Running a K155ID1 at 3.3V

[anton andersson]

Hi Everyone
I'm building a multiplexed Nixie clock with an Atmel XMEGA256A3B for
brains and a K155ID1 as cathode decoder/driver. The problem is that
the XMEGA is modern and wants 3.3V supply while the K155ID1 needs 5V.

The simplest solution I can think of is supplying both chips with 3.3V
and hoping for the best. Altough I don't read russian, I can tell from
the data sheet that the logic levels should match surpricinly well,
but I cannot tell whether the rest of the circuit functions properly
with a lower than expected supply voltage. I have limited capabilities
to test the circuit properly (with nixies, cathode driver and
microcontroller etc.) and therefore I wonder if anyone at Neonixie has
used the K155ID1 like this before. I guess that the question is pretty
generic really; is it possible to run a (early generation) TTL circuit
at 3.3V reliably? Keep in mind that I don't require very quick
switching, kilohertz are just fine here.

The next step would be to add a 5V-regulator just to feed the driver.
In that case, how/do I need to protect the outputs of the
microcontroller? I have looked into the 'standard' schematics for TTL
and it seems to me that a TTL input with a 5V supply would tend to
pull its input high, possibly destroying a 3.3V output. Is series
resistors sufficient or are transistors needed too?

Another thing I just spotted is that the XMEGA have very versatile
outputs, offering totem-pole, open-source and open-drain modes. Would
a wired-and or wired-or be helpful for level-shifting?

Kind regards
Anton

[jb-electronics]

Hi,

I have no idea about the interface of 3.3V and 5V, sorry, no experiences.

> The next step would be to add a 5V-regulator just to feed the driver.
> In that case, how/do I need to protect the outputs of the
> microcontroller? I have looked into the 'standard' schematics for TTL
> and it seems to me that a TTL input with a 5V supply would tend to
> pull its input high, possibly destroying a 3.3V output. Is series
> resistors sufficient or are transistors needed too?

If you plan on going through so much trouble, just replace the K155ID1
with ten transistors, they will work on 3.3V as well. The MPSA42 is
available in SMD as well, this way you will not lose too much space. The
K155ID1 is not that great a chip that it is absolutely necessary to
tweak every circuit into compatibility with that oldtimer, if you ask me ;-)

Regards,
Jens

[Adam Jacobs]

Agreed. Most of us use the 74141 for circuit simplicity.. if it is complicating your circuit, then it's a no-brainer to get rid of it. There are lots of alternatives.

-Adam

--
You received this message because you are subscribed to the Google Groups "neonixie-l" group.
To post to this group, send an email to neoni...@googlegroups.com.
To unsubscribe from this group, send email to neonixie-l+unsubscribe@googlegroups.com.
For more options, visit this group at http://groups.google.com/group/neonixie-l?hl=en-GB.

[Jonathan Peakall]

Disclaimer: I am not an EE and I don't even play one on TV. That said, I
have done similar things plenty of times, as my uC of choice is CMOS (3.3V).
I would supply the chip with 5V and then use a 1k resistors to the pins of
the uC. I have a bunch of stuff running like that and so far have never
blown a uC pin doing so. Some things won't work like that but I have never
trashed a part trying.

However, YMMV and don't cry if it doesn't work or blows a pin.

Jonathan

[Adam Jacobs]

Will a TTL part even recognize 3.3v as HIGH? I'm usually going the other way, running a strange 3.3v peripheral chip into a 5v CMOS microcontroller. a 5v AVR chip will definitely recognize anything above about 1v as high, but ancient TTL stuff.. Somebody on this list will know, not me.

In my opinion, mixing voltage needs in a design is bad design and should be done sparingly and only if it makes sense. For example: In the Uzebox, there is a 5v microcontroller, a 5v video encoder, & a SDcard. SDcards are 3.3v.
However, this works fine because SDCards are low current (the separate LDO linear regulator looks like a transistor). The 3.3v outputs from the SDCard need no voltage translation to the microcontroller inputs and the microcontroller outputs need simple resistor dividers to talk to the SDcard inputs.

Why are you using the particular microcontroller that you are using? Part of design should be to look at the voltage requirements of all of your devices and decide if the components you chose make sense. If the 74141 was the critical component, select a 5v microcontroller. If the 3.3v microcontroller is the critical component, select a nixie driver that will work at 3.3v. :)

-Adam

[Jonathan Peakall]

Well, most TTL parts I've tried will recognize 3.3V. Still not saying it's reccomended. I've never blown a 3.3V input with a 5V signal as long as the 1k resistor is in place. I wouldn't do a commercial design like that, but on the hobby level is a different matter.

