Digital to Analog converter concept problem

AI Thread Summary
The discussion centers on the derivation of the voltage output equation for a weighted resistor digital-to-analog (D/A) converter using a 4-bit resistive ladder. The initial equation presented is challenged due to the omission of the load resistance (RL), which significantly affects the output voltage (VA). Participants clarify that if RL is assumed to be zero, VA would be grounded, indicating that RL plays a crucial role in determining VA. The conversation concludes with the realization that for accurate D/A conversion, RL must be much larger than the resistors in the ladder to prevent loading effects that could distort the output voltage. The final derived equation for VA is confirmed to be (8V3 + 4V2 + 2V1 + V0) / 15, emphasizing the importance of considering RL in practical applications.
nil1996
Messages
301
Reaction score
7

Homework Statement


Hello PF :smile:
i am stuck with a concept of weighted resistor D/A converter.
For a 4-bit resistive ladder,

VA={V0/R0+V1/(R0/2)+V2/(R0/4)+V3/(R0/8)}/{1/R0+1/(R0/2)+1/(R0/4)+1/(R0/8)}

attachment.php?attachmentid=65838&stc=1&d=1390196087.png


I want to find out how the equation is derived but have no clue.

Thanks.
 

Attachments

  • Untitled.png
    Untitled.png
    12 KB · Views: 566
Physics news on Phys.org
Write the equations of this circuit and solve.
For example:
V0 = R0 i0 + RL iL
...
iL=i0+i1+i2+i3
 
You'll have a hard time doing that because your expression omits RL which is clearly necessary.
 
finally i have ended up with this:


VA/RL=V0/(R0+RL) + 2V1/(R0+2RL) +4V2/(R0+4RL) + 8V3/(R0+8RL)

Any way to get rid of RL??
 
nil1996 said:
Any way to get rid of RL??

No. Assume RL=0. What would VA be then?
 
rude man said:
No. Assume RL=0. What would VA be then?
thanks for the reply :smile:

if RL is zero then VA would be grounded and hence it will be zero.
 
nil1996 said:
thanks for the reply :smile:

if RL is zero then VA would be grounded and hence it will be zero.

Right. So VA depends a lot on RL, does it not.
So your given answer can't be right.

So now let's do it right:
Sum currents at VA to zero: (V3-VA)/(R0/8) + ... = VA/RL.
Fill in the blanks. and solve for VA.
 
rude man said:
Right. So VA depends a lot on RL, does it not.
So your given answer can't be right.

the equation --VA={V0/R0+V1/(R0/2)+V2/(R0/4)+V3/(R0/8)}/{1/R0+1/(R0/2)+1/(R0/4)+1/(R0/8)} is right from my textbook.

rude man said:
So now let's do it right:
Sum currents at VA to zero: (V3-VA)/(R0/8) + ... = VA/RL.
Fill in the blanks. and solve for VA.

i have got that equation on Post No. 4
 
here is the equation:

(V3-VA)/(R0/8) + (V2-VA)/(R0/4)+(V1-VA)/(R0/2)+(V0-VA)/(R0)=VA/RL
 
Last edited:
  • #10
nil1996 said:
the equation --VA={V0/R0+V1/(R0/2)+V2/(R0/4)+V3/(R0/8)}/{1/R0+1/(R0/2)+1/(R0/4)+1/(R0/8)} is right from my textbook.



i have got that equation on Post No. 4

Well, your textbook is right wrong. See my next post.
 
  • #11
nil1996 said:
here is the equation:

(V3-VA)/(R0/8) + (V2-VA)(R0/4)+(V1-VA)(R0/2)+(V0-VA)(R0)=VA/RL

That is wonderful!
Oops, you forgot the divide signs for several terms on the left ...
 
Last edited:
  • #12
rude man said:
That is wonderful!

My textbook says """ Resistance RL represents load to which divider is connected and RL is considered to be large enough so that it does not load the divider network.


For a 4-bit resistive ladder,

VA={V0/R0+V1/(R0/2)+V2/(R0/4)+V3/(R0/8)}/{1/R0+1/(R0/2)+1/(R0/4)+1/(R0/8)} """
 
  • #13
so does RL>>R0 affects the equation?
 
  • #14
nil1996 said:
so does RL>>R0 affects the equation?

If RL >> R0 then RL drops out, approximately.
So then, compute VA for RL >> R0 ...

I have the answer if you want to check yours ...

PS R0 drops out too ... which is obvious if you look at your given solution. Multiply numerator and denominator by R0.
 
Last edited:
  • #15
Can you explain me how RL drops out?
And what does it mean by ""RL is considered to be large enough so that it does not load the divider network.""
 
  • #16
nil1996 said:
Can you explain me how RL drops out?
And what does it mean by ""RL is considered to be large enough so that it does not load the divider network.""

Solve the equation in your post #9 for VA. Then let RL >> R0.
That will answer both your questions.

(The divider network is R0, R0/2, R0/4 and R0/8.)
 
  • Like
Likes 1 person
  • #17
solving that i get
(8V3+4V2+2V1+V0)/15=VA
 
Last edited:
  • #18
thanks a lot!:smile:
I got it.:thumbs:
(My textbook should have stated that it is appox.)
 
  • #19
nil1996 said:
thanks a lot!:smile:
I got it.:thumbs:

(My textbook should have stated that it is appox.)

Good work!
Yes, it should have.
A finite RL reduces the output voltage by a constant: R0/(15R0 + RL) instead of just 1/15.

The idea here is that in a real application there will be some finite value of RL; this computation shows that it should be >> R0 or the output voltage will be a function of RL which makes the d/a converter inaccurate since RL is usually not constant.
 
Back
Top