SINUS 5 WATT PLL

[sinus trouble]

Many thanks for all your support fellow Necks!
I can take criticism, its part of learning but it annoys me when someone who hasnt built anything other than a pile of b*llshit on a forum wants to dictate my intentions! Lets see what your "techincal team" can produce?? otherwise F*ck off!! the young internet users are waiting??

Thanks again to Mr Pjeva for all your assistance! Had your backin since day one bro! gonna give the PA a tweakin see if theres any improvements? il keep in touch geez!

[Elopid]

I can take criticism, its part of learning but it annoys me when someone who hasnt built anything other than a pile of b*llshit on a forum wants to dictate my intentions! Lets see what your "techincal team" can produce??

[koopa]

As per the norm another nice exciter. Your marrying of the pll and vfo is a nice combo and like teckings says your one watter is clean and works like magic.nice work Mr sinus

[sinus trouble]

Hello necks!
Here is the PCB and layout for the 5 watter!
PCB is complete and suitable for etching, layout has some minor issues but i will update soon as soo bare with me!

Psycho NRG 5 Watt PLL.png

Psycho NRG 5 Watt PLL Layout.png

[teckniqs]

Looking forward to the SO239 socket-free version.

[Dai Pole]

Could easily allow for socket and soldered coax connection by increasing length of the SO239 pad horizontally, and adding a ground pad to the right of it - maybe with thermal relief so it's obvious where to solder the braid. Unless the SO239 bolts would get in the way.

[Albert H]

Sinus - well done. Looks good!

One thing that Stephen Moss and I tried out was to etch the six coils for the VCO. We also used etched coils in the driver stage, but reverted to ordinary wound wires for the 2SC1947 end of the board. We also did away with the SO239 socket on the board. They were expedient in the early days, but later PCBs were adapted for (optional) direct coax connection.

Stephen was always wedded to the idea of using generic logic ICs for the PLL, on the basis that they would be available anywhere in the world. We did make a small run of boards with a TSA5511 PLL and a 33V charge-pump circuit for the varicap bias. We found that a bigger voltage swing would allow the VCO to cover the whole band without adjustment, and the driver stage was dimensioned so that it had more gain at 108MHz than it did at 88 MHz, so that the gain droop of the final stage could be compensated for. The output filter was dimensioned for cut-off at about 140MHz so any small amount of 1½f would be attenuated.

One thing to consider - the highpass filter from the output of the VCO stage is pretty crude, and could do with a serious upgrade to prevent ANY ½f getting through at all! This was something we considered, but never got 'round to it.

[sinus trouble]

Thankyou Mr Albert!

I suppose stripline inductors could have its merits, although wound coils take a bit more time? i really like having them to adjust if needed!
PLL aside, i think the VFO is an engineering masterpiece! its stability is incredible and tuning is very narrow, hardly any response from the trimmers until close to frequency!
I find the SAA1057 is far from perfect for this circuit, PIC is also susceptible to quick on/off supply surges, although not common in general use and resets after 30mins? it can cause lock problems!

The 33v tuning voltage is something ive been pondering on? the SAA1057 could be configured to do something similar? i might give it a go!

Also a bit of interesting info i got with the VFO? Not sure if it was written by stephen?

IMG-20160209-WA0001.jpg

[Albert H]

The 1057 will allow 30V - at least - for Vcc3. Have you got a 1k resistor on pin 1 of the 1057 (instead of the published 180ohms)? This prevents the PLL reducing the bandwidth of the mod by slowing down the loop filter!

[sinus trouble]

There is indeed a 1k on pin 1 sir!
I do have two more boards ready and shouldnt be too much trouble feeding an external 30v to Vcc3 for test purposes!
That aside, I still have a small issue with capacitance loading on the 1057 output amplifier (pins 5,6) which causes a very small swing back and forth of the carrier!
The audio input cap plays a big part in this instability! as suggested before a solution would be a seperate varicap? im dubious of modifying the whole tank circuit to accomodate another varicap so i need to look into other options :/

[Albert H]

I usually use two varicaps - sometimes more!

[Analyser]

There is indeed a 1k on pin 1 sir!
I do have two more boards ready and shouldnt be too much trouble feeding an external 30v to Vcc3 for test purposes!
That aside, I still have a small issue with capacitance loading on the 1057 output amplifier (pins 5,6) which causes a very small swing back and forth of the carrier!
The audio input cap plays a big part in this instability! as suggested before a solution would be a seperate varicap? im dubious of modifying the whole tank circuit to accomodate another varicap so i need to look into other options :/

What do you mean by a swing back and forth on the carrier, do you mean the frequency isn't stable? Does this happen with no audio or only when mod is applied?
Having seperate diodes is always a good idea but mainly becuase you can set different sensitivities and the audio level is not affected as much by the PLL loop voltage. If the audio is causing the loop to be unstable the cause is an under-damped loop filter and can normally be fixed by changing the loop filter values.
In FM broadcast applications you need to calculate the PLL for a small loop bandwidth, meaning that the PLL doesn't try and correct all but the very lowest audio frequencies; some people say 5-6Hz is desirable here but this leads to very long lock times. In professional applications they use a dual-speed PLL to get around this, it locks quickly but then switches to a different time constant to preserve the loop bandwidth.
In your case I would say about 20Hz should be ok, you'll get a lock in perhaps 10-15 seconds and the satbility and audio quality should be quite good.

