Searching \ for '[EE] Once again: Connecting IC with different Vdd' 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/power/decouple.htm?key=vdd
Search entire site for: 'Once again: Connecting IC with different Vdd'.

Exact match. Not showing close matches.
PICList Thread
'[EE] Once again: Connecting IC with different Vdd'
2004\12\30@134259 by Arkady Skorokhod

picon face
I know this was already discussed here but I still cannot get any clear understanding. Could somebody explain it to me, please.

I have two chips IC1 with Vdd @ 3.3V and IC2 with Vdd @ 5V. I want to connect an OUT pin of IC1 (let call it PIN1) to an IN pin of IC2 (PIN2). According to the datasheet for IC1 the maximal voltage for PIN1 cannot exceed 3.6V.
Question1: Can I safely connect directly PIN1 to PIN2?
Question2: Is the information sufficient or does the answer depend on what is behind PIN1 and PIN2 (are these TTL, CMOS and such)? (I suspect that the answer on question1 will be YES if PIN2 is CMOS, and NO if TTL (like LS series))
Question3 (more specific one): Let PIN1 be an OPEN COLLECTOR and PIN2 - a MCLR of a PIC. What is the answer then (on question1)?
Question4 (backward): Let PIN3 be some OUT pin of same IC2 and PIN4 an IN pin of IC1 (maximal voltage for PIN4 being still 3.6V of course). Let PIN3 be an OPEN COLLECTOR. Can I just connect the pins (adding a pull-up resistor to the 3.3V rail)? (I think - 'yes').

TIA.

2004\12\30@195235 by Russell McMahon

face
flavicon
face
I know this was already discussed here but I still cannot get any
clear understanding.
Could somebody explain it to me, please.

I have two chips IC1 with Vdd @ 3.3V and IC2 with Vdd @ 5V. I want to
connect an OUT pin of IC1 (let call it PIN1) to an IN pin of IC2
(PIN2). According to the datasheet for IC1 the maximal voltage for
PIN1 cannot exceed 3.6V.
Question1: Can I safely connect directly PIN1 to PIN2?
Question2: Is the information sufficient or does the answer depend on
what is behind PIN1 and PIN2 (are these TTL, CMOS and such)? (I
suspect that the answer on question1 will be YES if PIN2 is CMOS, and
NO if TTL (like LS series))
Question3 (more specific one): Let PIN1 be an OPEN COLLECTOR and
PIN2 - a MCLR of a PIC. What is the answer then (on question1)?
Question4 (backward): Let PIN3 be some OUT pin of same IC2 and PIN4 an
IN pin of IC1 (maximal voltage for PIN4 being still 3.6V of course).
Let PIN3 be an OPEN COLLECTOR. Can I just connect the pins (adding a
pull-up resistor to the 3.3V rail)? (I think - 'yes').

PIN1 IC has a maximum voltage it can be exposed to.
Whether using a hi/low driven pin or a pull up, the maximum voltage
seen MUST NOT exceed the maximum allowed operating voltagte during
operation or ther maximum allowed max voltage when not operating. Say
this voltage is 3 volts in each case.

PIN2 IC needs a certain hi level voltage to see a high level. This
will be in the spec sheet. IF the PION1 IC makes this much voltage it
will work. if not, it wont. Look at the spec sheet and see.

Interestingly, the high level; required by a "TTL" input is usually
LOWER than required by a CMOS input. Read spec sheet and see.

       RM


2004\12\30@210156 by Jose Da Silva

flavicon
face
On Thursday 30 December 2004 03:50 pm, Russell McMahon wrote:
> I know this was already discussed here but I still cannot get any
> clear understanding.
> Could somebody explain it to me, please.
>
> I have two chips IC1 with Vdd @ 3.3V and IC2 with Vdd @ 5V. I want to
> connect an OUT pin of IC1 (let call it PIN1) to an IN pin of IC2
> (PIN2). According to the datasheet for IC1 the maximal voltage for
> PIN1 cannot exceed 3.6V.
> Question1: Can I safely connect directly PIN1 to PIN2?

This depends on IC2.
Old CMOS chips would preferably have the inputs lifted as high as possible,
otherwise the chip has the PNP and NPN junctions both conducting.
If the input is a schottky type input, then you have a better chance of
saying yes. If IC2 specifications seem to be okay with it, then you are okay
with IC2.
If IC2 has an internal pull-up, you may have trouble with IC1.

