74LS47 substitution

[bitrex]

Is there a CMOS-type substitute for the 74LS47 with "open collector"
type outputs suitable for sinking around 20mA?

That uses less quiescent power than the 47's ~10mA. Supply voltage
either 5 or 3.3 volts, depending. It doesn't really matter if the chip
itself does a conversion from BCD to the appropriate outputs, as that
mapping can be done in software, but the chip would need to have at
least a 3 bit address space.

The '47 has an input that's suitable for dynamic dimming of the loads by
applying PWM, it would be cool if the replacement had something like that.

I think JL posted something along these lines in response to another
question of mine, but I can't find the thread.

[John Larkin]

A TPIC6595 might work. The data sheet says it needs 5 volts, but there
is an appnote that says it's OK at 3.3. We use them at 3.3.

The input is SPI serial, and the /G pin can do dimming.


--

John Larkin Highland Technology, Inc
picosecond timing precision measurement

jlarkin att highlandtechnology dott com
http://www.highlandtechnology.com

[bitrex]

John Larkin <jjla...@highlandtechnology.com> Wrote in message:

> On Wed, 15 Jun 2016 13:34:50 -0400, bitrex
> <bit...@de.lete.earthlink.net> wrote:
>
>>Is there a CMOS-type substitute for the 74LS47 with "open collector"
>>type outputs suitable for sinking around 20mA?
>>
>>That uses less quiescent power than the 47's ~10mA. Supply voltage
>>either 5 or 3.3 volts, depending. It doesn't really matter if the chip
>>itself does a conversion from BCD to the appropriate outputs, as that
>>mapping can be done in software, but the chip would need to have at
>>least a 3 bit address space.
>>
>>The '47 has an input that's suitable for dynamic dimming of the loads by
>>applying PWM, it would be cool if the replacement had something like that.
>>
>>I think JL posted something along these lines in response to another
>>question of mine, but I can't find the thread.
>
> A TPIC6595 might work. The data sheet says it needs 5 volts, but there
> is an appnote that says it's OK at 3.3. We use them at 3.3.
>
> The input is SPI serial, and the /G pin can do dimming.
>

Thanks! Will look into it...

--


----Android NewsGroup Reader----
http://usenet.sinaapp.com/

[bitrex]

John Larkin <jjla...@highlandtechnology.com> Wrote in message:

> On Wed, 15 Jun 2016 13:34:50 -0400, bitrex
> <bit...@de.lete.earthlink.net> wrote:
>
>>Is there a CMOS-type substitute for the 74LS47 with "open collector"
>>type outputs suitable for sinking around 20mA?
>>
>>That uses less quiescent power than the 47's ~10mA. Supply voltage
>>either 5 or 3.3 volts, depending. It doesn't really matter if the chip
>>itself does a conversion from BCD to the appropriate outputs, as that
>>mapping can be done in software, but the chip would need to have at
>>least a 3 bit address space.
>>
>>The '47 has an input that's suitable for dynamic dimming of the loads by
>>applying PWM, it would be cool if the replacement had something like that.
>>
>>I think JL posted something along these lines in response to another
>>question of mine, but I can't find the thread.
>
> A TPIC6595 might work. The data sheet says it needs 5 volts, but there
> is an appnote that says it's OK at 3.3. We use them at 3.3.
>
> The input is SPI serial, and the /G pin can do dimming.

The CD74AC238 appears to be rated for 50 mA through a pin, 200 mA
continuous to ground, at 1.5 to 5.5 volts supply, so that might
also be a possibility.

The edge case would be the "8"...

[John Larkin]

Do you need open drain? HC or AC logic will sink 20 mA, probably
better with Vcc=5.

The 6595 is a beast, designed for a lot of current.

