got TRISed properly and cleared, but
that would not explain why it started up right away after
the transistors were added.

Can't explain why the poor results though.  I'm wondering
if to switch off the off columns (2 of 3 at any time) by
setting those pins to inputs so they float, vs. setting
them to low.  I'll have to experiment with this tonight.

Cheers,
-Neil.



{Original Message removed}

2002\06\20@034307 by Alan B. Pearce

face picon face
>This reminds me of some articles about what could
>happen to you if electronic appliances would be
>virused. The Audi has a CAN bus and occasionally
>connects to computers which occasionally get updated,

Practically every modern mass produced vehicle has a CAN Bus and it is
through this that a skilled person takes about 1/2 minute to clock the
speedo. Service diagnostic machines plug into this also to diagnose engine
tuning problems etc.

>presumably via Internet, right ;-) ?

Not necessarily, see my previous comment, the activity is done using a small
handpiece that plugs into the bus.

--
http://www.piclist.com hint: To leave the PICList
KILLspampiclist-unsubscribe-requestKILLspamspammitvma.mit.edu


2002\06\20@141354 by Peter L. Peres

picon face
On Thu, 20 Jun 2002, Alan B. Pearce wrote:

>>This reminds me of some articles about what could
>>happen to you if electronic appliances would be
>>virused. The Audi has a CAN bus and occasionally
>>connects to computers which occasionally get updated,
>
>Practically every modern mass produced vehicle has a CAN Bus and it is
>through this that a skilled person takes about 1/2 minute to clock the
>speedo. Service diagnostic machines plug into this also to diagnose engine
>tuning problems etc.
>
>>presumably via Internet, right ;-) ?
>
>Not necessarily, see my previous comment, the activity is done using a small
>handpiece that plugs into the bus.

Yes but I was pointing out a virus migration path ... after all someone
DID program the handheld box with a software update downloaded from
somewhere trustworthy, didn't they ? ...

Peter

--
http://www.piclist.com hint: To leave the PICList
RemoveMEpiclist-unsubscribe-requestTakeThisOuTspammitvma.mit.edu


More... (looser matching)
- Last day of these posts
- In 2002 , 2003 only
- Today
- New search...