> Question2: Is the information sufficient or does the answer depend on
> what is behind PIN1 and PIN2 (are these TTL, CMOS and such)? (I
> suspect that the answer on question1 will be YES if PIN2 is CMOS, and
> NO if TTL (like LS series))

If Pin2 is TTL, I would think it generally more forgiving of the input being
only 3.3v.  LS even better.  If you have a breadboard, why not test IC2 with
pin2 at 5v and at 3.3v and at 3.3v-0.7v, take a look at the current draw on
the power lines, and also take a look at which way the current flows on the
input pin.

> Question3 (more specific one): Let PIN1 be an OPEN COLLECTOR and
> PIN2 - a MCLR of a PIC. What is the answer then (on question1)?

You got a problem then.
You clearly do not have a way for PIN1 to sink the voltage down to 3.3v
because it is open collector and you are going to use some sort of pull-up.
You may want to use a pull-up to the 3.3v instead of the 5v rail but then you
have a problem with the brownout circuit.
Go back to answer of question2. Breadboard test.

What happens to IC2 if the 3.3v drops to 0v?

You may want to look at capacitor coupling.

> Question4 (backward): Let PIN3 be some OUT pin of same IC2 and PIN4 an
> IN pin of IC1 (maximal voltage for PIN4 being still 3.6V of course).
> Let PIN3 be an OPEN COLLECTOR. Can I just connect the pins (adding a
> pull-up resistor to the 3.3V rail)? (I think - 'yes').

Sure.
>From the questions above, it appears it is some sort of monitor, so speed is
not a concern.
If you want to be hardware safe, but curent of little concern, I'd probably
put a pulldown resistor on PIN4, say 330k to GND, and a resistor between
PIN3 and PIN4 of 180k.

2004\12\30@214127 by Russell McMahon

face
flavicon
face
> On Thursday 30 December 2004 03:50 pm, Russell McMahon wrote:

No - none of that was mine. I was answering it but my answers are not
shown.


> Old CMOS chips would preferably have the inputs lifted as high as
> possible,
> otherwise the chip has the PNP and NPN junctions both conducting.
> If the input is a schottky type input, then you have a better chance
> of
> saying yes. If IC2 specifications seem to be okay with it, then you
> are okay
> with IC2.

All this is reasonably correct.

> If IC2 has an internal pull-up, you may have trouble with IC1.

True. You MUSt turm pullups in IC2 off to prevent them pulling IC1
above it's allowed voltage.

> If Pin2 is TTL, I would think it generally more forgiving of the
> input being
> only 3.3v.  LS even better.

True..

**** BUT ****

>If you have a breadboard, why not test IC2 with
> pin2 at 5v and at 3.3v and at 3.3v-0.7v, take a look at the current
> draw on
> the power lines, and also take a look at which way the current flows
> on the
> input pin.

I'm afraid that that is not good advice.
The way to do it is to look at the maufacturers data sheet and decide
what volateg levels IC2 requires.
trying to work it out by seeing what happens when you play or by
measurement MAY give a working result but you don't know if it will
work relaibaly or repeatably. you MUSt use the data sheet. use the
worst case values - ie ones most liable to make it not work.
eg IC2 Vin_hi may be specified as

   1.8v    minimum
   2.05 v typical
   2.2v    maximum

Here use 2.2v as it is the level at which all ICs should work.
Sometimes the manufacturers specs are wrong - but that's aniother
story.


{Quote hidden}

       Russell McMahon

2004\12\30@230037 by Jose Da Silva

flavicon
face
On Thursday 30 December 2004 06:41 pm, Russell McMahon wrote:
{Quote hidden}

I thought the above is a switching threshold spec. Not a static spec.
Sometimes the manufacturer's specs say schottky input, but still need to go
to 0v or 5v.

If the data sheet says nothing about "okay" to "hold" the voltage at 2.2v,
then the person asking needs to look further, and if the data sheet says
nothing, then time to pull-out a breadboard and a potentiometer and slowly
crank the voltage from 0v to 2.5v.

If the chip starts conducting current (like old CMOS chips), then the user
knows that that input cannot hold the voltage static at 2.2v or 3.3v, but
must let it rise to 5v within reasonable time period.

