Problem with voltage follower in R2R D/A converter

vsaar

New Member
Hi,

I'm testing an 8-bit R2R ladder based D/A converter and I can't get it to work properly. This will be a general purpose converter to be used with an ATmega16 development board. The resistor ladder itself works OK; the output voltage goes in even steps from 0 to +5V, even the 01111111 - 10000000 step is not bad because I carefully measured and picked those resistors to minimize uneven steps.

However, the voltage follower opamp works only from 0 to about 3.7V, but any input voltage above that makes the output to be 5V. The opamp is LMC660 which is said to be rail-to-rail, but is it really so? I think the problem would be solved by having at least 6.5V supply for the opamp (and no need for rail-to-rail then), but there's no more than 5V available. Also, I don't want to reduce the output range to overcome that problem. Is there any other opamp which would really work up to +V, or some other ideas?
 
The LMC660 is doing what is should do.
The output is R-R but the input is not!
The input will work to "0 volts" BUT It will not work above VCC-2.5 volts. In your case of VCC=5V so it will stop working at about 2.5 volts. The fact that it works to 3.7V is nice.

Look at:"Input common mode range"
 

Oops! I just learned something new...

However, I also found a simple solution to increase the supply voltage for LMC660. I found a schematic for Dickson charge pump, and luckily I had components for testing it. I managed to get 7.3V from it, which was enough to have LMC660 output go all the way to 5V. I think this works good enough for me, and I still can fit this to the perfboard I planned to use.

This is the schematic I found: https://en.wikipedia.org/wiki/File:Dickson_doubler.svg
For the "clock" I made a 555 based oscillator running about 700Hz, and those caps are 220uF.
 
Cookies are required to use this site. You must accept them to continue using the site. Learn more…