Minikits 1296 transverter circuit mods for Oscillator

[vk4apn]

FYI anyone that has built or uses the minikits transverter should look at the latest info. Just posted today.......................

http://www.minikits.com.au/eme65b.htm

A reasonable number of modifications, changes, additions are listed to improve the performance of the 576mhz osc section.

[VK2DAG]

The mod I did was to get some LOs from G4DDK http://www.btinternet.com/~jewell/2001info.html to replace both 23cm & 13cms. Should fix all the problems I have had....

[zl1ujg]

Hi,
Since I was mentioned in the notes. I did some testing a while back (mostly last year) and found some areas which looked as though they needed attention. Thats most of the info, compressed down. But ask a question if you are unsure of anything.
Kevin ZL1UJG

[VK2TDN]

The mod I did was to get some LOs from G4DDK http://www.btinternet.com/~jewell/2001info.html to replace both 23cm & 13cms. Should fix all the problems I have had....

hahahaha yup thats one way out of it

but for those of us that want to experiment ... I have discovered that those mods mentions of the '5EME site do work

over the last 24 hrs have done most of them on my EME65 LO and good success, its much more stable and easier to
get a clean signal out of 576MHZ

Thanks VK4APN for posting that

Also I'm going to possibly extend that LO to 639 MHz that I need for a 10GHz LO. Tried the 106.4990 MHz xtal out of the
non-working 10gig LO in the '5EME LO and it was producing a 639MHz signal ... just need to play with the wound
coil sizes and the final 2 striplines to get a better resonance at the required freq. but looks promising

cheers
Dave
VK2TDN

[VK2TDN]

Thanks for your efforts Kevin,

definately paid off.... it is one LO design that I would HIGHLY recommend access to a Specan for checking its
output. It takes very little change in tuning to go from a clean single peak output to a dirty one with many peaks
of either harmonics or spurii, and you cannot know that just using a freq counter and a power meter.

The mods I have currently done that have produced good results are the ...
1) ... changing of the 27pF to 10pF cap between the 2 transistors in the osc.
2) ... the back to back diodes (1N4148 (1N914) across the inductor
3) ... changing the 10 Ohm resistor to 100 Ohm that feed +V to the final transistor via the 1st stripline

I may yet change the 15pF interstage cap to a 6.8pF

I also notice, I have an earlir board design, dated 1998 compaerd to the 2005 one on the www mods page
my board has a trimmer cap (small blue one ) off one end of the xtal ... I cant tell what is on the board in
the photo looks like a diode ???
my board also has the large coil in the open (green former) rather than the metal cased one
also the photo shows an extra resistor to the right of that coil cant quite tell the value 560 Ohm ??

cheers
Dave
VK2TDN

[zl1ujg]

I also notice, I have an earlir board design, dated 1998 compaerd to the 2005 one on the www mods page
my board has a trimmer cap (small blue one ) off one end of the xtal ... I cant tell what is on the board in
the photo looks like a diode ???
my board also has the large coil in the open (green former) rather than the metal cased one
also the photo shows an extra resistor to the right of that coil cant quite tell the value 560 Ohm ??

Perhaps there has been a trimmer added to put it on frequency. I would check that the first coil resonates at the crystal frequency. Remove crystal and replace with 100pf and 50 ohms. Then resonate and listen on a fm receiver for oscillation. That first tuned circuit can be along way off and the circuit still oscillates. I also found another 1 nF on the cold side of the Tuning inductor, payed dividends.

What photo are you talking about, is that with the original circuit?

Kevin

[VK2TDN]

hey Kevin,

the photo's I am referring to are the ones on the VK5EME pages showing the mods we are discussing
http://www.minikits.com.au/eme65b.htm

so my older 1998 board has a trim cap off the xtal the 2005 board in the photo doesnt ... it looks
like a diode ?? and as mentioned above those photo's also show an extra resistor and a different
style inductor (canned)

dave

[zl1ujg]

the photo's I am referring to are the ones on the VK5EME pages showing the mods we are discussing
http://www.minikits.com.au/eme65b.htm

so my older 1998 board has a trim cap off the xtal the 2005 board in the photo doesnt ... it looks
like a diode ?? and as mentioned above those photo's also show an extra resistor and a different
style inductor (canned)

The resistor is the 560 ohm across the coil, while the next one is the 18 ohm, and then the bias resistors for the following multiplier.
I dont think there are any diodes in the later EME65B board,however there is a wire link just adjacent to the crystal. If you add a inductor across the crystal then the wire link will nedd replacement with a 220 to 1 nF capacitor as DC isolation.

[VK3TPR]

Interesting..... When I put back to back diodes across the coil I can no longer adjust the osc to the desired frequency, the slug is out the top of the coil and I'm still off frequency !

