Recently, I got a Marconi Instruments Model 6651 RF Plug-in (26.5-40 GHz), for the 6600 sweep generator. These were basically the 1st generation sweep generators extending into the millimeter wave region.

A number of plug-ins were available, mostly based on Watkins-Johnson BWO (backward wave oscillators), which are a particular type of electron tube.

These sweepers were first introduced in the late 1960s, early 1970s. New plug-ins became available as development of BWOs and similar sources advanced.

I have no price list, but surely these items came at a hefty price tag in these early days, even not quite cheap today.

My unit had a pretty “new” 1990 date-code BWO, I expect, it had been replaced by its former university owner at least once, also the cables and connectors of the BWO looks newer compared to the rest of the plug-in. Because of their nature, the BWOs only have a limited life-time, a few 1000 hours at best.

There are numerous warning labels, because to focus the electron beam, a strong magnetic field is needed.

The output is a gold-plated waveguide flange, for Ku-band waveguides.

The general layout of a BWO: there is an electron gun, producing, in our case, a hollow electron beam, with the necessary heater, kathode, anode. Then, the helix, and the collector (which is basically at helix potential or a little bit higher, say, 100 volts higher. The helix voltage is quite high, for exmaple, 1~1.5 kV.

While there is no information from Watkins-Johnson available online for any of these devices, a lot can be learned from related Hewlett-Packard microwave sweeper plug-ins (8697A), where (arguable) the same or very similar BWO has been used (some HP units may use Varian BWOs).

Unfortunately, after quite some effort with high voltages and pretty dangerous tests, it became clear that the tube barely gave some power. Maybe, because it had not been operated for a long time (some people and datasheets say that the BWO should be run at last once every 6 months, or similar, to preserve its function…).

But anyway, I have other, more stable and easy to use sources for this range, so I didn’t want to scrap the BWO without having a closer look at this. When do you normally get a chance to check out the secret workings of such marvelous devices!

First things first, hidden inside a heavy magnet, silicon rubber, and a waveguide coupler: the tube. A very mysterious piece of engineering, only a few companies ever mastered to produce these.

Also by the numbers and hand-writing, pretty clear that these units were all hand-made and assembled by engineers and highly skilled people.

The mount is all gold plated, a tapered wave guide eventually bent by 90°, so the microwave radian can be coupled from a small gap in the silver coating of the tube, to the outlet. This tube was meant to generate about 10 dBm.

The helix at the collector end.

You can see, the helix is kept centered by 3 glass (quarz?) rods, these rods are just held in place by the tension of the helix, but normally won’t move during operation and shipment. Surely, will all the glass-metal interfaces, mounts and connection points, it is no good idea to drop these units – some interfaces may break, or mis-align beyond repair. Furthermore, several Watts of powder will need to be dissipated, calling for a reasonably elaborate design of the mount not only to conduct the microwave radiation from the tube to the output, but also to dissipate the heat from the collector and helix.

All the high voltage socket part is embedded in two layers of silicon rubber. First they put some denser rubber around the socket to fix everything in place, probably run some tests, and then sealed it all by filling the further space with some pretty soft, highly filled silicon rubber. What they didn’t consider – the cable also has silicon isolation, and deteriorates in contact with other silicon materials. This may be one of the reasons for the unstable operation, because some wires were pretty close and the isolation seems very brittle.

Furthermore, I did some measurements of the critical parts, just in case I would even want to fabricate own BWOs, at least I have some working dimensions to start with.

The helix is normally made from flat molybdenum strip. At the bottom (gun end) it is mounted to a small ring. Scale is 210 px per 1 mm, in case you want to take dimensions…

A typical power supply – I am thinking about building a BWO power supply for some other old BWOs I have around (for other frequency regions), but it is fairly involved – the old design used high voltage 2 kV transformer with further high-voltage stable windings to provide the collector offset voltage, several expensive high voltage capacitors, and high voltage regulation vacuum tubes. I think it will be a good challenge for winter to build a 2 kV regulated and low-ripple power supply, solid state, probably with some series-connected MOSFETs.

Any of such future supplies will need a pretty special plug: heater voltage, anode and cathode voltages, helix voltage, collector voltage. At least I can go without any analog linearization voltage, if I manage to make the helix supply digitally controllable.

Inside the 6651 plug-in the BWO is connected to the board with some screw terminals. Seems to work well even at the high voltages. Surely there is no touch protection for these devices, just a few warnings not to get too close!