A US Navy WWII Encryption Device

[Puster Bill]

Hopefully, this will be of use to you guys who play SHIV as a way to communicate in a more immersive way.

I decided to built a strip version of the US Navy CSP-488 "Wheel Cipher". The CSP-488 was the Navy designation of the US Army M-94 wheel cipher. Any wheel cipher can be turned into a strip cipher quite easily by "unrolling" the wheels, and doubling the mixed alphabets of each wheel on the individual strips.

The CSP-488 wheel cipher was used up until 1942 or 1943, and it was replaced by the more secure CSP-845 strip cipher, which had 100 strips to choose from and 30 positions for enciphering.

I chose to make a strip version of the CSP-488 though, because you only have to worry about setting up and cutting out 25 strips, not 100.

Anyway, here it is:

http://img89.imageshack.us/img89/3792/dsc02550g.jpg

You can tell the device was first invented for use by the Army. If you look down the strip that is marked 'R17' at the top, you will notice that part of the alphabet reads "ARMYOFTHEUS". Wonder if that got on the nerves of Navy radiomen who had to use it to encipher!

How it works is that you set up the strips in a pre-arranged manner, according to a keylist. In this case, the 25 strips are ordered as follows:

YNXWSZJGFQRDVMTOHUPLEKICB

Once we've set the strips in the proper order, we slide the individual strips up and down until our desired plaintext is visible. Let's use 'SEND LAWYERS GUNS AND MONEY' as our plaintext:

http://img31.imageshack.us/img31/2729/dsc02543ycl.jpg

I put the plaintext at the top of the bottom retention strip. Admittedly, it's an unlikely message for a US Navy sub to send, unless of course it's the USS Sea Tiger. I padded it out with two X's to make the message exactly 25 letters long.

You can pull the ciphertext from any line above. Lets say you pick the line 10 down from the bottom of the top retention strip. It reads like this:

DMWQIJIVUKIFABOQDXGKBMIFG

An operator would seperate that into groups, probably 5 letter groups, so it would look like this when sent:

DMWQI JIVUK IFABO QDXGK BMIFG

For a message longer than 25 characters, you would repeat the same procedure for each 25 character 'chunk' of plaintext, but you would generally chose a different line of ciphertext for each 25 character block. It doesn't matter which one you choose, as they will all result in the same plaintext, which 'jumps out' at you because it's the only intelligible result.

Construction notes:

The strips were individually typed into columns in an EXCEL spreadsheet based upon the wheel information found at the USS Pampanito website:
http://www.maritime.org/csp488.htm
They were then printed out on heavy cardstock and seperated using my wife's cutting board, my pocket knife, and a metal yardstick as a guide for the knife blade.

The strip holder is a piece of heavy cardstock with two sets of parallel cuts to hold the strips.

The US Navy CSP-845 strip cipher is vertically oriented instead of horizontally like my device, but I find that the horizontal orientation is more intuitive, and you can use the retaining strips as a guide.

[Puster Bill]

Forgot to mention: I have the EXCEL file all ready to be printed out if anyone wants it. That way you don't have to re-type everything.

[Rockin Robbins]

Whew! Not exactly Enigma, is it? I wonder how many of those "secret" messages were read.

[Puster Bill]

Whew! Not exactly Enigma, is it? I wonder how many of those "secret" messages were read.

For the M-94/CSP-488, probably quite a number of them. The system isn't really that secure, and it was only used for low level tactical communications. Although, quite frankly, Japanese signals intelligence left a *LOT* to be desired, so maybe not.

Certainly, today, the M-94/CSP-488 isn't secure.

Now, the M-138-A/CSP-845 is a different story. With 100 strips, only 25 or 30 of which would be used at once, it probably gives about as much security as a 4 rotor Naval Enigma. It's telling, though, that it was considered only "medium" security.

I settled on an M-94 variant because the wheel alphabets are published, and there are fewer strips to cut out, which is the most tedious part of making this.

I'm thinking about building a wheel version. I'm going to check around at the local hardware stores and see what I can use as the wheels. If I manage to build a decent one, I'll post pictures and a how-to.

[Puster Bill]

I went back and looked at it, and the key space (ie., the number of possible keys) is much larger for the M-138-A than it is for the Naval Enigma, implying that it would be more secure.

You have to be careful with that number, though, because systems can have weaknesses unrelated to the number of total keys. Simple monoalphabetic substitutions have a total key space of 403,291,461,126,605,635,584,000,000, more than the Naval Enigma at
409,183,233,484,479,582,535,680 possible keys, yet even kids can solve them in a short amount of time.

