Problem measuring notch filter

I am new to this forum, so if this post blows up, please bear with me.

I am working with a 1 KHz passive notch filter. The design is based on the Hall Network as described in a paper by Kuhn (https://www.kennethkuhn.com/electronics/design_and_applications_of_the_hall_network.pdf) I am using the same component values as Kuhn, except I am using fixed resistors rather than potentiometers for fine tuning. I am getting some anomalous results using the FreqResponseChirp Test Plugin on the QA401 running V 1.920. I read the Notch Filters blog post, but did not find anything that pointed me toward a solution. See the attached screen shot. As a rough check, I collected some single point measurements using a function generator and DMM (HP33120A and HP34401A, respectively). The insertion loss at 50 Hz was -3.97 dBV rather than -10 to -11 dBV, and the shape of the notch was much more symmetrical, i.e., at 50 KHz the output was -1.16 dBV.

Any advice would be greatly appreciated.

Thanks,

ceulrich

Hi @ceulrich, how does the notch filter look if you use a series of discrete sines?

image

Hello Matt,

Thanks for the response.

See the attached.

Cheers,

ceulrich

Hi @ceulrich, since your sweep using a chirp and also using tones looks the same, I think it’s believable. If you still have low confidence, then you can do a sweep in loopback mode to convince yourself the QA402 response is flat.

There could also be some issues related to input and output impedance. The output impedance of the QA402 is 100 ohms, and the input impedance is 100k. What are the values you are using in your notch? Do you have a schematic?

Hello Matt,

The schematic is attached.

Yes, I did not think about the input impedance of the QA401. With this passive design, the 100K is a bit low to accurately characterize the shape of the left-hand side of the notch. However, at this stage I am only interested in defining the insertion loss for frequencies between 2K and10K, and it looks like the QA401 will be able to do this.

Thanks for the guidance.

Cheers,

ceulrich
Passive 1 KHz Notch Filter - Fixed Resistors 1

Maybe I’m thinking to simple, but why not do a simulation

Hi @FdW, Is your conclusion that the spice with the 100K load appears as you are seeing? It looks that way to me.

Hi Matt, I do not completely understand the question, but what I will say is.

I think that a simple simulation would have given a real impression of what to expect in the first place and would have removed to surprise of seeing something that could have been expected.

The other ting is, I do have a vested interest in the QA402 and other devices, so when the schematic was published I just simulated the notch-filter to do a reality/virtuosity check and then posted my result to share.

Any way, I would have thought that publishing a picture that matched reality (as was measured) with theory(as simulated) would ease any doubts that may exist on the work done and the equipment used.

For me it was just a good result that I was happy to show.

And I still do not understand your question.

In other words, my conclusion is that the measurement seems to indicate that reality matches prediction and that is a good thing I guess.

Hello FdW, Hello Matt,

FdW, Thanks for doing the simulation, it does clearly show where my “problem” lies. Unfortunately I do not have simulation software. When I got the funny looking notch, I just assumed I had a setting wrong somewhere, and didn’t consider a simple impedance mismatch. I agree with you, the QA401 – and now the QA402 – are great devices, particularly considering the prices, and the support. However, like so many things, it is only as good as the operator.

Cheers,

ceulrich

Thought you might find this listing of free simulation packages to be of interest…

Nice list

The one used for the simulation above is
LTspice - Wikipedia

User support at
Installing and using LTspice IV (now including LTXVII). From beginner to advanced. - diyAudio

Hello DanClark & FdW,

DanClark, thanks for the info.

I did manage to learn KiCad, so there is a chance I might also be able to handle a simulation program. I work mostly on a Mac, but I do keep a Windows 7 laptop around just for must have programs like REW and QA. Is there a simulation program you would recommend for a novice? My large screen desk machine runs OS 10.13.

Cheers,

ceulrich

I’d try LTSpice first since it supports OS X and also has a visual editor. Unless you’re doing digital circuits that’s probably your easiest option as it looks like you’re more OS X oriented (as am I)…

Seems like a good idea. as you see LTSpice in my choice also.
Here is the above notch filter, save all following as a ‘notch.asc’ file

Version 4
SHEET 1 880 680
WIRE -48 144 -64 144
WIRE -16 144 -48 144
WIRE 16 144 -16 144
WIRE 336 144 96 144
WIRE 432 144 416 144
WIRE 464 144 432 144
WIRE 480 144 464 144
WIRE -16 224 -16 144
WIRE 16 224 -16 224
WIRE 112 224 80 224
WIRE 176 224 112 224
WIRE 320 224 240 224
WIRE 352 224 320 224
WIRE 432 224 432 144
WIRE 432 224 416 224
WIRE 112 240 112 224
WIRE 320 240 320 224
WIRE 464 288 464 144
WIRE 480 288 464 288
WIRE 480 304 480 288
WIRE 480 304 464 304
WIRE -48 320 -48 144
WIRE 320 336 320 320
WIRE 320 336 208 336
WIRE 112 352 112 320
WIRE 208 352 208 336
WIRE 320 352 320 336
WIRE 464 352 464 304
WIRE -48 448 -48 400
WIRE 112 448 112 432
WIRE 112 448 -48 448
WIRE 208 448 208 432
WIRE 208 448 112 448
WIRE 320 448 320 432
WIRE 320 448 208 448
WIRE 464 448 464 432
WIRE 464 448 320 448
WIRE 208 464 208 448
FLAG 208 464 0
FLAG 480 144 out
FLAG -64 144 in
SYMBOL res 112 128 R90
WINDOW 0 0 56 VBottom 2
WINDOW 3 32 56 VTop 2
SYMATTR InstName R1
SYMATTR Value 233.7k
SYMBOL res 432 128 R90
WINDOW 0 0 56 VBottom 2
WINDOW 3 32 56 VTop 2
SYMATTR InstName R2
SYMATTR Value 4.99k
SYMBOL cap 80 208 R90
WINDOW 0 0 32 VBottom 2
WINDOW 3 32 32 VTop 2
SYMATTR InstName C1
SYMATTR Value 6.8n
SYMBOL cap 240 208 R90
WINDOW 0 0 32 VBottom 2
WINDOW 3 32 32 VTop 2
SYMATTR InstName C2
SYMATTR Value 6.8n
SYMBOL cap 416 208 R90
WINDOW 0 0 32 VBottom 2
WINDOW 3 32 32 VTop 2
SYMATTR InstName C3
SYMATTR Value 6.8n
SYMBOL res 96 224 R0
SYMATTR InstName R3
SYMATTR Value 27k
SYMBOL res 304 224 R0
SYMATTR InstName R4
SYMATTR Value 2.2k
SYMBOL res 96 336 R0
SYMATTR InstName R5
SYMATTR Value 6.4k
SYMBOL res 304 336 R0
SYMATTR InstName R6
SYMATTR Value 3.24k
SYMBOL res 192 336 R0
SYMATTR InstName R7
SYMATTR Value 238.7k
SYMBOL voltage -48 304 R0
WINDOW 123 24 118 Left 2
WINDOW 39 0 0 Left 0
SYMATTR Value2 AC 1
SYMATTR InstName V1
SYMATTR Value SINE(0 1 1k)
SYMBOL res 448 336 R0
SYMATTR InstName R8
SYMATTR Value 100k
TEXT -80 480 Left 2 !.ac dec 1k 1 1Meg