Searching \ for '[EE]: Interfacing to the ATA-interface' 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/index.htm?key=interfacing+ata
Search entire site for: 'Interfacing to the ATA-interface'.

Exact match. Not showing close matches.
PICList Thread
'[EE]: Interfacing to the ATA-interface'
2000\07\05@170134 by H.P. de Vries

flavicon
face
Hi.

Does anyone in here know how to interface 3.3V logic to 5V logic? I have a
PIC and some other stuff running at 3.3V , but need to interface with the
ATA-interface from my PC, and a LCD-display, which both need 5V logic
levels. Is there an easy way to do this? Or is it easier to run the PIC at
5V, and then convert the signals down to 3.3V?
Or is it possible (yes, I say possible, not allowed) to just put it all
together and operate the 5V logic devices at 3.3V ?

Hans

--
http://www.piclist.com hint: PICList Posts must start with ONE topic:
[PIC]: PIC only [EE]: engineering [OT]: off topic [AD]: advertisements

2000\07\05@172900 by Quentin

flavicon
face
IMO, since your PIC is the main part here (comm with PC and LCD at 5V),
I would say run it at 5V and convert down.The PIC will also still read a
3.3V input as Logic High.

Quentin

"H.P. de Vries" wrote:
{Quote hidden}

--
http://www.piclist.com hint: PICList Posts must start with ONE topic:
[PIC]: PIC only [EE]: engineering [OT]: off topic [AD]: advertisements

2000\07\05@174604 by H.P. de Vries

flavicon
face
Ok, that would be my first thought also. This also seems like the least
work, as I need to convert only about 5 lines, as opposed to the 24 5V
signals I had otherwise to convert.

But now, how do I convert the
voltages down in an efficient way? I don't like the idea of putting zeners
/w resistors on the outputs, since in that way I lose my low-impedance
output of the pic.
Did I mention I also have to drive an i2c-bus also at 3.3v? Now how do I go
about that? I'm no i2c expert at all, so this is one part I for sure won't
find an answer to myself easily.

<sigh>
Why don't all devices just run at 5V :-)

Hans
On Wed, Jul 05, 2000 at 11:26:48PM +0200, Quentin wrote:
{Quote hidden}

--
http://www.piclist.com hint: PICList Posts must start with ONE topic:
[PIC]: PIC only [EE]: engineering [OT]: off topic [AD]: advertisements

2000\07\06@025758 by Michael Rigby-Jones

flavicon
face
> -----Original Message-----
> From: H.P. de Vries [SMTP:spam_OUTH.P.d.VriesTakeThisOuTspamSTUD.TUE.NL]
> Sent: Wednesday, July 05, 2000 10:45 PM
> To:   .....PICLISTKILLspamspam@spam@MITVMA.MIT.EDU
> Subject:      Re: [EE]: Interfacing to the ATA-interface
>
> Ok, that would be my first thought also. This also seems like the least
> work, as I need to convert only about 5 lines, as opposed to the 24 5V
> signals I had otherwise to convert.
>
> But now, how do I convert the
> voltages down in an efficient way? I don't like the idea of putting zeners
> /w resistors on the outputs, since in that way I lose my low-impedance
> output of the pic.
> Did I mention I also have to drive an i2c-bus also at 3.3v? Now how do I
> go
> about that? I'm no i2c expert at all, so this is one part I for sure won't
> find an answer to myself easily.
>
The Philips book "The I2C bus and how to use it" has a section on dealing
with multi-voltage bus systems.  The answer is amazingly simple and very
elegant.  A discreet mosfet is simply inserted in each line between the
different voltage sections.  The source connects to the three volt bus,
drain to the 5 volt bus and gate to the 3.3v supply (Vdd).

The operation isn't immediatley obvious (at least it wasn't to me) as part
of the system relies on the internal diode accross a mosfet's drain/source,
but I can report that it definately works.

Cheers

Mike




{Quote hidden}

--
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

2000\07\06@140919 by Peter L. Peres

picon face
imho, cut your design into 3.3 and 5 V parts such that:

- you have no bidirectional lines crossing
- you have the fewest possible lines crossing
- you do not have fast data/clock lines crossing

then look for an interface chip that will solve your problem, or use
discretes. A 4 part/chip ancient CMOS chip exists that will do this, and
there are many modern variants with up to 32 x per package, but they all
require lots of power (even if only for switching).

