Am 20.05.2010 09:11, schrieb scott:

> Cool - is there also a name for when you connect a tiny speaker to two
> output pins and give one pin a square wave, and the other the inverse of
> the square wave - to make your speaker 2x louder than everyone elses? :-)

That would be differential audio (also frequently but incorrectly 
referred to as balanced audio).

Post a reply to this message

>> OK, so if I'm understanding this right... The 74126 (go look it up) 
>> contains 4 "buffers" - gates who's output is logically equal to their 
>> input. But each gate also has an "enable" pin. When the enable pin is 
>> high, the gate works like normal. When the enable pin goes low, it's 
>> like the output pin isn't connected to anything any more.
> 
> Yes exactly.
> 
>> And it seems that this allows you to connect several outputs together, 
>> forming a kind of wired-OR configuration, provided that at all times 
>> only one gate is "enabled".
> 
> Yep, you could make a simple address decoder logic circuit (ie have a 2 
> bit binary input to select which of the 4 buffers to enable) - that way 
> you would guarantee only 1 would be active at a time.  If you were to 
> connect a 2-bit counter to the 2-bit address lines you would then have a 
> 4-bit parallel to serial converter :-)

I was thinking more about routing bundles of signals from place to 
place. The "obvious" thing to do is use logic gates to combine the 
signals, but I guess using these weird 3-state systems allows you to 
save a few logic gates (and hence ICs, wiring, and ripple time).

Post a reply to this message

scott wrote:

> Load up the actual PDF spec that is linked from that page, it has more 
> detail, just stating the forward current like that means nothing.

Oh, right. I hadn't even noticed that hidden away in a corner there! I 
noticed some of the other pages had this, but in *light grey* it's not 
especially prominent.

> On page 3 there is the "Absolute max" table, the continuous current here 
> is listed as 20/25/30 mA depending on the colour.  That is the maximum, 
> above that they do not guarantee the LED will work.

I see.

> Now look at the lower part of page 3, some interesting charts there 
> (assuming the Bright Red colour, the others are on the next pages).
> 
> The one in the top right shows "brightness" against current.  You see 
> it's linear up to 10 mA and then flattens out.  This is saying there 
> isn't much point using more than 10 mA, it won't be very efficient.

Right. So it will light up with just 10 mA, just not quite at full 
brightness. (On the chart I'm looking at, 15 mA looks like it's 
meaningfully brighter, but any higher than that has negligable effect.)

> The one in the top left tells you what current will be drawn for a 
> particular voltage across the LED.  If you want 10 mA then you're going 
> to need to give the LED 2.1 V.  You also see here that even adding just 
> 10% to that voltage makes the LED current almost 3x higher!

Right. So I'm looking at 2.1 V or 2.2 V or thereabouts.

Now looking at <http://focus.ti.com/lit/ug/scyd013b/scyd013b.pdf>, on 
page 231 (which is actually page 236 of the PDF file), we see that for 
the 74HC00 I'm looking at using, we have Icc <= 0.02 mA, Iol = -Ioh <= 4 
mA, and tPLH = tPHL <= 27 ns. Now, if I actually knew WTF that means...

Post a reply to this message

Orchid XP v8 wrote:

> Hey, what do you *think* I've got a dad for? ;-)

Wait - DON'T ANSWER THAT! >_<

Post a reply to this message

Am 20.05.2010 10:06, schrieb Invisible:

> I was thinking more about routing bundles of signals from place to
> place. The "obvious" thing to do is use logic gates to combine the
> signals, but I guess using these weird 3-state systems allows you to
> save a few logic gates (and hence ICs, wiring, and ripple time).

For unidirectional point-to-point connections, 3-state logic doesn't 
help you much, as there's no sane way to properly identify the 
"floating" state (unless you care to use some kind of ADCs at the 
receiving end).

Post a reply to this message

> Right. So it will light up with just 10 mA, just not quite at full 
> brightness. (On the chart I'm looking at, 15 mA looks like it's 
> meaningfully brighter, but any higher than that has negligable effect.)

Yes, every LED datasheet has a brightness against current curve.  Even for 1 
mA or 0.1 mA it will light up, but obviously it won't be very bright!  Note 
that even for 0.1 mA you probably still need a least a volt or so, which is 
why LEDs don't light up at all below a certain voltage.

> Right. So I'm looking at 2.1 V or 2.2 V or thereabouts.
>
> Now looking at <http://focus.ti.com/lit/ug/scyd013b/scyd013b.pdf>, on page 
> 231 (which is actually page 236 of the PDF file), we see that for the 
> 74HC00 I'm looking at using, we have Icc <= 0.02 mA, Iol = -Ioh <= 4 mA, 
> and tPLH = tPHL <= 27 ns. Now, if I actually knew WTF that means...

