Searching \ for 'Looking for a SIP R/2R network' 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=looking+sip+r2r
Search entire site for: 'Looking for a SIP R/2R network'.

Truncated match.
PICList Thread
'Looking for a SIP R/2R network'
1996\10\23@235527 by dfr

flavicon
face
John Payson wrote:

> expensive and bulky.  For example, an 8-bit R/2R DAC with a cascading
> input requires sixteen discrete resistors or four 8-pin resistor packs.
> Using 16 resistors would take 0.4"x1.6" of board space (that's as big
> as a 32-pin skinydip!) while four resistor packs would take 0.4"x0.8"
> (still almost as big as a PIC!).

The reason that they're expensive is that one would use such a beast
only when extreme accuracy is required. Otherwise, use 0.1% 25ppm
resistors in close proximity to each other. Therefore, laser trimming is
used. I would expect better temperature tracking since the resistors are
fabricated at the same time, with the same material, and on the same
substrate.

One inexpensive (and low accuracy) approach I used recently turned 6
16C54A output ports into a DAC with R2^n ordering of port series
resistors fed to a load resistor and noise cap. This >1K source
impedance DAC was dual-slope ramped (up then down) into a pair of
comparators. One compared the ramp voltage to a temperature setpoint POT
while the other compared the ramp to a temperature sensor output. This
approach used 7 resistors and a cap for the DAC.

The two biggest sources of error are: The tolerances of the individual
resistors (and how well they track with temperature), and the variation
of source impedance of the port drivers themselves (50R or so). Of
course there are several other sources of error (VCC variation, etc).
The firmware detected when a comparator changed state on each slope...
The reading used was the average of the two points.

This is a very quick and inexpensive method of DAC implementation, and
works well for low slew rate, low accuracy signals. The above
implementation was for a temperature controller. The nice thing about it
was that I bit-banged 5V RS232 ASCII status messages on a spare output
port. They transmitted the setpoint and temperature readings 4 times a
second. During post-production test&repair an IC clip is fitted over the
PIC to send the text status messages to an old QUME terminal. It is the
ease of adding such additional-value features at little or no expense
that I appreciate the use of inexpensive microcontrollers like the PIC.
The software was written in MPC C.

--
Regards, Dana Frank Raymond

1996\10\24@132326 by Bob Blick

picon face
I have seen r2r resistor packs, they do exist, sorry I don't know who sells
them. The ones I saw were manufactured by Dale, and they were in "parallel
port sound adapters". Chances are you could call Dale, get a part number,
make a special order for 1000 and wait 8 weeks only to be told the delivery
date has been extended to August, 1997, or that the part number you gave
them was invalid. :-<
>
>> expensive and bulky.  For example, an 8-bit R/2R DAC with a cascading
>> input requires sixteen discrete resistors or four 8-pin resistor packs.
>> Using 16 resistors would take 0.4"x1.6" of board space (that's as big
>> as a 32-pin skinydip!) while four resistor packs would take 0.4"x0.8"
>> (still almost as big as a PIC!).
>

More... (looser matching)
- Last day of these posts
- In 1996 , 1997 only
- Today
- New search...