top of page

Finished: -

ME R001-1

1.0-2.0 GHz travelling wave tube amplifier

After the big success of the HP 8614A project, and having had a fun hands-on introduction to RF electronics, it is time to explore this field a bit further. To use the signal generator, it might come in handy to have a decent amplifier to do some interesting experiments with things like bidirectional couplers, filters, cavities, waveguides, etc. Sadly, these amplifiers are very rare and incredibly expensive. This means it might be a lucerative option to build one from scratch. After Googling for a while for different amplification techniques at these frequencies, there was one that got my gears turning: the Travelling Wave Tube, or TWT for short. I found an eBay listing in America of an old TWT within a useful frequency range, so I decided to impulse buy it and get it delivered. 

The word "expensive" is a relative term. Whilst this TWT is relatively cheap compared to the other semiconductor based options (which went into the thousands of euros), it is always frustrating to see import taxes and shipping be almost as expensive as the object itself. This is one of those things you have to accept living in a country where it is strictly illegal to hand down old lab equipment from a company to their employees. Rather than giving the stuff away, it gets locked up in garbage bins, to be sent off to a landfill and spoil the environment...  

But I digress. I think about a 100 euros in total is a reasonable price to pay for such an integral part of the project. Whilst the tube was on it's way to cross the Atlantic, let's first get a bit more acquainted with a TWT.

What's a TWT?

Like mentioned earlier, this is a device that is capable of amplifying RF signals. But this is a very broad way of putting it so let's nuance it a little more. To give some general figures, the average TWT is lucerative to operate at frequencies between 0.3 - 100 GHz, and has capabilities to amplify a weak RF signal with as much as 40-70 dB

I'm not a talented author, and neither am I well acquainted (yet) with the TWT. Therefore I think is better to really read the description on THIS website. If you're not interested in the beautifully written and illustrative in-depth explanation of the talented author of that website, then you're welcome to stay with me and read my summary. 


Firstly, a bunch of electrons are boiled off a heated cathode through thermionic emission. By generating a big potential difference between the cathode and anode, these electrons get accelerated to a similar velocity as the RF signal propagates through the helix. The beam of electrons is kept centered through the tube with the help of permanent magnets. A signal is fed into a waveguide, coupling with the helix. The induced RF signal in the helix bunches the electrons of the beam in small clusters. Then through some very cool black magic and whatever other brilliant 1900s engineering (really, just read the description from Radartutorial...), the amplified signal can be picked up with a waveguide from the other end of the helix. The electrons then strike the collector and can be taken out of the system. 

What's needed to drive our TWT?

Now we roughly know what to expect from a TWT, let's try to figure out some numbers so we can design a proper power supply to drive it. The seller did not seem to have a lot of information about this device, and the description was very... descriptive?

Aha, it's a Watkins & Johnson Travelling Wave Tube of the type "WJ 3500". And it operates in the 1.0 - 2.0 GHz range... And I get my money back if it doesn't work. Seems fair to me!

Trying to Google for a datasheet of this particular TWT ended up in a lot of frustration and sadness. I think I've gotten really spoiled by the HP/Agilent community by their amazing efforts to keep all documentation alive and easily accessible. For Watkins & Johnson, however, the fanbase seems to be about non-existent and extinct. Thankfully I found three different sources that will greatly help to figure out what the power requirements of TWT are, and what kind of performance I can expect from it. Namely:

1. Actually flipping through the seller's pictures of the listing

Evolution gave us eyes, but what use do they have if we don't actually open them once in a while. This list of voltages really helps to compare this TWT with other datasheets I did manage to find from W&J. To clarify what I initially thought what the sticker meant is as follows:

  • Cathode (K)                                           0 V               (Yellow) 

  • Filament (Ef)                                          6.3 V            (Brown)

  • Anode (Ew or Eb????)                          844 V           (Orange/Red)

  • Helix (Ew or Eb????)                             844 V           (Orange/Red)

  • Collector 1 (theres two Ec's?)             -9.3 V           (Green)

  • Collector 2 (this one makes sense)     592 V           (Blue)

As we will see in head numero 2. my expectations were silly. Though, would it really be that much to ask to actually come up with logical abbreviations??

