Tuning a R-2R Ladder

A

ADWSystems

Member

• Mar 26, 2017

• #21

That's what it was pushing on the bench. Not much more I can say.

 

ronsimpson

Well-Known Member

Most Helpful Member

• Mar 26, 2017

• #22

If I was doing this (not that we really know what you want), I would use two 8 pin ICs and two resistors.
The DAC is any 6 or 8 pin serial DAC. The op-amp simple.
First make any voltage from 0 to 5 volts in 1mV steps.
Amplify by 4.
Now you have 0 to 20 volts in 4mV steps.
(edited)Should say DAC in the picture.

 

Last edited: Mar 26, 2017

A

ADWSystems

Member

• Mar 26, 2017

• #23

Yep. Been there done that (except I used a DAC). Prototype I eluded to a few posts back, is currently in the trash. I could not get under 10mV error at each 1 volt step. All of the readings were close to 50mV error. The error was not linear but absolute per step (ie., varied by step, but always under the actual reading).

While we're talking ADC/DAC. Any ideas on how to keep the 1mV resolution from a 0-5V ADC that must read a 0-15V input signal? I'm thinking I need to subtract the "excess" voltage because a resistor divider will squash the resolution.

(yes, I don't have all the details. I'm not in the office and I'm working on the weekend)

 

Last edited: Mar 26, 2017

ronsimpson

Well-Known Member

Most Helpful Member

• Mar 26, 2017

• #24

Thanks for the proof reading.
It is hard to substract voltage.
Have you thought about a diff amp. With 10V (ref) on one input and you unknown voltage on the other input, the output could move 10 to 15V down to 0 to 5V. You will not see the 0 to 10V input range.

 

A

ADWSystems

Member

• Mar 26, 2017

• #25

I had thought of that. I'm concerned about what might happen when Vx gets near Vref. IIRC, they can't get too close together. In the past they stopped at some voltage apart, I think it was CMRR. If I am remembering that correctly, then I might end up with some type of automatic "gain" (offset) control in three or four stages where only the middle 4-4.5V of the 5V range is used.

Oh what a tangled web we weave.

 

P

Pommie

Well-Known Member

Most Helpful Member

• Mar 26, 2017

• #26

I don't understand the problem. A 14 bit DAC can be had for under $5 and will easily give the accuracy you need with Ron's suggestion above.

Mike.

 

AnalogKid

Well-Known Member

Most Helpful Member

• Mar 26, 2017

• #27

ADWSystems said:

Yep. Been there done that (except I used a DAC). Prototype I eluded to a few posts back, is currently in the trash.

Click to expand...

What are the part numbers of the DAC and opamp? Schematic?

ak

 

ronsimpson

Well-Known Member

Most Helpful Member

• Mar 26, 2017

• #28

DAC: MCP4725, MCP4726, MCP4921 There are many others like these.

I looked for a op-amp that has a supply voltage of greater voltage 20V. I want a R-R output so you can use a 20V supply and the op-amp output can go from 0 to 20V. (most amps can not do that) The input range needs to be 0 to 5V. (most will not go to 0) The first amp I found is OPA192. There others but you can look for them. .
OPA192 supply as high as 36V, output is rail to rail (R-R), input is R-R

 

A

ADWSystems

Member

• Mar 27, 2017

• #29

I was looking at the MC34072. Yes? No?

 

ronsimpson

Well-Known Member

Most Helpful Member

• Mar 27, 2017

• #30

ADWSystems said:

MC34072

Click to expand...

This tell us the inputs will work all the way down to "ground" but must stay away from "20V". This is good! (Assuming the power supply pins are connected to ground and 20V)

So the example here show a (+5, -5V) and a (+15, -15V) supply. With 2k and 10k loads. We can learn that the output can not get to the (+) and (-) supply. 1.4 away from plus and 0.4 away from the negative supply. (that is why I looked for a R-R amp)
In this case; to have an output go from 0 to 20V you need a supply larger than 0 to 20V. (maybe -5V and 25V) or (-3.3 and +25)

 

A

ADWSystems

Member

• Mar 27, 2017

• #31

In a diff amp usage, does Vicr=Vcc-1.8 and Voh=13.6 (ie 1.4 below the rail) matter? Both cannot/neither can reach the rail by almost the same amount.

On the flip side, Vicr=Vee(0V) but Vol=0.1-0.7V is more "unfortunate" (not sure of the right word).