Author Topic: questions about tank circuits  (Read 858 times)

0 Members and 1 Guest are viewing this topic.

Offline greenelephantTopic starter

  • Newbie
  • Posts: 1
  • Country: gb
questions about tank circuits
« on: March 24, 2024, 10:12:57 pm »
In reference to the three jpeg files attached, how can 3 tank circuits be modified to allow a lower limit of 100KHz? can the first one with capacitor c20 be modified to allow 100khz by changing c20 to 2500 picofarads? I appreciate everyones feedback?
 

Offline jonpaul

  • Super Contributor
  • ***
  • Posts: 3366
  • Country: fr
Re: questions about tank circuits
« Reply #1 on: March 24, 2024, 10:58:52 pm »
Check Pi and Tee filter designs.

Use  filter scaling laws but  note effect on  Zo/Zin in the application

Jon
Jean-Paul  the Internet Dinosaur
 

Offline szoftveres

  • Regular Contributor
  • *
  • Posts: 115
  • Country: us
Re: questions about tank circuits
« Reply #2 on: March 24, 2024, 11:39:33 pm »
The first circuit is a BPF, the other two are LPF, not sure if your question makes much sense. What are your specs?
« Last Edit: March 24, 2024, 11:43:47 pm by szoftveres »
 

Offline vk4ffab

  • Regular Contributor
  • *
  • Posts: 230
  • Country: au
Re: questions about tank circuits
« Reply #3 on: March 25, 2024, 02:30:23 am »
In reference to the three jpeg files attached, how can 3 tank circuits be modified to allow a lower limit of 100KHz? can the first one with capacitor c20 be modified to allow 100khz by changing c20 to 2500 picofarads? I appreciate everyones feedback?

The quick and dirty method is to open up LT Spice, simulate the filters and play with values till it looks kind of right for your application. The second method is to learn a bunch of math and run the calculations to design the filter, or use a suitable calculator tool for the topology.

 

Offline RFDx

  • Regular Contributor
  • *
  • Posts: 149
  • Country: de
Re: questions about tank circuits
« Reply #4 on: March 25, 2024, 04:13:13 am »
can the first one with capacitor c20 be modified to allow 100khz by changing c20 to 2500 picofarads?

No, it's not enough to adjust only C20. Changing C18/C21 to 33nF, L4 to 820uH and C20 to 2.2nF retains the passband ripple and return loss and extends the lower corner frequency to less than 100kHz.
 

Offline T3sl4co1l

  • Super Contributor
  • ***
  • Posts: 21696
  • Country: us
  • Expert, Analog Electronics, PCB Layout, EMC
    • Seven Transistor Labs
Re: questions about tank circuits
« Reply #5 on: March 25, 2024, 05:17:50 am »
The characteristic frequency of a network is of the form w_0 = 1/sqrt(L*C), and characteristic impedance Zo = sqrt(L/C).  Exact element values vary above or below these, according to the filter function designed, and how the elements are interconnected; but the overall network will average (that is, geometric average) to these values.

Assuming the designs are correct, and appropriate for whatever impedance system they're embedded within, frequency involves rescaling everything proportionally.  As frequency goes down, inductance and capacitance go up.

Impedance matching is paramount.  None of the three listed, indicate what impedance they are designed for, so they aren't very useful.  You don't mention what impedance you're designing for either, so I'm afraid we cannot draw any meaningful conclusions.

The second one, for example, seems to be designed for around 280Ω.  But this depends on the desired Q, the peaking in the transition band, and could be meant for other values.

Tim
Seven Transistor Labs, LLC
Electronic design, from concept to prototype.
Bringing a project to life?  Send me a message!
 

Offline p.larner

  • Frequent Contributor
  • **
  • !
  • Posts: 690
  • Country: gb
Re: questions about tank circuits
« Reply #6 on: March 26, 2024, 08:19:23 am »
search for tonne filter software,comes with a lot of arrl antenna books.
 

Offline vk6zgo

  • Super Contributor
  • ***
  • Posts: 7589
  • Country: au
Re: questions about tank circuits
« Reply #7 on: March 26, 2024, 09:45:13 am »
The characteristic frequency of a network is of the form w_0 = 1/sqrt(L*C), and characteristic impedance Zo = sqrt(L/C).  Exact element values vary above or below these, according to the filter function designed, and how the elements are interconnected; but the overall network will average (that is, geometric average) to these values.

Assuming the designs are correct, and appropriate for whatever impedance system they're embedded within, frequency involves rescaling everything proportionally.  As frequency goes down, inductance and capacitance go up.

Impedance matching is paramount.  None of the three listed, indicate what impedance they are designed for, so they aren't very useful.  You don't mention what impedance you're designing for either, so I'm afraid we cannot draw any meaningful conclusions.

The second one, for example, seems to be designed for around 280Ω.  But this depends on the desired Q, the peaking in the transition band, and could be meant for other values.

Tim

If you are using a "lowish" impedance, high values of inductance start to look like wire wound resistors, that is, the resistive components of your inductors become high enough to turn your "LC" network into a LCR network, with a quite different Q & passband.

Many years ago, using filter charts & formula I built a notch filter for just above 500kHz at an impedance of 50 \$\Omega\$.
The charts & formula were happy, but the real world result was nothing like the prediction.

That was how I learnt why many filters were designed for higher impedances, with matching sections at the input & output.
 


Share me

Digg  Facebook  SlashDot  Delicious  Technorati  Twitter  Google  Yahoo
Smf