The DXR-700 Conversion / Transverter Project Update  

PLL Update ( More Simple mod to get on 5616 or thereabouts...).

The worst bit of the original mod was to move the diodes and add loops.
After greater understanding of the interdigital filter oscillator, I decided to have another go at lowering the PLL locking frequency.

With simple mods, I have managed to get it to lock with the advantage of lower phase noise and stability.

You will need some Copper shim strap (thin), two  0p5  0805 SMD caps, small piece of overhead projector acetate sheet, nail polish, side cutters and some patience.

1. Join the outer thin tracks so that the whole length of the line is increased to the max length. Do it for both sides.
   
   
   
2. Cut a small piece of copper shim strap to fit the thicker track in the middle. This piece should be able to cover the gaps to extend the overall length o fthe track.
3. Apply a small dab of nail polish to glue down a small piece of acetate that will be slightly larger than the copper shim overhang, this is mainly for insulation purposes.
4. Solder the foil down to just cover the slits in the thick track and extend up to the 2nd hole set in the ground plane pattern.
5. Now add 0.5pF  0805 SMD caps to the other end of the filter section. There will be pads on the opposite side to the VCO transistor to easily fit the cap there. Solder the hot side of the cap first to anchor it and then the cold side as the ground plane soaks up heat and is harder to solder.
   On the VCO transistor side, add a 0.5pF across the cap that feeds the base of the transistor.

With the PLL set to run at 5616MHz, it should now lock with about 700 odd millivolts on the VCO varicap control voltage test point.

It should also run at 5760 as well with the voltage now up to about 1.2 volts or so.

Setting up the transmitter section can be done by setting PLL to 5760 and peaking up the output section. Peaking is somewhat a custom affair but some guidelines can be seen in the photos.
Silver paint can be used to do this. (Beware that the paint has to dry before it becomes conductive). Tuning can be done by scraping off small bits at a time with a tooth pick to tune it.
In some other places, there are small bits of wire or copper strip to tune as well.

Hot spots that can be tuned can be found by using a small piece of TC wire poked in the end of the wooden cotton bud stick or similar. The TC wire should be small, about 5mm long approximately. Touch filter stubs to see what tunes to get a peak. Some will dip as they are de tuned.

Using One Local Oscillator.

I have decided to have a go at using only 1 PLL to drive both sections of the transverter, this involves installing two SMA sockets - one in both the TX unit and the RX unit.
The TX PLL will be used as the common LO source.

Using this simple procedure gives several advantages:-

Use only 2 micros.

Use 2 voltage regs instead of 4

Don't have any PLL phase beating issues as can happen with two PLLS.

Phase noise / low level modulation of VCO less prevalent.

Needs only 1 control board.

The inside of the cavity. Note how simple it is now. Power saving due to switching off lots of bits must be a bonus!! Note the LO Feed joiner that carries about 1.8GHz from PLL to pre tripler sections.

The PLL output is picked off at a  well buffered point. The best point is after the ERA device that amplifies the signal taken off at the pick up loop that feeds back to the PLL prescaler unit.

Drill a small pilot hole in the middle of the large  ground plane area that is near the VCO output and the VCO buffer MMIC. Do this for both units. Using an SMA socket that has a long centre pin such that is used when a socket may pass through a thick side wall of a chassis.

The TX unit socket in my case, protrudes enough to solder to. The RX unit may need the casting milled down enough to allow the pin to show through the other side.

TX side.
Couple the socket to the pickup loop MMIC output via an 22pF smd cap via a small coax to the sma socket. I get approx -5dBm at this point.

RX Side.
Remove the coupling cap from the VCO that feeds the VCO buffer. Tip it on the side so one end is upward to allow the connection of a small piece of kynar wire to the SMA socket.

Disable the TCXO and PLL supplies by cutting the appropriate tracks as shown in the photos.

Peaking For Max Output

Peak up the filters and tripler chains for both the TX and RX units to get approximately -7dBm at the LO pickup point. There should be a 10dB or greater improvement in output of the LO chain at 5616MHz. I think that it was at about -27dBm in an un altered state. In reality, the mixer most likely gets about +7dBm directly.

Peaking up generally needs a spectrum analyser to start and then move to a power meter that can read low level signals. It's easier to see small increases on a meter than peering at a graticule scale.  I used a Marconi  6924 meter to do fine tuning adjustments. review spectral output when done.

	TX Chain - first device after filter seems to have the habit of oscillating so loading stub seems to fix this.
	Post Mixer TX filter
	LO Filter on RX
	Local Oscillator filter on TX

 

 

HOME