Thursday 24 May 2018

Anglian 3L 2m Transverter from G4DDK - Part 3

Finally ready to try driving the Anglian from the rig I needed to connect it up.   Originally I was going to use the Kuhne transverter interface for the IC7300 but after some consideration decided to actually use a combiner along with a 20dB 30W attenuator though there is obviously a loss incurred in the RX path switching between the Anglian and Iceni will however be easier.

I bought the attenuator from eBay from for the bargain price of £25 these things are now cheap as chips from China and when I tested the unit with a VNA gave 20dB of attenuation at 50ohms from 1MHz to 600MHz which is as far as I bothered but the device will apparently work up to 3GHz but I couldn't vouch for that.

The combiner is a Merrimac 2 way 0 deg inductive unit from rf-microwave.com which cost just over £40 pounds which is cheaper than the mini-circuits equivalent unit and more than up to the job.

Merrimac combiner with 20dB 30W attenuator
With my Yaesu FT991 set to give 40dBm of drive which  will provide 20dBm to the Anglian once passed through the attenuator I measured and output as 39dBm so some additional attenuation is required between the Anglian and the PA.

With an extra 16dB of attenuation placed between the Anglian and the PA I measured just under 37dBm out which is good enough to drive my little Microset SR100 amplifier.   

I couldn't measure the 2nd and 3rd harmonic product with my equipment but I did a quick check with my SDR and I couldn't see anything on the SDR# waterfall.

Anglian 3L 2m Transverter from G4DDK - Part 2


Well blow me I went to align the TX chain on the Anglian and low and behold no output, this turned out to be IC7 being grounded once this was resolved the TX chain worked as expected giving out 20dBm for 0dBm of drive.

I've had all manner of issues with this Anglian 99% caused by the builder.   I've come to the conclusion that solder paste though easy to use can cause issues if a little too much is used.   I never had this issue with the Iceni however did spend more time cleaning the board and that seems to have paid dividends with the Iceni, so mental note to self don't skimp on making sure any solder paste residue is removed.

I also had a replacement QH40A heater from Kuhne arrive which I duly fitted and low and behold the exact same issue in that as the heater heats up the frequency slowly drifts up to the correct frequency and then suddenly drop by around 500Hz and then slow drifts back up again before repeating.   I may ask Sam if he has another crystal I could buy just to experiment with however as I have a GPSDO there is no real need, even without the heater after a 10 minute warm up period the unit only drifts by a few cycles.

Sam also does an 8W Class AB amplifier which has good linearity when limited to 5W PEP for the Anglian which is simple to build.  The unit draws around 1.5A with a supply voltage of 12.8V.

Sadly I omitted to take any photographs of the PA but you can almost see the unit in the pictures of the completed transverter below.

The last thing to do is plug it all in and test the output from the PA doesn't exceed 5W PEP.

Front Panel
Rear Panel
The inside.





Thursday 17 May 2018

Anglian 3L 2m Transverter from G4DDK - Part 1

With the success of building the Iceni transverter from G4DDK I decided I'd also purchase and build Sam's Anglian 3L 2m transverter as well which I'd heard glowing reviews for.

The Anglian 3L is of a similar design as the Iceni using much of the same circuitry and form factor.

Here's a list of specifications from Sam's website.

Parameter Performance
RX noise figure 1.6-1.8dB
RX gain 25-26dB
RX IIP3 Better than 0dBm
TX gain 20dB (35dB with optional TX IF stage)
28MHz Drive required for +20dBm 144MHz output 0dBm (-15dBm with IF amplifier)
Power output (Saturated/P1dB) +22/+20dBm
Harmonic output (2nd/3rd/higher) at saturation/td> -40dBc/-50dBc/<-60dBc
Image frequency output suppression <70dBc
LO suppression <70dBc
Transmit sideband noise Better than -140dBc/Hz at 20kHz offset

The kit went together fairly easily but you must double check both the parts list and circuit diagram as well as the assembly documents to get it right.

The Anglian 3L uses the same PCB as the original Anglian 3 with either some components changed, added or even removed so you can't just follow the silkscreen print on the board.

I found a few omissions and errors with the documents but Sam is always willing to aid assistance and be told of mistakes so he can fix them and in fact has a searchable PDF if you ask him.

The completed module can be seen below, sadly I omitted to take any photographs during the build.

Anglian 3L after much fault finding.

On first power up I found there was nearly 750mA of current draw instead on the 250mA the build document said there would be.  I removed both TR4 and TR5 from the unit so that both the receive and transmit chains would be isolated from the rest of the circuit and plugged the unit back in and measured a current draw of 70mA with a 12V supply which is in agreement with what Sam said the unit would draw.

Next I replaced TR4 and again measured the current draw which has now returned to the 750mA I originally measured which indicated the error was in the receive chain so I removed TR4 again and then replaced TR5 and measured the current draw again which was now 80mA proving the error was in fact in the receive chain, I then refitted TR4 again.

Knowing the problem was in the receive chain I slowly isolated each part of the circuit until I ended up with only the PSA4-5043 MMIC IC3 left in the circuit, sadly removing it without a reflow hot air gun is a pain in the backside and in fact I ended up destroying it so it was back to Sam to ask for a replacement which he kindly sent out.   I couldn't see any shorted tracks under the microscope and I can only assume that I had a short on the underneath of the IC3 which I couldn't see, anyway with the replacement fitted the unit now had a current draw of 260mA hooray!

