how to improve the CMRR of opamp?

Question

1) Kindly provide any other opamp having higher CMRR, but with similar specification of ADA4895. 
 2) Kindly provide methods to improve CMRR of opamp apart from resistor matching.

Hooman · Answer

Hi rmurthy720 , 
 The guaranteed DC CMRR of 100dB for ADA4895 is hard to beat for most high frequency devices I searched for. However, I've shown a list of comparable ADI devices you could look through:

Part # 
 Feedback Type 
 CMRR (min) 
 Temp Range 
 # of Amps 
 GBP (typ) 
 Slew Rate (typ) 
 Vos (max) 
 VNoise Density (typ) 
 Ibias (max) 
 Iq/Amp (typ) 
 Vs span (min) 
 Vs span (max) 
 Price (1000+) 
 Associated Model 
 Package Area

(dB)

(Hz) 
 (V/us) 
 (V) 
 (V/rtHz) 
 (A) 
 (A) 
 (V) 
 (V) 
 ($ US) 
 
 (mm&sup2;)

LTC6229 
 Voltage 
 100 
 -40 to 125&deg;C 
 2 
 890M 
 500 
 95&micro; 
 880p 
 2.5&micro; 
 16m 
 2.8 
 11.75 
 $2.37 
 (LTC6229IDD#PBF) 
 9mm&sup2; (10-Lead DFN (3mm x 3mm w/ EP)) (LTC6229HDD#PBF)

LTC6228 
 Voltage 
 100 
 -40 to 125&deg;C 
 1 
 890M 
 500 
 95&micro; 
 880p 
 2.5&micro; 
 16m 
 2.8 
 11.75 
 $1.58 
 (LTC6228IDC#TRPBF) 
 4mm&sup2; (6-Lead DFN (2mm x 2mm w/ EP)) (LTC6228HDC#TRPBF)

LT1222 
 Voltage 
 100 
 -40 to 85&deg;C 
 1 
 500M 
 200 
 300&micro; 
 3n 
 300n 
 8m 
 5 
 36 
 $4.25 
 (LT1222CN8#PBF) 
 31.00979mm&sup2; (8-Lead SOIC (Narrow 0.15 Inch)) (LT1222CS8#PBF)

ADA4895-2 
 Voltage 
 100 
 -40 to 125&deg;C 
 2 
 1.5G 
 943 
 350&micro; 
 1n 
 16&micro; 
 3m 
 3 
 10 
 $3.21 
 (ADA4895-2ARMZ) 
 15.965mm&sup2; (10-Lead MSOP) (ADA4895-2ARMZ)

ADA4895-1 
 Voltage 
 100 
 -40 to 125&deg;C 
 1 
 1.5G 
 943 
 350&micro; 
 1n 
 16&micro; 
 3m 
 3 
 10 
 $1.89 
 (ADA4895-1ARJZ-R7) 
 9mm&sup2; (6-Lead SOT-23) (ADA4895-1ARJZ-R7)

AD829 
 Voltage 
 100 
 -55 to 125&deg;C 
 1 
 750M 
 230 
 1m 
 1.7n 
 7&micro; 
 5.3m 
 9 
 36 
 $3.65 
 (AD829ARZ) 
 31mm&sup2; (8-Lead SOIC) (AD829JRZ-REEL)

AD8099 
 Voltage 
 98 
 -40 to 125&deg;C 
 1 
 3.8G 
 470 
 500&micro; 
 950p 
 13&micro; 
 15m 
 5 
 12 
 $2.00 
 (AD8099ACPZ-REEL) 
 9mm&sup2; (8-Lead LFCSP (3mm x 3mm w/ EP)) (AD8099ACPZ-REEL7)

LT6230-10 
 Voltage 
 95 
 -40 to 85&deg;C 
 1 
 1.45G 
 250 
 500&micro; 
 1.1n 
 10&micro; 
 3.3m 
 3 
 12.6 
 $1.50 
 (LT6230CS6-10#TRPBF) 
 8.12mm&sup2; (6-Lead TSOT-23) (LT6230CS6-10#TRMPBF)

LT1226 
 Voltage 
 94 
 0 to 70&deg;C 
 1 
 1G 
 400 
 1m 
 2.6n 
 8&micro; 
 7m 
 5 
 36 
 $3.28 
 (LT1226CN8#PBF) 
 31.00979mm&sup2; (8-Lead SOIC (Narrow 0.15 Inch)) (LT1226CS8#PBF)

AD8021 
 Voltage 
 86 
 -40 to 85&deg;C 
 1 
 1G 
 130 
 1m 
 2.1n 
 11.3&micro; 
 7.8m 
 4.5 
 24 
 $1.31 
 (AD8021ARMZ) 
 31mm&sup2; (8-Lead SOIC) (AD8021ARZ-REEL7)

LTC6254 
 Voltage 
 85 
 -40 to 125&deg;C 
 4 
 720M 
 280 
 350&micro; 
 2.75n 
 3&micro; 
 3.3m 
 2.5 
 5.25 
 $3.55 
 (LTC6254CMS#PBF) 
 19.7911mm&sup2; (16-Lead MSOP) (LTC6254CMS#TRPBF)

To improve CMRR: If there is anything that can be done to reduce the CM voltage variation that the device experiences, that would of course reduce the error voltage. 
 Also, I've not tested this but if you were to introduce a diff pair in front of an Op Amp (in a composite amplifier configuration), and used precision / tracking / matched resistors) this would boost the differential gain and should reduce CMRR. However, something tells me that you give up many things (e.g. offset error, bandwidth, noise, etc.) when you do something like this as a penalty. So, I'm not sure how practical this approach may be, but thought I mention it anyway. 
 Regards, 
 Hooman