[1000mp] Good reading regarding IF tap topic (I'm not the author)

[Ross Stenberg]

Application Note for SoftRock v6                 21 January 2007

Using a SoftRock v6 as a DSP back-end for an Yaesu FT1000/1000D

By Steve Ireland, VK6VZ

There has been some strong interest in using a Softrock as a digital signal
processing back-end to an FT-1000/FT-1000D, connected before the 8.215MHz
2nd IF crystal filter.
Back in August 2006, Tony KB9YIG built up a v6 SoftRock for me to experiment
with my FT1000 as an IF signal processor, in order to see how well this
worked in practice.  I am a very keen 160m weak signal CW DX operator and
the idea (inspired by Alex VE3NEA) was to have a very sensitive bandscope
that could 'see' weak signals appearing across the CW portion of 160m at
sunset/sunrise times, coupled with the ability to use this function on the
other lowbands and to have a complimentary transmit function.
It would also enable me to compare the FT1000's famous analog receiver (with
500Hz crystal filters in both the second and third IFs), plus a brilliant
audio peaking filter, plus IF shift and width) with the Softrock backend,
which was to use Rocky and KGKSDR software with digital filtering that is
variable down to several tens of kilohertz. 
Some of you will be aware that Phil VK6APH and myself write a monthly column
in the Radio Society of Great Britain's RadCom magazine about software
defined radio and we are intending writing one on using SoftRocks as digital
back ends in this manner in the near future. 
The crystal Tony used for the FT1000 8.215 MHz IF application was a 11.0 MHz
crystal, used with 1/3 sub-harmonic sampling.  The center frequency is
about:
 3* (11.0 -0.003 ) / 4 = 8.248 MHz, i.e. 33 kHz above the IF frequency.
The modified SoftRock v6 is used with a M-Audio Delta 44 soundcard that can
sample at 96 kHz, so the IF tuning with the SoftRock is from 8.2 MHz to
8.296 which gives excellent coverage of the FT1000 2nd IF passband.  The
soundcard runs from a Pentium IV personal computer, with a clock speed of
3.4GHz For those who are interested in converting a SoftRock v6 for this
kind of application, I think Tony still has a few more of the 11 MHz
crystals and the associated components.
Tony's completed SoftRock was mounted in a diecast aluminium box by me and
then Phil VK6APH worked out how to interface it to the FT1000 and modified
the latter.  Phil had to drill a hole in the rear of the case of the FT1000D
in order to mount an RCA socket for the connection between the FT1000 2nd IF
and the SoftRock antenna socket - something which may put a few people off
this modification.  

PLEASE NOTE THAT THE SURGERY TO THE FT1000 WAS QUITE DIFFICULT TO PERFORM
AND REQUIRED THE USE OF A VERY FINE POINT SOLDERING IRON.  IT IS NOT FOR THE
INEXPERIENCED OR FAINT HEARTED AND REQUIRED THE SUBTLE SKILLS OF DR
FRANKENSTEIN (VK6APH) AND NOT THOSE OF IGOR (VK6VZ).  IT IS POTENTIALLY A
REALLY GOOD WAY TO BREAK YOUR BELOVED/EXPENSIVE FT1000.

Neither VK6VZ or VK6APH except any responsibility for any damage caused as
the result of attempting to carry out/carrying out this modification!  Note
that the process that will be described below assumes that the W8JI
modification to the FT1000 noise blanker has been carried out - this ensure
that the noise blanker is totally switched off when the NB is switched off
on the FT1000 front panel.

To power the SoftRock, we used the 13.8V available at an RCA socket on the
rear of the FT1000.  

After experimentation, Phil found the best place to get the IF output (for
the SoftRock) from the FT1000 was by 'tapping' an IF buffer amp (Q2003) at a
low impedance point (one of the gates, to be precise) with a 22pf capacitor,
following this with a 10k potentiometer/47 ohm attenuator - see the diagram
nearby.

With the antenna socket of the Softrock connected to the new FT1000 IF
output, the SoftRock connected to the PC's sound card in the usual manner, a
resonant antenna connected to the FT1000 and the latter tuned to the centre
of the 40m band, this level of attenuation gave a noise floor on the KGKSDR
bandscope of about 10dB above the base line noise (on the KGKSDR bandscope)
when there was no antenna connected to the FT1000.

