1

I find there are at least two methods to specify the nonlinearity of a programmable gain amplifier (PGA). For PGA204, it is given as '% of FSR', but for PGA280, it was given as 'ppm'. It seems the latter are more common. Then why use '% of FSR'? Are there some reason to specify it like this? What does 'FSR' refer to, input or output?

Null
  • 7,448
  • 17
  • 36
  • 48
diverger
  • 5,828
  • 5
  • 43
  • 79

1 Answers1

1

What does 'FSR' refer to, input or output?

Gain non-linearity (whether as ppm or as a percentage of FSR) refers to the output signal non-linearity compared to the input signal should that input signal have been amplified perfectly by the same amount.

Both methods of specification are directly interchangeable so I don't see it (personally) as a big deal. If you are looking for why someone should use % FSR then it's not a sound engineering reason - it boils down to whatever the guy writing the specification is most comfortable with. There is no engineering reason to specify it one way or the other.

Looking at the two data sheets for gains of unity, the PGA204 specifies 0.0004% of FSR and the PGA280 specifies 1.5 ppm. If the output signal level is the same (say 10 V) then 1.5 ppm = 15 uV whereas 0.0004% of 10 V is 40 uV.

Andy aka
  • 434,556
  • 28
  • 351
  • 777
  • Can you give some informations about how to exchange the two? Or a simple example is ok. – diverger Jan 21 '16 at 13:01
  • @diverger I've just added it! – Andy aka Jan 21 '16 at 13:01
  • You mean the ‘FSR' just the desired output voltage level when give it a dc signal? – diverger Jan 21 '16 at 13:04
  • If refers to full-scale resolution and if the amp is capable of producing 10Vp-p (this of course is power supply dependent) then 10 V is the figure to use. Between -5V and +5V (10 Vp-p) the expected non-linearity is 15 uV for one device and 40 uV for the other device. – Andy aka Jan 21 '16 at 13:19
  • Yes, even a sine wave are combined by a series voltage levels, so any point will have a gain error, and the worst one is 40uV, right? – diverger Jan 21 '16 at 13:22
  • Yes, the worst one is the PGA204 but, like you rightly pointed out, most devices specify PPM and PPM does't require FSR to be stated. If the signal level were much smaller than the full-scale you could argue that the ppm-specified device is far better BUT I expect that both would be pretty similar. Maybe look for distortion figures in the data sheets and make comparisons. – Andy aka Jan 21 '16 at 13:35
  • Yes, but the B version of pga204 has same worst nonlinearity with 280. Sure 280 has better typical value. What strange is 204 B version is more expensive than 280. – diverger Jan 21 '16 at 15:48
  • Prices are no measure of quality - sometimes prices are used to imply quality to unsuspecting buyers! – Andy aka Jan 21 '16 at 15:50