[razark]

Whew! Not exactly Enigma, is it? I wonder how many of those "secret" messages were read.

The question isn't how many were read, it's how many were read in time to be effectively used. A system like this could be useful for short, quick messages and carried by an individual on the field. A tabletop system like Enigma can't exactly be slipped into a pocket.

Enigma doesn't exactly have a reputation of being secure, either. The M-94 could just be read quicker.

Razark

[Puster Bill]

The question isn't how many were read, it's how many were read in time to be effectively used. A system like this could be useful for short, quick messages and carried by an individual on the field. A tabletop system like Enigma can't exactly be slipped into a pocket.

Enigma doesn't exactly have a reputation of being secure, either. The M-94 could just be read quicker.

Razark

That's a large part of it. A system just has to be secure *ENOUGH*, at least from a military standpoint.

Actually, you could use an Enigma today in a very secure manner by simply doubly enciphering like the 'OFFIZIER' messages, but without the tell-tale indicator and without transmitting the wheel settings (they'd have to be part of the pre-agreed upon key settings). That would pretty much stymie any potential attack that I am aware of.

[Puster Bill]

PREHN JXYCZ LHCGX DXKEQ NLYNV
GUNQP KYDDR UWVOC SETYO SJFJJ

For those who haven't built one yet, here is an electronic version:
http://members.aon.at/cipherclerk/VirtualM94.html
Key is the same as the previous message.

[razark]

PREHN JXYCZ LHCGX DXKEQ NLYNV
GUNQP KYDDR UWVOC SETYO SJFJJ

For those who haven't built one yet, here is an electronic version:
http://members.aon.at/cipherclerk/VirtualM94.html
Key is the same as the previous message.



TRLBZ WNQNM VKPBR YWSLJ APHPD
EBXRY FIRNK MVRSE UICWV ZYARS

Thanks for the link to the electronic version.

[Puster Bill]

You're welcome, but one operational point: Generally, you wouldn't end a message with a group full of X's, you just use them (or some other rare letter like J or Z) to pad the last group out to five.

I would suggest that you build a device like mine, though, because while the electronic version is convenient, it's like kissing your sister. It just doesn't feel the same.

I'm thinking, if there is some interest, in making up an improved version that can be printed on cardstock and then cut out. I'd make it vertically oriented like the M-138-A/CSP-845, and print something on it like this:

CONFIDENTIAL

Model CSP-488 Mk II Cipher Device
Office of Naval Operations
Code and Signal Section
United States Navy
1942



This would give the flavor of using a historically correct cipher device, a kind of cross between a CSP-488 and a CSP-845. Instead of an electronic analog, though, you be using a physical device. It adds a little something to game play, kind of like how I prefer to use a physical whiz-wheel to do my targeting calculations instead of an electronic version, or how I use an actual E-6B and physical sectional maps when playing with flight simulators.

[Rockin Robbins]

Love the link to the electronic version! I've had a lot of fun playing with the electronic Enigmas, but you're right. For this one, holding it in your hand is the best.

Wouldn't it be great if radio transmissions were in code and the game had an electronic code machine to decode messages? Especially if you could turn it on and off!:O:

[Puster Bill]

Love the link to the electronic version! I've had a lot of fun playing with the electronic Enigmas, but you're right. For this one, holding it in your hand is the best.

Wouldn't it be great if radio transmissions were in code and the game had an electronic code machine to decode messages? Especially if you could turn it on and off!:O:


The ancient (1992) computer game Das Boot had just such a facility. Although it was supposed to be an Enigma, in actuality it was a simple 3 letter polyalphabetic cipher. It was kind of fun though, especially since you could decode enemy traffic and try to intercept the ships mentioned.

Unless, of course, you mean *Morse* code, like this:

http://www.youtube.com/watch?v=BxISYBvCbkc - Me sending a U-boat short signal

http://www.youtube.com/watch?v=GEXR-OEuBdk - Me sending a properly formatted U-boat message (not a kurzsignale).

That would be ubercool.

[Rockin Robbins]

Awesome! A code within a code on the morse code there! CW is a lost art. When I was a teenager I learned it pretty fluently but I've lost it. If you're not a HAM operator the skill atrophies. It was a real treat to hear your video!

[Puster Bill]

Awesome! A code within a code on the morse code there! CW is a lost art. When I was a teenager I learned it pretty fluently but I've lost it. If you're not a HAM operator the skill atrophies. It was a real treat to hear your video!

