CAD – Target Enigma

[David Jupp]

On David Jupp Said:

Dave,

I modelled a simple wheel with the wrapped 26 letters.

Your 4mm character height with 4.23mm line spacing gave a slightly enlarged gap between Z and A

If I did the sums to another decimal place 110/26 = (approx.) 4.231  than the result was near perfect.

Thanks David.  From your comment and Jason showing earlier how Alibre handles embossing on a curve, I believe SE makes comparatively heavy weather of this problem.  Might be doing it wrong but seems SE requires user to:

• create a tangent plane on the target surface
• draw a line on the plane on which to align the sketch later.  (Not strictly necessary but it helps!)
• create a text profile that’s then snapped to the line on the plane (several options)
• use another tool to wrap the text profile onto the curve.  SE seems to have a limited range when wrapping a vertical list round a tight curve.
• know that the usual Extrude tool doesn’t work properly, and he has to use ‘Normal Cutout’ or ‘Normal Protrusion’

Though I like SE it has it’s quirks!

Dave

[David JuppParticipant @davidjupp51506]

Dave,

I found in Alibre, that it did help a lot if the middle of the text sketch was close to the tangent point of the disc.  If I for example had the top of the A at that location, the wrap failed.  All CAD systems have quirks.

My main point was that you have to be a bit more precise with the sums when working out the line spacing, if you want nice even gaps.

[SillyOldDufferModerator @sillyoldduffer]

More done on the lamp assembly:

 

Pleased to say laying out QWERTZUID was straightforward.  After I realised this is a job for a fixed font!  Adjusting for a flat surface can be guesstimated and then tuned by eye.  Laying out vertically on a curve needs the maths and I shall have to have another go.

Wasted loads of time on the MES10 thread so the bulb isn’t properly modelled.  Have to pay for the standard IEC60061-1 and I haven’t been able to determine the thread form.  It’s rounded, not a recognised V, square, Whitworth, Sellars, or BS fastener profile.   Also seems to be a weird mix of metric and Imperial measure.  E10 is 10mm diameter (tap 8.8mm) and 14tpi. Maybe mixed because it was defined in the USA by electricians who have always been metric.

Dave

[SillyOldDufferModerator @sillyoldduffer]

Ordered a book earlier and it’s just arrived: “Inside Enigma” by Prof Tom Perera.  Lots of pictures, and it’s interesting to see what I got right and wrong.  Give myself 5/10, not bad for a freelance based on limited information, but my version wouldn’t impress a collector!   My rotors are on the left whilst the real one has them centred, and other details.

Going to carry on with what I’ve got for the time being, but this new information means there will have to be a second attempt.   What I’ve done isn’t time wasted and I’m not tempted to try and reuse components from the prototype.  The whole lot can go into my electronic junkbox.   Next time I’ll get the scale right, and because I’ve honed my Solid Edge skills, should be much quicker – the experience will pay off!

Mentioning re-use reminded me that I’ve been able to some parts repeatedly for different purposes, best example being the lamp-plate:

It appears 3 times on the lamp assembly, and the same layout is used for the letters too.

I think the same part can be used in the keyboard, which is next on my list.

Bother with Windows today.  Took an upgrade a few days ago and it’s been slightly flaky since.  Boot problem, failed to wake up from hibernate, lost one of my two screens, and thinks something is running at close down when it isn’t. The missing screen problem took an hour to fix.  Annoying!

Dave

 

 

 

[SillyOldDufferModerator @sillyoldduffer]

Onwards and upwards.  Added the keyboard and a few keys, plus the paddle that advances the rotors.

The key is quite simple.  Pressing it passes a rod through a hole in the Copper plus volts plate and then, by pressing on the plate underneath, starts the rotor moving.  Finally a wider section of the rod completes the circuit by hitting the volts plate.  May be too simple!

Section of a key:

Next image shows a near collision, ringed in red, which ought to be fixed, even though the 3V lamp has an exaggerated globe:

General view:

Learned more about Solid Edge, nothing exciting, and got into trouble by trying to create an assembly that depended on itself.   Assemblies within assemblies within assemblies; my poor head!

Dave

[SillyOldDufferModerator @sillyoldduffer]

Realised this morning my key design causes avoidable production and assembly time wasting.   As conceived I used the same end-cap at both ends, with a securing ring:

Works in theory but!!!  First problem is the securing ring and end cap provide no way of gripping them for tightening.  So turn the ring into a nut and add flats to the end-cap:

But that leaves an even worse problem.  Using the same end-cap at both ends means each and every key has to be individually assembled on the mounting plate.   Better if keys can be assembled separately as units that just drop into the mounting plate and are secured with a nut.   Rather than having a tray of key parts, the assembler has a bix of ready to plug in keys.   Design change!   The lower end-cap is replaced by a plug, marked ‘A’ below:

Changing the key sub-assembly automatically updates the main model:

Design requires a certain amount of thinking ahead.  Experience helps in cases like this: yes the design works and can be made as described, but it’s flawed.   Why did I miss it?  Partly inexperience with practical mechanical assembly problems, and partly brain fade.

Car headlamps may have become a maintenance problem as a result of the same design oversight. So awkwardly placed in a tight space that replacing a bulb needs tiny hands with a 270° swivelling wrist and gorilla-on-steroids muscles!

Dave

 

 

[SillyOldDufferModerator @sillyoldduffer]

Are problems part of the fun?  I’ve smacked into a couple today working on the plugboard (Steckerbrett), and am bruised.

26 normally closed socket pairs at the front of the machine allow the operator to reroute up to 13 letters, massively increasing the number of possible encryptions.

The letters and numbers on the front panel tell the operator which letters are which.   For reasons unknown the secret key sometimes used the letter and sometimes it’s number equivalent.  I is the 9th letter in the alphabet, A 1, B 2 etc..

Inserting a plug lifts a shorting bar that normally bridges a pair of sockets.

 

Here’s a view of the spring loaded shorting bar, my version, not a copy of the original:

The plug has insulated tips, so lifting the bar breaks the contact.  Except as I’ve modelled it, the sockets are still connected via the springs!  They need to be insulated.

Bad news – I have interferences and misfits:

First the springs extend beyond the plugboard – they won’t fit in the box!

 

And the nuts that fix the socket to the front panel collide.  It can also be seen that neighbouring shorting bars will collide.  It’s too cramped.

Oh dear, should have taken up knitting!

Dave

 

 

[noel shelleyParticipant @noelshelley55608]

Very interesting Dave, thank you. I had always been curious as to the use of 3 and 4 rotor set ups. It seems that the later Kriegsmarines 4 rotor machine had a position on the 4th rotor that would make it act as a 3 rotor machine, so being able to communicate with army or Luftwaffe. Noel.

[noel shelley]

On noel shelley Said:

Very interesting Dave, thank you. I had always been curious as to the use of 3 and 4 rotor set ups. It seems that the later Kriegsmarines 4 rotor machine had a position on the 4th rotor that would make it act as a 3 rotor machine, so being able to communicate with army or Luftwaffe. Noel.

Tempting to have a go at a 4 rotor machine except they aren’t well documented.  And I’m in enough trouble already! Not many 4 rotor machines survived the war, either going down with the ship, or destroyed by intact units after the surrender.

The Wehrmacht and Luftwaffe had a lot more 3-wheel machines and more machines survived because their collapse was much more chaotic.

What little I have on the 4-wheel machine shows several differences – wasn’t just an extra rotor tacked on to an Army machine.  One of the difference must be how they were able to decode three rotor messages as well as being extra secure for the Navy.

Dave

[SillyOldDufferModerator @sillyoldduffer]

No good deed goes unpunished.  I modelled an overcomplicated lid yesterday and wasted this afternoon producing a simpler version.  The ten minute estimate turned into a 90 minute slog and then testing the assembly to ensure the lid opened and closed led to all sorts of shenanigans.  Zooming in finally showed the problem was a tiny collision, fixed by shaving 0.1mm off the edge.  Solid Edge is supposed to beep when parts collide; either I have to turn it on, or I’m deaf.

With a lid, it’s possible to change the lamps without taking the machine to bits!

The space in the top right corner is for a 4V battery.  No idea what it looked like.

Also added:

• a lever operated cam (top left) which squeezes the rotors together to make electrical contact, and allows them to come loose so the they can be lifted out and changed.  (The Army enigmas mostly came with a set of 5 rotors, of which 3 were in the machine, order varied daily.)
• The brown reflector is notched to stop it rotating.
• on the right of the rotor assembly is the input box.  I guessed round, the book shows the original was a box, ho hum.
• Did a proper model of a 3V torch bulb, and it’s not any better than my first bodge.  Pretty though!

New items on the TODO list!

• ‘Improved’ the plate that steps the end wheel when keys are pressed and got it the wrong way round!  Should turn the rightmost wheel, not the left.
• Found a bezel is fitted under the lid to channel light into the letter display.  Has to be added and it has interesting curves.
• The plugboard socket and plug have to be plugged in the right way round.  Done by making one pin/socket bigger than the other so they only fit one way.  More work!
• In the original the bulbs screw down onto sprung contacts.   My version, where they screw down on a hard plate would work electrically, but many filaments would break if the machine were dropped.
• Tackle the odometer mechanism.  I have an idea now: a pin on the side of each rotor engages a lever once per rotation, and the lever clicks the ratchet round.
• I’ve caused mild problems by not planning the assembly tree more carefully.  Most of the model is assembled off the rotor shaft and it’s now apparent the rotors should be a sub-assembly, and it’s more logical to assemble off the box frame. Although an assembly can start with almost any part, the build is more stable if it’s done foundations up, i,e with a fixed part that doesn’t change much and provides a solid reference.  As is I’ve got a few parts in the ‘wrong’ sub-assembly, making it slightly awkward to navigate.
• Freelancing at reduced scale continues to cause trouble.  Couldn’t find the machine’s actual dimensions so the next version will assume the keyboard is about 250mm wide.  In theory action should follow on the heels of decision, but I’m dithering about starting again.

Dave

[SillyOldDufferModerator @sillyoldduffer]

Not much done today because of a computer problem, but here’s the latest:

Gone as far as I think wise with this version, and, starting with the scale being wrong, fixing it is time wasted.  Better to start again and apply lessons learnt.   I’m now planning the Mk2.

Had the Mk1 been good enough for production, CAD offers multiple ways of doing that:

• Generate CAM for subtractive processes like CNC, and find someone to make them.
• Generate CAM for additive processes like 3D printing, welding and casting.   Can’t weld or cast myself but I own a 3D printer and can braze/solder small parts.
• Generate 2D drawings, 49 so far, and make the parts using time-honoured methods in my traditional workshop.  CAD helps!  Creating 2D drawings from 3D models is fairly quick once you get the hang of it, certainly faster than 2D CAD and far, far faster than a manual drawing board.   Have to admit though that CAD generated drawings aren’t as good looking as hand drawn plans.  They’re not works of art though, only needed to provide dimensions in the workshop.

A good thing about CAD is how much of the Enigma was modelled without me measuring anything!   Almost all the dimensions and proportions flow from the diameter of the rotor ring, not from endless calculations, dozens of 2D-drawings, and a notebook full of dimensions.   Not having to track dimensions rigorously is one of the things that derails 3D CAD learners with strong 2D skills.  In 2D, designers have to track dimensions, in 3D they don’t, and believe it’s essential wastes time galore.  Important to think in 3D which is a different skill.

My biggest mistake was to start by modelling at reduced scale.  With hindsight, I’d have done better to model at full-size.  And scale down from that later if a miniature is to be made.

Not finished yet!

Dave

 

 

 

[SillyOldDufferModerator @sillyoldduffer]

Computer problems galore, but I’ve inched forward slightly.

I need 26 Single Pole Double throw switches, each operated by pushing a spring loaded key.   I think they can be made without much bother:

Wonder if paxolin stripboard is springy and tough enough.  If so, easily made with a Dremel and some soldering!  Otherwise, Beryllium Copper, Phosphor Bronze or wiper blade spring strip.

Thinking about the Mk 2 full-size model, I shall probably go with this assembly structure:

Dave

 

 

 

 

 

 

 

[Ian PParticipant @ianp]

Fibreglass would be better that Paxolin but as the pictures do not give any idea of size its to hard advise.

I think that when you started this thread you said it was more of an exercise than an actual build project but if it is  something you are going to build, then how close to the original design and materials are you aiming for?

I’ve not looked at how the switches are positioned but based on your switch diagram picture it that are all mounted in one plane then why not make your sandwich from three PCB’s, the NC and NO being achieved two different spacer thicknesses. The middle wiper PCB being a double sided one.

Ian P

[duncan webster 1Participant @duncanwebster1]

Miniature microswitches, fairly cheap and snap action.

You used to be able to buy little brackets to panel mount them as push buttons, including a little bush and captive button, but I haven’t seen them for years

[Ian P]

On Ian P Said:

Fibreglass would be better that Paxolin but as the pictures do not give any idea of size its to hard advise.

Your wish is my command, 45l x 15w x 13.5h, in mm:

 

I think that when you started this thread you said it was more of an exercise than an actual build project but if it is  something you are going to build, then how close to the original design and materials are you aiming for?

Intended to be buildable, so I’m trying hard not to fake anything, but it’s a freelance.  Starting out I had very little information about the innards, so lots of guesswork based on blurred photos.  Since then got a copy of ‘Inside Enigma’, which has more photos.   Still not completely clear and no drawings – I don’t even know the size of the box.

Not trying for a perfect replica, impossible without more data, but I’d like to keep it fairly authentic with 1920’s technology.  Might veer away by 3D printing much of it in plastic, cos that’s easy.   May never get built because I’m unwell.

I’ve not looked at how the switches are positioned but based on your switch diagram picture it that are all mounted in one plane then why not make your sandwich from three PCB’s, the NC and NO being achieved two different spacer thicknesses. The middle wiper PCB being a double sided one.

Good idea.  My notion that the holes drilled in stripboard will make it springier than PCB is a wild untested guess!  The original is made of metal insulated with what looks in a B&W photo to be Paxolin or Bakelite.

Ian P

Duncan’s microswitch suggestion is tempting, and so is replacing the plugboard parts with 1/4″ stereo jacks and sockets.  All options are valid, I’m designing to please myself!

Thanks for the suggestions!

Dave

[Grindstone CowboyParticipant @grindstonecowboy]

Just a thought – you’ll have to make at least two, so you can send messages to someone 🙂

Rob

[Ian PParticipant @ianp]

Not trying for a perfect replica, impossible without more data, but I’d like to keep it fairly authentic with 1920’s technology.  Might veer away by 3D printing much of it in plastic, cos that’s easy.   May never get built because I’m unwell.

Not sure how to respond to the last bit except to say get well soon.

So, another question. How far does the switch moving part have to travel? I presume these are not latching switches and that there is one switch per alphabet key. I know fibreglass (especially a whole PCB ‘comb’ structure) is way out of authentic, as is 1/4″ jacks and microswitches. I suggested FR4 board material because Paxolin or SRBP is brittle by comparison. However I dont have a clue about Enigma but sort of assumed that pressing each letter keys moved a rotor, oh, but then I remember rotors having to be set each day so I’m definitely confusing myself.

Are the plugboard jumpers single core wires or do you need multipole connections?

The switch in your drawing looks to have quite a lot of parts, with contacts fastened (how?) to the blade. If there are twenty odd switches each with there will be quite a lot of nuts and bolts!

Ian P

[Ian P]

On Ian P Said:

 

So, another question. How far does the switch moving part have to travel? I presume these are not latching switches and that there is one switch per alphabet key.

Correct!

 

I know fibreglass (especially a whole PCB ‘comb’ structure) is way out of authentic, as is 1/4″ jacks and microswitches. I suggested FR4 board material because Paxolin or SRBP is brittle by comparison.

Which may be a good reason for avoiding stripboard!  The advantage of stripboard is it comes with copper tracks and pre-drilled holes.  I’d bet money fibre glass makes a better spring but it’s harder to drill and cut and there are 78 of them!

 

However I dont have a clue about Enigma but sort of assumed that pressing each letter keys moved a rotor, oh, but then I remember rotors having to be set each day so I’m definitely confusing myself.

That’s right.  Pressing a key steps the rightmost rotor once through 360/26 degrees.  When the rotor has stepped a rotation if moves the next rotor one step, like an odometer (my next challenge!)

This part view shows my first answer, which is wrong:

 

When a key is pressed a rod passes through a hole in the Copper ground plate, not touching it, and pushes the ratchet drive plate down.  The drive plate turns the rotor by engaging a pawl with its’ ratchet wheel.    The rod is wider near the end of travel, and in this design, makes electric contact with the ground plate.  This is wrong, see circuit diagram below.

Are the plugboard jumpers single core wires or do you need multipole connections?

The switch in your drawing looks to have quite a lot of parts, with contacts fastened (how?) to the blade. If there are twenty odd switches each with there will be quite a lot of nuts and bolts!

Multipole! Yes, indeed lots of parts.  Each key contains 7 parts, and 26×7 = 182!  Each switch contains 9×26=234 parts. The contacts would be soldered to the blade.  Easy if the blade is copper, harder if spring steel.  Building an Enigma, or two as Rob suggests, calls for production technique because there are so many bits!

