What is this module?

On Sun, 24 Jan 2021 15:34:39 +0100, Piotr Wyderski



That is very strange. It looks labor intensive.

This is more normal for core memory:

https://www.dropbox.com/sh/zwv8c15rh5t791f/AADT2pjOBm1cONXXPe3I_FI8a?dl=0

snipped-for-privacy@highlandsniptechnology.com wrote:



Yes, it is. This NASA note:

https://core.ac.uk/download/pdf/80668483.pdf

at page 4 says:

"There are many predictions that other technologies are to replace  
magnetic cores, but magnetic-core technology is so widespread and  
established that it is difficult for other technologies  to replace it.

The core suffers, however, from the limitation of only DRO operation.  
unless more sophisticated geometries are used such as the multi-  
aperture devices (MAD) or Biax. The multi-aperture  devices use the  
technique of two or more holes whose axes are parallel. can be set up to  
give NDRO operation, although a prime current cycle is required after  
every interrogate pulse. A memory of this type is very expensive, and  
requires relatively intricate wiring and has severe temperature  
limitations. Two devices (ref. 3) using this technique are the  
transfluxor and the Shmoo element. The currents re- quired for these  
devices are high, 900 mA for clear and 550 mA for set, with a minimum  
read-modify-write time of 4 to 5 usec."

So it appears that this Soviet module might indeed be an extremely rare  
form of RAM based on transfluxors. No idea why, as BIAXes were known  
back then.



If it is from a military device, as advertised, then the required effort  
would not be an issue. I can't find a photo right now, but once I saw a  
PCB with a number of toroids. Each toroid had a big hole in the PCB for  
the core and a a number of banana-shaped slots for the windings. The  
windings were wound through these PCB slots, making a flat and robust  
integrated structure. I think they did it manually, albeit the reason  
for going there is unclear to me.



Yes, a typical coincident current RAM from that era. The read-outs were  
destructive, as they currently are in FRAM.

    Best regards, Piotr

On Sun, 24 Jan 2021 19:18:37 +0100, Piotr Wyderski



One of our instruments uses a FRAM to save the current front-panel
setups, because the write endurance is huge, 1e14 cycles. So we can
write to it often and not do a lot of thinking.

CYPRESS             FM25L16B-G

If read is destructive, it's transparent to us. But we only read a
saved setup once, at powerup.

snipped-for-privacy@highlandsniptechnology.com wrote:



It is transparent, but it consumes the write cycles the device should  
endure. It is worth to keep in mind not to be surprised one day.


So this is not an issue in this application.

    Best regards, Piotr

On Sun, 24 Jan 2021 21:11:38 +0100, Piotr Wyderski



That chip, at the maximum possible SPI data rate of read/write cycles,
would last 85 years. We write about once per second, which should last
about 3 million years. More actually, since we use rolling,
checksummed buffers and only one has to work.

On Sunday, January 24, 2021 at 6:34:46 AM UTC-8, Piotr Wyderski wrote:


The biax had orthogonal fields for the two holes, and the three-hole example
has a thick wire making circumferential field, with the thin wire sensing radial field,
also at right angles.   I've mainly seen three-wire toroid core memories, the
Soviet vintage item is certainly odd.   Maybe it was easier to produce a cylinder
with three axial  holes, than a rectangular biax item?

whit3rd wrote:



They knew the biax technology:

https://www.ebay.pl/itm/Very-RARE-USSR-Soviet-Military-BIAX-Ferrite-Magnetic-Core-Memory-Module-MFDZU-3/254702613297

which makes the engineering choices behind the first module it even  
stranger.

    Best regards, Piotr