[bitrex]

On 06/15/2016 04:39 PM, John Larkin wrote:
> On Wed, 15 Jun 2016 16:02:30 -0400 (EDT), bitrex
> <bit...@de.lete.earthlink.net> wrote:
>
>> John Larkin <jjla...@highlandtechnology.com> Wrote in message:
>>> On Wed, 15 Jun 2016 13:34:50 -0400, bitrex
>>> <bit...@de.lete.earthlink.net> wrote:
>>>
>>>> Is there a CMOS-type substitute for the 74LS47 with "open collector"
>>>> type outputs suitable for sinking around 20mA?
>>>>
>>>> That uses less quiescent power than the 47's ~10mA. Supply voltage
>>>> either 5 or 3.3 volts, depending. It doesn't really matter if the chip
>>>> itself does a conversion from BCD to the appropriate outputs, as that
>>>> mapping can be done in software, but the chip would need to have at
>>>> least a 3 bit address space.
>>>>
>>>> The '47 has an input that's suitable for dynamic dimming of the loads by
>>>> applying PWM, it would be cool if the replacement had something like that.
>>>>

>>>> I think JL posted something along these lines in response to anotherh

>>>> question of mine, but I can't find the thread.
>>>
>>> A TPIC6595 might work. The data sheet says it needs 5 volts, but there
>>> is an appnote that says it's OK at 3.3. We use them at 3.3.
>>>
>>> The input is SPI serial, and the /G pin can do dimming.
>>
>> The CD74AC238 appears to be rated for 50 mA through a pin, 200 mA
>> continuous to ground, at 1.5 to 5.5 volts supply, so that might
>> also be a possibility.
>>
>> The edge case would be the "8"...
>
> Do you need open drain? HC or AC logic will sink 20 mA, probably
> better with Vcc=5.
>
> The 6595 is a beast, designed for a lot of current.

I don't think so. The 74LS47 outputs don't pull down much lower than 0.3
volts.

While it looks like most HC logic can sink 20 mA, the problem is it
seems most are rated for only 50mA continuous through GND, and the
SOIC-16 packages are rated for only 500mW continuous dissipation.

The CD74AC238 can sink 200mA to ground, and is rated for 1.1 watts
dissipation at 30 degrees, so even if I have to derate it a bit for
140mA passing through the device when all segments are lit, it seems it
should be okay.

[Tauno Voipio]

How about the ULN2001 family, 500 mA / pin?

--

-TV

[Peter Heitzer]

bitrex <bit...@de.lete.earthlink.net> wrote:
>Is there a CMOS-type substitute for the 74LS47 with "open collector"
>type outputs suitable for sinking around 20mA?

Although kind of overkill. Perhaps a small µC. The AVRs for example
can sink about 20 mA/pin with a total of 200 mA for VCC and GND.
Of course the outputs are not open collector. If you need OC and more
than one digit to drive a combination of µC and one OC driver can be
the solution with the least components. The µC could also do PWM and
multiplexing.

--
Dipl.-Inform(FH) Peter Heitzer, peter....@rz.uni-regensburg.de

[bitrex]

On 06/16/2016 04:31 AM, Peter Heitzer wrote:
> bitrex <bit...@de.lete.earthlink.net> wrote:
>> Is there a CMOS-type substitute for the 74LS47 with "open collector"
>> type outputs suitable for sinking around 20mA?
> Although kind of overkill. Perhaps a small µC. The AVRs for example
> can sink about 20 mA/pin with a total of 200 mA for VCC and GND.
> Of course the outputs are not open collector. If you need OC and more
> than one digit to drive a combination of µC and one OC driver can be
> the solution with the least components. The µC could also do PWM and
> multiplexing.
>

Yeah, there's going to be a uC involved one way or the other...I'm
trying to figure if it would be better to just use an AVR with more pins
vs an 8 pin AVR and an external switch IC.

An ATTiny85 that costs under a dollar in quantity plus an external
switch that costs 25 cents in quantity is a better value than something
like a Mega168 that costs $1.65 in quantity, as the board size would be
pretty much fixed by larger components either way.

There are other weird AVR devices that seem to use a "Tiny" core but
have more pins, but I'm taking advantage of the Arduino toolchain and
AFAIK they're not supported, and I don't really want to take the time to
write an implementation.

