Betty Blitzkrieg - An Audio Destruction Device based on 1940s Military junk

A Historian, I am not!

A few months back a friend of mine gave me a load of old valve-related electrical junk after me telling him of my recent project plans for Studio Silo. Amongst the junk was a curious black box, half gutted, but still holding onto a few of it's remaining valves, transformers and a beautiful panel meter. Upon further inspection I managed to get a date on the meter; "1946". I also managed to realise that this was not something meant for operation on the ground. It had a 500Hz, 180V input Transformer. Hmm! This could mean only one thing, Aviation! From googling I realised that most aircraft power systems run at 500Hz AC, meaning that all wound components such as transformers etc can be made smaller in size. Makes sense I guess! The unit had a brass badge stating "C.M.R. Lock Unit - Serial 246". This unfortunately didn't give me any leads through Google as to what this device's purpose was or what craft it may have been from. All screws are imperial thread, so this must have been from either a US or British Military Aircraft. Investigations are ongoing as to the plane that it may have once been in service aboard...

Inside the dusty lump were a number of EF50 metal-can valves, one 6V6, and a pair of EA50 diode valves. Also a pair of leaf switch relays and the mains transformer mentioned previously...
Anyway!

So the building started!

...with plans to use at least one of the metal-can EF50 tubes in the new concoction and maybe to try out the EA50 diodes in some way. Maybe some of these valves still work eh! Hmmm

First I built a power supply in the tight middle section using an old Tandberg Radio transformer for the HT voltage (about 260Vdc) and used a... Laptop Charger for the 12V heater supply. Yes! Quite pure DC and stable, don't judge me! :-) With the power supply functioning I started on building a gain stage around one of the EF50 tube sockets, to see if any of them still actually worked. Standard pentode gain stage with stepped resistors on a rotary switch for anode and screen just for the fun of it. And sure enough, all the tubes amplified audio, but some not so cleanly as others. Two of them I deemed acceptable so the project was still on track. After messing around with the EF50 I decided to get a proper schematic scribbled of what I wanted to achieve and get some direction. Here is the one of the (messy!) drafts along with a panel layout:

I wanted to provide for both Hi-Z (Jack) and Balanced Lo-Z (XLR?) inputs as this will be used in Studio and could also find it's way onto a mucky stage somewhere. The first gain stage is an EF86 in Pentode Common Cathode mode. From there the signal splits and goes to a buffer/phase splitter for the clean path to the output buffer (discussed further down), and also to the ancient EF50 via the incorrectly named 50K "Gain" pot. The EF50 is also in Pentode CC mode but with switchable Resistors for anode and screen, giving various gain settings and curves (Values will be in the final schem. if anyone is curious). Here the bias is also adjustable via the 10K pot. After the EF50 I decided to add a diode LED clipper which I saw used in a HiWatt design. As you can see the arrangement gives a few volts of threshold before clipping as you have 3 LEDs in series on each half of the cycle. Variable R and switch in series with this to GND to vary the amount of clipping. It really does fuzz up the sound nicely, creating a very thick distortion, especially when the whole unit is at full tilt :-). Proceeding the clipper I then tried a number of tone stacks, to no excitement. It isn't detailed here but I finally decided on an FAC control as seen on Orange Amps. It basically has a chain of caps around a rotary switch and as you rotate it the signal gets put through more caps, adding up to less and less capacitance at each click. It basically creates a High-Pass filter, which is exactly what I needed because the clean path gives all the original low end, while the distorted path contains mostly added harmonics and noise.

Above is a later rendition of the clean path and the mixer/OP buffer. In the clean path you can see that I had to invert the signal from the EF86 so that it was in phase with the input and also in phase with the distorted signal. While I was at it I decided to build a phase-splitter so that I could switch between inv. and non-inv. clean signal (might chuck this out in the final stages). Finally the clean and distorted signals each go through 100K level pots and are then mixed at the input to the Cathode Follower buffer. After this the signal goes to the Bypass relay, which is one of the original 1940s leaf relays from the original guts. This initially caused big issues due to the leafs having capacitive coupling, and the input signal coupling with the output and causing nasty oscillations.

I am now adding IP and OP transformers to allow balanced I/O...

A final addition to the unit before I put the covers on, is to try and create some crude form of noise gate on the distorted signal path. This would allow high gain overdrive on the likes of vocals etc with the ability to kill the noise and hiss between actual bursts of wanted signal. I am planning to try either silicon diodes (as Peavey once used) or better still the EA50 tube diodes which were in the unit already. A great thing would be if I could control the threshold of the tube diodes, thereby controlling the threshold of the "gate". You might notice a small sketch in the top right corner of the first notebook scan which shows a possible arrangement. We'll see!

Final instalment coming soon, showing the finished unit...