Maker Pro
Maker Pro

Driving transistor from HC273

J

Jonathan Kirwan

Jan 1, 1970
0
Follow up. 1800 led drivers @ 10 % duty cycle > 1800 * 10% * ( 6-1 mA ) saving on the
PSU.

That's nearly an AMP !

I had misread that, thinking he was talking about making that many
units that included this circuit. But now re-reading it, I see it the
way you just mentioned. He's planning an array.

Jon
 
J

Jonathan Kirwan

Jan 1, 1970
0
I had misread that, thinking he was talking about making that many
units that included this circuit. But now re-reading it, I see it the
way you just mentioned. He's planning an array.

Makes me wonder if there are LEDs with different voltages (colors),
since with that many it might pay off dissipation-wise to use separate
switching supplies for each color/voltage.

Here's one possible arrangement that would provide about 60mA with a
Vcc of 5V.

: Vcc Vcc
: | |
: | |
: \ |
: / Rca_N |
: \ 47k |
: / |
: Rcb_N | |<e Qcol_N
: COL_N----------------/\/\-+-----| 2N3906
: 100 |\c
: |
: |
: | COL_N
: +----------------------------+-------------------------------------+------------,
: | | | |
: | | | |
: --- --- --- |
: \ / DN1 \ / DN2 \ / DNZ |
: --- --- --- \
: | | | / Rcc_3
: | | | \ 47k
: | | | /
: | | | |
: | | | |
: | | | |
: Vcc Vcc | | | |
: | | | | | gnd
: | | | | |
: \ | | | |
: / Rca_2 | | | |
: \ 47k | | | |
: / | | | |
: Rcb_2 | |<e | | |
: COL_2----------/\/\-+-----| 2N3906| | |
: 100 |\c | | |
: | | | |
: Qcol_2 | | | |
: | | | | COL_2
: +----------------------------+-------------------------------------+-------------------,
: | | | | | | |
: | | | | | | |
: --- | --- | --- | |
: \ / D2A | \ / D2B | \ / D2Z | |
: Vcc Vcc --- | --- | --- | \
: | | | | | | | | / Rcc_2
: | | | | | | | | \ 47k
: \ | | | | | | | /
: / Rca_1 | | | | | | | |
: \ 47k | | | | | | | |
: / | | | | | | | |
: Rcb_1 | |<e | | | | | | |
: COL_1---/\/\-+-----| 2N3906| | | | | | gnd
: 100 |\c | | | | | |
: | | | | | | |
: Qcol_1 | | | | | | |
: | | | | | | | COL_1
: +----------------------------+-------------------------------------+--------------------------,
: | | | | | | | | | |
: | | | | | | | | | |
: --- | | --- | | --- | | |
: \ / D1A | | ^ \ / D1B | | ^ \ / D1Z | | ^ |
: --- | | | --- | | | --- | | | \
: | | | | | | | | | | | | / Rcc_1
: | | | | | | | | | | | | \ 47k
: | | | | | | | | | | | | /
: +------+------+------' +------+------+------' +------+------+------' |
: | ROW_A | ROW_B | ROW_Z |
: | | | |
: | | | |
: Rrow_1 |/c Qrow_1 Rrow_2 |/c Qrow_2 Rrow_Z |/c Qrow_Z gnd
: ROW_A----/\/\------| 2N3904 ROW_B---/\/\----| 2N3904 ROW_Z----/\/\----| 2N3904
: 3300 |>e 3300 |>e 3300 |>e
: | | |
: | | |
: | | |
: | | |
: | | |
: \ \ \
: / Rset_1 / Rset_2 / Rset_Z
: \ 22 \ 22 \ 22
: / / /
: | | |
: | | |
: | | |
: | | |
: gnd gnd gnd

Jon
 
J

Jonathan Kirwan

Jan 1, 1970
0
Slight change, for somewhat better clarity:

