Inverter-ized R209 Mk.2

[G6Tanuki]

The vibrator-transformer in my R209.2 having seemingly developed a shorted-turn [the thing was drawing something like 4Amps whereas it should be 1.4A, the transformer was getting worryingly-hot, the voltage-regulator wasn't strikinmg reliably and tbere ws a nasty rough warble when receiving CW/SSB stations] I decided to do a rework.

At this year's Newbury rally I bought one of the ubiquitous Chinese DC-DC converters. Yes I know these have a reputation for being a bit 'rough' but it couldn't possibly be worse than the imminent barbecue that was the old transformer/vibrator.

So, out went the vibrator, its transformer, and quite a bit of the complicated filter-network that it needed; this gave space for the new module. I fitted a length of aluminium angle to the PSU chassis, then cut a section of single-sided PCB material slightly larger than the inverter to serve as a backplane as close as possible to the underside of the module, and found that if I used 6BA 'thick' brass nuts they were the perfect thickness to separate the module from the PCB. Feedthrough capacitors on both the input and output, accompanied by small ferrite cored yellow suppressor chokes and some little red ceramic capacitors. The 'pigtail' leads for the input- and output-earth connections are the braid from some of the thin coax used to connect WiFi antennas to their routers, sleeved in heatshrink.

Fitted to the R209, HT set to 105V via the inverter's multi-turn pot. Voltage-regulator strikes and stays-on; current consumption a nice stable 1050mA. With the metal shield/cover fitted and the radio reinstalled in its diecast case, tuning around without an antenna fitted showed a few spot-frequencies on the lowest band where there was a bit of a 'buzz' detectable; this went way-down when I connected a suitable dummy-load resistor across the antenna/earth terminals, and is totally masked by antenna-noise when using only a few feet of wire.

SSB/CW reception is now possible on 14MHz too!

So I think it's a successful reworking. One thing I did notice is that - despite being insulated from the switching-transistor in the usual way - the heatsink [which is not connected to the PCB's earth] seemed 'hot' to RF; I suspect that this is where the vestiogial noise is being radiated from. Perhaps grounding it would be worth trying? Or, given that there's not the slightest sign of the transistor getting at all warm, the heatsink could be dispensed-with?

Conclusion: an inverter upgrade is perfectly feasible and doesn't need to cause significant RFI issues if you put a bit of thought into things.

[ex seismic]

I use a buck converter to drop 12v down to 6v for my Mk1, still on the original supply. I do not notice any ill effects.

[Lloyd 1985]

I recently got one of those modules to use with my R1155, and without any modifications it doesn't half kick out some hash!! I'll try what you have done and see how it behaves.

Regards,
Lloyd

[Bazz4CQJ]

Interesting! I bought two of those a while ago after JoeBog1 wrote about using them. I got as far as checking them out for things like current rating, but have not yet used them. Far down my list of 'on-going' projects is the Acorn valve GDO (copy of the "Megacycle Meter"), and I was thinking of using one of them with that, so making it portable.

B

[G6Tanuki]

Sure, they can be noisy - but my way of thinking was that if the designers of the R209 PSU could deal with the noise from a spark-transmitter switching an amp or so through an inductor, sitting a few inches away from the RF stages, I could do the same with said inverter-module.

Modern components [like the bolt-down feedthrough capacitors and ferrite-cored chokes I used] are far better at decoupling RFI than the old metal-cased paper capacitors and air-wound inductors the R209's designers had to use.

A bit of understanding of how the hash gets out, and a sprinkling of modern parts, and all can be tamed.

In times-past I used some really rather nice Erie 'feedthroughs' that were essentially a pi-filter. 1000pf 1KV rated capacitors at each end and a 100uH inductor, all mounted into a silver-plated tube with a 3/8-inch-UNF nut-and-bolt fixing to ground. Great for bypassing the leads between transmitterrs and the screened neter-enclosures. They would have been brilliant for suppressing modern inverter PSUs. Even in the 1980s they cost around £40 each! I dread to think what they'd cost today.

If I was to re-do my R209 inverter design, I'd consider also mounting a similar chunk of PCB material to that which I used as the base, but positioned over the upper side of the components, so the module was esentially sandwiched between two grounded plates.

That might cut the hash down even further.

And I'm wondering about the underlying wisdom of having one of the yellow wire-ended ferrite chokes sitting so close to the module's ferrite-transformer too - could this be causing unwanted coupling of energy in the transformer back into the input/output leads?

[2000 type]

Would it fit inside the original vibrator can? That way you would achive screening against radiated hash and preserve the look of the R209.

[cmjones01]

Originally Posted by G6Tanuki

One thing I did notice is that - despite being insulated from the switching-transistor in the usual way - the heatsink [which is not connected to the PCB's earth] seemed 'hot' to RF; I suspect that this is where the vestiogial noise is being radiated from. Perhaps grounding it would be worth trying? Or, given that there's not the slightest sign of the transistor getting at all warm, the heatsink could be dispensed-with?

If the converter isn't working hard, you could probably remove the heatsink or fit a smaller one. Earthing it may also make an improvement. If it's not earthed, there will be capacitive coupling in to it from the switching transistor's drain terminal which will be merrily waggling up and down with a sharp-edged waveform of the output voltage at several tens of kilohertz, probably with a whole bunch of ringing too. That would explains its RF hotness.

Chris

[G6Tanuki]

As an experiment I am going to try fitting a couple of little wire links between the inverter's heatsink and ground to see if it reduces the level of hash still further.

Like you say, since the inverter is only needed to produce a couple of Watts of output [and it's supposed to be able to produce more like 50W]I could probably remove the heatsink altogether.

[G6Tanuki]

Originally Posted by 2000 type

Would it fit inside the original vibrator can? That way you would achive screening against radiated hash and preserve the look of the R209.

Alas it's too big to fit inside the can.

Truth is, 'visuals' aren't ever going to be significant on a R209; the vibrator [or inverter] lives inside a steel shielding-case that's held in place with a bunch of screws, and the radio's diecast case itself is held together by about 20 nuts/bolts and a rubber gasket so it is airtight/waterproof.

If you can see the vibrator there's something missing!

[livewireless123]

Hello, I did something similar to that. I have a 62 set and the rotary converter
was worn out.I transplanted the inverter circuit from a Pye cambridge AM10 into
it and it has been working well ever since,that was about 1985.
I managed to shoe-horn all the bits into the convertors screened box and it is completely noise free with full HT level. It's nice to keep these old rigs going.

[G6Tanuki]

Yes there was actually an official inverter modification for the WS62, using a pair of TO3 Germanium transistors (OC28/OC35 types) , we had one so equipped when I was in the school CCF back in the early 1970s.