The 10k variable resistor was set between one third and a half of its value,
in order to provide the correct level signal into the SoftRock.  The
adjustment isn't critical but care should be taken so as not to overload the
SoftRock/soundcard combination.

The adjustment is easiest carried out by using the Rocky software and a
signal generator.  Making sure that the FT1000 Noise Blanker is switched
off, put a signal of around S7 (on the FT1000 'S' meter) into the FT1000
antenna input (on the frequency the FT1000 is tuned to) and adjust the pot
until the point that the level of the signal shown on the Rocky bandscope no
longer increases.

Now, switch on the FT1000 noise blanker and check the signal level is the
same as before on the Rocky bandscope.  If not, you may need to adjust the
pot slightly until this situation is achieved (i.e. that the signal level on
the Rocky bandscope looks the same with the noise blanker switched on or
off).

The SoftRock v6 can be overloaded with signal from the FT-1000, so care
needs to be taken to adjust the gain of the SoftRock in KGKSDR and Rocky so
that this does not happen.

 


Conclusions

The results are just terrific - we tested the FT1000/SoftRock combination on
40m and could see +/- 30kHz from the frequency that the FT1000D was tuned
to, using Rocky or KGKSDR as our bandscope.

Given the narrow nature of the section of 160m where CW operation takes
place (1800KHz to 1835kHz) this enabled us to easily see all of this part of
the band at once. 

The SoftRock backend also gives a useful second receiver for the FT1000D.
You can see on the Rocky or KGKSDR bandscope where the main FT1000 receiver
is tuned, since the 8.215MHz crystal filter leaves a 'suck out' in the
passband (i.e. in the form of a depression in the level of the trace) on the
Rocky and KGKSDR bandscopes.

In Rocky, we set the LO frequency to 0 in the Settings dialog (many thanks
to Alex VE3NEA for this advice), while in KGKSDR we had to set the LO
frequency to 0.001KHz because it wouldn't accept a 0KHz LO.  We also found
that in KGKSDR, the bit depth needed to be set to 16-bits for acceptable
performance.

I can say conclusively that a weak SSB and CW signals are generally more
readable/more comfortably readable listening through the digital filters of
the SoftRock/Rocky or KGKSDR/Delta 44 than through the crystal filtering of
the FT1000 when there is QRN/static crashes present.  As the FT1000 receiver
is widely regarded as one of the best lowband weak signal receivers, owing
to its excellent cross mod performance and filtering, this says a lot for
how far digital filtering and processing has come.

I found KGKSDR is the best software to use for SSB for this application,
whilst Rocky is the king for weak signal CW.  I was 'seeing' and then
copying several 160m CW signals from the USA that were barely audible at
sunset today on the SofRock back-end - I never would have found them by
tuning up and down the band on the FT1000. 

What people should understand is that the bandscope that you get from
Rocky/KGKSDR etc is able to see signals that you can hardly hear (and in
some cases, not hear).  Apparently, those bandscopes used in the current
generation of HF transceivers seem fine at seeing strong signals, but not
weak ones. 

Frankly, the results are so good from the FT1000/SoftRock combo and it
appears such an aid to an active DXer and contester like me, that until I
used the combo to win a major contest, it was very tempting to keep it a
secret ;-).  The only downside so far is that SoftRock acts as a monitor
receiver when I transmit and I can hear myself coming out the speaker
delayed by a few hundred milliseconds, which is terribly distracting (!).

As a result, I need to find a way to mute the SoftRock when I transmit,
which shouldn't be too difficult. 

It would be appreciated if people didn't email me directly for further
information, as my salt mine job often requires 12-hour days at present and
my kids need the small amount of time I have outside this.  As I learn more,
I will (of course) post it on the SoftRock 40 reflector. 

Phil VK6APH and I will write up a general piece on using software defined
radios in this manner for RadCom magazine, which should appear in our column
in early 2007. 



Vy 73

Steve, VK6VZ