high precision bridge Measurement 25uV max

Hi Bob,

thank you for your good suggestion, I read datasheet of external PGA that ADS1259 suggest to use as external PGA ,in the datasheet called it has 4.5 μVPP noise and also its offset is 3uV if G=128, and this parameter will done with very carefully PCB design, with this background that discrete circuit has many problem, so I prefer to use an ADC with internal PGA but with this features that my project need I couldn't find any suitable ADC with internal PGA.

my first step in this project is finding good way for acquisition data:

a: use high precision ADC+ internal PGA (if it is possible this is best way)

b: find another way, like using external PGA, that is has many problem like noise,

which way do you suggest? do you have better solution? 

any help would be greatly appreciated,

Sincerely yours,

Siyavash.

Hi Siyavash,

I would not use the PGA280 for the gain stage with the ADS1259 for the situation you are using.  The PGA280 is meant for high voltage inputs.

As I mentioned earlier, it would be better to ask the precision linear forum for the best amplifer for your configuration.  Initial offset voltage should not be a big concern, however drift performance should be a consideration.  Noise is the biggest issue. Perhaps an INA333 would work.  I know there are some newer quiet op amps as well, I just don't recall the numbers off the top of my head.

Best regards,

Bob B

Hi Bob, 

I ask about which PGA, OPAMP or INAMP is suitable for my work in  precision linear forum but no body answered to my question!

I read ADS1240 - ADS1241 and in "page 21- FIGURE 9. Block Diagram for a High-Precision Weigh Scale" that is used two OP-AMP OPA2335. so I understand that for high precision application I need to use an external differential amplifier, but before using OPA2335, it used EMI-filter and I don't have good technical documents about EMI-filter, I know TI have special IC for EMI filter like tpd3f303 in SIM CARD INTERFACE. has TI EMI filter IC for this application? what about technical documents? in my typical work I use single Ferrite Bead as EMI filter. 

And my other question is in datasheet of ADS1241 called even in the best condition (PGA=1) ENOB is =21 and it isn't 24 bits, so what is 24 bits resolution? 

thank you for spending time to solve my problems,

Best regards,

Siyavash.

Hi Siyavash,

The resolution of an ADC refers to the smallest analog incremental change that can be measured (LSB size). What is confusing is even though the converter has the ability to measure these small changes, the noise of the conversion process reduces the effectiveness so that the result may be less than the optimal resolution ( which is ENOB compared to resolution).

As I said in my first response, noise is going to limit your results.  If you have significant noise outside of the conversion noise, the effective resolution will decrease further.  If you use an external amplifier, I would use the ADS1259 as I doubt the ADS1241 will be good enough.  Even then, you may not be able to achieve your goals.  You are trying to read very, very small voltage changes.  This requires very low noise components, ADC and amplifier.  It also requires low noise signal prior to amplification.  You can use ferrites for EMI filtering, but this may not be enough.  I'm having difficulty suggesting what components are necessary as I have not built anything to level of what you are attempting to do.  It will take a lot of effort.

Best regards,

Bob B

Hi Bob,

after many search I found ADS1282 31-bit ultra high precision ADC + internal PGA,  if its ENOB be 27 bit my work will be done, but I don't see any refer to this parameter in ADS1282 datasheet. what about ADS1282EVM? does the ADS1282EVM have ENOB about 27-bits?

I think with this consideration I can't do this work with 2 layer PCB, for having good shielding & etc, I should use multilayer PCB, is it true?

Best wishes,

Siyavash.

Hi Siyavash,

The ADS1282 is an excellent device, but it is primarily designed for AC input signals so the datasheet does not show DC parameters.  With increased data rate, the level of noise near DC is quite high due to quantization as compared to devices such as the ADS1232 and ADS1259 where this noise is filtered at the lower data rates.  To get lower noise for the ADS1282 you will need to post process the data.  The simplest way is by averaging. 

With the ADS1282EVM I collected some data in the past, and then averaged.  I used a running average of 20 samples in the average using a data rate of 250 sps which brought the throughput to about 12.5 sps. I was able to achieve 22.54 bits ENOB with shorted inputs.  The noise free resolution was 19.82 bits.  This may be improved with other post processing methods.  With the ADS1259 at 10sps using sinc2 filter the ENOB is 23.4 and noise free bits at 21.6.  So increased resolution does not guarantee increased ENOB.

The total number of bits necessary is based on the reference size relative to the input signal.  If you increase the input signal with external gain, then you may not need to resolve to as low a value as you might think.  You just need to have enough signal to noise margin at the low input levels to overcome the noise of the conversion process.  Also keep in mind that the shorted input case is the best the converter can do, and when a signal is applied the conversion noise may increase.  As the input signal increases, the noise increases and becomes the worst as the input signal approaches the value of the reference.

Your last comments are very important.  PCB layout must have adequate ground and return signal paths.  This usually requires a multi-layer board.  But it also requires good layout practices.  Analog and digital signals must never cross paths.  EMI/RFI noise pickup must be minimized (no antenna like traces) and trace size and spacing must be considered to reduce crosstalk noise pickup.

Best regards,

Bob B