opamp as virtural ground

Hello.

I'm building a virtual ground for a customer IC using an opamp (LT1802). The circuit that it interfaces to is an NMOS, whose gate samples at 10MHz and outputs a current between 100uA and 200uA. The virtual ground connects to the drain of the NMOS. I'm also using a feedback resistor of 10KOhm to act as a I-to-V converter.

My problem is.. I need a capacitor of about 10pF in parallel with the feedback resistor in order to be stable. However, the settling of this configuration is too slow. The voltage output cannot track the samples at 10MHz. From the LT1802 documentation, I see that a pole is formed by the capacitance at the noninverting input as well as the feedback resistor. It is this pole that determines the stability. However, I have little control of this pole except to use the parallel capacitor which lowers the bandwidth.

My question is.. how can I make it stable and fast at the same time?

Thanks for your time.

Scott

Reply to
Scott
Loading thread data ...

It is not very clear what are you trying to do. It would be helpful if you provide a link to schematics.

The GBW of LT1802 is only 80 MHz.

How about the precompensation of the signal at the input?

Vladimir Vassilevsky DSP and Mixed Signal Design Consultant

formatting link

Reply to
Vladimir Vassilevsky

As Vladimir suggested, post a schematic. I/V conversion via opamp becomes unstable when there is too much capacitance from IN- to GND. One method would be a cascode up front.

The LT1802 doesn't have a lot of slew rate, might not be the right fit here. One thing you can do is lower the 10K to a few kohm until you get enough bandwdith plus some margin. Then add a regular 10-20dB amplifier to scale up the voltage.

--
Regards, Joerg

http://www.analogconsultants.com/
Reply to
Joerg

Thanks Vladimir and Joerg. The schematic is a basic opamp I/V converter:

formatting link
where R=3D10K and IN- connects to the D of an NMOS, whose S is GND and G is a 10MHz-sampled signal.

I will first try a lower R and use another opamp in the LT1802 to amplify the voltage. Any suggestion on a drop-in replacement of LT1802 with better GBW and slew rate?

Best regards,

Scott

Reply to
Scott

Be careful there little boy, your virtual ground effect can change your output.

...Jim Thompson

--
|  James E.Thompson, P.E.                        |    mens |
|  Analog Innovations, Inc.                         |     et      |
 Click to see the full signature
Reply to
Jim Thompson

I suppose you're talking about this: View in a fixed-width font such as Courier.

. . . R . .----/\\/\\---. . | | . | |\\ | . Iin>------+-----|-\\ | . | | >--+-->

. |-' .-|+/ . 10MHz>----| | |/ . _ _ |> | . .. _| |_| |_.. | | . '-+-' . | . com .

You have added quite bit of stray with the NMOS Coss and then there is a fundamental time constant associated with its channel turn-off. As JT pointed out, you will have a residual output due to opamp offset of (1+R/Rds)xVos, where Rds is channel resistance, and this effect can be quite large. You might consider chopping the output...

Reply to
Fred Bloggs

Good idea.

How much of GBW and slew rate do you need exactly?

Vladimir Vassilevsky DSP and Mixed Signal Design Consultant

formatting link

Reply to
Vladimir Vassilevsky

Maybe you can find a FET with less capacitance.

Depends on your supply voltage. In the 5V range there are lots, for

+/-12V the selection thins out. You'd have to find one that is unity-gain stable. Either via mfg web sites or via Digikey. The latter is what I usually do. It's more work but then you can restrict the listing to what is really in stock so you know you can actually buy them right away.
--
Regards, Joerg

http://www.analogconsultants.com/
Reply to
Joerg

ElectronDepot website is not affiliated with any of the manufacturers or service providers discussed here. All logos and trade names are the property of their respective owners.