getting gain on ADC signals

[powerfly]

what about these:

ebay.co.uk/itm/INA333-Low-Power-Zero-Drift-Precision-Instrumentation-Amplifier-Module-NEW/253056645871?_trkparms=aid%3D111001%26algo%3DREC.SEED%26ao%3D1%26asc%3D20160908105057%26meid%3D954fda06e6084541992c947f4bea3988%26pid%3D100675%26rk%3D1%26rkt%3D15%26sd%3D253056645871%26itm%3D253056645871%26pmt%3D0%26noa%3D1%26pg%3D2380057&_trksid=p2380057.c100675.m4236&_trkparms=pageci%3Aa187e060-1bf7-11ea-9b39-74dbd180afaf%7Cparentrq%3Af443886216e0ad3c58d19fdaffbaf376%7Ciid%3A1

the thing is again, the bandwidth looks like a limiting factor:

http://www.ti.com/lit/ds/symlink/ina333.pdf (if you just find the bandwidth it claims a max of 100x amplification at 3.5kHz I believe) but I might just go with it and see if it's good enough as I haven't found anything else even though it doesn't seem ideal.

[Kleinstein]

The INA333 is relatively slow and also quite noisy, though as an AZ type essentially without 1/f noise. This can be a problem for a fast application as the noise BW is relatively high for a strain gauge use.  If only a quarter bridge is used the drift will be large and thus no AZ type really needed and DC performance limited anyway.

 With a typical maximum of 1 mV/V and maybe 5 V at the bridge, the full scale range of the sensor would be some 5 mV. Resolving 10 Bits would mean measuring to some 5 µV. 5 V at the bridge would already be high for only 120 Ohms.

For 10 kSPS samplig the BW would likely be at some 3 kHz (up to 5 kHz). So the needed noise level for 10 Bit performance would be at around 
5 µV / Sqrt(3 kHz) = 90 nV/sqrt(Hz).  So at best the INA with a 50 nV/sqrt(Hz) noise level would give an 11 bit noise-level.

[powerfly]

The INA333 is relatively slow and also quite noisy, though as an AZ type essentially without 1/f noise. This can be a problem for a fast application as the noise BW is relatively high for a strain gauge use.  If only a quarter bridge is used the drift will be large and thus no AZ type really needed and DC performance limited anyway.

 With a typical maximum of 1 mV/V and maybe 5 V at the bridge, the full scale range of the sensor would be some 5 mV. Resolving 10 Bits would mean measuring to some 5 µV. 5 V at the bridge would already be high for only 120 Ohms.

For 10 kSPS samplig the BW would likely be at some 3 kHz (up to 5 kHz). So the needed noise level for 10 Bit performance would be at around 
5 µV / Sqrt(3 kHz) = 90 nV/sqrt(Hz).  So at best the INA with a 50 nV/sqrt(Hz) noise level would give an 11 bit noise-level.

Do you recommend something other than the INA 333? I saw that Marco recommended the INA 163 or 217 as alternative options

[SiliconWizard]

The INA333 and INA163 are in a completely different league. The latter is indeed much faster and lower noise, but it has a typical Iq of 10mA vs. the 50µA of the INA333, and a much higher input bias current.
So it really all depends on your exact requirements.

[Kleinstein]

I would think about AD8421 or AD8221 as more intermediate versions.
An INA would be needed in combination with a single ended ADC, like the filter + MUX + SAR ADC option. If separate ADCs with differential input (like most SD type ADCs) are used, one may be better of with a special fully differential amplifier or 2 OPs and 3 resistors to build just the first half of the INA.