: Vcc Vcc
: | |
: | |
: \ |
: / Rca_N |
: \ 47k |
: / |
: Rcb_N | |<e Qcol_N
: COL_N----------------/\/\-+-----| 2N3906
: 100 |\c
: |
: |
: | COL_N
: +----------------------------+--------- . . . -------------------+------------,
: | | | |
: | | | |
: --- --- --- |
: \ / DN1 \ / DN2 \ / DNZ |
: --- --- --- \
: | | | / Rcc_3
: | | | \ 47k
: | | | /
: | | | |
: | | | |
: | | | |
: Vcc Vcc | | | |
: | | | | | gnd
: | |
: \ | . . .
: / Rca_2 |
: \ 47k | . . .
: / |
: Rcb_2 | |<e . . .
: COL_2----------/\/\-+-----| 2N3906
: 100 |\c | | |
: | | | |
: Qcol_2 | | | |
: | | | | COL_2
: +----------------------------+---------------- . . . ------------+-------------------,
: | | | | | | |
: | | | | | | |
: --- | --- | --- | |
: \ / D2A | \ / D2B | \ / D2Z | |
: Vcc Vcc --- | --- | --- | \
: | | | | | | | | / Rcc_2
: | | | | | | | | \ 47k
: \ | | | | | | | /
: / Rca_1 | | | | | | | |
: \ 47k | | | | | | | |
: / | | | | | | | |
: Rcb_1 | |<e | | | | | | |
: COL_1---/\/\-+-----| 2N3906| | | | | | gnd
: 100 |\c | | | | | |
: | | | | | | |
: Qcol_1 | | | | | | |
: | | | | | | | COL_1
: +----------------------------+----------------------- . . . -----+--------------------------,
: | | | | | | | | | |
: | | | | | | | | | |
: --- | | --- | | --- | | |
: \ / D1A | | ^ \ / D1B | | ^ \ / D1Z | | ^ |
: --- | | | --- | | | --- | | | \
: | | | | | | | | | | | | / Rcc_1
: | | | | | | | | | | | | \ 47k
: | | | | | | | | | | | | /
: +------+------+------' +------+------+------' +------+------+------' |
: | ROW_A | ROW_B | ROW_Z |
: | | | |
: | | | |
: Rrow_1 |/c Qrow_1 Rrow_2 |/c Qrow_2 Rrow_Z |/c Qrow_Z gnd
: ROW_A----/\/\------| 2N3904 ROW_B---/\/\----| 2N3904 ROW_Z----/\/\----| 2N3904
: 3300 |>e 3300 |>e 3300 |>e
: | | |
: | | |
: | | |
: | | |
: | | |
: \ \ \
: / Rset_1 / Rset_2 / Rset_Z
: \ 22 \ 22 \ 22
: / / /
: | | |
: | | |
: | | |
: | | |
: gnd gnd gnd

Jon
 
P

Pooh Bear

Jan 1, 1970
0
It's nearly 10% of the power going up in needless heat for heavens sake !
he's going to be using about ten amps to drive the LED's, he didn't
say anything about a power problem, and he wants to use 2N2222's, so
that's how it goes.

I very much doubt he *wants* to use 2N2222s at all. Probably he just found them in a text
book and hasn't looked at sensible modern alternatives.

Jeez ! 2N2222 is in TO-18 ( un required ) and costs a fortune too ! A very bad choice.
Around 10x what you need to pay.

Graham
 
P

Pooh Bear

Jan 1, 1970
0
Anthony said:
Off of the 11A continuously drawn by the LEDs. Man....this thing is
really gonna put out some heat.

The led supply *could* be lower voltage since the transistors are open
collector to save some dissipation. 2V5 for example.

Graham
 
P

Pooh Bear

Jan 1, 1970
0
But they *aren't* dirt cheap John.

Hermetic TO-18 encapsuation costs an arm and a leg.

He only needs epoxy encapsulation so a more logical choice would be the 'near
generic' 2N3904 or BC548


Some textbooks sem to be suck in a time warp with regard to their examples of
suitable devices. Things have moved on ! Having a larger beta is very useful
and makes design less cricitical.

Graham
 
P

Pooh Bear

Jan 1, 1970
0
Jonathan said:
Yup. A year ago, I bought 1000 to-92 pn2222a for $8. Less than 1
cent each. Cheap.

Aren't they in epoxy though ? Not 2N2222s at all !

Graham
 
A

Anthony Fremont

Jan 1, 1970
0
Pooh Bear said:
The led supply *could* be lower voltage since the transistors are open
collector to save some dissipation. 2V5 for example.

