Virtual Wartime Bletchley Park
by Tony Sale

Trying some runs on the Bombe

(You will need to print these pages because you cannot run the Bombe from within the HTML page.)

It is useful to see how the current flows within the Bombe and how a stop is reached. We will use the first menu in the US Report on the British Bombe, fig 9 on archive page 12 (report page 7). Load first the text data file brpfig9t.txt. This causes a stop after each menu item.

To do this exit the HTML page and launch bombe.exe by double clicking on it in the folder listing where you have unzipped bombe.zip.

Now click on 'Setup' in the top left hand Setup window. The Setup Text window appears. Click on 'Search' on its menu bar. The Search window appears and scroll down the list of text files in the lower left hand window.

This is the menu we are going to run first. Double click on the name brpfig9t.txt and it will be transferred to the input field in the Setup text window. Click the OK button and the contents of the text data file will be transferred to the Bombe Setup form. Note that both the fields 'stop-at-menu-item:' and 'stop-each-letter:' are set to 'Y'.

Now click OK on the button in the Bombe Setup window and then click 'trace' in its menu bar.

The Registers window appears top right and the Results window appears bottom right. The Registers window shows the registers of the diagonal board which are actually in use for this menu. All the alphabet letters are set in each register and the wheel start letters are set to AAA in the Results window.

By clicking on 'Next' in the menu bar in the Results window, the simulator steps through each item in the menu calculating where the current would flow in the real Enigma drums on the real Bombe. When the 'current' arrives in a register the letter at that point is replaced by blank. Thus with successive clicks on 'Next' the 'current' flow through the Bombe can be seen gradually spreading through the registers.

The calculation proceeds until no more 'current' flows, i.e. it has propagated everywhere it can. It takes one last iteration through all the menu items to make sure this has been reached, that's why nothing seems to happen for a large number of clicks at the end.

Once the current flow has stabilised, the state of the registers has to be investigated to see if this, is a possible 'drop' to make the Bombe stop. The first requirement for this is that only one letter is left in the 'input' register. In the Setup form the 'input-at:' letter is 'N' so it is the N register which has to be checked. You will see both 'P' and 'Y' still standing so this is not a possible stop.

Continue clicking on 'Next' and the wheel start increments to BAA, (remember that on the Bombe it is the left hand wheel which goes fastest!), and the Registers are reset to full house.

So now the menu is being run against new wheel start positions. Carry on clicking on 'Next'. The next two, BAA and CAA do not clear so no stops either. At the wheel start 'DAA' the O register clears completely so this isn't a stop. (remember just one letter must be left).

You have now seen the way the 'current' propagates through the wiring in the Bombe and can now see why simulations of the Bombe are so slow. The current has to be calculated for every possible connection one after the other by the modern PC. The current flowed in parallel through all the connections in the original Bombe at nearly the speed of light. No competition!!

Now click on the 'Finish' button in the Setup window and click on the 'stop-at-menu-item:' field. It turns yellow and now delete the 'Y' and enter 'N'. (must be capital letter!). This will cause to simulation to stop on completion of all the 'current' calculations for a set of wheel starts.
Click on the 'OK' button on the Setup window and then click on 'trace' again on the menu bar. Now the simulator will repeat the calculations for wheels 'AAA' and by clicking on 'Next' you can move on to 'BAA', 'CAA' etc. with the final states of the registers showing each time.

At 'DAA' the 'O' register is filled, at 'EAA' both 'S' and 'X' are left in the 'N' register, but 'FAA' looks more promising. Only one letter in 'N' and in fact only one letter in each of the registers. Now clicking 'Next' causes the stop information to appear in the Results window. However it is not a valid stop since the Checking Machine would show more than one register letter steckered to 'E' and this is not allowed.

Moving on further by clicking 'Next' reveals another possible stop at 'IAA' but again this has repeated letters, but further on at 'LAA' a stop is revealed with no contradictions. It could be the correct stop for this menu.

However what you have just done reveals the need for the Machine Gun. This was an attachment to the Bombe which checked to see if, at a possible stop, the same letter remained in more than one register. i.e. on the display (and on the actual diagonal board) it scanned vertically for any duplications. It did this with Post Office Uniselector switches which make a load clattering noise as they scanned round, hence the name.

You can show the machine gun working by clicking 'Finish' on the Setup window then clicking on the machine-gun: field and changing 'N' to 'Y'. Now click 'OK' and 'trace' to restart the simulation. Clicking 'Next' will repeat the wheel start sequence but this time only the 'LAA' stop will come through to the Result window showing that the Machine Gun has rejected all stops with duplicate Stecker letters.

Carrying on will reveal possible stops at 'CBA' and 'QBA' etc.

Fig 9 is a very weak menu with no closures so lots of possible stops.

So now lets move on to other menus.

| next page | intro page |

This page was originally created by the late Tony Sale
the original curator of the Bletchley Park Museum

Virtual Wartime Bletchley Park by Tony Sale

Trying some runs on the Bombe

Virtual Wartime Bletchley Park
by Tony Sale