Sonsivri
 
*
Welcome, Guest. Please login or register.
Did you miss your activation email?
May 22, 2019, 11:55:44 23:55


Login with username, password and session length


Pages: [1]
Print
Author Topic: Low pass filter with no phase shifting  (Read 1166 times)
0 Members and 1 Guest are viewing this topic.
an007_rld
Junior Member
**
Offline Offline

Posts: 36

Thank You
-Given: 30
-Receive: 23


« on: March 16, 2019, 08:32:20 20:32 »

Hi All,

I need an analog 1'st order low pass filter with frequency corner at 2KHz and with 0deg phase shifting for frequencies up to 120Hz. Any idea of implementation?

Thanks,
-an
Logged
Sideshow Bob
Cracking Team
Hero Member
****
Offline Offline

Posts: 622

Thank You
-Given: 204
-Receive: 642



« Reply #1 on: March 17, 2019, 03:28:08 03:28 »

Hi All,

I need an analog 1'st order low pass filter with frequency corner at 2KHz and with 0deg phase shifting for frequencies up to 120Hz. Any idea of implementation?

Thanks,
-an
Do you want this to be done in an analog or digital setting. Also what are you trying to achieve?
Logged

I have come here to chew bubblegum and kick ass... and I'm all out of bubblegum
optikon
Cracking Team
Hero Member
****
Offline Offline

Posts: 729

Thank You
-Given: 771
-Receive: 1886


« Reply #2 on: March 17, 2019, 07:31:27 07:31 »

Hi All,

I need an analog 1'st order low pass filter with frequency corner at 2KHz and with 0deg phase shifting for frequencies up to 120Hz. Any idea of implementation?

Thanks,
-an

As stated, your requirements cant be met (analog filter)

You can have a very high order LPF design at 2kHz and have "small" (not zero!) phase at 120Hz

*OR*

You can have a single order filter but you will have push the cutoff frequency much higher frequency (> 2kHz)  to achieve very small phase at 120 Hz

You can eliminate the concept of a zero phase shift right now, its non causal and not physically realizable. You can however, design to an arbitrarily small phase.

Once you mention one of these tradoffs, a good suggestion can be made.


« Last Edit: March 17, 2019, 07:33:37 07:33 by optikon » Logged

I can explain this to you. I can't comprehend it for you.
an007_rld
Junior Member
**
Offline Offline

Posts: 36

Thank You
-Given: 30
-Receive: 23


« Reply #3 on: March 17, 2019, 09:56:17 09:56 »

Thanks Optikon for your answer,

The LPF is used to attenuate the clock of an high order switching capacitor filter (generate variable frequency, crystal controlled sine wave from square wave); I can compensate for phase in digital domain but the application is multichannel, so it's very difficult to replicate the same analog filter on each channel due to the components tolerance. Also, a multichannel DDS gets too expensive.

Anyway, I'll try to push higher as much as I can the rolloff frequency...

Thanks and most appreciate your suggestions.
-an
Logged
Signal
Active Member
***
Offline Offline

Posts: 182

Thank You
-Given: 99
-Receive: 76



« Reply #4 on: March 17, 2019, 12:20:07 12:20 »

In case you need a set of identical RC filters, adding R-trimmer is not very difficult though requires some extra work.
Still do not get it clear. Do you suppose to use multichannel DDS or not? If DDS single when thing becomes multichannel?
Logged

Give a right name to a right game and play it right
optikon
Cracking Team
Hero Member
****
Offline Offline

Posts: 729

Thank You
-Given: 771
-Receive: 1886


« Reply #5 on: March 17, 2019, 06:11:20 18:11 »

Thanks Optikon for your answer,

The LPF is used to attenuate the clock of an high order switching capacitor filter (generate variable frequency, crystal controlled sine wave from square wave); I can compensate for phase in digital domain but the application is multichannel, so it's very difficult to replicate the same analog filter on each channel due to the components tolerance. Also, a multichannel DDS gets too expensive.

Anyway, I'll try to push higher as much as I can the rolloff frequency...

Thanks and most appreciate your suggestions.
-an


If you can use LTspice or other spice sim, you can try an ideal square wave clock fed to a single pole LPF.

You can then try push pole location according to

1) High as possible until you do not get enough clock attenuation.
2) Low as possible until you see too much phase at 120Hz

If you can do more than 1 order, it helps. butterworth or other types can give very strong attenuation roll offs

Logged

I can explain this to you. I can't comprehend it for you.
an007_rld
Junior Member
**
Offline Offline

Posts: 36

Thank You
-Given: 30
-Receive: 23


« Reply #6 on: March 17, 2019, 07:58:14 19:58 »

I will not use DDS because the project requires 12 channels to be synchronized  and 3pcs 4 channel DDS makes the project very expensive.
The idea is to filter 12 synchronize square waves (8 order switched caps + 1pc LPF to attenuate the clock - clock ratio to rolloff is 100) and transform them in sine wave.
Logged
Sideshow Bob
Cracking Team
Hero Member
****
Offline Offline

Posts: 622

Thank You
-Given: 204
-Receive: 642



« Reply #7 on: March 18, 2019, 02:57:23 02:57 »

It is a filter topology named allpass or all-pass. If all your frequencies are fixed, once and for all MAYBE and a very big emphasize on maybe. You can use allpass filters to tune the signals into equal phase. However this is not something I even remotly have done my self. What I would have done. Is to use the most out of phase signal as a reference, and then tried to tune the rest of the 11 signals into this phase with allpass filters, although this would probably require multiturn pots. And will also require a lot of tuning work. Once again this is just an idea from the top of my head
Logged

I have come here to chew bubblegum and kick ass... and I'm all out of bubblegum
SBW
Newbie
*
Offline Offline

Posts: 7

Thank You
-Given: 0
-Receive: 0


« Reply #8 on: April 05, 2019, 04:54:17 16:54 »

You do not have to use expensive DDS chips.  You could implement your own multi-channel DDS in a small FPGA. 

For minimal hardware, you can implement your own DACs using delta-sigma modulation.  With 32x oversampling and a fourth order filter you can get the quantization noise down to 16-bit equivalent.  (Getting true 16-bit performance requires additional tricks, like running two outputs as out-of-phase return-to-zero and summing them to minimize the data-dependent non-linearity.)

Logged
Pages: [1]
Print
Jump to:  


DISCLAIMER
WE DONT HOST ANY ILLEGAL FILES ON THE SERVER
USE CONTACT US TO REPORT ILLEGAL FILES
ADMINISTRATORS CANNOT BE HELD RESPONSIBLE FOR USERS POSTS AND LINKS

... Copyright 2003-2999 Sonsivri.to ...
Powered by SMF 1.1.18 | SMF © 2006-2009, Simple Machines LLC | HarzeM Dilber MC