Among discretes, open collector digital transistor arrays are popular with
me. One of those and a SIPP resistor array and you have 8 or 9 wires
converted.  Only works with slow signals, but phantom power is not a
concern. Most other drivers have a problem with phantom power (i.e. what
happens if one of the supplies is missing or if you power up or down in
the wrong sequence).

Peter

--
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

2000\07\06@145649 by Peter L. Peres

picon face
>i2c bus at 3.3V

If you use bit banged I2C then use a separate data in and out pin and you
are set. The data out only needs a shottky diode, the input a
(relatively large) resistor. The clock driver can have just a series
resistor. Most I2C components clamp the I2C pins efficiently to Vcc and
GND and are not bothered by this mode of operation. However, check ;-)

Peter

--
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

2000\07\07@091120 by H.P. de Vries

flavicon
face
On Thu, Jul 06, 2000 at 08:59:10PM +0300, Peter L. Peres wrote:
> >i2c bus at 3.3V
>
> If you use bit banged I2C then use a separate data in and out pin and you
> are set. The data out only needs a shottky diode, the input a
> (relatively large) resistor. The clock driver can have just a series
> resistor. Most I2C components clamp the I2C pins efficiently to Vcc and
> GND and are not bothered by this mode of operation. However, check ;-)
>

I saw another design which also uses just a resistor in series of the I2C
data-line. I plan on using the I2C feature of the 16c87x -series. Is this
also possible with just a series-resistor? The design I saw had a 100-ohms
resistor in series.

About the open-collector transistor solution : It should be something like :

               |Vdd (3.3v)
              [R]
               |_________ out
signal (5V)   |/c
    --[R]----|
           b |\e
               |
               |------- gnd

(Great ASCII , isn't it?)

Can I use this up to speeds of 1MHz, or do I need to search for special
transitors then?

Anyway, I don't think 3.3V inputs to the pic would really need converting.
Doesn't it recognise 3.3v as high anyway? The only thing I then needed to
convert was the input to the pic, as it probably won't really like the
over-voltage on its inputs.

Any suggestions?

Hans

--
http://www.piclist.com hint: The PICList is archived three different
ways.  See http://www.piclist.com/#archives for details.

2000\07\07@135057 by Oliver Broad

picon face
>> Does anyone in here know how to interface 3.3V logic to 5V logic? I
>> have a PIC and some other stuff running at 3.3V , but need to interface
>> with the ATA-interface from my PC, and a LCD-display, which both need
>> 5V logic

The simplest level translator I've seen (other than a resistor) is to use a
bus switch device. The first one I located in Farnell was the Fairchild
FST3244WM, though there are probably others. The devices consist of banks
of N type FETS as pass gates. The usefull part is that the on resistance
rises rapidly when a certain voltage is reached, and by running the device
off slightly less than 5v (1 diode drop I believe but I've lost the circuit
I saw) any signal will be clamped at around 3V. Below this the device has
low on resistance and therefore will pass fast signals. Because the device
is purely a switch it is fully bidirectional and should not be affected by
floating signals. I haven't tried this, it came from a circuit ideas
booklet I got a couple of years ago.


RE LCDs. My experience has been that the HD44780 LCD driver will run from
3.2v, although such use appears undocumented. The snag is the display will
be barely visible. There is a TRICK to get round this. Vo (contrast) CAN BE
TAKEN BELOW GROUND! VCC-Vo determines contrast, so a weak negative supply
restores contrast. Below 3V a test circuit still appeared to work (an old
16c84 seems nearly unstoppable) but the display faded anyway, so the logic
is probably giving up.


Oliver.

--
http://www.piclist.com hint: The PICList is archived three different
ways.  See http://www.piclist.com/#archives for details.

2000\07\07@191159 by Isaac Wagner

flavicon
face
Pericom also makes chips that can be used in mixed voltage systems. We use
the 74LPT541 in some of our designs. The only limitation is that they are
unidirectional. Other than that they work great.

-Isaac Wagner


----- Original Message -----
From: "Michael Rigby-Jones" <mrjonesspamKILLspamNORTELNETWORKS.COM>
To: <.....PICLISTKILLspamspam.....MITVMA.MIT.EDU>
Sent: Thursday, July 06, 2000 12:55 AM
Subject: Re: [EE]: Interfacing to the ATA-interface


> > {Original Message removed}

More... (looser matching)
- Last day of these posts
- In 2000 , 2001 only
- Today
- New search...