Author Topic: help me to improve the filter response  (Read 3199 times)

0 Members and 1 Guest are viewing this topic.

Offline xzswq21Topic starter

  • Frequent Contributor
  • **
  • Posts: 295
  • Country: 00
help me to improve the filter response
« on: December 06, 2018, 07:13:23 pm »
Hi guys
actually I have designed and built a 50MHz low pass filter and someone tested it with a VNA but I'm wondered why the VSWR in the low frequency is not good!
red line is the response:
Insertion Loss is very good and very similar to the software simulation:
 

now VSWR:
 

as you see the VSWR at the low frequency is not very good (red line). it should be like the blue line based on the software simulation.
what do you think about the filter?
I have used CS series from Coilcraft inductors (0603) and TDK COG type capacitors (0402)

my friend connected two 1 meter cable at the both connector of the filter. he told me I calibrated the setup.

Thanks
« Last Edit: December 06, 2018, 09:41:58 pm by xzswq21 »
❤ ❤
 

Online tggzzz

  • Super Contributor
  • ***
  • Posts: 19281
  • Country: gb
  • Numbers, not adjectives
    • Having fun doing more, with less
Re: help me to improve the filter response
« Reply #1 on: December 06, 2018, 07:25:27 pm »
There's a similar question, and a suggested solution at https://www.eevblog.com/forum/beginners/3db-and-an-lpf/msg1997762/#msg1997762
There are lies, damned lies, statistics - and ADC/DAC specs.
Glider pilot's aphorism: "there is no substitute for span". Retort: "There is a substitute: skill+imagination. But you can buy span".
Having fun doing more, with less
 

Offline xzswq21Topic starter

  • Frequent Contributor
  • **
  • Posts: 295
  • Country: 00
Re: help me to improve the filter response
« Reply #2 on: December 06, 2018, 07:29:12 pm »
There's a similar question, and a suggested solution at https://www.eevblog.com/forum/beginners/3db-and-an-lpf/msg1997762/#msg1997762
Thanks
but I don't have any problem with f3dB here. my problem is about VSWR at low frequencies from 10MHz to some MHz.
❤ ❤
 

Offline T3sl4co1l

  • Super Contributor
  • ***
  • Posts: 21609
  • Country: us
  • Expert, Analog Electronics, PCB Layout, EMC
    • Seven Transistor Labs
Re: help me to improve the filter response
« Reply #3 on: December 06, 2018, 09:23:34 pm »
Well, the VSWR would about account for the insertion loss, wouldn't it?  Just eyeballing it.

Does your simulation include component and layout parasitics?

This looks like a pretty high order filter, for which very precise values will be required, and parasitics will matter despite the relatively low center frequency.

Tim
Seven Transistor Labs, LLC
Electronic design, from concept to prototype.
Bringing a project to life?  Send me a message!
 
The following users thanked this post: xzswq21

Offline xzswq21Topic starter

  • Frequent Contributor
  • **
  • Posts: 295
  • Country: 00
Re: help me to improve the filter response
« Reply #4 on: December 06, 2018, 09:50:03 pm »
Well, the VSWR would about account for the insertion loss, wouldn't it?  Just eyeballing it.

Does your simulation include component and layout parasitics?

This looks like a pretty high order filter, for which very precise values will be required, and parasitics will matter despite the relatively low center frequency.

Tim

yes, Actually I have calculated mismatch loss from the below link:
https://www.microwaves101.com/calculators/872-vswr-calculator
the VSWR @10MHz is 1.7 and so ML=0.3dB
could you please help me a little?
I have simulated the circuit by ADS. but I have used two connectors form "Linx Technologies " company and a Chinese company.
I have used ITEQ IT-180A material for the PCB.
can I send you the layout?

Thanks
« Last Edit: December 06, 2018, 09:53:38 pm by xzswq21 »
❤ ❤
 

Online tggzzz

  • Super Contributor
  • ***
  • Posts: 19281
  • Country: gb
  • Numbers, not adjectives
    • Having fun doing more, with less
Re: help me to improve the filter response
« Reply #5 on: December 07, 2018, 12:04:44 am »
There's a similar question, and a suggested solution at https://www.eevblog.com/forum/beginners/3db-and-an-lpf/msg1997762/#msg1997762
Thanks
but I don't have any problem with f3dB here. my problem is about VSWR at low frequencies from 10MHz to some MHz.

The 3dB point is irrelevant.

The similarity is much more fundamental: component tolerance, and monte-carlo analysis in Spice. Plus, depending on component values, parasitics.
« Last Edit: December 07, 2018, 12:06:39 am by tggzzz »
There are lies, damned lies, statistics - and ADC/DAC specs.
Glider pilot's aphorism: "there is no substitute for span". Retort: "There is a substitute: skill+imagination. But you can buy span".
Having fun doing more, with less
 

Offline xzswq21Topic starter

  • Frequent Contributor
  • **
  • Posts: 295
  • Country: 00
Re: help me to improve the filter response
« Reply #6 on: December 07, 2018, 07:15:20 am »
The 3dB point is irrelevant.

