What crystal filters for the K3 are linear phase delay (constant group delay) ?

[Larry Lopez]

I'm wondering if any digital filters exist for the K3.
I'm wondering which common filters work best on the K3.

I'm interested in JT65 and JT9

linear phase delay and constant group delay
means that the signal of different frequency
take the same time to go through the filter
so that the waveform is preserved.

this is needed to get best results.
for narrow filters this is more serious
since the group delay changes are
at the skirts which are close together.

I've haven't been able to find such filter.

I've been using the filters on my FT1000D
and I do notice the 250 Hz one doesn't
seem to help JT65HF as much I would
have expected.


Any ideas,
Larry

[Jim Brown-10]

On 4/21/2014 1:08 PM, Larry Lopez wrote:
> I've been using the filters on my FT1000D
> and I do notice the 250 Hz one doesn't
> seem to help JT65HF as much I would
> have expected.
>
>
> Any ideas,

Read previous posts about this from W4TV and myself this morning.

73, Jim K9YC
______________________________________________________________
Elecraft mailing list
Home: http://mailman.qth.net/mailman/listinfo/elecraft
Help: http://mailman.qth.net/mmfaq.htm
Post: mailto:[hidden email]

This list hosted by: http://www.qsl.net
Please help support this email list: http://www.qsl.net/donate.html
Message delivered to [hidden email]

[Larry Lopez]

Thank you Jim,
I read all your posts.

I'll quote your quote of K1JT again here:

    K1JT says that it is best to run the radio broadband and let the WSJT
    software provide the needed selectivity in its decoding algorithm. The
    reason is simple -- phase shift can cause decoding errors, filters in
    the radio create phase shift.

    73, Jim K9YC

But what about:

very weak signals reducing the bandwidth increases the signal to noise ratio.
very strong signals reducing the bandwidth can prevent the DSP chain from
desensitizing the receiver and/or saturating the a/d converter.

Of course reducing the bandwidth tends to "phase shift can cause decoding errors".

Which is why I'm asking if there are any constant group delay filters out there.

Larry

[Larry Lopez]

Now with EME this isn't much of a problem I would think.

If no strong signals are present then the DSP wold be able
to shrink the bandwith just fine.

Even though a wide filter lets in a lot of noise
the DSP can get rid of it.

[Larry Lopez]

This is wrong:

  Even though a wide filter lets in a lot of noise
  the DSP can get rid of it.

The WSJT software is a DSP.

It has it's own filters and the DSP on the K3
is not needed.

So the only job of the K3 is to produce a clean
wideband signal for the WSJT.

You would still have to prevent overload
by what ever means.

Larry

[Jim Brown-10]

On 4/21/2014 1:59 PM, Larry Lopez wrote:
> But what about:
>
> very weak signals reducing the bandwidth increases the signal to noise
> ratio.
> very strong signals reducing the bandwidth can prevent the DSP chain from
> desensitizing the receiver and/or saturating the a/d converter.

Joe is well aware of the issue. Hiss answer that the phase distortion
issue trumps S/N ratio. You may want to study what he has had to say on
the issue. Joe reads the WSJT list.

73, Jim K9YC
______________________________________________________________
Elecraft mailing list
Home: http://mailman.qth.net/mailman/listinfo/elecraft
Help: http://mailman.qth.net/mmfaq.htm
Post: mailto:[hidden email]

This list hosted by: http://www.qsl.net
Please help support this email list: http://www.qsl.net/donate.html
Message delivered to [hidden email]

[Larry Lopez]

I've already recanted on S/N ratio.
I was an idiot.
I kept skipping the part where the PC is also a DSP.
Sven's post on the other thread also brought my attention to it.

The only thing left is strong carriers overloading the receiver.
I'll go find WSJT and learn some more.

Thanks again Jim.

Larry N2CS

  Joe is well aware of the issue. Hiss answer that the phase distortion
  issue trumps S/N ratio. You may want to study what he has had to say on
  the issue. Joe reads the WSJT list.

  73, Jim K9YC

[Joe Subich, W4TV-4]

 > Joe is well aware of the issue. Hiss answer that the phase distortion
 > issue trumps S/N ratio.

K1JT is correct as it relates to EME where one does not generally have
strong interfering signals.  Remember the detection bandwidth in WSJT-X
is on the order of a few Hertz - not the 2.8 to 5 KHz audio bandwidth
that would be the case with human decoding of CW or SSB.

Note that K1JT also recommends operating without AGC if possible.
However, when we get to HF with its mixture of weak and strong signals,
narrow filters are sometimes necessary in order to prevent blocking
(AGC and/or ADC overload).

