Hi Ken,
First read your homework assignment:
http://www.qth.com/inrad/roofing-filters.pdf
http://www.sherweng.com/Dayton_2004/Dayton_2004-Sherwood.pdf
http://www.qth.com/inrad/QSTReview.pdf
http://www.qth.com/inrad/RadComReview.pdf
Quick answers to your questions:
>In order for us to have an intelligent discussion about roofing filters,
we need to agree on a definition of roofing filter. I don't recall
hearing that term prior to a few years ago. By the usage I have seen
here on this email reflector it seems that "roofing filter" refers to a
filter in the first IF of a receiver.
1. Yes, generally the **first** filter following the 1st mixer.
>It is unclear to me whether that
would mean that the two pole, 15 kHz BW monolithic crystal filter in my
Omni VI is the roofing filter or eight pole 2.4 kHz BW filter is the
roofing filter. Or perhaps that term does not even apply to the Omni VI
without modification.
2. The roofing filter in an unmodified Omni is the
15 kHz filter. I don't know whether that filter is
bypassed by the Inrad mod or whether it is still inline.
In the case of Orion, the 20 kHz filter is still inline
when the 500 and 250 positions are activated, but as we
learned that configuration (including following amplifier
stage) had some IMD problems versus going directly through
the stock 1000 Hz or the optional Inrad 762 600 Hz. Maybe
someone else can explain details of Inrad's mod for the Omni.
One thing is clear...it makes a dramatic improvement!
http://www.qth.com/inrad/kits.htm#tentec
>In order for a filter to be a "roofing filter"
does it have to follow the first mixer (be in the first IF) and precede
any gain stages in the first IF?
3. Yes, as stated in #1 above.
>Does it have to be fairly narrow, like
on the order of a couple times the bandwidth of the mode being used, or
less? Or can it be any filter in the first IF?
4. It can be any bandwidth, but it works best
if it is matched to the passband of the desired signal,
i.e. 500-600 Hz for CW, 2.4 kHz for SSB, 6 kHz for AM or
15-20 kHz for FM.
>Certainly bandpass filters in the RF stages before the
first mixer do not qualify as roofing filters, yet they
do perform a similar function, in that they limit the
amount of total power that the following stages have to deal
with, without producing IMD. What IS a roofing filter?
5. True, they perform a similar function and
as Jerry said this would be ideal since it would
prevent unwanted IMD products from ever reaching
the 1st mixer. However it not be cost-effective to
have banks of 500 Hz filters covering 14.000-14.0005,
14.0005-14.001, etc. Inrad does have some front-end
filters but they are typically too wide to be effective
roofing filters. They are useful in multi-station
environments to allow receiving on separate areas of
a band while another station is simultaneously
transmitting.
http://www.qth.com/inrad/front-end-filters.htm
>I don't recall hearing the term used until receivers were being made
that used IF DSP.
6. The term has been used for years. Sherwood
Engineering first offered the 600 Hz roofing filter for the
Drake R4C in the 1970's.
>Is the term only applicable to receivers using IF DSP?
7. No. I believe the reason it is associated with
DSP rigs may be one manufacturer's claim that DSP radios would
not need crystal (roofing) filters. This was clearly wrong
and the IC-7800 was the first Japanese rig to offer a 6 kHz
roofing filters (below the more typical 15-20 kHz) in order
to improve its IMD performance. Unfortunately they forgot
about phase noise which becomes another big problem at close
signal spacings. Unless Icom improved the 7800's phase noise,
the recent addition of a 3 kHz filter (in response to Yaesu's
3 kHz in the 9000) may have little effect since any improvement
will be swamped out by phase noise problems.
>I know that Inrad makes a roofing filter modification for the Omni VI
rigs. Is that an incorrect usage of the term?
8. Yes it is. It replaces the 15 kHz filter as noted in
#2 above.
>All filters in all stages of a receiver perform essentially the same
functions. That is to limit the bandwidth that following stages have to
handle, and when before a mixer to reject image response, or when before
a ADC in a DSP receiver to reduce aliasing responses. So what makes some
filters "roofing filters" and others not?
9. Simply put, a roofing filter is the FIRST filter
following the 1st mixer. If it is appropriate for the
mode selected (see #4 above), it will reduce IMD products
passing into subsequent stages of the receiver.
10. An anti-alising filter for DSP has a different
purpose but can be fulfilled by bandpass filters in a
receiver with multiple IFs like all current DSP rigs. Even
if we ever get 24-bit ADCs at 60 MHz sample rates, which
would allow direct sampling of HF signals up to 30 MHz,
we would still need a 30 MHz low-pass anti-aliasing filter
ahead of the DSP, so any claims that DSP rigs will not need
analog filters will still be untrue. More info here:
"Most often, an anti-aliasing filter is a
<http://encyclopedia.thefreedictionary.com/low-pass+filter>low-pass
filter. However, this is not a requirement. The Shannon-Nyquist
sampling theorem states that the sampling rate must be greater than
twice the bandwidth, not maximum frequency, of the signal. For the
types of signals that are bandwidth limited, but not centered at
zero, a
<http://encyclopedia.thefreedictionary.com/band-pass+filter>band-pass
filter would be used as an anti-aliasing filter. For example, this
could be done with a
<http://encyclopedia.thefreedictionary.com/Single-sideband+modulation>single-sideband
modulated or frequency modulated signal. If one desired to sample an
FM radio broadcast on channel 200, then an appropriate anti-alias
filter would be centered on 87.9
<http://encyclopedia.thefreedictionary.com/megahertz>MHz with 200 kHz
bandwidth (or pass-band of 87.8 MHz to 88.0 MHz), and the sampling
rate would be no less than 400 kHz. (In this case not the audio of
the broadcast is sampled, but the actual transmission signal itself,
which is not very common.) "
http://encyclopedia.thefreedictionary.com/Anti-aliasing%20filter
Hope this helps and I'm sure others will jump in (or on)
if I made mistakes.
73, Bill W4ZV
_______________________________________________
TenTec mailing list
TenTec@contesting.com
http://lists.contesting.com/mailman/listinfo/tentec
|