The similarity is much more fundamental: component tolerance, and monte-carlo analysis in Spice. Plus, depending on component values, parasitics.


I think the component tolerance doesn't have a significant roll for VSWR at the low frequencies.
I think DCR of the Inductors or ESR of the capacitors both have more significant roll in my problem.
as you know:
Z=sqrt{(R+sL)/(G+sC)}
at low frequencies: Z=sqrt{R/G}
at high frequencies: Z=sqrt{L/C}
❤ ❤
 

Offline xzswq21Topic starter

  • Frequent Contributor
  • **
  • Posts: 295
  • Country: 00
Re: help me to improve the filter response
« Reply #7 on: December 07, 2018, 08:29:42 am »
this is the layout:

if you see 8mils clearance between the lines and the ground guards bcoz I have calculated the co-planer lines with a tool.
Do you see any critical issues?
is it necessary to remove ground layer around the lines?
is there any problem with the direction of C4 and C7 capacitors? I could have placed C4 and C7 like C5 and C6 capacitors.
plus you see a square instead of a line between inductors and capacitors. is it necessary to modify it?
« Last Edit: December 07, 2018, 08:36:57 am by xzswq21 »
❤ ❤
 

Online tggzzz

  • Super Contributor
  • ***
  • Posts: 19281
  • Country: gb
  • Numbers, not adjectives
    • Having fun doing more, with less
Re: help me to improve the filter response
« Reply #8 on: December 07, 2018, 09:16:55 am »
The 3dB point is irrelevant.

The similarity is much more fundamental: component tolerance, and monte-carlo analysis in Spice. Plus, depending on component values, parasitics.


I think the component tolerance doesn't have a significant roll for VSWR at the low frequencies.
I think DCR of the Inductors or ESR of the capacitors both have more significant roll in my problem.
as you know:
Z=sqrt{(R+sL)/(G+sC)}
at low frequencies: Z=sqrt{R/G}
at high frequencies: Z=sqrt{L/C}

Your intuition may be right, but you could have mentioned it in your first post.

For such a small circuit you could easily do a spice analysis including the parasitics you mention - which would turn "think" into "know". And if you find that's not the case, then try a monte-carlo analysis
There are lies, damned lies, statistics - and ADC/DAC specs.
Glider pilot's aphorism: "there is no substitute for span". Retort: "There is a substitute: skill+imagination. But you can buy span".
Having fun doing more, with less
 
The following users thanked this post: xzswq21

Offline cdev

  • Super Contributor
  • ***
  • !
  • Posts: 7350
  • Country: 00
Re: help me to improve the filter response
« Reply #9 on: December 09, 2018, 11:06:40 pm »
I would never put inductors in a line like that unless I want them to influence one another. If you can stagger their placement so the 3 inductors are in a zig zag, at right angles to one another, the real world performance might be closer to the ideal performance.

Sometimes it doesn't matter, sometimes it does. But its what I do now.
« Last Edit: December 10, 2018, 02:55:12 am by cdev »
"What the large print giveth, the small print taketh away."
 
The following users thanked this post: xzswq21

Offline xzswq21Topic starter

  • Frequent Contributor
  • **
  • Posts: 295
  • Country: 00
Re: help me to improve the filter response
« Reply #10 on: December 10, 2018, 06:46:26 am »
I would never put inductors in a line like that unless I want them to influence one another. If you can stagger their placement so the 3 inductors are in a zig zag, at right angles to one another, the real world performance might be closer to the ideal performance.

Sometimes it doesn't matter, sometimes it does. But its what I do now.

could you please draw your idea on a paper? I got your idea but not completely.
is it necessary to put a keep-out line around the SMA's grounds at top layer? (the grounds will connect to each other at the bottom side)
« Last Edit: December 10, 2018, 07:07:31 am by xzswq21 »
❤ ❤
 

Offline T3sl4co1l

  • Super Contributor
  • ***
  • Posts: 21609
  • Country: us
  • Expert, Analog Electronics, PCB Layout, EMC
    • Seven Transistor Labs
Re: help me to improve the filter response
« Reply #11 on: December 10, 2018, 03:17:38 pm »
They're plenty far enough apart not to do weird things, assuming probably the worst case air core wirewound inductor in that package.

Tim
Seven Transistor Labs, LLC
Electronic design, from concept to prototype.
Bringing a project to life?  Send me a message!
 
The following users thanked this post: xzswq21

Offline cdev

  • Super Contributor
  • ***
  • !
  • Posts: 7350
  • Country: 00
Re: help me to improve the filter response
« Reply #12 on: December 10, 2018, 04:10:56 pm »
I am going to defer to T3sl4co1l's much better informed position, under the caveat that you're likely using small SMD-mounted inductors.

