[DSP-10] CW transmit has a raspy sound

[Lee Scott - AA1YN]

At 04:58 PM 3/31/2006 -0800, you wrote:
>Hi Lee,
>
>I must confess to being somewhat overwhelmed!
>
>Can we focus for a moment on one item.  How about your own 
>DSP-10?  Picture #6 of the 124 MHz pix (I will call it 124-6) shows a 
>reasonable VCO spectrum, free running. Pix 124-10 shows that osc (with a 
>transistor change) locked and it is terrible.  To me, this says it is a 
>PLL problem, i.e.
>* The component values in the loop filter.
>* Noise on the 10 MHz reference.  <-  Unlikely as per your tests.
>* Noise on the power supply coming from U103
>* The wave form of the signal coming back to the PLL from R112
>
>But, not a VCO problem, is it? The PLL should reduce the close in noise on 
>the VCO.
>
>The noise modulation required to mess up the phase is in the microvolt 
>range and can't be measured directly. We must deduce its level from the 
>PLL spectrum.
>
>But, before trying to suggest more avenues, is this the right problem 
>description---your DSP-10 gets noisy when the loop is closed?
>
>Lots of fun!!
>
>73, Bob

Bob,

I am trying to focus on my own DSP-10.  I just compared NG4C's J310 osc to 
mine to see if what I was seeing on mine was like his. There is only one 
photo of NG4C's and all the others are mine.

First off, I was under the assumption that a PLL could not make a dirty osc 
look clean which is why I was focusing on the J310. So much for knowing a 
little about a lot of things and a PLL expert I am not. (though maybe after 
this is done) I am also not an expert on using a spectrum analyzer though I 
am getting familiar with one fast. I am finding all of this lots of fun.

I will agree when the PLL is brought into play, the noise gets worse and 
not better.

When you speak of modulation, I assume you mean modulation on either the 
power supply, the loop filter line, and/or the F_IN line. I think any of 
these could cause what we are seeing.

Since we can't bypass the loop filter line as that would change the filter, 
all we can do is make sure the proper components are in place ( done that ) 
and the components are ok. (I can vouch for C113).  I can try replacing the 
components as I believe I have some spare. - will try that later. I do not 
have the knowledge to interpret the SA and determine values to adjust in 
the loop filter.  If there are changes needed, you would have to recommend 
values to try.  I have retouched the solder joints on the loop components 
and have just replaced C113 incase I destroyed it when I shorted it out. 
When I first replaced C113, I accidentally grabbed the 10000pf which made 
it a bit worse.  Then after realizing it, I replaced it with 1000pf. (.001uf)

I will agree that noise on the +5v line is very detrimental.  To that end, 
I've placed a  .1uf cap on the board at the point where the 10v power 
splits off for U105 and U106.  I also placed a .1uf cap on the VP line of 
U104.  It cleaned up the trace somewhat with the PLL operating. I have not 
replaced U103 yet but I have spares of most everything including a couple 
of LMX1501s. I do not have a spare MC145170 but if someone has a spare or 
two kicking around I wouldn't object.

With two very different 10MHz sources, the spectrum display is the save so 
I have to assume there is no phase jitter on the two sources. I do have an 
HP 10544A oscillator I can also try if I get the chance to build up the box 
for it.

If this is any hint to the problem - I have noticed that over a period of 
time the signal gets better than worse.  In watching a faster less 
resolution scan on the SA, there seems to be a cycle of about 5 seconds in 
length where the noise peaks.

As for wave forms, I've placed some on my website. The link is at the top 
of the SA page.

I also don't mind keeping this open to the group as others might at least 
enjoy reading if not learning.


Lee Scott - AA1YN http://www.aa1yn.com
Home of the VHF & Up Register http://www.aa1yn.com/vhf
Home of the VHF Talk chatroom http://vhftalk.aa1yn.com
Hooksett, NH
FN43gc52