[DSP-10] CW transmit has a raspy sound

[Bob Larkin]

Hi to All!

Lee, the SA photos that you posted with C113 shorted, look OK. These look 
to be well within the range that will be cleaned up by the PLL. They look 
comparable to what I see and mine cleans up fine with the PLL!

The level's of the LO at the TUF-1 is fine. I believe it may be a few dB 
over the +7 dBm nominal, but this should be OK. If one wanted to reduce 
this a bit, the DC power feed resistors to U106 would be worth playing 
with, but i doubt that one could see any difference afterwards.

So, the problem would seem to be with the PLL. Possible areas might be:
* The component values in the loop filter.
* Noise on the 10 MHz reference.
* Noise on the power supply coming from U103
* The wave form of the signal coming back to the PLL from R112

Let me add a couple of words on the latter item.  I originally did not have 
C146 in the circuit, and would see noise on the PLL, not unlike what you 
see. Experimentally, C146 provided a major change, even though it has 
little effect on the 126 MHz level. C125 and C146 are separated physically 
by perhaps 35 mm and act as a LPF on that line. I speculate that a sine 
wave makes the LMX1501 happier than a distorted wave form. You might try 
adding a small cap of a few pF across C146.

Good Luck!

73, Bob  W7PUA

At 11:55 AM 3/29/2006 -0500, you wrote:
>Bob,
>
>I've done a little more -
>
>Note that in all I've done so far, the PLL was not running - ie it was not 
>programmed. As a matter of fact, the KDSP-10 is not even connected.
>
>As soon as I read this reply from you, I realized by shorting C113, you 
>eliminate the PLL as a source of phase noise all together.  I am in the 
>process of building a DSP-10 for NG4C and it is at the stage where I can 
>operate the 124MHz osc so I shorted C113 on both DSP-10s.  I've posted the 
>results on my website.  I also shorted R152 on mine so it would eliminate 
>any noise coming from the varicap diodes but only saw a slight improvement.
>
>At this point, the only components I think I can suspect is the two 
>capacitors C117 and C118.  I changed the values of these as the oscillator 
>would not operate at the proper frequency.  I chose a value of 5pf for 
>C117 and 22pf for C118.  I kept the ratio of the two the same.  Perhaps 
>this is the problem??? I used some Panasonic capacitors. They are NPO 50v 
>ceramic caps.  Do they need to be something special or a better grade? 
>Should I have just reduced maybe C118 and kept C117 at 10pf? What values 
>did you end up using on your circuit?
>
>I remember the discussions on this reflector before about the osc being 
>too low in freq and I know others have reduced these caps to get the 
>circuit to tune properly. Since some others are having the same complaints 
>about the quality of the signal, I would suspect they are having the same 
>problem. I am willing to give anything a try.
>
>Lee - AA1YN
>
>At 08:50 PM 3/28/2006 -0800, you wrote:
>>I posted the second picture of the 126 MHz oscillator, free-running, at
>>http://www.proaxis.com/~boblark/127mz2.jpg
>>You can see that it has picked up some 5 kHz modulation , but otherwise 
>>is clean.  The 5 kHz is likely stray pickup from the PLL IC that is still 
>>running. The output of the phase detector is removed by a wire across 
>>C113. The frequency at this voltage has moved down to about 123 MHz.