[Adrian Pardini]

On 31/07/2011, anton andersson <anta...@gmail.com> wrote:
> I have limited capabilities
> to test the circuit properly (with nixies, cathode driver and
> microcontroller etc.) and therefore I wonder if anyone at Neonixie has
> used the K155ID1 like this before. I guess that the question is pretty
> generic really; is it possible to run a (early generation) TTL circuit
> at 3.3V reliably? Keep in mind that I don't require very quick
> switching, kilohertz are just fine here.

Hi Anton. Most ttl devices interpret any level above 2.5V as '1', you
should be fine feeding 5 volts to the K155 and 3.3 to the xmega. I
don't know if the K155 pulls it's inputs high but a 1k resistor can be
used to limit current through the pin.

regards

--
Adrian.
http://ovejafm.com
http://elesquinazotango.com.ar
http://www.elarteylatecnologia.com.ar

[threeneurons]

The western 74141 and its Russian equivalent (K155ID1) are truly old
fashion TTL parts. If you ever look at an old TTL databook, there are
some common TTL specs:

Input logic-0: 0.8V or less
Input logic-1: 2.0V or higher
Output logic-0 (tied to 10 like inputs): 0.4V maximum
Output logic-1 (same): 2.4V minimum

TTL inputs naturally float high, to a level of ~1.6V. It registers
this as a logic-1, but its prone to noise, hence it they like you to
at least drag it up to 2.0V. The current draw is also asymmetrical. To
pull it high (logic-1), only ~40uA (0.00004A) flows into the input,
but drag it down to a logic-0 (low) and it will draw 1.6mA (0.0016A)
from the input. Roughly 40x more low, than high. This also makes
putting 1K resistors in series with their inputs. 1K*1.6mA translates
to 1.6V, so the output of your uC, low (0V) or high (3.3V), may just
look like a logic-1 to the TTL device (74141). The best thing is to
use two supplies: +5V & +3V, and make sure the +3V turns ON first.
Since your using a large uC, use one of its IO bits to turn on the +5V
supply. Before that supply is turned ON though, make all the outputs
between the uC and the 74141, low.

[David Forbes]

On 7/31/11 1:11 PM, anton andersson wrote:
> Hi Everyone
> I'm building a multiplexed Nixie clock with an Atmel XMEGA256A3B for
> brains and a K155ID1 as cathode decoder/driver. The problem is that
> the XMEGA is modern and wants 3.3V supply while the K155ID1 needs 5V.
>

Supply the CPU with 3.3V and the TTL chip with 5V, and wire the signals
direct. It will work just fine, as TTL is really a 3.3V logic family,
with a 5V power supply.
--
David Forbes, Tucson AZ

[Quixotic Nixotic]

This simple circuit lets you level shift data in both directions:

http://www.rocketnumbernine.com/blog/wp-content/uploads/2009/04/level-
converter-mosfet.jpg

You may not even need the pullups. For I2C use you may need a lower
value of pullup on that side.

John S

[anton andersson]

Hi!
Thanks for all usefull advice! =D
Even though I now know perfectly well how to interface 5V devices with
3.3V ditto I have simply decided not to. As Adam Jacobs said, the
K155ID1 should be used to simplify the designs, not introduce further
complexity.

The benefit of the K155ID1 is that it offers BCD to Decimal decoding
with high voltage tolerant outputs. Thus only four pins are required
on the µC.
Four pins for the anodes and four for the cathodes will fit neatly
into an 8-bit port, which will in turn allow for elegant code.

However, as I don't like to design and route a separate 5V net just
for 1 chip I have decided to look for something completely different.
A document written by Jason Harper that I read some time ago put me on
the right track. Even though the article doesn't say it explicitly I
beleive that it is a form of Charlieplexing.

The XMEGA, and other microcontrollers, offer not just two different
modes as one usually thinks. Apart from being logical true and false a
port may also become high impedive, in practice almost disconnected.
Every cathode still needs a transistor of their own, but placing them
in a matrix and connecting their bases and emitters in a most clever
way it is possible to activate them individually using only 4 pins on
the µC. Heck, this is even better than the K155ID1 since up to 12
cathodes can be steered this way, at any voltage the designer chooses
(as long as the BE-junction is doesn't suffer from breakdown...) IMHO
the circuit is most beautiful when Nixies with dual decimal commas are
used, since the decoder is only utilized fully then.

Curious readers will find the article at this site: http://nixietube.info
Thanks Jason!

Kind Regards
Anton

[Adam Jacobs]

They can be a bugger to debug, though. :)
Good luck with your design, Anton. Sounds like you've got a plan of
action. Let us know if we can help.

-Adam

On 8/2/2011 12:19 PM, anton andersson wrote:
> The XMEGA, and other microcontrollers, offer not just two different
> modes as one usually thinks. Apart from being logical true and false a
> port may also become high impedive, in practice almost disconnected.
> Every cathode still needs a transistor of their own, but placing them
> in a matrix and connecting their bases and emitters in a most clever
> way it is possible to activate them individually using only 4 pins on

> the �C. Heck, this is even better than the K155ID1 since up to 12