G1HBE

Better than Nothing.

 

Jan 2008: Having recently met an old friend and swapped some gear, I've found myself in posession of an elderly dual-band multimode rig, the Kenwood TS780. It's faulty, but if I can make it work I've got a plan to convert the 2 metre section to 4 metres. I'll use the 70cm half as is.
The 780 works on 2m as follows: incoming sigs at 144MHz are mixed with the rig's 114 - 116MHz LO to produce an IF of 30MHz. After this, it's simple fixed conversions down to 10.695MHz and 455KHz. The TX process is the same in reverse.
The accepted way of transverting from one band to another is to add an extra mixer and add or subtract a fixed amount from the incoming and outgoing signals. Going from 144MHz to 70MHz can be tricky, as the 74MHz signal is quite close to the required signal and its second harmonic is not far from 144MHz. As these signals are added into the signal path, they can cause all manner of woes.
So here is my cunning plan:
Generate a 74MHz signal with a crystal oscillator as before, but mix it with the 114MHz LO to produce a tunable 40MHz LO signal. Both these signals should be well filtered to ensure no harmonic mixing can take place. Now introduce this signal to the rig's own RX and TX mixers where it will be mixed with the IF of 30MHz to make 70MHz.
Of course the rig's front end and TX strip will need to be rebuilt or replaced with 70MHz ones, but that's half the fun. I'll keep this page up to date so you can all chortle at my weedy efforts.

 

 

 

Here's the old lady on the bench, ready for fettling. The reported fault was 'no receive or very deaf' and 'transmits for short period then stops'. On test, it  received very well on both bands, but the TX power level varied a lot . I spotted a patch of burnt board at the 70cm RF module's output pin. After scraping away the carbonised board, the 70cm TX seems OK. I also replaced a stripped nut on the heatsink, which was causing heat build-up and power drop.

Just look at those innards, eh? To me, this generation of sets (1980's) will always be 'proper' radios. Well-built, understandable and fixable.

 

 

 

 

 

 

 

 

 

 

 


Early Feb 2008: Here's the 74MHz crystal osc and LPF in a tin box, and the stuff outside the box is the mixer (my friend the SA612) and a 40MHz BPF.   The rig's 114 - 116MHz local osc goes in one end, 74MHz goes in from the box, and 40MHz tunable comes out via the filter. And it's clean down to about -60dBc.

 

 

 

 

 

 

 

 


 

 

4 Feb 2008:  RX front end. Out came the 2m helicals and in went a pair of homebrew 2-section bandpass pairs.  I also replaced the local oscillator 'cleanup' tuned circuit which you can just see near the left-hand end of the board. The original was tuned to ~115MHz, whereas the new LO frequency is ~40MHz.

It's reasonably sensitive, hitting smooth noise at around -114dBm.

 

 

 

 

 

 

 

 5 Feb 2008: Here's the TX mixer. This takes the ~40MHz from the LO mixer (above) and the 30MHz IF signal from the rig and produces 70MHz for transmit.  The mixer is all nice and balanced, so most of the unwanted signals are well suppressed, leaving just the wanted signal (70MHz) and the unwanted (10MHz). The 70MHz one is amplified up to 60mW and according to my analyzer it's clean to beyond -60dB.

 

 

 

 


 

 

Progress report: The 74MHz crystal is now in place in the LO mixer, and the 780 actually receives on four metres using its own local osc, no signal generators or any artificial aids. Now it's all that time-consuming detail like: antenna change-over relay, rebuilding the 70cms PA, tracing switched supplies and so on.

Another update. A busy night saw me finding both original faults. The varying TX power was caused by a temperature-sensitive mosfet in the 30MHz TX IF amplifier. In went a good old 40673 and the prob went away. The deafness on RX (which I'd not heard) suddenly made itself felt when I was diddling around the RX IF input. It turned out to be a cracked copper track near the IF input socket.

 

 

 

 


I decided to go for a Mitsubishi RF brick instead of the completely home-brewed PA which had been causing a certain amount of heart-ache.  It's one of their new mosfet ones, which is capable of 30 watts output. However in this rig it will be turned down to 10 watts due to power supply limitations.

