TLV9302: 6th order active notch filter has only -20dB

Abderrezak HADJI

Part Number: TLV9302
Other Parts Discussed in Thread: TLV9304, INA149

Hello Dear TI,

We are building a 50Hz-6th order notch filter around TLV9302. We have used Filter Design Tool to setup the design parameters. I selected an 6th Bessel with following parameters:

Order: 6
Ao: 1.000 V/V
Fc: 50 Hz
BWp: 30 Hz
Rp: 0.01000 dB
BWs: 3 Hz
Asb: -40.00 dB

After getting resistors vales we build our board (see picture for routing).

First results of prototype are big surprise. We have only -22dB at 50Hz. We are expecting much more ~-35db. Can you help with this topic?

4 months ago

0 Caro Walter 4 months ago

TI__Mastermind 18557 points

Hello,

Thanks for posting on e2e! Could you please share the output report from the filter tool?
If this is not available, could I please see the schematic?

All the best,
Carolina

0 Abderrezak HADJI 4 months ago in reply to Caro Walter

Prodigy 40 points

Hello,

Please find attached the report.

One more important point, we want a minimal attenuation of -30dB at 47.5H and 52.5Hz

filter-design-report-73.pdf

0 Abderrezak HADJI 4 months ago in reply to Caro Walter

Prodigy 40 points

In analog applications journal slyt613 it is said "The Q of the notch is dependent on the gain of the amplifiers as opposed to component matching." I guess OL gain. How this is possible? Does TLV9304 fit our requirement?

0 Caro Walter 4 months ago in reply to Abderrezak HADJI

TI__Mastermind 18557 points

Hello,

Thank you for your response, I will run a few simulations and have my suggestion by end of day.

All the best,
Carolina

0 Caro Walter 4 months ago in reply to Caro Walter

TI__Mastermind 18557 points

Hey Hadji,

Can you tell me more about the supply voltages provided to the op amp and initial input source (voltage and frequency range)?

I ran simulations on the provided details and got as expected around -39dB.

Bessel_6_order_TLV9302.TSC

All the best,
Carolina

0 Abderrezak HADJI 4 months ago in reply to Caro Walter

Prodigy 40 points

Hey Carolina,

TLV are supplied with a well filtered analog supply of +/-10V. Input signal (coming out from INA149 for 50Hz+/-5% current measurement) is +/-5V. We have to filter the fundamental before data acq. I agree that simulation works fine. I got same results using the TLV model under LTSpice. I think that the main problem come from capacitors tolerances. At 1uF the best available tolerance is 10%. Can you scale for us the design? Or give me the procedure how the capacitor is selected in the tool? Our requirement is an attenuation at least -30dB between 50Hz+/-5% and not more than -0.1dB at 250Hz. Please help us with support this is an urgent topic for our client.

0 Caro Walter 4 months ago in reply to Abderrezak HADJI

TI__Mastermind 18557 points

Hello,

I recommend leveraging PSpice for TI to run a monte carlo sim while taking into account the component tolerances. I couldn't find a ti resource for the equations of the filter but here is a non-TI resource, https://www.changpuak.ch/electronics/Bainter_Notch_Filter.php (since not TI I cannot support further with this calculator).

All the best,
Carolina

0 Abderrezak HADJI 4 months ago in reply to Caro Walter

Prodigy 40 points

Hello,

Can you share how the algorithm is done in the tool to calculate parameters from given filter type.

0 Caro Walter 4 months ago in reply to Abderrezak HADJI

TI__Mastermind 18557 points

Hello,

Unfortunately, this is not available.

All the best,
Caro

0 Caro Walter 4 months ago in reply to Caro Walter

TI__Mastermind 18557 points

Hello,

Although the algorithm for the tool is unavailable, we do have simulation models available that may be leveraged to see the circuit response.

I created a PSpice for TI project and ran a Monte Carlo simulation.

This is with a sample of 15:

The tolerances may be adjusted in the parameters of the "Stage1" page:

I also ran a new filter design to include values that would be closer the 10% accuracy: filter-design-report-40.pdf

I see very skewed behavior depending on the passive values, I also couldn't meet the goal listed below even with 0 tolerance passives and an ideal op amp.

Abderrezak HADJI said:
not more than -0.1dB at 250Hz

I recommend adjusting the values in the filter designer tool and simulating the tolerance available with the project provided. Bainter.zip

Good luck in your design

All the best,
Carolina

Amplifiers

Amplifiers forum

TLV9302: 6th order active notch filter has only -20dB