With the unit now seeming to be working the next thing to do was to set the oscillator to 116MHz with an output of around 20dBm, well blow me there was no output from the oscillator a quick check with the voltmeter showed there was no voltage on the collector of TR2.  After a bit of head scratching I realised R13 was missing from the board,  the silkscreen was marked L8 and not R13 so it got omitted during the initial construction, I said double check the circuit against the silkscreen as there is a few of these changes that are easy to miss if you're not careful.

I could not get the oscillator on frequency it was also over 200Hz off and had a very sharp adjustment so I ended up removing L27 to give a little leeway in the adjustment and managed to get the oscillator bang on frequency.

On a side note I had a Kuhne QH40A heater attached to the crystal.  The heater keeps a constant temperature of 40C, though the crystal is cut for a temperature inflexion point of 25C, now this worked great with the Iceni but caused horrendous drift on the Anglian and the only reason I can think of the heater was actually heating above 40C so I've removed it for the time being.

Sam recommends the QH40A with both the Iceni and Anglian and it worked really well with the Iceni but if you suffer frequency drift with the QH40A fitted don't assume it's the Anglian drifting but irregularities with the heater instead.  I did have a word with Kuhne regards reducing the voltage the heater to see if that would lower the temperature but they say it won't make a difference so I need to order another.

Next thing to do is align the receiver this involves adjusting L14, 15, 16 and 20 using the coilcraft trimming tool included with the kit.   Now the build document said to set a receiver to 29MHz and connect the Anglian then try and equalise the noise across 28 - 30 MHz by adjust the L14 - 16.   I found to my horror that I appeared to have no receive output from the receiver, bangs head on wall, bother!

To workout where the fault  I injected a 144MHz signal into the Anglian's RF_RX port and with an oscilloscope 1x probe connected an RTL2823U dongle and SDR# set to 144MHz traced the signal through to the mixer I then switched the SDR# frequency to 29MHz and traced though to the LPF where I discovered yet another short this time on C24 I removed and refitted it which resolved the problem.

Even though the receive chain was now fully operational I still couldn't hear an difference in noise level when adjusting the inductors and in the end came to the conclusion that either my hearing is a little dodgy, the fact I couldn't turn the AGC off on the Yaesu FT991 receiver I have or even both.

To get over this problem I again connected a signal generator to the RF_RX port injected a 144MHz into the Anglian I then adjusted the coils for maximum signal which I then repeated at 143 and 145MHz until I had a equalised the read signals across the band.   The assembly document said the inductors would probably be a few turns below the top of the can but you can see in the picture below mine didn't end up like that at all.

The three inductors in line are for the RX alignment.
I maybe able to redo the alignment once I have the Icom 7300 out of the box but until then I'll leave the adjustment as is.

Just to be certain everything up to this point appeared to be working I connected an antenna to the Anglian and tuned in the GB3VHF beacon on 144.4285MHz and ran WSJT up to make sure the continuous tone the beacon transmits between the JT65B and Morse sequences fell on the 1.5kHz mark on water fall and I was pleasantly surprised to see it was with a few cycles of where it should have been, though bear in mind I did use a GPSDO due there being no heater currently fitted so it should have been pretty much spot on.

The next thing to do is to the align the transmit chain and write part two. ;-)

Tuesday 15 May 2018

Cheap Chinese 70cm 100W Linear Amplifier - Part 3

I've had quite a bit on lately finishing the Iceni and Anglian transverters by G4DDK but I managed to find some time to do a bit more to the 70cm PA.

Using the same setup as the last time I tested the harmonic output I got the following results.

2nd Harmonic was at -38dBc originally measured at -40dBc

and the

3rd Harmonic was at -68dBc originally measured at -53dBc

So not the big improvement I was hoping for and in fact 2dB worse for the second harmonic.

I tried running the unit with a lower drive giving an output of 50W and in fact the 2nd harmonic got worse as well as at various other lower drive levels.

I've yet to decide how to rectify that small problem I could use a small mini-kits LPF from Australia or the cheap way a 1/4 wave stub notch filter cut for the 2nd harmonic.

The other thing I've discovered is to not pick the PA up after running a CW test as the heatsink gets mighty hot. ;-)

Monday 7 May 2018

M0CVO HFC-3015V Vertical Short Antenna for 40m - 10m Update

I finally got the HFC-3015V installed today, apparently the hottest day this year so far,  I tried installing the antenna without placing any wood within the PVC pipe where it bolts to the mast and found it to be impossible to fix securely without it crushing the tube so it definitely requires the extra wood inside.

The installation instructions say to fit with a counterpoise if going to used on 17, 20 and 40m bands however doesn't give any recommendations on length or amount of counterpoises required.   I ended up fitting two, approximately 4 metres long.

The antenna is fitted with an LDG RT-100 ATU on a short 0.5m RG58 patch lead and feed back to the shack with some Ecoflex 10 I had spare.

I've not had a good chance to try the antenna out yet but checked to see how it would match, on 20m it gave an VSWR of 1.6:1 which wasn't too awful however on 40m I got a VSWR of 1:1 which was quite impressive now this doesn't of course mean it's going to radiate well but least it's matching on the two bands I use the most.  

More to come once I've used the beast for awhile.