[bitrex]

It's a good chip, it doesn't look like it draws any current whatsoever
unless the Darlington outputs are switched on...only 25 cents in
quantities of 100 surface mount.

But it needs one address line for each output, and I was hoping to cut
costs by using an el cheapo 8 pin AVR and an el cheapo driver, instead
of shelling out for an AVR with more pins but also more features and
code space which I don't need for this. If I wanted to use 7 pins it
would make more sense to just use a uC with more pins, as a poster
mentions below I think even the AVRs in SMT can handle 20mA per pin and
200mA to ground (though driving the outputs of the display directly from
the processor close to the limits makes me feel a little uncomfortable.)

[bitrex]

On 06/16/2016 04:31 AM, Peter Heitzer wrote:

> bitrex <bit...@de.lete.earthlink.net> wrote:
>> Is there a CMOS-type substitute for the 74LS47 with "open collector"
>> type outputs suitable for sinking around 20mA?
> Although kind of overkill. Perhaps a small µC. The AVRs for example
> can sink about 20 mA/pin with a total of 200 mA for VCC and GND.
> Of course the outputs are not open collector. If you need OC and more
> than one digit to drive a combination of µC and one OC driver can be
> the solution with the least components. The µC could also do PWM and
> multiplexing.
>

The other issue with using the uC pins to drive the display directly is
that there aren't enough hardware PWM channels to dim every segment
common anode. If the display I had to use was common cathode I could
simply implement dimming by using an external pass transistor on the
cathode and PWMing that to ground and driving the segments with a "high"
logic voltage, but unfortunately I don't have that option here, so I
think I'd have to implement some scheme using a filter and regulator on
the supply and manipulating its reference, but it would likely have to
be an LDO, which would probably cost around the same as an appropriate
driver IC in a SOIC-16.

[bitrex]

Actually, if I decided to go that route I could probably just chop the
supply with PWM and a two transistor switch like so:

http://www.avrfreaks.net/sites/default/files/Common_Anode.jpg

[bitrex]

Oops, I made a mistake in my power dissipation calculation, what a
dummy. It's only the voltage drop across the pass device that matters,
so even a 500mW dissipation should be good.

[bitrex]

It looks like the best "bang for the buck" right now is to simply use an
AVR with a "Tiny" core that has 12 IO pins instead of 6, and then PWM
the supply with a two transistor switch.

I didn't realize that AVRs could sink as much current to/from the
supplies as they appear to be able to.

The ATTiny84 has two PWM channels and is just over a dollar at Mouser, a
PNP and NPN for the switch is basically free.

[Peter Heitzer]

The only supported "Tiny" device is the Tiny85, but the Tiny2313 is not
very different to the Tiny85.
http://playground.arduino.cc/Main/ArduinoOnOtherAtmelChips
mentions the 2313. So it should be possible to use the 2313 with the
toolchain. The 2313 costs about the same as the 85 but gives you 18
usable pins vs 6 pins on the Tiny85.

[bitrex]

Looks like setting it up for the Tiny84 is straightforward these days:

http://highlowtech.org/?p=1695

The Tiny84 only has 12 pins, but even that is more than I need. I'd
prefer the larger code space, 8k vs 2k for the 2313. They're both about
the same price in quantity at Mouser.

[Peter Heitzer]

bitrex <bit...@de.lete.earthlink.net> wrote:

>Looks like setting it up for the Tiny84 is straightforward these days:

>http://highlowtech.org/?p=1695

>The Tiny84 only has 12 pins, but even that is more than I need. I'd
>prefer the larger code space, 8k vs 2k for the 2313. They're both about
>the same price in quantity at Mouser.

That makes sense. If you program the uC in C 2 KiB can be used up
easyly if you happen to use "printf" or its friends. You also could
use linear code more easyly if you need speed for some tasks.

[rickman]

There are about a million I2C/SPI LED drivers out there that have built
in PWM dimming. Have you looked at any?

