Shop OBEX P1 Docs P2 Docs Learn Events
[solved] No fun in MDAC land. — Parallax Forums

[solved] No fun in MDAC land.

frank freedmanfrank freedman Posts: 1,983
edited 2012-05-16 00:19 in Propeller 1
I seem to be missing something maybe obvious here playing with an LTC8043 MDAC. The prop driver is not a problem, the chip can be made to put out a current proportional to the digital value in. With a +9V reference, I can go from about 0 uA to around +635uA. My problem seems to be converting this current into a useful voltage range using a single supply op-amp. Specifically an NTE 928. After exploring multiple pages and methods, none seem to really explain the topic well enough for me to use. Can this even be done using a single supply opamp?<br>
<br>
Frank

Comments

  • pik33pik33 Posts: 2,398
    edited 2012-05-14 23:13
    If you put 1 k resistor, you get 625 mV from 625 uA, so single opamp in classic non inverted mode is sufficient to convert it to some volts.

    [FONT=courier new]
from dac                         +------------------------+
             up to 625 mv here   |                        |
---------------+-----------------| + input                | 
               |                 |                        |
              +-+            +---| - input                |     0..6.25 V here
              | |            |   |                 output |--+--------------   
              | |1k          |   |                        |  |
              +-+            |   |                        |  |
               |             |   +------------------------+  |
               |             |                               |
               _ gnd         +-----------| 9k |--------------+ 
                             |
                            +-+
                            | | 1k
                            | |
                            +-+
                             |
                             - gnd

[/FONT]


    NTE 928 seems to contain two opamps :) and ltc8043 contains a 10k feedback resistor which you can use in inverting opamp config like in PDF attached to ltc8043, and some opamp schematics:

    cds.linear.com/docs/Datasheet/lt8043.pdf
  • Christoph_HChristoph_H Posts: 31
    edited 2012-05-15 00:19
    Frank, if I understand the topology correctly the IOUT pin has to be at the same level as the GND pin. Otherwise the output current wouldn't be in a linear relation to the given digital input code.
    To keep IOUT at GND level you have to sink the same current that IOUT is sourcing; therefore the connected op amp output would have to go below ground.

    Generating the missing negative supply for your op amp would seem to be the easiest option if you have to keep this DAC.
  • Mike GMike G Posts: 2,702
    edited 2012-05-15 05:57
    Current-to-voltage converter
  • Peter JakackiPeter Jakacki Posts: 10,193
    edited 2012-05-15 07:16
    The non-inverting configuration will work but fails to keep the output at ground which the inverting configuration does through it's virtual ground at the inverting input. How about a compromise and use the non-inverting configuration but use a shunt resistor of around 10R and adjust the gain of the amplifier accordingly. I'm not sure what bandwidth you are after but a better opamp would be in order in this higher gain configuration as the NTE928 is just a very basic LM358.

    @Mike G: yes of course, but all the opamp circuits shown assume you have a negative supply (notice they don't show the supply rails). How to do it with a single supply for both this MDAC and the opamp is the trick.
  • Mike GMike G Posts: 2,702
    edited 2012-05-15 10:58
    Single to voltage converter single supply
  • frank freedmanfrank freedman Posts: 1,983
    edited 2012-05-15 11:08
    Thanks Peter, that has been the thing with nearly all sites, they don't show the supply connections so I am left guessing. More fun tonight? We shall see. May get out the dual supply to get it going, then migrate to single supply.... Just fire up my trusty Heathkit triple power supply.......

    Thx,

    Frank
  • Peter JakackiPeter Jakacki Posts: 10,193
    edited 2012-05-15 18:47
    Mike G wrote: »
    Single to voltage converter single supply
    Sorry Mike, I'm not quite sure what you are getting at as you are supplying links as if they are the solution (they're not) but not saying anything to elucidate us further as to how any of this could apply in this particular circumstance (remember the MDAC topology).

    @Frank: There won't be any problem with going with the non-inverting single supply solution as long as you use a very low value shunt to equalize the two arms of the MDAC's resistor ladder. The only problem is due to the higher gain and the slew-rate limitations of the LM358 which will limit the bandwidth but that's easily fixed by using a better opamp if necessary.
  • frank freedmanfrank freedman Posts: 1,983
    edited 2012-05-16 00:19
    IT'S ALIVE!!!!!! IT'S ALIVE!!!!!!!!!! IT'S ALIVE!!!!!!!!!
    A huge (pun intended) thanks to Peter and PIK33. PIK33's circuit proved I was on the correct path with non-inverting input, and Peter supplied the missing link with his comment regarding the shunt. I tested using a 100 ohm and a 1k ohm shunt and currently the 1k seems to be giving me a nice response without too overdriving the opamp. I think that Christoph_H was getting toward the same concept. Virtual ground was not quite. Once I am really driving this circuit, I may have to review the shunt R choice in light of Cristoph_H's comment regarding linearity. No luck using an inverting configuration. So for now I will stick with what works.

    As to the NTE, it was cheap and available when I was at fry's electronics and since I bought 3 of them when I built the IKALogic R2R DAC a few months ago (already that long?) I just used it rather than getting better from digikey. 3 because you always need a spare of something, and I did not know if I would blow up one or two when I hacked the original design into a 12 bit converter instead of 8.

    Now I can get back to the PASM output for this device and really have some fun with the multiplying function. Yeah, I could have just dumped the LTC8043 on the to lazy for that pile and gone to a voltage output simple (MCPxxxx) DAC, but where would the fun be in that? Once the PASM driver is done, I will have to revisit (correct again Peter) the datasheet library to determine a more appropriate opamp for the anticipated bandwidth and drive though this will still remain a two quadrant multiplier configuration for now. Maybe try for four later w/ dual supply opamps etc.

    Thanks again y'all,

    Frank

    P.s. Not my intent to stir potential contention, but I typically ask for help when all normal avenues are unhelpful, and even google failed to come up with the final shunt suggestion for this. Heck, even one of the best references and sources, MCP's Bonnie Baker did not mention this issue with MDACs. then again maybe she did and it was missed, I returned to amazon the ebook version of her Bakers Dozen book because the Kindle format was flat near useless due to unreadable diagrams and other limitations as reviewed on amazon. Yes, I will be buying the hardcopy version unless I find her publisher releases a well put together PDF version of the book.

Sign In or Register to comment.