Author Topic: can low pass filter change signal width?  (Read 2253 times)

0 Members and 1 Guest are viewing this topic.

Offline elcrniTopic starter

  • Regular Contributor
  • *
  • Posts: 98
  • Country: at
can low pass filter change signal width?
« on: April 25, 2021, 05:11:07 pm »
hi all,

So i have a device that should serve a purpose of making a radio signal more resilient to all sorts of noise but after a lot of tests i've discovered that my output signal is "wider" that it should be.
How it works, i have a DCF77 (atomic clock time transmission) receiver that connects to this sort of filter and amplifier to remove any noise and to make a stable signal output. DCF77 is a 1Hz signal and every second 100ms and 200ms pulses should be received. This is of course never exactly 100 and 200 ms but should be very close.
when i connect DCF receiver alone i get pulses 95-205ms and 195-205ms and this helps my MCU easier to recognize and decode BUT, then when i connect the filter device, my output is on average 130ms and 230ms, so the signal is now wider.. very stable but too wide.

I am trying to figure out what is making this signal wider, not so much to save my project but i really want to know how this works and why.

Anyone has any clues?

Schematic attached.
2021-04-25-19-00-51" border="0

many thanks,
Alek
 

Offline fourfathom

  • Super Contributor
  • ***
  • Posts: 2004
  • Country: us
Re: can low pass filter change signal width?
« Reply #1 on: April 25, 2021, 06:15:56 pm »
What does your filter circuit look like?  If you run a pulse through a low-pass filter you will reduce the amplitude and stretch out the duty-cycle.  With enough filtering you will transform a narrow pulse into a sine wave with 50% duty cycle.
We'll search out every place a sick, twisted, solitary misfit might run to! -- I'll start with Radio Shack.
 

Offline elcrniTopic starter

  • Regular Contributor
  • *
  • Posts: 98
  • Country: at
Re: can low pass filter change signal width?
« Reply #2 on: April 25, 2021, 06:48:02 pm »
thanks fourfathom!
What does your filter circuit look like?
on above attached schematic, filter is R6 and C5, both seem rather high in values, to my untrained eye...
many thanks,
Alek
« Last Edit: April 25, 2021, 06:53:16 pm by elcrni »
 

Offline bsfeechannel

  • Super Contributor
  • ***
  • Posts: 1668
  • Country: 00
Re: can low pass filter change signal width?
« Reply #3 on: April 25, 2021, 07:44:44 pm »
A low pass filter will invariably widen your pulses if the decision point on the pulse regeneration is closer to ground (as in the case of the figure below).



R7 is pulling up the negative inputs of IC1(pins 2 or 6), which will shift the decision point away from 1/2 UB+ provided by R5 and R4.
« Last Edit: April 25, 2021, 08:52:40 pm by bsfeechannel »
 

Offline fourfathom

  • Super Contributor
  • ***
  • Posts: 2004
  • Country: us
Re: can low pass filter change signal width?
« Reply #4 on: April 25, 2021, 09:38:08 pm »
R7 is pulling up the negative inputs of IC1(pins 2 or 6), which will shift the decision point away from 1/2 UB+ provided by R5 and R4.

R7 may have that effect, or it may have no effect at all, it depends on the characteristics of the pulse source.

Your 100K / 0.1uF filter gives the 100ms and 200ms pulses enough settling time to where the distortion with a 50% comparator threshold should be minimal.  But this will only be the case if the input signal is a 0V / 5V signal.  If it is anything else then your 2.5V threshold will give you pulse shrinkage or expansion.
We'll search out every place a sick, twisted, solitary misfit might run to! -- I'll start with Radio Shack.
 

Offline elcrniTopic starter

  • Regular Contributor
  • *
  • Posts: 98
  • Country: at
Re: can low pass filter change signal width?
« Reply #5 on: April 25, 2021, 09:52:26 pm »
Many thanks guys!

Bellow is the schematic of the module "DCF1" that connects to the "filter module".