--

Rick C

[bitrex]

On 06/16/2016 07:11 AM, Peter Heitzer wrote:
> bitrex <bit...@de.lete.earthlink.net> wrote:
>
>> Looks like setting it up for the Tiny84 is straightforward these days:
>
>> http://highlowtech.org/?p=1695
>
>> The Tiny84 only has 12 pins, but even that is more than I need. I'd
>> prefer the larger code space, 8k vs 2k for the 2313. They're both about
>> the same price in quantity at Mouser.
> That makes sense. If you program the uC in C 2 KiB can be used up
> easyly if you happen to use "printf" or its friends. You also could
> use linear code more easyly if you need speed for some tasks.
>

I have a slight pathology in that I like using C++, even for "small"
projects. avr-g++ compiles pretty efficient code, so long as one avoids
using a ton of "virtual" methods, and initializes all memory required
for the system at startup and then doesn't use the "new" operator to do
anything but place or copy objects onto pre-allocated memory.

I think the moderate increase in code size is worth the advantages
gained in portability/maintainability.

[Peter Heitzer]

rickman <gnu...@gmail.com> wrote:
>On 6/15/2016 1:34 PM, bitrex wrote:
>> Is there a CMOS-type substitute for the 74LS47 with "open collector"
>> type outputs suitable for sinking around 20mA?

>There are about a million I2C/SPI LED drivers out there that have built
>in PWM dimming. Have you looked at any?

I think the OP is looking for a part costing less than a dollar. The
drivers I found are rather costly.

[bitrex]

On 06/16/2016 10:38 AM, Peter Heitzer wrote:
> rickman <gnu...@gmail.com> wrote:
>> On 6/15/2016 1:34 PM, bitrex wrote:
>>> Is there a CMOS-type substitute for the 74LS47 with "open collector"
>>> type outputs suitable for sinking around 20mA?
>
>> There are about a million I2C/SPI LED drivers out there that have built
>> in PWM dimming. Have you looked at any?
> I think the OP is looking for a part costing less than a dollar. The
> drivers I found are rather costly.
>
>

Yup. Definitely looks like the cheapest way is to drive the segments of
the common anode display directly to ground right from a uC, if it can
take the punishment.

If there were more displays to drive instead of instead of just one, it
might be worth looking at using a dedicated driver chip for each
display. Or who knows, it might still be cheaper to use a uC for each
display, slaved together on a serial bus.

[rickman]

MCUs often have a total sink rating which can limit how many loads you
drive at the same time.

Not sure what you've found before, but a quick search turned up some TI
parts, TLC6C598 - 8 bits, $0.34 and TLC6C5912 - 12 bits, $0.50 at 1k
quantity. These are basic shift registers with a clear and buffer
register (open drain output) with a blanking control which can be used
for PWM. 30 mA per output up to 40 volts. Supply current in uA.

--

Rick C

[WangoTango]

In article <%kg8z.2$e9...@fx01.iad>, bit...@de.lete.earthlink.net
says...

I had to update a product that used two of them recently.
What I opted to use was an AS1109. It uses an SPI style interface
instead of a parallel BCD input and you will need to use a look up table
for the decoding, but that's all pretty trivial.
It has a single input for dimming and you can daisy chain them together
so you don't need multiple /CS lines. It also has the advantage of a
built in CC driver, so you don't need to put down individual resistors
for each output and a dimming input that handles all the output
channels.
Don't know if that will work for you, but it's worth a look.

http://www.digikey.com/product-search/en?mpart=AS1109-BSOT&v=961

[bitrex]

rickman <gnu...@gmail.com> Wrote in message:

The TLC6C598 is a cool part...if at some point I have to control
more than one display it would make sense to use that for each
and drive them from a single uC,

I was surprised though that most AVR uCs in SOIC can sink a total
of 200 mA to ground, at usually 20 mA or over per pin. The
packages have a total dissipation rating of 500 mW at 30C, so if
the output low voltages are only a few hundred mV I think I
should come in under that, even with 7 segments.