Dunno, what Icc means here, maybe the current consumption of the device 
itself?

From this page:

http://www.kpsec.freeuk.com/components/74series.htm

The HC series can sink and source up to 4 mA on the output pins if you want 
the signals to still be valid (eg for input to further logic gates).  If you 
are just using them to drive LEDs then apparently up to 20 mA is OK.

Post a reply to this message

Am 20.05.2010 10:20, schrieb Invisible:

> Right. So I'm looking at 2.1 V or 2.2 V or thereabouts.

Make sure to leave some margin for tolerances; resistors may have 
tolerances up to 10% (though 5% seem to be more common nowadays), and 
your power supply might be more (or less) than 5V as well. You don't 
want to kill your LED just because its actual "lethal" forward current 
happens to be pretty close to nominal, while the power supply and 
resistor parameters happen not to be.

Post a reply to this message

Am 20.05.2010 10:20, schrieb Invisible:

> Now looking at <http://focus.ti.com/lit/ug/scyd013b/scyd013b.pdf>, on
> page 231 (which is actually page 236 of the PDF file), we see that for
> the 74HC00 I'm looking at using, we have Icc <= 0.02 mA, Iol = -Ioh <= 4
> mA, and tPLH = tPHL <= 27 ns. Now, if I actually knew WTF that means...

 From the values specified there, I'm pretty sure Icc in this case means 
the maximum current that the IC will conduct from Vcc to GND under worst 
conditions.


As for Iol and Ioh, these are apparently the maximum currents that the 
IC can conduct from Vcc to output pins (in case of positive values), or 
from output pins to GND (in case of negative values), though I wouldn't 
be too sure whether these are...

(1.a) per-pin or (1.b) per-IC values;

(2.a) maximum values at which output voltage levels are still guaranteed 
to be within proper range, (2.b) absolute maximum values the chip is 
guaranteed to be able to supply, or (2.c) absolute maximum values the 
chip is guaranteed to survive.


tPLH and tPHL are probably signal propagation delays for low-to-high and 
high-to-low transitions, respectively.


I'd suggest you search <http://focus.ti.com/general/docs/dsnsuprt.tsp> 
for a product-specific datasheet, which will give much more detail, such as:


Absolute maximum ratings (i.e. values that, when exceeded, may cause 
permanent damage):

Continuous output current: +/- 25 mA
Continuous current through VCC or GND: +/- 50 mA


That is, for instance, if you plan to drive LEDs with 15 mA through one 
of these ICs, you /must not/ connect two in parallel to any output pin, 
and you /must not/ try to drive more than 3 LEDs in total per IC (unless 
you want to risk chip destriction).


Elextrical characteristics (i.e. values that are guaranteed by the 
manufacturer):

Voh (output voltage when pin is high) may (or may not!) drop as low as 
3.84 V (given a Vcc of 4.5 V) when sustaining a current of -4 mA (so I'd 
expect even lower values when trying to draw -15 mA)

Vol (output voltage when pin is low) may (or may not!) rise as high as 
0.33 V (again given a Vcc of 4.5 V) when sustaining a current of +4 mA 
(so I'd expect even higher values when trying to draw +15 mA)

That is, driving the LEDs in low-active configuration between Vcc and 
output pin is likely to give you a somwehat more predictable voltage.

Post a reply to this message

Invisible wrote:
>>> OK, so if I'm understanding this right... The 74126 (go look it up) 
>>> contains 4 "buffers" - gates who's output is logically equal to their 
>>> input. But each gate also has an "enable" pin. When the enable pin is 
>>> high, the gate works like normal. When the enable pin goes low, it's 
>>> like the output pin isn't connected to anything any more.
>>
>> Yes exactly.
>>
>>> And it seems that this allows you to connect several outputs 
>>> together, forming a kind of wired-OR configuration, provided that at 
>>> all times only one gate is "enabled".
>>
>> Yep, you could make a simple address decoder logic circuit (ie have a 
>> 2 bit binary input to select which of the 4 buffers to enable) - that 
>> way you would guarantee only 1 would be active at a time.  If you were 
>> to connect a 2-bit counter to the 2-bit address lines you would then 
>> have a 4-bit parallel to serial converter :-)
> 
> I was thinking more about routing bundles of signals from place to 
> place. The "obvious" thing to do is use logic gates to combine the 
> signals, but I guess using these weird 3-state systems allows you to 
> save a few logic gates (and hence ICs, wiring, and ripple time).

Like this?

(Seems to work in simulation...)

Post a reply to this message

Attachments:
Download 'logic1.png' (2 KB)

Preview of image 'logic1.png'

> Like this?
>
> (Seems to work in simulation...)

Careful, go much further and you're going to end up designing a RAM chip :-)

Post a reply to this message