[TenTec] 9MHz 500HZ filter

[Bill Tippett]

         Long but I'm responding to comments in different posts:

N4VZ wrote:
 >One could easily add a simple (cheap) resistive pad at the output
of ones homebrew 4 pole filter to bring the level back down to that
expected by the receiver.  In doing so, one might consider equalizing
the loss to that of the 250Hz filter, which has a couple of dB greater
loss than the 500Hz filter.  Presumably, this would improve the
IMD degradation problem.  Then, one could leave the 1000Hz filter
in place and put the new 600Hz filter in the 500Hz slot.

         Good thinking!  I already tried a 6 dB pad on my Inrad 400.  In
the meantime, TT apparently adjusted the gain compensation in DSP for the
500 Hz roofing filter position.  I removed my pad after I noticed the
gain dropped in a recent firmware revision.  I'm guessing they actually
decreased the DSP gain because of my apparent decrease in background
noise (ignoring S-Meter readings) when in the 500 position.  I had
already compensated with my 6 dB pad but then they apparently duplicated
that with the DSP gain change, so I removed my pad.  In case anyone is
totally confused by this here is the way Orion's IF gain works:

20k-1k 9.001500 filters:  filter > amplifier > DSP including amplifier

250/500 9.000750 filters:  20 k filter > amplifier > 250/500 filter > 
amplifier > DSP including amplifier.

The second amplifier stage in the First IF is ~12 dB, which is too high
for the 500 Hz filter, but is OK for the 250 Hz filter.  Since the second
amplifier's gain cannot be changed, TT has apparently now decreased the
gain at the following DSP stage.

BUT...I still don't want the extra amplification stage contributing to IMD 
problems (even if gain were perfectly compensated).  This is why
I want a low insertion loss 600 Hz filter in the 1000 position which
needs only ONE stage of amplification.  The 600 Hz filter's insertion
loss should be similar to the 1000 Hz (both are 4-pole and I estimate ~4
dB insertion loss).

 >Of course, for those interested in my original post about creating
useful first IF filters for the OMNI VI, you would want to get on
board Bill's train for the special order, since the offset is
critical to make the passbands of the two IFs align.

         I believe Barry N1EU is correct about Omni VI.  Orion uses 9.001500
for the 1000 and higher positions, not 9.000750 which is used in the
250/500 positions (which is what I believe the Omni VI requires).  Thus I
don't believe 9.001500 filters will work in Omni VI.  Besides, Omni VI
probably needs a better shape factor than Orion does to precede its
following IF stages.

 >If you get a center frequency
of 9.001500 with a narrow bandwidth, 400, 500 or 600 HZ,
you are assuming you want a CW tone of 1500Hz with
a BFO of 9 MHz.  The correct center frequency for
the narrow filter is 9.000750, which results in a
default CW pitch of 750 Hz, +/- the adjustment.

         BUT...you forget that Orion can compensate for any Pitch offset
from 100 Hz to 1200 Hz via software.  Pitch, UCW/LCW, etc. are all
perfectly maintained because it is taken care of in software which
Omni cannot do.  HOWEVER, Orion's software ASSUMES the 1000 Hz and
above positions are using a CF of 9.001500, so anything else requires
using PBT to correctly center.  If you don't maintain the correct CF's
in Orion's IF, you will need to adjust PBT if you go between 1000 and
higher positions to the 250/500 positions which is very inconvenient.  If
you don't believe me, try plugging an Inrad 400 or TT 500 into the 1000
slot and you will find you cannot use the radio in LCW (works for UCW
only because of limited PBT adjustment range).  Do you really want to be
adjusting PBT by 750 Hz every time you switch between the 1000 and up 
positions and 500/250 positions?

 >In summary, I believe a better choice for a special
order frequency is 9.000750, which would accomdate
both Orion and OMNI VI users.

         Definitely not true for Orion users.  As I said before, this filter
is ONLY for Orion but there is nothing to prevent anyone from making
a separate order for Omni users.

 >BTW, will the kit include the four pins for mating with the socket?

         The $24 kit does NOT include the PC board with 4 pins.  W2VJN will
supply blank TT-compatible PC boards with pins for $10 additional.  This
is why I suggested ~$40.  $24 for the kit, $10 for the board plus First
Class mail from Inrad to me (10 each) and from me to you (1 each).

         Lest we forget, the whole point of this exercise is the following:

1.  AVOID completely the second stage of amplification in the First IF.
2.  Achieve a ~900 Hz -20 dB BW instead of ~1500 Hz BW with the 1000 Hz.

This should improve Orion's already excellent BDR and IMD at 1 kHz signal
spacings (see below), and maintain current 1 kHz performance at even
tighter spacings, with no loss in functionality of PBT, UCW/LCW etc (i.e.
everything should work just as it does now with the 1000 Hz filter but
simply with a tighter bandwidth).

         As I told someone off-list, I would probably move my 1000 filter to
the 1800 position in CW contests, just to have the added flexibility.  I
may permanently move my 1800 to the 2400 position anyway since 1800 is
what I always use in SSB for contests or DX-ing.

         So far I have 3 definite's and 2 possible's.  Also I make no
guarantees about anything since I have not already tried this.  However, 
the ~$40 cost seems like a reasonable risk to me.

                                                 73,  Bill  W4ZV

Page 16 http://www.arrl.org/members-only/prodrev/pdf/lab/orion_etrr.pdf

Notes:
1. On the Orion, the DSP bandwidth was set to 500 Hz, but the first IF 
filter was set to 1000 Hz bandwidth. Although the
optional 500 Hz IF filter (designed for use in earlier Ten-Tec 
transceivers) was installed, it was not used in the ARRL Lab
measurements because the dynamic range was degraded (by 3-5 dB) when it was 
selected. The 500 Hz filter has some
additional insertion loss, and there is an additional amplification stage 
that is used in the receiver to compensate for this loss.
Apparently, that amplification stage is responsible for the degraded 
performance.


Pages 19 and 20, Swept Graphs at 1 kHz spacings:

Pre-amp IMDDR3          BDR
OFF             84 dB           119 dB
ON              91 dB           119 dB