73,

    ... Joe, W4TV


On 4/21/2014 5:28 PM, Jim Brown wrote:

> On 4/21/2014 1:59 PM, Larry Lopez wrote:
>> But what about:
>>
>> very weak signals reducing the bandwidth increases the signal to noise
>> ratio.
>> very strong signals reducing the bandwidth can prevent the DSP chain from
>> desensitizing the receiver and/or saturating the a/d converter.
>
> Joe is well aware of the issue. Hiss answer that the phase distortion
> issue trumps S/N ratio. You may want to study what he has had to say on
> the issue. Joe reads the WSJT list.
>
> 73, Jim K9YC
> ______________________________________________________________
> Elecraft mailing list
> Home: http://mailman.qth.net/mailman/listinfo/elecraft
> Help: http://mailman.qth.net/mmfaq.htm
> Post: mailto:[hidden email]
>
> This list hosted by: http://www.qsl.net
> Please help support this email list: http://www.qsl.net/donate.html
> Message delivered to [hidden email]
>

______________________________________________________________
Elecraft mailing list
Home: http://mailman.qth.net/mailman/listinfo/elecraft
Help: http://mailman.qth.net/mmfaq.htm
Post: mailto:[hidden email]

This list hosted by: http://www.qsl.net
Please help support this email list: http://www.qsl.net/donate.html
Message delivered to [hidden email]

[Larry Lopez]

It turns out that the Elecraft 2.7 KHz 5 pole filter is the winner
according to:

http://www.qsl.net/gm3woj/k3filtertestssn2203.pdf

which makes me want to make my own filter.
Still, I'm going to keep the 400Hz 8 pole filter which
is coming with my S/N 27xx Elecraft Radio !!!

[Larry Lopez]

And just for the record.

If you are going to optimize for bandpass shape
then you are going to have non constant group delay.

If you add more crystals you can have both and the
filters would be more expensive.

It's a compromise and we demand this compromise.

A phase linear filter looks bloody awful in terms
of bandpass shape.  No sharp sides.

But these days the DSP can provide some of the
attenuation so perhaps a compromise can be made.

Which is why I like the 2.7 KHz 5 pole filter.

[Larry Lopez]

Some of the wider filters in that paper are pretty good.
The narrowest filters are not so good.
The flat part of the passband has constant group delay.
The skirts are awful

For narrow filters it starts going weird.

Larry Lopez [via Elecraft] wrote:

It turns out that the Elecraft 2.7 KHz 5 pole filter is the winner
according to:

http://www.qsl.net/gm3woj/k3filtertestssn2203.pdf

which makes me want to make my own filter.
Still, I'm going to keep the 400Hz 8 pole filter which
is coming with my S/N 27xx Elecraft Radio !!!

---

If you reply to this email, your message will be added to the discussion below:

http://elecraft.365791.n2.nabble.com/What-crystal-filters-for-the-K3-are-linear-phase-delay-constant-group-delay-tp7587633p7587774.html

To start a new topic under [K3], email [hidden email]
To unsubscribe from Elecraft, click here.
NAML

[Jim Brown-10]

On 4/23/2014 10:13 PM, Larry Lopez wrote:
> Some of the wider filters in that paper are pretty good.
> The narrowest filters are not so good.
> The flat part of the passband has constant group delay.
> The skirts are awful

Nothing at all surprising. In general, phase distortion (group delay)
goes along with variation in amplitude response, steeper skirts results
in more phase distortion in the passband. That means it's greatest with
steep filter skirts, AND extends within the passband from the filter
skirts. DSP signal processing is a simulation of analog circuits -- when
you do filtering in DSP, it's a simulation of an analog filter, and it's
far easier to build a filter in DSP with low phase distortion than it is
with physical components.

In general, when phase shift is an issue, we want to use DSP for the
narrow filtering and fairly wide analog filters (the crystal roofing
filters) to protect the DSP from overload. And when phase shift is an
issue, wider bandwidth is better.

The current wisdom from those who "get it" is to use a 400 Hz roofer for
most narrow digital modes (RTTY, JT65) and set DSP for about the same
bandwidth. K1JT, the Nobel laureate who gave us the WSJT modes, says to
run the RX broad and let his software do the filtering.

73, Jim K9YC
______________________________________________________________
Elecraft mailing list
Home: http://mailman.qth.net/mailman/listinfo/elecraft
Help: http://mailman.qth.net/mmfaq.htm
Post: mailto:[hidden email]

This list hosted by: http://www.qsl.net
Please help support this email list: http://www.qsl.net/donate.html
Message delivered to [hidden email]

[Jim Brown-10]

On 4/23/2014 10:47 PM, Walter Underwood wrote:
> The usual approach is to choose a filter with steep skirts that is wider than you need, so you can use the middle of the passband where the phase response is not so bad. Then do additional filtering in DSP, where you can have good amplitude and phase performance at the cost of delay. Better filters, more delay.
>
> There is no free lunch, it's physics.

Well said.

73, Jim K9YC
______________________________________________________________
Elecraft mailing list
Home: http://mailman.qth.net/mailman/listinfo/elecraft
Help: http://mailman.qth.net/mmfaq.htm
Post: mailto:[hidden email]

This list hosted by: http://www.qsl.net
Please help support this email list: http://www.qsl.net/donate.html
Message delivered to [hidden email]

[Larry Lopez]

I ordered from Inrad 6KHz AM filter.
I bought on ebay a Astron RS-35M Power Supply
I figure even if it has something wrong with it,
it will be easy to fix.

Thank you for all the help.
It's been mind bogglingly interesting.

Larry