[sinus trouble]

Yes id say the carrier without audio isnt stable however small!
its hard to explain? but ill try lol
The carrier has a very slow gentle rock back and forth on frequency? if a 10uf audio decoupling cap is in circuit? (C12) the tuned indicator on my radio receiver blinks on and off at approx 1 second intervals? As i reduce this capacitor in value, tuned indicator becomes more and more steady? If i remove this capacitor completely? tuned indicator is solid!
Pira do mention that high cap loading can cause instability
When audio is applied,there doesnt seem to be a battle between audio and PLL and its sounds great!

[Albert H]

Some people say 5-6Hz is desirable here, but this leads to very long lock times. In professional applications they use a dual-speed PLL to get around this, it locks quickly but then switches to a different time constant to preserve the loop bandwidth.
In your case I would say about 20Hz should be OK, you'll get a lock in perhaps 10-15 seconds and the stability and audio quality should be quite good.

The "Pira" circuit with the 1057 and the 16F84 is a dual-speed PLL if you use the original firmware. Some people trade off the dual-speed for a lock indicator (it's one of the "test" modes of the 1057), but I found that this lead to a noisy carrier and a less stable lock. The "original" circuit for this PLL has a time constant of about 2 seconds at first - to bring it into lock - then about 18 seconds to let the bass through!

The 1057 and the 16F84 solution is actually very elegant and just uses one 8-way DIP switch to cover the whole band. I found that the VCO that you use it with does matter - the NRG two transistor balanced job works very well, and with the right varicaps and 30V on the Vcc3 pin of the 1057, you can tune the whole band without any adjustment. The only major change I made was to use a second dual varicap for the modulation (effectively in parallel with the tuning one), with its DC bias provided by a low-noise op-amp. The op-amp also provided a "monitor" take-off point for the processed audio, which could be very useful when setting up links.

[Analyser]

Yes id say the carrier without audio isnt stable however small!
its hard to explain? but ill try lol
The carrier has a very slow gentle rock back and forth on frequency? if a 10uf audio decoupling cap is in circuit? (C12) the tuned indicator on my radio receiver blinks on and off at approx 1 second intervals? As i reduce this capacitor in value, tuned indicator becomes more and more steady? If i remove this capacitor completely? tuned indicator is solid!
Pira do mention that high cap loading can cause instability
When audio is applied,there doesnt seem to be a battle between audio and PLL and its sounds great!

Sinus, is there a circuit diagram for your board anywhere I can have a look at? Having to change audio coupling caps to maintain stability is fairly undesirable IMO. Perhaps a simple change can improve things, although you might need to add an opamp and second varicap.

[sinus trouble]

I dont have a schematic as yet but i will make one!
I have no doubt that a second independant varicap would be the best solution as you and Albert suggest, it does make sense!
Anyways il post a diagram in the next day or two and id be most grateful for any ideas you have

[Shedbuilt]

Does reducing to 0.47uF (like I think you did with the 1W design you shared), cure this ?
From a quick look at the component values (assuming they're otherwise the same as the 1W), I don't think the audio coupling / DC blocking capacitor should need to be any bigger than this, and would expect that you could probably go a bit lower without noticeable bass loss. I do agree that separate varicaps is probably a better solution.

[sinus trouble]

You are spot on Mr Shed
The 1Watt is identical and suffers from the same problem, using a 0.47uf significantly improves stability with no noticeable deterioration of audio quality?
Infact the bass still dominates with such a low value cap and a flat audio input, a bit more high frequency may not be such a bad thing?
Maybe I could also perform a sweep test to see if any frequencies are lacking?

[Shedbuilt]

One thing I can't see on your PCBs, is the pull up resistor for the output of the PLL PCA (R8 in the Pira schematic). This could well be problematic. You really, also need more of a loop filter. In an ideal world, this should be designed for the PLL output characteristics, and VCO characteristics, but I think "Berlin" proposed a loop filter design he'd used with SAA1057 before. If you combine the PLL with this kind of audio chain, into a single varicap, however, there will always be interaction between the loop filter and audio chain. If you go with Berlin's values, you will probably need to increase R8, and decrease C12 correspondingly - to avoid long settle times. Again, separate varicaps will avoid this direct interaction.

[Albert H]

An active loop filter (using an op-amp or two) is easier to configure than a passive one. It does necessitate the use of a few more components, but the results will be worth it!