2021-04-25-23-48-22" border="0

I think the signal HIGH is whatever the voltage is supplied to the module DCF1, in this case that should be 2.5V...

thanks,
Alek
 

Offline fourfathom

  • Super Contributor
  • ***
  • Posts: 2004
  • Country: us
Re: can low pass filter change signal width?
« Reply #6 on: April 25, 2021, 10:18:12 pm »
But this will only be the case if the input signal is a 0V / 5V signal.  If it is anything else then your 2.5V threshold will give you pulse shrinkage or expansion.

My mistake above, for some reason I though that the "+UB" voltage was 5V.  It's actually 2.5V, and your threshold is 1.25V.  The transistor driver output of your source should generate a 0/2.5V signal, with the rise-time slightly slower than the fall time, due to the 10K pull-up.  This wouldn't be enough to account for your 30ms pulse-width expansion.

Can you remove the filter capacitor to see if that is indeed causing the problem?
We'll search out every place a sick, twisted, solitary misfit might run to! -- I'll start with Radio Shack.
 

Offline elcrniTopic starter

  • Regular Contributor
  • *
  • Posts: 98
  • Country: at
Re: can low pass filter change signal width?
« Reply #7 on: April 25, 2021, 10:39:00 pm »
Can you remove the filter capacitor to see if that is indeed causing the problem?

Yes, i can easily remove it, getting it back may be a bit tricky though, soldering 0402 by hand is not one of my strengths :-)

Any way to test this capacitor theory without actually removing it?

Many thanks,
Alek
 

Offline fourfathom

  • Super Contributor
  • ***
  • Posts: 2004
  • Country: us
Re: can low pass filter change signal width?
« Reply #8 on: April 26, 2021, 12:41:33 am »
Can you bridge R6 (100k) with a 1K or 100 Ohm resistor?  This won't stress the driving transistor but will reduce the filter time-constant; it will be about 0.1ms on the positive edge, and much less than that on the negative edge. I am wondering if that filter capacitor is actually 0.1uF, or perhaps the resistor values aren't as shown on the schematic.

One other possibility is that the opamps you are using as comparators, while plenty fast in the linear region, may have different delays coming out od positive and negative saturation.  This is just a wild guess though, I don't know enough about those amps and the specs don't tell me enough.
We'll search out every place a sick, twisted, solitary misfit might run to! -- I'll start with Radio Shack.
 

Offline bsfeechannel

  • Super Contributor
  • ***
  • Posts: 1668
  • Country: 00
Re: can low pass filter change signal width?
« Reply #9 on: April 26, 2021, 03:53:04 pm »
Any way to test this capacitor theory without actually removing it?

Do you have a scope?
 

Offline elcrniTopic starter

  • Regular Contributor
  • *
  • Posts: 98
  • Country: at
Re: can low pass filter change signal width?
« Reply #10 on: April 26, 2021, 06:29:11 pm »
Can you bridge R6 (100k) with a 1K or 100 Ohm resistor? 

i can try! :-)


Any way to test this capacitor theory without actually removing it?

Do you have a scope?

Yes, i do.

Thanks,
Alek
 

Offline elcrniTopic starter

  • Regular Contributor
  • *
  • Posts: 98
  • Country: at
Re: can low pass filter change signal width?
« Reply #11 on: April 26, 2021, 08:54:38 pm »
ok, i have bridged R6 (100k) with 1k and then with 100Ohm, not a huge difference but pulse width now dips much closer to both 100ms and 200ms although still goes to as high as 130 and 230.
So if i take the average value, its more or less the same, maybe slightly better, it just get much closer to target values than before.

when i connect everything, first 1-2 pulses get to almost exact values, the they get wider... capacitor?

Many thanks,
Alek
 

Offline bsfeechannel

  • Super Contributor
  • ***
  • Posts: 1668
  • Country: 00
Re: can low pass filter change signal width?
« Reply #12 on: April 26, 2021, 09:55:49 pm »
Do you have a scope?

Yes, i do.

Cool. Then, please, poke around with your scope probe along the signal path and post the waveforms so that we can try to figure out what is going on with the circuit.
 


Share me

Digg  Facebook  SlashDot  Delicious  Technorati  Twitter  Google  Yahoo
Smf