Searching \ for '[EE] Multiplexing non-TTL signals' 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=multiplexing+non
Search entire site for: 'Multiplexing non-TTL signals'.

Exact match. Not showing close matches.
PICList Thread
'[EE] Multiplexing non-TTL signals'
2010\01\31@232646 by Vitaliy

face
flavicon
face
Hi List,

Say we have a circuit with a number of different signal lines (CAN, RS232,
etc) that we'll call A, B, C, D. Now imagine that we have a connector with
13 pins (1 thru 13). Finally, we have a requirement to switch any line to
any pin (e.g., B can go to any pin, 1 thru 13).

If the lines were TTL, one could use a 1-to-13 multiplexor on each pin (and
the software would ensure that no two signals are wired to the same pin).
However, what do you do if the lines can have different voltages and won't
tolerate a high series impedance (>10 Ohms)?

Reed relays and FETs (perhaps wired through multiplexors to reduce required
PIC pin count) naturally come to mind. Unfortunately, using discrete
components means excessively high BOM and assembly costs.

Any other ideas?

Thank you in advance,

Vitaliy


'[EE] Multiplexing non-TTL signals'
2010\02\01@000114 by ivp
face picon face
> Any other ideas?

Op-amps ?

2010\02\01@012006 by Richard Prosser

picon face
Vitaliy,
What about analogue mux's ? You may have to use something better than
the 4051/52/53 series to get a low on resistance though. Otherwise as
long as the voltage is within range it should work OK.  Maybe parallel
up devices to get a lower resistance ?

RP

On 1 February 2010 17:01, ivp <spam_OUTjoecolquittTakeThisOuTspamclear.net.nz> wrote:
>> Any other ideas?
>
> Op-amps ?
> -

2010\02\01@013146 by Sean Breheny

face picon face
Analog devices makes sub 1 ohm analog switches but I'm sure they are
costly. There might even be a cross-point version with lots of inputs
and outputs.

You might also be able to use high-resistance (about 100 ohm) cheap
analog switches followed by analog buffers (fast op-amps, for example)
on the outputs. There might even be such a beast available. Do you
need each pin to be able to be either input OR output, or is that
fixed?

Sean


On Mon, Feb 1, 2010 at 1:20 AM, Richard Prosser <.....rhprosserKILLspamspam@spam@gmail.com> wrote:
{Quote hidden}

>> -

2010\02\01@013333 by Wouter van Ooijen

face picon face
> Any other ideas?

Make clear to the one who created this requirement that it is
excessively expensive!

Maybe you can find a cheap driver chip that can do everything, and put
such a chip on every pin. Do the routing at the input side of the driver
chip. The receiver can probably tolerate an analog mux in series.

--

Wouter van Ooijen

-- -------------------------------------------
Van Ooijen Technische Informatica: http://www.voti.nl
consultancy, development, PICmicro products
docent Hogeschool van Utrecht: http://www.voti.nl/hvu

2010\02\01@040510 by Alan B. Pearce

face picon face
>However, what do you do if the lines can have different
>voltages and won't tolerate a high series impedance (>10 Ohms)?

You mean apart from why you need such flexibility ... ??

I guess you could use something like a bunch of DG201 type multiplexors,
provided you kept things like the RS232 lines within +/-12V. Using suitable
series TVS clamps should be able to limit input voltage excursions enough to
protect them.

2010\02\01@135520 by Vitaliy

face
flavicon
face
ivp wrote:
>> Any other ideas?
>
> Op-amps ?

Care to elaborate? Example circuits? :)

Vitaliy


2010\02\01@151812 by Vitaliy

face
flavicon
face
Richard Prosser wrote:
> What about analogue mux's ? You may have to use something better than
> the 4051/52/53 series to get a low on resistance though. Otherwise as
> long as the voltage is within range it should work OK.  Maybe parallel
> up devices to get a lower resistance ?

Yeah, we looked at analog muxes, unfortunately they were designed with
slightly different applications in mind. Paralleling them up is a
possibility, but one of the things we're after is minimizing the part count.

Vitaliy

2010\02\01@152525 by Vitaliy

face
flavicon
face
Sean,

You may be on to something, turns out that Analog has a bunch of switches
that we may be able to use.

Some of the lines are bidirectional.

Vitaliy


{Original Message removed}

2010\02\01@153127 by Vitaliy

face
flavicon
face
Wouter van Ooijen wrote:

>> Any other ideas?
>
> Make clear to the one who created this requirement that it is
> excessively expensive!

