Op Amp ADC signal conditioning

I am going to assume you are working in the 2.5--5.5 range of supply voltage

They have PDIP packages of this type and you can find them all over and they are rail to rail with a good output drive capability..

--
I'm confused.

What do you mean by "offset"?

ning

d

ply

Chapter 4 in the book "Op Amps for Everyone" is about single supply Op Amp design techniques. The chapter also suggest some rail to rail Op Amps to use for off setting and scaling a signal. The book is available on the web at:

Howard

What's wrong with a voltage divider? What do you actually need for output?

Tim

-- Deep Friar: a very philosophical monk. Website:

I may not have termed it correctly for op amps but in PLC programming we get an a/d and scale and offset the reading to real units. For example converting from Deg C to Deg F the scale would be 1.8 and the offset would be 32.

To get high resolution from the A/D over a wide range of voltage, I plan to offset the ranges using a differencing amp. Output of the amp will be 0-5V with 0-5V IN and no offset, 0-5V Out with 5-10V in with 5V offset, 0-5V Out with 10-15V in with a 10V offset, etc. For the offsets perhaps an analog output, digital pot, or reference voltages with a resistor voltage divider selectable with analog multiplexer. Of course it might be easier to just use a higher resolution A/D converter. If nothing else I'll need the amplifier to amplify the current sense resistor.

RogerN

"RogerN" wrote in news:5_2dnfxHQPzsE3TRnZ2dnUVZ snipped-for-privacy@earthlink.com:

Is it single ended or differential? How many bits?

If you have a 16b ADC you should look at low offset choppers. One is the AD8628ARZ low noise low drift low intial offset.

If its 12b or less you could use maybe the OPA2727AIDR

TI has some good calculators to use.

Fully Differential Amplifier calc

and

Op Amp Gain and Zero Shift Status: calculator.

I'm wanting:

0 to 5V OUT with 0 to 5V IN and 0V offset, 0 to 5V OUT with 5-10V IN with a 5V offset, 0 to 5V OUT with 10-15V IN with a 10V offset, 0 to 5V OUT with 15-20V IN with a 15V offset, and so on up to about 50V input. that way I can select my range and still get high resolution from my A/D. Maybe there is a better (easier) way?

RogerN

"RogerN" wrote in news:64edncjDGKJOVXTRnZ2dnUVZ snipped-for-privacy@earthlink.com:

Yea factor your 50V by ten using a resistive divider. Then buffer that to your ADC.If you can keep it in the 5V range their are more op-amp's and ADC's to choose from.

Many ADCs require the signal to be supplied by a low impedance source for optimum performance .

Howard

fer

\Many ADCs require the signal to be supplied by a low impedance source \for optimum performance . \ \Howard

I'm hoping the op-amp to the A/D and a capacitor at the input will do a sufficient job. This is for battery charging, reading voltage and current, so I don't need anything high speed, I probably need some filtering to get the SMPS noise out or averaged.

RogerN

Wouldn't a negative constant current source do the trick? If there's a

100K Ohm series resistor on the input then 0µA, -50µA, -100µA, -150µA.

If you don't need full precision at the highest frequencies you can add dithering noise to the input. Precision will increase as you average more noisy values together.

--
I will not see posts or email from Google because I must filter them as spam

But! How is this going to fix the windowing gain that he is looking for? All his is looking for is to have a 0..5 V out with a min and max voltage range to drive the ADC.

I've made such a window gain circuit before by using a diode from another base reference for the external reference to work against. This allowed for the op-amp to rest at 0 output at anything below the fixed minimum fixed reference.

I really don't think he is simply looking for a range scale! otherwise I would've thought he'd simply specify 0..10,0..15,0..20 etc...

Just my thoughts..

Why not scale down with a voltage divider or less-than-unity-gain amplifier. For current sense you could use a (get ready) "current sense amplifier". ;-) ADI has a bunch of them:

Forgot to add, what precision, tenths or nanovolts??

Tim

--
Deep Friar: a very philosophical monk.
Website: http://webpages.charter.net/dawill/tmoranwms

Oh, well hell, you hardly even need 8 bits, do you? A nice long divider with bypass cap will keep CMOS ADCs happy. Any uC has a 10 or 12 bit unit onboard, which more than accounts for the 5/40th divider you'll need.

10 bits of 40V is slightly less than 40mV. You can get a bit more with dithering, which can be done subtractively through a spare PWM for even better. Bandwidth literally doesn't matter, so you could run the whole thing slow and filter the piss out of it. You could also shut off the uC core (and most peripherals besides PWM) during the settling time to save even more current. And turn everything off for maybe a whole second between readings, just because the sample rate can be pitiful.

Tim

--
Deep Friar: a very philosophical monk.
Website: http://webpages.charter.net/dawill/tmoranwms

t

Just get an accurate input resistor and use a current-to-frequency converter type of ADC. Then, a long gate time gets you part-per-million resolution, while a shorter gate time gets you the appropriate resolution for that 0...5V range. To do it with a single-chip micro's ADC, you have to go through a lot of hoopla to keep the input potential in the ADC range and that means multiple calibrations (for the different ranges and offsets).

RogerN schrieb:

Hello,

there is a better and easier way, just take an ADC with 4 bits more for the input from 0 to 50 V. It will give you better results than your 10 selectable ranges with different offsets.

Bye

Jamie wrote in news:IZkGo.35739$ snipped-for-privacy@newsfe10.iad:

Actually when he shows the same 0-5V output that is exactly what I think he is looking for.That must mean his refrence is 0V and 5V full scale. After reading his intentions; yep scale the the input using a resistive divider.

The most accurate charge requirements I have are for Lithium batteries, I'm hoping to be accurate to maybe 0.02V on the lowest range. One time I heard the charging tolerance for LiPo batteries was around 1%, not sure it that's still true or if they are asking for better these days.

RogerN