2. HP 489A Service manual

By sheer coincidence, whilst reading up on TWTs, I found a video of a teardown of a similar HP device in a different frequency range. The TWT in that one looked an awful lot like the one I ordered and after doing a ton of digging, I found out the 489A actually used the same tube! How nice, we get to see something HP in this project too! Let's have a quick look at the schematic of the HV powersupply:

Now these engineers did have access to the original datasheet of our TWT. This means some things can be cleared up:

  • Cathode (K)                                           0 V               (Yellow) 

  • Filament (Ef)                                          6.3 V            (Brown)

  • Collector                                               844 V           (Red)

  • Helix                                                      844 V           (Orange)

  • Grid                                                       -9.3 V           (Green)

  • Anode                                                   592 V           (Blue)

Okay, this needs some clarification. We suddenly have a grid in the equation?? In the schematic, that particular terminal seems to go off to some other board. Let's quickly have a look what the HP engineers had to say about this:




That's nice, this TWT actually has a built-in modulation grid. According to the modulator description, it can be easily seen the grid can be driven by a frequency of about 0 - 500 KHz for small amplification, and 0 - 100 KHz for a lot of amplification. Besides, the least negative voltage that is effectively used during modulated operation is -40V and the most being -300V. In the calibration section of the manual, it's mentioned that the listed voltage on the case of the TWT is actually the maximum voltage that's allowed to be put on the grid. I know this all sounds confusing, to me it is as well. The manual for this device is very poorly constructed compared to most other HP manuals I've seen. I'll clear everyting up more once we're building the modulation electronics.

For the rest, in the calibration a bunch of other interesting adjustments are given for the accuracy of the regulation and the ripple. This will come in handy later when constructing our own datasheet of the TWT, but I don't really think it is necessary to really mention about that here. I copied the figures one-to-one from the manual.

We've seen a lot of voltages, but what about current? Using the schematic for the powersupply, we can get a good indication of what we have to deal with. The collector, helix, and pretty indirectly the anode as well, are fed by the same supply line, which is regulated by two 8068 beam pentode HV regulating tubes (V1 and V2). In the datasheet can be found that the comfortable operating cathode-to-anode current is about 36 mA per tube. This means that both elements are expected to draw roughly 75 mA together rounded up to a fun number. When the helix starts to draw too much current, a relay will shut down the high voltage to prevent damage. 

Finally is mentioned that this particular amplifier must not be operated without a 50 ohm load on its output. The max power on the input is 100mA, and at 1mW input power, the amplifier provides 1W output. The signal gain is a whopping 30dB, and over its entire bandwidth, at 1mW there's less than 6dB variation. Looks like a champion to me. Let's take this data and find some TWT's to compare values with to fill in the final gaps.

3. Early 60's Watkins & Johnson catalog

Finally I managed to find a couple old catalogs from W&J's trending products. I decided to pick the oldest catalog and find the best matching tube to see whether the found values are somewhat of the same order of magnitude. This is what I found:




I find the cold-war sentiment in this datasheet pretty endearing in an odd way. Quite some data is dedicated to applications in extreme flight related conditions. Let me mention, this tube is not even under the head:

Exactly, there was a whole market for these TWT's to be put in missiles! In hindsight it makes sense with this booklet coming from a propaganda filled world. It's these little things that remind me of the fortune and liberty that I get to grow up in a relatively peaceful time with all the opportunities within reach to understand all parties in a conflict (although we still suffer from the lack of ability of building bridges).

But from this fighterjet-grade TWT we can derive that the estimations made above are about in the ball-park of the average TWT. No grid seems to be mentioned though, it seems to be some special feature in the one that is going to be used in this project. 

The finished datasheet can be found HERE on the downloads page (computer access only). 

Regulated high voltage powersupply

Now knowing the average current ratings and voltage ratings to operate this TWT, a design can be made for the powersupply. We could simply copy HP's schematics, but that won't be very interesting. Besides, I don't have a similar transformer at my disposal for this project. 

I set as a restriction to build the HV part with tubes. I think this actually makes some sense as tubes are inherently designed to work in high voltage circuitry. This also opens up to easily use voltage regulator tubes, which are more temperature stable than high voltage zener diodes. This also allows me to use a high voltage reference without having to either amplify the reference voltage, or severely step down the feedback. (You can see I am just making a ton of excuses to justify playing with tubes)Another restriction is to use only European tubes. This sounds a bit odd, but American-exclusive tubes are pretty rare and very expensive in my area. European ones are still widely in circulation here, and if I don't have the correct ones in my collection already, I can simply buy tested or NOS ones for dirt cheap. I really only use the American ones in exceptional cases, like if one burns up in an existing piece of HP measurement equipment.

The selection of tubes I ended up making is a set of OA2 tubes as 150V regulators, PY80 tubes as HV rectifier diodes, PL36 tubes as regulator tubes, and maybe the ECC82s as feedback amplifiers (although for this one I also have some other options in mind like something fancy such as the E88CC or the ECC83S).

This project ground to a halt because of the difficulty finding a suitable transformer! This project will continue as soon as I find a good transformer to continue the build!!! 17-07-2022

bottom of page