[DSP-10] Using dspx with dsp10

[Bob Larkin]

Hi Dave,

I'm running a bit behind, as usual!  I have both the EZ-Kit and the DSPx 
versions of the DSP-10. At this point, I am in the middle of comparisons 
between them as well as some improvements.

First, I believe the only differences between the two, using the issued 
software, is the CODEC initialization and calls. None of the functional DSP 
program differs.  So, I use the same UHFA.EXE for the PC and am able to 
control either.

I will attempt to insert comments after yours:

At 07:12 AM 2/13/2004, you wrote:
>Anyone succesfully got a dsp-x interfaced to the dsp-10?
>
>I have; but it only works "sortof", as the following explanation details:-
>
>The DSPX is hooked up as per the diagram on the dspx web site to the dsp10
>board
>
>The dsp-10 board was previously working perfectly with the ez-lite dsp
>
>I downloaded it with uhf3.exe dated 05 27 03 using ezfast (from the dsp-x
>web site)
>
>i am using uhfa.exe dated 01-22-04 as the front panel
>
>If i launch uhfa with no uhfa.cfg file it receives and transmits
>
>With no rf input I get the following signal levels against rf gain
>
>RFG    dbm
>100    -145
>94    -138.9
>88    -132.6
>82    -127.1
>76    -120.8
>70    -115.1
>64    -108.8
>
>which is going the wrong way?

I think this is right!   As you back down the RF gain, the signal level 
that corresponds to the noise level keeps getting bigger.  UHFA.CFG has a 
place for the step sizes, if you know what these are. The defaults that you 
got are 6 dB steps.  When you change the RF gain, the S-Meter/Power 
indicator.  If the steps, in the rfgdb entry of the configuration file are 
correct, the indicated power of a signal, well above the noise, will remain 
about constant.



>Also on a loud signal, ssb or cw,  received with rfgain=100 (received
>signal =-75dBm the front panel says) it is so distorted as to be
>unreadable. If I change to rfgain = 94 it is fully understandable, but it
>says received signal =-72dBm (lowering rf gain increases signal level?)

The present dynamic range should allow signals to about -56 dBm without 
overload, at full RF gain. Minimum RF Gain allows roughly 36 dB stronger 
signals to about -20 dBm.


>On transmit the locals complain that the FM is too wide (especially
>compared to the ez-lite board). This is at all Mic gain settings that
>produce deviation. SSB is ok, just "sounds electronic"! The DSPX board
>gives around 0.5V peak to peak on a good whistle on SSB, on FM it gives a
>constant amplitude (as one would expect on FM) of 3.2V peak to peak. On
>the debug scale on the uhfa screen, the A/D level meter just hits red on
>both FM and SSB. On SSB the  D/A level meter just hits red on speech
>peaks; on FM every bar, including all the bars is lit and doesnt vary. The
>RF level is correct, hitting 30mW. Comparing the o/p spectrum of the DSP10
>transceiver on my HP8536E against both my IC706 and VX1R it does fill more
>bandwidth.....i dont have a deviation meter

With the present software, it is important that in SSB transmit the last 
(right-most) red bar only light only occasionally.  The control of the 
overload is "manual."  Because of the nature of voice, this will 
not  produce an average power very close to the peak of FM or CW.  In 
addition, I did not do a very good job of handling overload, and it is 
possible, at overload, to have "Wrap-around" distortion, that is offensive 
to the ear and splatters.  The good news is that I have been playing with 
that area of the DSP program and have a version that is MUCH better.  The 
new one does clipping and filtering. There is an indicator to show the 
amount of clipping and tests so far show this to prevent splatter, get 
quite a bit more output and sound good on the air.  I need to test this on 
the DSPx and also get a hex file together, so it may be a few days,.

The FM has similar problems, plus I got the wrong pre-emphasis/deemphasis. 
I was out-of-date and used 75 microseconds, which is not enough treble 
boost on transmit or bass boost on receive. This is all fixed in the new 
one, along with the handling of overload.

Also, remember the microphone equalizer (SCRL F5). This is very powerful 
tool for improving the sound of the signal.



>If i exit uhfa to dos of course it writes a uhfa.cfg file
>
>If I then launch uhfa again i get No transmit on any mode...(the LED on
>the RF board does come on on TX) the locals dont get to complain about
>deviation!

This leaves me puzzled.

>However the need to change RF gain to 94 to hear a loud signal dissapears,
>the loud signal is perfectly copyable at rfgain = 100, even though the
>input signal level is still the same!
>
>With no input i get the following signal levels against rf gain
>RFG    dbm
>100    -194
>94    -188
>88    -182
>82    -176
>76    -170
>70    -164
>64    -158

There is a configuration variable gain_adj that might cause this, but the 
DSP-10 never changes this internally. Only for operator with a test editor!

>Which probably breaks Boltzmanns KTB formula!
>
>If i exit uhfa to dos of course it writes the uhfa.cfg file again and on
>restarting UHFA i go back to scenario 1:-
>
>((With no rf input i get the following signal levels against rf gain
>
>RFG    dbm
>100    -145
>94    -138.9
>88    -132.6
>82    -127.1
>76    -120.8
>70    -115.1
>64    -108.8
>
>etc
>
>And so the cycle repeats forever
>
>Anyone any ideas on whats happening?

Dave, if you are still seeing this pattern, I would be interested in 
looking at the configuration files. Maybe one after each of the 
states.  The pattern does not fit what I am familiar with, but  maybe 
comparing the files would give a hint.

>Dave
>
>ww2r


Good Luck and I will proceed with the new program changes--- maybe this 
will help!!

73,
Bob   W7PUA