Analog Filter Wizard

40kHz BP.json.txt
{
	"Name": "40kHz BP",
	"Description": "Center 40kHz\nPass 60kHz\n4 order",
	"Tool": "FW",
	"Version": "1.1",
	"Design": {
		"visitedTabs": [
			"filter-type",
			"specifications",
			"components",
			"tolerances"
		],
		"filterType": "bandPass",
		"specifications": {
			"gain": 26,
			"gainUnit": "dB",
			"passbandAttenuation": -3,
			"passbandFrequency": 60000,
			"stopbandAttenuation": -40,
			"stopbandFrequency": 550000,
			"filterResponseValue": 0.15,
			"filterResponseSlider": "32",
			"centerFrequency": 40000
		},
		"components": {
			"vsPlus": 5,
			"vsMinus": -5,
			"optimization": "RecommendedSpecificComponents",
			"optimizationPreference": "specific",
			"compensateForGbw": true,
			"stages": [
				{
					"stageLetter": "A",
					"componentSizing": 56,
					"gain": 161,
					"gainEnabled": true,
					"implementation": "deliyannisFriend",
					"opAmps": "AD8039"
				},
				{
					"stageLetter": "B",
					"componentSizing": 56,
					"gain": 161,
					"gainEnabled": true,
					"implementation": "deliyannisFriend",
					"opAmps": "AD8039"
				}
			],
			"recommendedAmps": [],
			"sortOrder": [
				0,
				1
			]
		},
		"tolerances": {
			"resistorTolerance": "1%",
			"capacitorTolerance": "1%",
			"inductorTolerance": "5%",
			"opAmpGbwTolerance": "20%",
			"resistorPreferredSeries": "E24",
			"capacitorPreferredSeries": "E24",
			"inductorPreferredSeries": "E12"
		},
		"nextSteps": {}
	},
	"CreatedDate": "2020-05-07 09:40:55 AM",
	"UpdatedDate": "2020-05-07 09:40:55 AM"
}

The gain feedback resistor goes to the amplifier positive input. Where do I go wrong?

  • 0
    •  Super User 
    on May 7, 2020 12:16 PM 9 months ago

    Hello nielsudk,

    > The gain feedback resistor goes to the amplifier positive input. Where do I go wrong?

    The program realize the circuit with the "Deliyannis-Friend Filter". The positive feedback is used to increase Q. See page 14 in the PDF.

    https://hasler.ece.gatech.edu/Courses/ECE3400/Proj4/CMOSActiveFilters_Temes01.pdf

    Nevertheless the calculated magnitude response is far off from the target, because you have used resistors from the series E24.

    Theoretically calculated filter by the Analog Filter Wizard

    Active filter circuit created from Filter Wizard.

    You should use 1% resistors to get closer to the nominal ferquency response. Therefore select the E96 series for the resistors.

     

    The resistor values in the positive feedback should be kOhms and not Ohms.

    Corrected: See the follow-up answers.

    Helmut

    design.zip

  • 0
    •  Analog Employees 
    on May 7, 2020 12:55 PM 9 months ago in reply to helmuts

    Hi Helmut,

    Thanks for the excellent analysis.  I have nothing to add - you've covered it all.

    Regarding the small resistors - when using the "Pick for me" modes (Low Power, Voltage Range, and Low Noise) the tool should give bigger values than the resistor values in the saved file.  But in "I want to choose" mode, the tool gives a lot of leeway to adjust these resistors with the feedback resistor slider.  (Called gain resistor slider, because they set the gain in the low-pass and high-pass cases, but not really in our bandpass case here.)  , I wonder if while you were learning about these feedback resistors, you adjusted the feedback resistor slider to the left?  If not, do you remember how to reproduce the small resistor behavior, starting from loading the tool from scratch?

    Matt

  • 0
    •  Super User 
    on May 7, 2020 1:37 PM 9 months ago in reply to Matt.Duff

    Hello Matt,

    Thanks a lot for your help with the slider for "Gain Resistors". I had overlooked this feature.

    I just tried with another browser and the slider for "Gain Resistors" has been on the very left side - lowest resistor values.

    Best regards,
    Helmut

  • Thank you Helmut and Matt

    Positive feedback.... I should have known

    I remember the feedback resistor low values were there before touching the slider. I also remember the slider could change the values, but they were still in the low Ohm range and never in kOhm. Strange