Inter-inductor interactions would likely not be a problem unless they were closer, he says, and if the scale of your circuit is the one pictured, he's likely right.

For an HF circuit that uses larger hand wound inductors like I often make, try to offset them by 90 degrees.

The size of the wavelength relative to the size of the inductors matters there too.
"What the large print giveth, the small print taketh away."
 
The following users thanked this post: xzswq21

Offline eb4fbz

  • Regular Contributor
  • *
  • Posts: 178
  • Country: es
Re: help me to improve the filter response
« Reply #13 on: December 10, 2018, 04:44:06 pm »
Can you post component values?
« Last Edit: December 10, 2018, 04:56:14 pm by eb4fbz »
 

Offline dmills

  • Super Contributor
  • ***
  • Posts: 2093
  • Country: gb
Re: help me to improve the filter response
« Reply #14 on: December 10, 2018, 05:55:33 pm »
What does the VNA say about what the impedance is at low frequency?

SWR is all very well, but where on the SWR circle are you, knowing what it is reporting for R_jX at low frequency would be useful? Stick it in smith chart more and look.

My immediate thought is that you probably have a reasonable amount of stray cap from the internal nodes to ground just due to the copper areas on the PCB, but without some better modelling it is hard to tell.

Regards, Dan.
 
The following users thanked this post: xzswq21

Offline xzswq21Topic starter

  • Frequent Contributor
  • **
  • Posts: 295
  • Country: 00
Re: help me to improve the filter response
« Reply #15 on: December 10, 2018, 06:29:18 pm »
SWR is all very well, but where on the SWR circle are you, knowing what it is reporting for R_jX at low frequency would be useful? Stick it in smith chart more and look.
My immediate thought is that you probably have a reasonable amount of stray cap from the internal nodes to ground just due to the copper areas on the PCB, but without some better modelling it is hard to tell.
Actually my friend only sent me return loss and insertion loss. so I should tell him to send me more details in the next test.
❤ ❤
 

Offline xzswq21Topic starter

  • Frequent Contributor
  • **
  • Posts: 295
  • Country: 00
Re: help me to improve the filter response
« Reply #16 on: December 10, 2018, 06:36:29 pm »
For an HF circuit that uses larger hand wound inductors like I often make, try to offset them by 90 degrees.
The size of the wavelength relative to the size of the inductors matters there too.
if you are interested to turn the inductors or something similar:
as you see the Layout B has the best response to reject the IMAGEs near Fn
but Layout A has the best response for very high frequencies
and layout C...

« Last Edit: December 10, 2018, 06:39:06 pm by xzswq21 »
❤ ❤
 

Offline cdev

  • Super Contributor
  • ***
  • !
  • Posts: 7350
  • Country: 00
Re: help me to improve the filter response
« Reply #17 on: December 10, 2018, 07:38:22 pm »
Build it on some piece of scrap PCB - its fairly easy to try out all the different possibilities quickly at lower frequencies.
"What the large print giveth, the small print taketh away."
 
The following users thanked this post: xzswq21

Offline markus_jlrb

  • Regular Contributor
  • *
  • Posts: 140
  • Country: de
Re: help me to improve the filter response
« Reply #18 on: December 12, 2018, 03:12:25 pm »
@xzswq21,

I would like to recommend to replace the inductors with
0 ohm resistors in order to verify and exclude the influence
of your pcb layout and pcb material and connectors.

If you get a proper SWR without any L's & C's in the
corresponding frequency range of interest you could
look as already mentioned for the influence of the
other filter parts.

Markus
 
The following users thanked this post: xzswq21

Offline cncjerry

  • Supporter
  • ****
  • Posts: 1270
Re: help me to improve the filter response
« Reply #19 on: December 14, 2018, 09:57:05 pm »
I am far from the expert on this subject, but when I measure SWR I use a calibrated 50ohm termination on one filter port and hook it to a calibrated reflection bridge.

Jerry
 

Online ConKbot

  • Super Contributor
  • ***
  • Posts: 1380
Re: help me to improve the filter response
« Reply #20 on: December 27, 2018, 07:26:38 pm »
How was the setup calibrated? Mechanical SMA standards? Or the quick-cal feature some of the field-fox units have?
If the latter, I find them a bit touchy over what load "standard" you use during the quick cal. For example, I've done work with MMCX connectors, and a mediocre MMCX load (1.3:1) gave worse results then adapting back to SMA and using a cal standard or high quality termination. Minicircuits SMA terminators ( https://www.minicircuits.com/WebStore/dashboard.html?model=ANNE-50L%2B )  have great return loss in the sub-GHz range, and I like to use them for field-fox quick-cals when breaking out the standards for a proper OSL isn't needed.
 
The following users thanked this post: xzswq21


Share me

Digg  Facebook  SlashDot  Delicious  Technorati  Twitter  Google  Yahoo
Smf