Here it is on test. The block itself is screwed tightly down onto the heatsink, and you can see the decoupling components, bias supply (just a pot!) and the LPF, which is near the relay at the bottom.

 

 

 

 

 

 

 

 

 

 

 

 

 

 


4m output spectrum from 60 to 80MHz at 10 watts. The worst of the sproggies is at 4.8MHz LF at 64dB down, which is fairly decent.

Power levels on 4m are 10w/3w and on 70cm are 10w/2w, which will do for me.

 

 

 

 

 

 

 

 

 

 

 Finally a shot from 10MHz to 250MHz to show the harmonic output. The second harmonic is -57dBc and there's no sign of the third. That'll do. I'll run the rig for a while with the covers off to see if anything untoward turns up, but other than that it's done! I've had several QSO's on the rig now, and no-one has mentioned anything I need to worry about.

 

 

 

 

 

 

 

Oh-oh....

 

Feb 2009: After a period of faithful service, the 780 is showing signs of trouble. The 4m part is fine, but a few days ago the 70cm bit started acting up. I first noticed this while I was operating on 70cm FM while running my scanning receiver on UHF. With the PTT pressed, the scanner would start producing odd noises and as this had not happened before I stoked up the spectrum analyzer to see if anything funny was coming out of the 780.

Oh dear. spikes everywhere. Sometimes just for a second, sometimes for several seconds. On SSB, there was 'spitching' and nastiness to be heard up and down the nearby spectrum. I mentioned this to Ross G6GVI, who has had a similar problem with his old 780. He cured his by fitting extra earthing braids around the PA block and slightly reducing the supply voltage, but mine didn't respond to either of these.

I'm beginning to think that the PA block is slowly passing away. Not surprising really, as when I got this rig I noticed that one of the block mounting bushes was stripped and this had been causing overheating. There were signs of charring of the PA PCB around the output PIN diode too.  Turning the power down to below 5 Watts stops the sproggies, so that's what I've done for now.

Where to go from here? Well, 5 watts is only 3 dB lower than the full rated 10 watts, so it probably doesn't matter much. But it would be nice to have the rig working properly and to banish the worry over splashing sproggies all over the place. A new Mitsubishi mosfet module will knock me back about £40 I think, then I'd have to gut the old PA compartment and make a new PCB to fit. I can feel a lie-down coming on...

Well it won't do itself, so I've made a start. I've ordered a mosfet PA block from GH Engineering and manufactured a PCB for it to live on. The pictures tell the story, really. 

 

 

 

The two coaxes are for testing the LPF, the one on the left coming from the signal generator and the one on the right going out to the speccy analyzer.  The 'scope probe was looking at the output from the detector.

 

 

 

 

 

 

 

 

 

 

 

 

 23 Feb: The PA arrived today. My careful measurements paid off, as the leads lined up with my tracks! After attaching a dummy load and a feed for the speccy analyzer, with no RF drive  I slowly twizzed up the bias pot for 150mA. Then,  watching the ammeter and the analyzer, I turned up the drive from the sig gen. Ten watts out was achieved with just 20mW input, and the block was drawing about 2.3 amps.

 

 

 

 

 

 

 

 

All back in place with the lid on! The two pots on the left are for setting the high and low power levels and the board on top is the receive pre-amp. This rig cost me nothing (thanks Darren!) and has cost me about £90 for two PA rebuilds. It works a treat and it can be fixed by ordinary blokes like me without fuss.

Another job done cheap and very enjoyable it was too!

 

 

 

 

 

 

 


 

 

More faults!

 

June 2013. I think the old lady is trying to tell me something now. A few months ago, one of the local ocillators developed a habit of wobbling and jumping in frequency, and as the fault was on the big IF board I didn't fancy getting involved. Luckily, I discovered that turning the RIT ON stopped it, so I've been using it like that for a while. Now another fault has come along - no TX output in USB mode. Sometimes it works, sometimes it doesn't. After cleaning the mode switch and wiggling connectors to no avail, I've decided it's time to whip out all the add-ons I've put in and give the old girl a dignified burial.

So I'm looking around for a newer old banger to do up!