LOL, unfortunately this decision had been made a long time ago, with no
input from us. :)

We're still in the exploratory phase of this project, in reality the
requirements are a bit more lax (each line must be switchable with a subset
of the pins, although it depends on the pin).


> Maybe you can find a cheap driver chip that can do everything, and put
> such a chip on every pin. Do the routing at the input side of the driver
> chip. The receiver can probably tolerate an analog mux in series.

That's what we're leaning toward, especially after I checked out the
switches that Sean suggested.

Vitaliy

2010\02\01@153729 by Vitaliy

face
flavicon
face
Alan B. Pearce wrote:
> >However, what do you do if the lines can have different
>>voltages and won't tolerate a high series impedance (>10 Ohms)?
>
> You mean apart from why you need such flexibility ... ??

:-)


> I guess you could use something like a bunch of DG201 type multiplexors,
> provided you kept things like the RS232 lines within +/-12V. Using
> suitable
> series TVS clamps should be able to limit input voltage excursions enough
> to
> protect them.

100 Ohms Drain-Source resistance is about 10 times greater than what is
acceptable for this application.

Vitaliy

2010\02\01@160521 by ivp

face picon face
>> Op-amps ?
>
> Care to elaborate? Example circuits? :)

Er....um.....I'm exercising my right to remain silent

So sorry, was thinking buffers not switches. BTW, presumably
these do have to be electronically-controlled selectors ?

wbr

2010\02\01@180926 by Vitaliy

face
flavicon
face
ivp wrote:
>>> Op-amps ?
>>
>> Care to elaborate? Example circuits? :)
>
> Er....um.....I'm exercising my right to remain silent
>
> So sorry, was thinking buffers not switches.

Heh. It's alright. :)


> BTW, presumably
> these do have to be electronically-controlled selectors ?

Preferably. Mechanical user selectable option is on the table, but is far
far less attractive.

Vitaliy

2010\02\02@001204 by William \Chops\ Westfield

face picon face

On Jan 31, 2010, at 8:24 PM, Vitaliy wrote:

> Say we have a circuit with a number of different signal lines (CAN,  
> RS232,
> etc) that we'll call A, B, C, D. Now imagine that we have a  
> connector with
> 13 pins (1 thru 13). Finally, we have a requirement to switch any  
> line to
> any pin (e.g., B can go to any pin, 1 thru 13).

cisco serial interfaces have a connector that supports EIA-232, X.21,  
V.35, EIA-530, EIA-530A, and EIA-449 in DTE and DCE configurations,  
depending on a magic cable to identify which one you want.  I believe  
it makes extensive use of driver chips that are individually disable-
able, rather than trying to mux the analog side of the drivers.  It's  
not a particularly large circuit (given modern-sized chips.)
  No CAN.
  I don't know a lot about it.
  I believe it's the subject of patent or two, so even if I
   did know about it, I couldn't talk about it.

But looking at one might give you some clues...

BillW

2010\02\02@100540 by Geo

picon face
Vitaliy wrote:
> Say we have a circuit with a number of different signal lines (CAN, RS232,
> etc) that we'll call A, B, C, D. Now imagine that we have a connector with
> 13 pins (1 thru 13). Finally, we have a requirement to switch any line to
> any pin (e.g., B can go to any pin, 1 thru 13).
>
> Reed relays and FETs (perhaps wired through multiplexors to reduce required
> PIC pin count) naturally come to mind. Unfortunately, using discrete
> components means excessively high BOM and assembly costs.
>
> Any other ideas?
>  
Rectangular pogo pin array and plastic punched cards?
Similar to the Mullard valve tester where siganls routed to various pins
and levels set.





2010\02\08@182816 by Larry G. Nelson Sr

picon face
I have done this type of things for metrology applications on
industrial lasers and ion beam implanters. Very low leakage required
and able to survive a high voltage discharge. We test at 25KV air and
contact discharge for ESD.


At 04:05 AM 2/1/2010, you wrote:
> >However, what do you do if the lines can have different
> >voltages and won't tolerate a high series impedance (>10 Ohms)?
>
>You mean apart from why you need such flexibility ... ??
>
>I guess you could use something like a bunch of DG201 type multiplexors,
>provided you kept things like the RS232 lines within +/-12V. Using suitable
>series TVS clamps should be able to limit input voltage excursions enough to
>protect them.
>
>

More... (looser matching)
- Last day of these posts
- In 2010 , 2011 only
- Today
- New search...