LMH6624: LMH6624: LMH6624MAX NOPB amplifier noise reduction when used to measure the current with shunt.

Hello Anurag,

   Thank you for the PCB snippets. Please share the schematic as well. Are you still seeing the same issue with oscillations?

Thank you,
Sima 

Please see schematic. Yes. Same issue is there. 

ce.pdf

Hi Anurag,

what OPAmp do you use? The LMH6624 as in the title of your thread? Or the LMH6609 like in the schematic?

What role play R35 and R36? These resistors cause a heavily unwanted coupling of all OPAmps via the supply voltage lines. You should remove them. Or do you want to add RC filters in the supply voltage lines of OPAmps? Then spend a 10...22R resistor into the supply voltage line of each OPAmp separately.

Kai

LMH6624MAX NOPB --> in use.

R35, R36 are 0E currently.

Hi Anurag,

in order to operate stably the LMH6624 needs a minimum gain of 10V/V. But in your circuit the OPAmps are running at a gain of 6V/V...

Kai

Hi Kai,

I have updated in circuit as per communication from below and tried but not worked.

LMH6624: LMH6624MAX NOPB amplifier noise reduction when used to measure the current with shunt. - Amplifiers forum - Amplifiers - TI E2E support forums

Hello Anurag,

   The attached ce.pdf file above with the updated schematic shows a gain of 5V/V, but Kai's simulation on the previous thread has simulations for a circuit with a gain of 10V/V (RF = 1k, RG = 100 or RF = 500, RG = 50). 

Thank you,

Sima

Hi Sima,

RF = 500, RG = 50 I have tried as per trailing references. Not working. 

Is there any required modification in layout or in schematic?

Else I do try it again.

Hi Anurag,

there are several issues with the layout:

1. The blue +5V copper track on the bottom side of PCB unfortunately ruins the solid ground plane and ruins the supply voltage decoupling of OPAmps. This may result in a catastrophic failure. See how the HF decoupling currents have to flow arround the gap in the solid ground plane. The decoupling situation couldn't be worse:

Route the +5V copper track close to the -5V copper track or move it to the top of PCB. But keep it on the top plane so that the solid ground plane remains intact. Or use a 4-layer board.

2. The decoupling caps are too far away from the OPAmps and have too few vias at their ground terminals.

3. The ground terminals of R8, R9, etc, have too few vias.

4. You need way more ground vias connecting the ground fills on the top plane with the solid ground plane in the bottom plane.

5. The ground plane and the ground fills are not removed from the feedback components, the -input and output copper tracks of OPAmp.

6. The copper track between the +input and R25 is way too long.

7. There are way too many unused components in your layout. Remove C1, C2, C3, C4 and C5. Also remove R26, R28, R30, R32, R34 and R20, R27, R21, R29, R22, R31, R23, R33. They are completely unnecessary.

Proper example layouts have been discussed in these threads:

https://e2e.ti.com/support/amplifiers-group/amplifiers/f/amplifiers-forum/1194027/lmh6629-high-speed-transimpedance-amplifier-for-single-photon-detection

https://e2e.ti.com/support/amplifiers-group/amplifiers/f/amplifiers-forum/1033505/opa847-in-this-tia-circuit-can-i-use-series-resistors-dividing-voltages-to-couple--vbias-to-the-pin-s-anode-when-the-dc-voltage-source-is--5v-and-the-anode-voltage-needs-to-be-1v

With a bit luck, the blue +5V copper track is the major issue here and after fixing it the circuit may work 

Kai

Hi Anurag,

as a quick fix, cut the blue +5V track at every decoupling cap (C12, C14, C16, C19, C20, C22) and behind R35 and make the ground plane continuous and solid again by soldering copper foil over the whole gap. If you don't have copper foil then at least solder lots and lots of wires over the whole gap.

Then feed 1...4R7 resistors from R35 (R35 = 0 !) to each decoupling cap.

Since your overall gain is extremely high (10000V/V...100000V/V ?) you urgently need pi-filters in the supply voltage lines of every single OPAmp.

I would try to run the circuit with a way smaller gain first. At the beginning I would start with two LMH6629 in a row, each with a gain of 10V/V. This already may be a challange. Use pi-filters in the supply voltage lines and use a solid ground plane.

Kai