Peter

[zl1ujg]

Hi Peter,I was in Glen Waverley a few months ago. There could be two reasons why the crystal oscillator is off frequency. It is quite probable that the oscillator is not tuning to the correct frequency. If you replace the crystal with a dummy crystal 47 ohms and a 1 nF capacitor. The circuit should resonate around the crystal frequency. This could be picked up with a FM receiver. The addition of the diodes adds about 4 pf extra. I would try around 10 pF (8.2 to 18pF) across the tuned circuit to try resonance first.
The crystal oscillator will oscillate, even with the LC circuit many 10's of MHz off, due to lots of gain around the circuit.
Even if the crystal oscillator is optimally tuned, you may experience being off frequency. The reason for that, is when crystals are manufactured, they typically may be +/- 10ppm (or more) off frequency (calibration tolerance). This is effectively +/-1 kHz or more 100 MHz. One may tune them a little with series capacitance to raise the frequency, and a little inductance to drop the frequency.
It is important to add a resonating inductor across the crystal (ensure DC blocking cap is fitted). The inductor cancels the capacitance across the crystal plates and can improve stability.The tuning of the oscillator will become more even both HF and LF of the marked frequency, due to cancellation of the adjacent parallel resonance (of the crystal), by the addition of the inductor. Try 390 nH or some adjacent value (330 or 470 nH)
Kevin
ZL1UJG

[VK3TPR]

Thanks Kevin,

I have settled for the external voltage regulator mod only for the moment and the LO now seems quite stable with very little drift.

I have two of these oscillators for two transverters, one 23cm, one 13cm.

I will perhaps do the other mods later but at the moment enjoying making a few contacts on 13cm.

Peter

[VK4ADC]

Hi guys

I know I just put these same comments in the EME65 LO topic but just found this one too, so posting it here as well.

Just making a start at some 1296 gear and bought an EME65B LO kit plus a EME72B transverter kit from Mark VK5EME at Minikits during this last week.

I have encountered a few "funnies" along the way so have created a new web page on my site : http://www.vk4adc.com/23cm-EME72-tvtr.php - to record the progress.
Anyone with (or buying) these kits might be interested in a quick look as they may encounter the same issues.

I will shortly be starting construction of a conventional yagi style (rather than loop-yagi) antennas so will probably document that on a new page on the site as it progresses too.

73 Doug VK4ADC

[zl1ujg]

Hi Doug, and others.

Yes I have come across the extra capacitors needed. I didnt put an extra one on the base, but made comment that a chip is probably needed there. I think the second one on the collector tuned circuit cold side, is probably needed either because of circuit radiation, as some circuit components are are a distance away from the original decoupling cap, or that the original component is not that effective.
The BFR92A's have lots of gain into the microwave region so care is needed.
I made comment previously that the LO helical filter can on the transverter should be soldered to ground on the top earthplane. The holes for the filters are not plated thru, and the LO filter can connections has no vias thru to the groundplane so this is a must. The other filters have vias, but soldering them should help as well.

Kevin
ZL1UJG

[VK2TDN]

Hi Doug,

interesting reading on your page
and further to Kevin's comments about soldering the LO helical can on the top side. I have also made it
practice to do the same with all grounded component leads that go through the cct brd. You wanna try that on
the LO board as well for all the caps and resistors etc. From your photo's you havent done that yet
I am a fervent believer in having top and bottom groundplanes connected where-ever (and as much as ) possible.
It will ALWAYS improve stability of ccts operating in the UHF and uWave bands.

cheers
Dave
VK2TDN

[VK4ADC]

Dave, Kevin

I have now soldered the L1 coil can to the top of the EME65B PCB plus added another 5 x 1nF SMD bypass capacitors to the top of the PCB. With the plated-through vias, I wouldn't have thought there was much to be gained by soldering the other "grounded" component leads both sides. Updated details plus a photo of capacitor positions on the page at http://www.vk4adc.com/23cm-EME72-tvtr.php.

The LO seemed unstable again when plugged into the LO port of the main EME72B PCB but have solved that initially by terminating that with a 47 ohm resistor. I may increase that value and see what effect it has on LO / ERA3 buffer/multiplier stability in comparison to LO injection level. It can't be the shielding between the ERA3 & the MGF1302 / instability noted on the Minikits EME72B page notes and requiring the RF absorb material because the receive RF preamp is currently bypassed awaiting another device to arrive.

The extra-bypassed LO now seem seems unconditionally free from spurious regardless of where the tuning adjustments are.

The transmit side of the EME72B is currently producing about 25mW / +14dBm with 0dBm drive at the 145.1 I.F. with the double termination of the LO, thus dropping injection level.

73 Doug VK4ADC