Page 1 of 1

Making a saber secure module live outside the radio.

Posted: Mon Mar 07, 2005 9:20 pm
by mancow
I'm still working on this.
I tried duplicating the hookup circutry the best I could by looking at the saber service manual and the kvl schematics/pinouts.

I'm still getting a keyfail error.

If anyone has the time to take a look at the lines going into this thing in the manual and could post some ideas I would appreciate it.

I'm hung up wondering whether or not the thing needs to have the radio data lines running into it as a clock for timing etc.....

I think the next step will be to take an operating radio and start disconnecting the various line traces to determine what is needed or not.

mancow

Posted: Wed Mar 16, 2005 7:30 pm
by willbartlett
Neat project, I'll dig up the manual I have and see if it looks doable. trying to encrypt the telephone?

Will

Posted: Wed Mar 16, 2005 9:14 pm
by mancow
Nah, just trying to adapt it other radios but the telephone idea is a good one.

I got the keyfail error as expected but the kvl acts like it's not erroring right off the bat at times, as if it's making it part way through the process.

I'm thinking there needs to be some sort of clock on one of the lines to cascade the data through the various components.

I believe it has to talk to the micro p. before it will work. Also, there doesn't seem to be a logic line to select encrypt / clear. It's all done serially over the data bus and that complicates things greatly.

As far as spectras go I guess that's what the carrier board is all about in addition to the CVSD stuff it takes care of.

I was just hoping to find a way to feed the saber module the minimum of whatever it needs to survive on its own.


mancow

Posted: Thu Mar 24, 2005 7:58 pm
by kf4sqb
I don't know anything about the Saber secure module in particular, but I know a little about digital electronics. It does have to have a clock signal from somewhere. If it doesn't have one on-board, it could very well be getting one from the radio. I would also suspect that, if it does have one on-board, it would still be looking for a clock signal from the radio as well, as they would have to be sync'ed together for proper operation.

Check where it connects to the Saber, find the clock signal, and see what "frequency" it is. It will be a square wave signal, and may or may not always be present. If you can find it, you may be able to build a simple oscillator to generate the proper signal to feed to it, and that may solve that problem. However, you may have to extract the clock signal from whatever radio you plan to hook it to. I'm not real knologeable on encryption, but it may have to be clocked the same as the radio it is hooked to. Good luck!

DES and DES-XL modules....NTN-5833A, NTN-5835A

Posted: Sun Mar 27, 2005 1:28 am
by AEC
Yes, there are not one, but TWO clock oscillators in the modules, one for encrypt and one for decrypt. (both are cylindrical in shape)
The 'board' is actually a flexible circuit folded three times so it all fits inside the casing, and each 'side' has its own header that snaps into each other to make one solid 'block' that the casing slipped down over and soldered to two ground pins.

Good luck with your 'outside-the-box' encryption project.

73!

Posted: Tue Mar 29, 2005 5:30 am
by K4RUR
Hi,
I was intrigued by this subject, so I did some studying of the Saber schematic I have and the secure module does get connected to the main CPU. It is connected to a RESET line on the CPU. I haven't looked at it with an Oscope to see if it is a constant data flow, but it is connected to the CPU. If you need the pin numbers, let me know.


Take care,
Lowry K4RUR

Re: DES and DES-XL modules....NTN-5833A, NTN-5835A

Posted: Sun Jan 07, 2007 4:26 am
by radio-link
AEC wrote:Yes, there are not one, but TWO clock oscillators in the modules, one for encrypt and one for decrypt. (both are cylindrical in shape)
73!
Two oscillator are right, but encryption and decryption happen with the same clock, I guess one clock is for the encryption timing, the other for the internal clocks for communication to the CPU and for the memory.

Is this project still in progress? I am highly interested in this topic, and I have access to some fancy stuff at work, including sophisticated logic analyzers and digital scopes. The problem is, I have no schematics from inside the saber or systems saber radios.

Posted: Tue Mar 06, 2007 10:38 pm
by Zap
I've always though it interesting that the keyloader and the spectra use the same hybrid module.
KVL uses it for...??? but the spectra uses it for encrypt/decrypt the audio.

As you already looked at the keyloader schem, perhaps comparing that to the spectra may add more the that advill headache ;)

Posted: Tue Mar 06, 2007 11:32 pm
by radio-link
Zap wrote:I've always though it interesting that the keyloader and the spectra use the same hybrid module.
KVL uses it for...??? but the spectra uses it for encrypt/decrypt the audio.

As you already looked at the keyloader schem, perhaps comparing that to the spectra may add more the that advill headache ;)
The kvl needs the hybrid to encrypt the test tone, to show that the key is transferred correct, and maybe also to encrypt and decrypt the non-volatile keys in the EEPROM.

Ralph.