I was a Morse interceptor in the Army, and found I loved code so much that when I got out, I got my ham license.

I don't get on the air as much as I would like, but I manage to make a contact here and there. Contacted a couple of Swedish stations a week ago.

Given my signals intelligence background, though, you can see how this kind of stuff interests me. I'm still trying to find the alphabets for the M-138-A/CSP-845 device though. If anyone one knows where I can find it, let me know.

[CDR Resser]

Kim Ronhof over at subsowespac.com, and the Pacific Thunder site built an electronic cipher program for communications within the Pacific Thunder Campaign. He might have some of the information that you are looking for. Sorry, I don't know his email address.

Respectfully Submitted;
CDR Resser

[Puster Bill]

Kim Ronhof over at subsowespac.com, and the Pacific Thunder site built an electronic cipher program for communications within the Pacific Thunder Campaign. He might have some of the information that you are looking for. Sorry, I don't know his email address.

Respectfully Submitted;
CDR Resser


Thanks. I took a look at what they had at subsowespac.org, and it appears to be an electronic version of the SIGABA/CSP-888/889 cipher machine. The CSP-888/889 was a rotor machine, essentially a more complicated version of the German Enigma.

If I were going to use a simulation of a CSP-888/889, I'd probably use the one found at Frode Weierud's 'CryptoCellar':
http://cryptocellar.org/simula/sigaba/

The Navy also used the M-209, which they designated as CSP-1500. Dirk Rijmenants (of Enigma Simulator fame) has a nifty M-209 simulator here:
http://users.telenet.be/d.rijmenants/en/m209sim.htm

The CSP-1500 was actually the replacement for the CSP-488 that I implemented here. It was somewhat more secure, but could still be broken if enough plaintext could be had, or if two or more messages had keys that overlapped. Apparently, it was used to communicate with units that didn't have the (presumably more secure) CSP-845 strip cipher.

I have neither the time, the skills, nor the money to build a SIGABA or an M-209, though. Nor do I have the money to buy them (M-209's occasionally come up for sale). I do have the time and patience to make strip ciphers, though.

By the way, I re-did the CSP-488 Mk II into a vertical orientation, and I used my wife's paper cutter to cut out the strips, which made it *SO* much easier. I'll post a pic when I get a chance. I now have an EXCEL file for the strips, and a Word document for the strip holder.

[Puster Bill]

OK, here is the new and improved vertically oriented CSP-488 Mk II:

http://img26.imageshack.us/img26/44/dsc02551a.jpg

I may improve on it yet again. My thinking is that I should cover the center of the strip holder, between the two retention strips on either side, with an inkjet transparency that has a number of vertical lines down it, to correspond the columns generated by the strips, and perhaps to designate them with a letter or a number to make it easier to read the columns.

Either that, or I may just run a piece of thread down the center for alignment purposes.

[Puster Bill]

MSG 213 Z 210255Z AUG 42 GR22
bt
YDVEV BJGWZ ROUJQ ISVZA PTIDC PBBJG OWIAB ZGQEM BJWMI SJRMT
OQSDH OZXSS EUNVR NHYSV SOXPA IOYPR JQBXV DKULG OVZNW JIVTX
YXDYN VVMCT
ar K

[Puster Bill]

OK, I'm going to ask for some help from anyone near Pensacola Florida.

The US Navy used a strip cipher during WWII called the "CSP-642". I was a lower-level cipher similar to the ones I posted above, except that there were 30 strips, and they used 25 to 30 of them at a time, depending on the particular key for that day.

The Japanese captured examples of it at Wake and the Aleutians, but even though they devoted considerable effort in breaking it, they only had spotty success. They recognized that it was used by US subs, though, by the indicators used: BIMEC and FEMYH.

I have been unable to find any images of it online.

There is an example of this cipher at the Naval Security Group Command at Corry Station. It's called the "Wenger Command Display". It's open to the public, the information necessary to visit personally is here:
http://ehistory.osu.edu/wwii/interactive/wenger/index.cfm

If anyone can get me some good, clear pictures of it, I can make up the files necessary and then upload them.

I *COULD* make a notional example with my own mixed alphabet strips, but I'd rather it be based on the real thing if possible.

Thanks in advance.

[Puster Bill]

OK, I got good photocopies of an original CSP-642 US Navy strip cipher. When I get the chance, I'll make both a detailed copy, and an easy to build analogue, and I will post the images and instructions here.

[Hitman]

In case it's a different one, here's also one offered by the guys at Subsowespac: http://www.subsowespac.org/silent_hunter_mods.shtml#21 :up: