DSP-10 2 Meter Transceiver

Below are the spectrial measurements of the local oscillators on my DSP-10. This is to try and trace down why the transmitted signal sounds raspy. Both oscillators were disconnected from their mixers and a coax pigtail with an SMA connector was connected to the ouput of the oscillators. This provided a 50 ohm termination via the measurement equipment. The internal 10mHz osc was used in all cases.

The output levels of both oscillators were measured using an HP 435B power meter with a 300mw head. Neither output seemed abnormal for the TUF-1 mixers. The spectrum analyzer is an HP 141T with an 8555A RF Section. The output of the 19mHz oscillator while not totally clean, looked ok to me. The 124mHz oscillator on the otherhand, appeared to be quite dirty.

For waveforms, click here.

Click on the thumbnail photos for a higher resolution image.
 

 

19MHz osc

I started with an unlocked measurement which from what I understand should be about as noisy as a PLL can be. The photos show the settings of the 141T.
 
I adjusted the PLL control voltage (output of U6) to 0.75v
 
I adjusted the PLL control voltage (output of U6) to 4.6v
 
I adjusted the PLL control voltage (output of U6) to 8.9v
 
  

124MHz osc

Again, for the 124MHz osc, I started with an unlocked measurement. Looked like the output of an RC oscillator - kind of what I would expect.
 
This is the same analyzer settings as above but the osc is locked.
 
  This is almost the same analyzer settings the 19MHz above with the control voltage (junction of R104 and R105) set to 0.75v. The sweep is 5KHz/div instead of 2KHz/div. Unfortunately it still looks like an RC oscillator. This is not what I expected.
 
As above with the control voltage (junction of R104 and R105) set to 4.5v.
 
I am in the process of building a DSP-10 for NG4C. Since I can't program the PLL at this moment, I didn't look at its 124MHz osc. As Bob pointed out to me, by shorting C113, we eliminate any possible phase noise coming from the PLL circuit. This is a little better but I don't feel it is acceptable.
 
This is the same setup on my DSP-10. C113 shorted. Still pretty bad.
 
In looking at the circuit, I figured I could also short R152 and this should eliminate any noise coming from the varicap diodes. C113 is also shorted. Just a hair bit better but still pretty bad.
 
I removed the input pin on U105 and connected the coax pigtail to the junction of C119 and U105. This is what the mmic would see. The power level is -3.0 dBm.

If the output of U106 is to be +7dBm, then the output of U105 should be -1.0dBm and the output of the osc (the point we are measuring here) should be -23dBm. Just as a try, I pushed the input pin of U105 back down to the board and saw no change in the signal except for a slight frequency shift. The output level stayed the same.
 
I built a seperate osc based on the same original components except for the J310 where I used an SOT-23 version. This osc had considerably less noise than the one on the DSP-10. I decided to replace the J310 on the DSP-10 and came up with not quite as good results as the test oscillator but better than the original. Here is a photo of the test osc I built along with it's output SA. Here is the photo of the replacement J310 on the DSP-10.
 
I unshorted R152 and C113, reconnected the KDSP and booted up. Here is the output now of the 124MHz osc. The PLL locked ok but YEECCH !

This is with the 10pf to gnd on the input of U105 and a 6.8pf cap on the PLL Fin line where it taks a via to the bottom of the board. Nothing I did wanted to make it any cleaner. It loked like this prior to the additions. Maybe the replacement J310 wasn't as good an idea. Next - I'll fiddle with the loop filter components - maybe there is something wrong there.
 
  Conclusions so far:

1) We've eliminated the PLL as the source of the phase noise and is thus strictly in the oscillator.

2) The only components I used in the osc which were not from the kit is the two capacitors C117 and C118 on both DSP-10s. I replaced them with a 5pf and 22pf capacitors made by Panasonic. This was required to get the oscillator to operate at the proper frequency. In choosing these values, I kept the ratio of the two approximately the same.

As far as the value of C119, I did use a 3.3pf cap on the test board with the same +12dBm output. On the DSP-10, I tried a 2.2pf and then a 1.1pf capacitor and could not reduce the output of U106 by more than 0.5dB so I placed the 3.3pf back and added a 10pf from the input of U105 to ground in hopes of using it as a voltage divider. There is a slight crook to the sine wave output of U105 on my o'scope and it slightly cleaned up that signal but I guess not enough. I may get back to this later. It doesn't seem to effect the output of U106 as that looks more like a sine wave.

| AA1YN Home | VHF Home | WSJT Home |