Ian P

The unchecked wiring diagram shows why my design above is wrong!  Pressing a key turns all the lamps on!  A  SPDT switch is needed:

Enigma was conceived just after WW1 when the Germans found that the British were reading their secret diplomatic telegrams.  See Zimmerman Telegram.  As is common with inventions, at least 3 others had much the same idea at the same time.  Scherbius was first to patent and history made his machine famous.   Few have heard of Typex or ECM.  No drama, no excitement because the Germans couldn’t break them.

Thanks for asking the question.  Tackled Solid Edge’s 2D drawing tool to do the wiring diagram.  Time well spent!

Dave

[Grindstone Cowboy]

 

On Grindstone Cowboy Said:

Just a thought – you’ll have to make at least two, so you can send messages to someone 🙂

Rob

Just write down the scrambled output and then type it back in. If you’ve got it right you should get back the original message. Of course to be truly authentic there should be some Morse code involved..

There have been computer simulations of Enigma machines, no idea how they operate, but I know a man who will. I’ll point him towards the forum

For the springy switch things I’d look for hard bronze. Used to be available as draught excluder, but the march of progress stopped that

[duncan webster 1]

On duncan webster 1 Said:

[…] There have been computer simulations of Enigma machines […]

I cast caution to the wind today, and spent all of £1.99 on the one that I mentioned earlier.

MichaelG.

.

Ref. https://www.model-engineer.co.uk/forums/topic/cad-target-enigma/#post-778754

[John HinkleyParticipant @johnhinkley26699]

Dave,

I have been silently following your progress with interest and have been playing with Alibre Atom and Fusion 360 to try to replicate the letter drums.  Unfortunately, Atom doesn’t have the extrude or wrap facilities that the Expert or Pro versions have, but I have been able to model the drums in Fusion.  It requires the plane upon which the letter list is placed to be offset – by an non-specific amount – from the circumference (not tangential) to produce the required result.  At least the way I do it, it does!  I used Bold Arial font, Height = 10mm and line spacing = 3.296mm, if it’s of any interest.

By the way, all the pictures of Enigma machines that I have seen have the letters wrapped “the other way around” the drum, the reverse direction to that shown in your CAD shots.

See the picture here the National Museum of Computing

 

John

Back to watching silently ……..

[SillyOldDufferModerator @sillyoldduffer]

Thanks, letters the wrong way round added to my TO DO list.  Would have been sent to the Russian Front for making that mistake in WW2!

Offsetting the wrap in Solid Edge as you suggest seems to break it.  I’ll try again though – experimenting with SE wrap had me thoroughly confused, so worth trying again now it nearly works.

Cheers,

Dave

[SillyOldDufferModerator @sillyoldduffer]

Spent a few hours on Spring Design this morning.  Conclusion, spring design is hard! Baffled. Due to ignorance I’m probably asking SE to design a spring that falls outside ‘best practice’.

Having failed to learn basic springs, I had a go at the mechanism needed to turn Enigmas rotors.

Enigma rotors move like an odometer, counting in base 26.  The rightmost rotor is stepped 1/26 of a revolution every time a key is pressed.  And every full revolution of a rotor turns the next rotor along 1/26 of a revolution.

Disaster, designing mechanical movements clearly isn’t my forte!  Latest attempt:

The pin on the right hand rotor lifts an arm.  The arm translates the motion to the left where it lifts a pawl engaged with a ratchet wheel:

It works except the design can only turn one adjacent rotor, and cannot be replicated to drive a third.  Also discovered my understanding of Assembly constraints is imperfect.

Aargh!

Dave

 

 

 

[noel shelleyParticipant @noelshelley55608]

Study an old speedo or counter to see how to get the rotors to turn in 1s 10s and 100s ? Noel.

[noel shelley]

On noel shelley Said:

Study an old speedo or counter to see how to get the rotors to turn in 1s 10s and 100s ? Noel.

Good idea, except I don’t have either!  Googling failed too in that the mechanism in the pictures I found on the web aren’t clear.  The information is out there, I’ve failed to find it.

Quite interesting what happens when attempting to design a mechanism from scratch rather than copying an existing example.  Whoever invented the odometer was a clever chap!

Triggered an idea though – I used to have a rotating number stamp.   Wonder where it is?  Also remembered one of my clock books had something, bingo!  It’s the Geneva Mechanism.  Image from Wikipedia:

Thanks,

Dave

 

 

 

[Author]

Posts