I was referring only to the power dissipated by the LEDs themselves.
Worst case with all 1800 on and drawing an average of 6mA (60mA @
10%duty) comes out to 10.8A. As you say, the OP should really consider
using as low voltage supply as possible so that he/she isn't burning up
another 10-15A in current limiting resistors. I'm guessing, but I
figure this device will not be battery operated. ;-)
 
J

Jonathan Kirwan

Jan 1, 1970
0
Aren't they in epoxy though ? Not 2N2222s at all !

yeah. Eternally damned TO-92's. I think all BJTs should be at least
in a nice TO-5. Maybe a TO-18, if the manufacturers beg forgiveness
while offering them. TO-92s can't be decently labeled. But then, I'm
a hobbyist.

Jon
 
J

Jonathan Kirwan

Jan 1, 1970
0
Jeez ! 2N2222 is in TO-18 ( un required ) and costs a fortune too ! A very bad choice.
Around 10x what you need to pay.

Ah, you can still get them in TO-18 for less than 5 cents each. About
5X, not 10X. ;)

And you can actually _read_ their number, then, because it's nicely
visible; hold them in your hands without fearing losing them in the
rug; and know that you've got some real metal there -- the feeling of
substance! Not to mention that the leads exit the package in the
sensible triangular arrangement with the emitter nicely tagged, and
not the disgusting linear layout that can confuse you because there
are some BJT layouts that differ from each other in that packaging and
none of it, sensibly.

It's a veritable crime that their R(theta_JC) is no better than for
those little TO-92's, though. :)

Jon
 
J

John Fields

Jan 1, 1970
0
It's nearly 10% of the power going up in needless heat for heavens sake !


I very much doubt he *wants* to use 2N2222s at all. Probably he just found them in a text
book and hasn't looked at sensible modern alternatives.
 
L

logjam

Jan 1, 1970
0
I'm going to use the ULN2803. More expensive, but better in the long
run.

Take a look at these schematics and try to find something wrong. I
know of a few things I need to do, like put a pull down resistor on OE
(not /OE). This is my first high current design. I don't want any
bugs since I'll need 30 boards. :)

There are 19,008 LEDs in the array. The whole LED display will consume
138 amps MAX, and of that at least 276 watts MAX will be heat.

I have a 150A 5v supply. ;)

http://www.stockly.com/images2/060121-Root_1_1.PDF

http://www.stockly.com/images2/060121-DriverBlock_1_1.PDF
 
L

logjam

Jan 1, 1970
0
So, I got the part number wrong. N2222 type. The data sheets I've
found list it up to 500ma collector current, have I been reading them
wrong?

I chose that transistor because like other people said, through surplus
places they are under a penny a piece (in quantities of 2000). I've
decided to use the ULN2803, so its over. ;) But I would love to get
comments on the above schematics. Tear them apart just as much as this
transistor debate. ;)
 
L

logjam

Jan 1, 1970
0
192x99

8x11 is a good character font 24x9 character display, or bitmap
graphics.

I've built a single character as well as a 26x22 array. For the bigger
array I have around 10,000LEDs soldered so far. Only 9000 to go. :)
 
J

John Fields

Jan 1, 1970
0
The led supply *could* be lower voltage since the transistors are open
collector to save some dissipation. 2V5 for example.
 
J

John Fields

Jan 1, 1970
0
But they *aren't* dirt cheap John.

Hermetic TO-18 encapsuation costs an arm and a leg.

---
But, as I indicated earlier, we all know that he's talking about the
chip, not the package.
---
He only needs epoxy encapsulation so a more logical choice would be the 'near
generic' 2N3904 or BC548

---
He stated that he wants to use the 2N2222 and he only needs the
package to dissipate Vce(sat) times 60mA, so why would a more
"logical" choice be a 'near generic' substitute when he can get what
he want for less than a US penny a piece?
--
Some textbooks sem to be suck in a time warp with regard to their examples of
suitable devices. Things have moved on ! Having a larger beta is very useful
and makes design less cricitical.

---
Not true.

Taking unwise advantage of a greater beta by forcing the base to
live at a higher impedance than it should is a recipe for disaster.

If you haven't yet discovered that, then I propose, politely and
respectfully, that you're not as knowledgeable as you propose to be.
 
J

Jonathan Kirwan

Jan 1, 1970
0
99x192 33x576 32x594 man there's a bunch of combinations for such an
odd number.

Yup, but quickest to just ask with one example. I was curious.

Jon
 
Top