new interesting CMOS op-amps

[Dominique]

Spec-sheet-wise, these new CMOS op-amps come close to good JFET input op-amps, they seem worth to take a closer look at!

Did anyone try one of them?

I mean Linear's LTC6241, Maxim's Max4477/4489 and Analog's AD8656.

Cheers,
Dominique

[Dominique]

I just want to add that they're meant for low supply voltages, 6V to 12V (in the case of LTC6241HV) absolute maximum!


PS. I couldn't find the "edit" button to edit my post anymore!

[peranders]

They might be useful. AD8656 has many good properties on the paper. 2.7 nV/Hz, 0.0007% distortion, 50 MHz Bandwidth. The only but may be the rather high 1/f corner for the noise but it starts from a vey low level.

It's a huge difference between AD8656 and CA3140 (or was it 3160?)

The edit time is 30 minutes.

[nuvistor]

Also consider Nat Semi LMV751 (lowest noise, single only) and LMV771/2/4 (single/dual/quad). LMV751 noise is close to OPA627/637 and better than OPA132, AD8610. Output current, slew rate and bandwidth are less than modern jfet opas but by the numbers good enough for a low level preamp stage with 0.5Vrms out at 20kHz. These low-noise CMOS opas are probably intended for mic amps in portable devices and will probably be passed over by performance-obsessed DIYers, no one at head-fi or headwize mentions them.

[peranders]

Originally posted by nuvistor
...no one at head-fi or headwize mentions them.

... doesn't mean anything really.

[amplifierguru]

Hmm,

Could be interesting for a chip driven amp?

Hi Peranders,

bit of understatement -

....."The only but may be the rather high 1/f corner for the noise but it starts from a vey low level."

it's 4nv at 1KHz corner !!!!!

There goes the noise figure. Have to A weight it. OK!

Cheers,
Greg

[nuvistor]

Typical 10Hz, 1kHz en, nV/rtHz, and slew rate in V/us, GBW in MHz for the opas mentioned above (with 5V rail for CMOS):

OPA en10Hz en1kHz SR GBW

AD8656 30 4 11 28
LTC6241 15 8 10 18
MAX4475 21 5 3 10
MAX4488 21 5 10 42
(Avcl=5)
LMV751 17 6.5 2 4.5
OPA132 23 8 20 8
OPA627 15 5.2 55 16

My point is that en for most of these is competitive with the better modern JFET opas, even at low frequencies, with SR and GBW "good enough" for audio given the 5V p-p output available. The 6V to 7V max rail and the generally moderate max output current needs to be considered.

I would seriously consider the LMV751 or LTC6241 for 1st stage and MAX4488 for 2nd and 3rd stage of a battery-powered MM phono preamp with passive RIAA eq like Hagerman Bugle. En is close to OPA627/637, with Iq <1mA for LMV. <2mA for LTC and <2.5mA for MAX.

[amplifierguru]

Hi Nuvistor,

In 70's I was doing MM phono preamps using a diff'l pair of PN4250's driving a feedforward LM301 from memory. I did numerous 2SK389 + op amp and discrete ones before my MC/MM LT1028 which was DC coupled and gave the lowest noise.

I saved the OPA627/37's for line level and front of power amps where 5nV is nothing. Appreciate that the disc surface noise dominates all after eq.

Cheers,
Greg

[nuvistor]

Originally posted by amplifierguru
Appreciate that the disc surface noise dominates all after eq.

Cheers,
Greg

Yep, 5nV for MM is more than enough given disc noise and eq, currently I prefer a tube MM phono pre with higher noise. Given the eq, en below 500Hz is most important.

When using op amps at line level I prefer bipolar types like LM6171 with volume control incorporated into an inverting feedback network ala Baxendal, but the source impedance has to be 600 ohms or less, at higher source impedance requiring noninverting feedback I prefer opamps with cascoded JFET input like OPA627/637.

[jcx]

don't forget the ad743/745 for the lowest monolithic fet op amp noise;

3.2 nV/sqrt(Hz) @ 1KHz, 5.5 @ 10 Hz

2.8 V/uS unity gain (743) 4.5 MHz GBW

12 V/us decomp Av 5 (745) 20 MHz