2004\12\31@001630 by Russell McMahon

face
flavicon
face
> I thought the above is a switching threshold spec. Not a static
> spec.
> Sometimes the manufacturer's specs say schottky input, but still
> need to go
> to 0v or 5v.
>
> If the data sheet says nothing about "okay" to "hold" the voltage at
> 2.2v,
> then the person asking needs to look further, and if the data sheet
> says
> nothing, then time to pull-out a breadboard and a potentiometer and
> slowly
> crank the voltage from 0v to 2.5v.
>
> If the chip starts conducting current (like old CMOS chips), then
> the user
> knows that that input cannot hold the voltage static at 2.2v or
> 3.3v, but
> must let it rise to 5v within reasonable time period.

You're correct. That's a fair test when you don't know anything about
the IC.
If at all possible though the manufacturer's spec should be consulted.
Any reputable manufacturer will give a spec for switching levels.


       RM

2004\12\31@102841 by Arkady Skorokhod

picon face
Thank you all replied for the explanation.

One thing's still not so clear to me. Assume that the lower of the two
voltages on the OUT pin is sufficient to drive the IN pin high. Now I should
be concerned about this OUT pin not to be exposed to high voltage possibly
seen on the IN pin. If the datasheet for the second part does not say
explicitly what is the input like (if there is a clamp diode, pull-up
resistor and so), then I CANNOT connect the two ICs directly, right? Or
should I just measure the voltage present on a free-wheeling IN pin, and if
it is lower then the voltage allowed for the OUT pin of the other chip it is
still OK to connect them?

2004\12\31@104455 by olin_piclist

face picon face
Russell McMahon wrote:
> I have two chips IC1 with Vdd @ 3.3V and IC2 with Vdd @ 5V. I want to
> connect an OUT pin of IC1 (let call it PIN1) to an IN pin of IC2
> (PIN2). According to the datasheet for IC1 the maximal voltage for
> PIN1 cannot exceed 3.6V.

That doesn't matter, since PIN1 is an output.  It will be driven by IC1 from
0 to 3.3V.

> Question1: Can I safely connect directly PIN1 to PIN2?

Yes.  PIN2 can tolerate 0-5V, and PIN1 will only produce 0-3.3V.

> Question2: Is the information sufficient or does the answer depend on
> what is behind PIN1 and PIN2 (are these TTL, CMOS and such)?

Yes, no.

> Question3 (more specific one): Let PIN1 be an OPEN COLLECTOR and
> PIN2 - a MCLR of a PIC. What is the answer then (on question1)?

Still yes.  There is nothing inherently unsafe about connecting the pins.
However, the pullup voltage must not exceed the maximum allowable operating
voltage of PIN1.  Note that open collector outputs are often rated for
higher voltages than the chip supply voltage.

> Question4 (backward): Let PIN3 be some OUT pin of same IC2 and PIN4 an
> IN pin of IC1 (maximal voltage for PIN4 being still 3.6V of course).
> Let PIN3 be an OPEN COLLECTOR. Can I just connect the pins (adding a
> pull-up resistor to the 3.3V rail)? (I think - 'yes').

Yes, and this arrangement will even work too.


*****************************************************************
Embed Inc, embedded system specialists in Littleton Massachusetts
(978) 742-9014, http://www.embedinc.com

2004\12\31@161059 by Russell McMahon
face
flavicon
face
> One thing's still not so clear to me. Assume that the lower of the
> two
> voltages on the OUT pin is sufficient to drive the IN pin high. Now
> I should
> be concerned about this OUT pin not to be exposed to high voltage
> possibly
> seen on the IN pin. If the datasheet for the second part does not
> say
> explicitly what is the input like (if there is a clamp diode,
> pull-up
> resistor and so), then I CANNOT connect the two ICs directly, right?

An input pin will "generally" be high impedance and an output pin low
impedance. How high and how low will vary in practice but it must be
true in relative terms for the output to drive the input. As long as
the output dominates the input you will generally be OK.

If there are pullups on the input they should be turned off if
possible. As long as the input pullup value is relatively high (say
100k or more) you will PROBABLY be OK but it IS as you suggest,
possible for a pullup to pull your output to an unallowed voltage
level. If this is a real issue and you can't turn pullups off you
could add an external pulldon such that the open circuit voltage does
not exceed allowed voltage.

eg if you have a 3v and 5v system and a 100k pullup, then adding a
100k/2 x 3 = 150k or less pulldown will never allow the input to go
above 3v. This will usually not be needed.

Clamp diodes are not a problem as they only operate when the input
goes outside its rails. With a 3v output driving a 5v input this will
never happen.



       RM

More... (looser matching)
- Last day of these posts
- In 2004 , 2005 only
- Today
- New search...