First things first. Before you can do anything, you need a power supply. So that's the unit to start with. Fortunately, all documentation is on the internet, including the schematic diagrams.
The schematic diagrams include a partslist, which makes it easier to find the necessary components. One component however, caused me a headache: the transformer. First, let's have a look at the power supply schematic.
Power Supply Schematic. Click here for a full size version.
The power supply is where I'm going to leave the original design the most. As you can see, the circuit at the transformers secondary is nothing special. A Wheatstone rectifier brigde, where in the original Selenium rectifiers were used (real OM's will remember the typical smell when they went up in smoke) and some capacitors, inductors and resistors. I will use modern diodes for the rectifiers.
When receiving, the receiver current flows through R21 and that makes the negative bias for the receiver valves. Receiver current is about 25mA which makes -12.5V bias voltage.
The biggest problem is the primary side of the transformer. The original transformer was a special one: on the primary side it could be jumpered to support 105-115-125-135-205-215-225-235V using 4 headers, and it supported battery supply using a 6V battery which was the common voltage for car batteries those days. The mechanical vibrator, which had to handle 10A, took care of the high voltage, while the valve filaments were directly connected to the battery. Switching between AC and DC was done again with a header that had to be plugged in straight or 180 degrees, thus selecting the wanted supply.
If you look at the power supply, you will notice that there is no + or - indication near the supply terminals. That was not necessary, because the converter was completely mechanical and also the filaments don't care about polarization. So the transformer caused the most headaches. For two reasons: availability - but I could always consider to wind the transformer myself - and the 6V supply. 6V is original, but nowadays unusable. And yet I want to be able to use the set on DC, but on the more common 12VDC. So I had to find a solution for that.
Fortunately there are modern "vibrators": 12-230V inverters. And they are cheap. I bought a 180W inverter from dealextreme for just over 20USD including shipping. For the transformer I picked the N7/1+2 manufactured by Welter from Germany. Not cheap, but I did not want to gamble by using a fleemarket-type with unknown specifications. The transformer is exactly 10cm wide and resembles the original pretty well. After the first building efforts, the transformer, inverter and elco's are in position, as you can see here:
Setup of the power supply for the replica
Next I had to make the sunk part with the switch, fuses and connectors. I decided to make that as a small box underneath the main lid. But it was not possible to fold the aluminium at such short distances. Challenge number one.
Also, in the original are 38 holes for convection cooling in 5 rows of 8-7-8-7-8 holes respectively. Probably done by a hole-press, and I don't have one. Trying to drill the holes with an ordinary drill will not give a good result because the drill will run away and the holes will not be in a row. A column drill would be ideal, but I don't own one. But Hugo PA2HW does. So we spent an evening with Hugo and Mans PA2HGJ drilling all the holes, starting with a 2mm drill and increasing in steps of 0.5mm until the holes were 4mm diameter. Because I feared that the paint would reduce the hole diameter, once at home I increased the holes to 4.5mm diameter. And then the box had to be fixed.
I used 10mm aluminium L-profile, glued to the box with special Bison 2-component metal glue.
L-profile glued to the box.
The grey stuff is glue. According to specification, the glue can handle 180kg/cm2 and with a length of 20cm that makes 3600kg per side - making a total of 7200kg. Enough to lift 6 medium sized cars, so I won't tear that apart pulling a plug, I suppose. Now the box is ready to be glued to the lid:
Box ready to be glued.
Box glued to the lid.
The glue remains soft for about an hour and after that, it is rock solid. And then it is ready for a black hammer finish. But that is meant for steel and does not stick very well to aluminium. Therefore, I had to use a primer for non-ferro metal.
Supply with primer.
Using hammer finish is an experiment in itself: according to the instructions the hammer finish has to be applied in two layers, the second 4 hours after the first. If you do not apply the second layer within 4 hours, you may not apply the second layer for a week. Don't know why, but inexperienced as I am, I will follow the instructions...
After the paint has dried, the switch, fuses and supply connectors are fitted into place. As you can see, there are no separate connectors for AC and DC: the 7-pin MIL-connector supports both, connecting the output of the 12-230V inverter to the transformer primary by jumpers in the plug. I use all 7 pins: the center pin is earth, 2 pins are used for 12V in, 2 pins for 230V in and 2 pins for 230V out. In the supply lead for 12V, the battery is connnected to the 12V in pins, and 230V out is jumpered to 230V in. The mains supply lead uses only the 230V in pins, of course. The primary supply selection jumpers are also missing, because the transformer only has a 230V primary winding. Not even 2x115V: just a single 230V. The finger plates are missing at this moment: they will follow later in the project.
Litterature describes that a battery had to be at hand at all times. This because a favourite method of the German radio detection service localizing a illegal transmitter was to cut off the power to a street or block. If the transmission disappeared, they knew in which area to search. According to the B2 manual, you had to switch off the set, reverse the AC/DC header and switch on again. Then you could continue on battery power. If that was a sensible thing to do when the Germans were so close, can be questioned. With my power supply, I have to plug in the other connector and I can continue on 12V. Or even better: I would put a power supply in parallel with the 12V battery and run from the buffered DC. When the enemy cuts off the power, I just continue on the battery without interruption. Anyway, nowadays with those sophisicated directionfinders you'll probably be localized within a meter in seconds...
Now the painting is finished, it is time to assemble the supply. I did that in two parts. First, the transformer part.
The two elco's are soldered on a small PCB which in turn has been screwed to the chassis. On the PCB are 4 diodes type BY550-800 and they are capable of withstanding 800V reverse, 5A continuous and 60A repetitive peak. That should be enough for the purpose. Furthermore you see the decoupling capacitors and the choke for the 500V supply. The second part is the lid itself:
On the frontpanel is the DPDT switch that switches both a 230V and the +12V lead. Of course there are 2 fuses: T0,5A for the mains and T6,3A for the 12V. In the original there was 10A for the 6V supply (60W). Considering the efficiency of the inverter, I calculated the fuse for about 75W. Should be enough. The choke you see here is in the 230V supply (see original). There also is a 0,006uF capacitor to earth in the original schematic. But that is not how I am going to do that. Depending on the position of the mains plug (which, in contrary to the UK, has no defined position in NL) the 230V is connected through the capacitor (0,0047uF in my case) to the chassis and that can be quite uncomfortable. The capacitor is now connected across the mains just after the choke to keep mains interference out and leaking RF to the mains.
I wired the two parts laying side by side so future maintenance remains possible. Disadvantage is that a lot of wire has to be sandwiched when folding the two parts together. But it all fitted and then nuts an bolts were used to fit the parts together.
The completely wired power supply
At first I put a 1800 Ohm 10W resistor in series with the 230V. The main reason was that I did not "form" the elco's in advance and I wanted to prevent blowing things up before ever using it. The voltage stabilised at about 400V and that was fine. After removing the resistor the voltage was about 510V and that is what I expected with 2x180V secondary. The receiver voltage was also fine.
Time for the DC test. After connecting the 12V supply it remained quiet for a while. Obviously the inverter first checks the type of load and that is quite an inductive load. After about 2 seconds it started en then you hear the humming of the transformer (must sound very similar to the old mechanical vibrator HI). Whatever the output of the inverter is, it probably only read about sinewaves, but is not even close. The filament voltage was just over 6V on the DVM. The high voltage, without load of course, slowly exceeded 700V! This may be caused by the peaks as a result of the square wave drive. How that will work out when fullly loaded, we'll have to see...