I built a system for extracting data from Priam hard drives

[Jeff Flitton]

Hi all,

I recently borrowed a Priam 3450 hard drive from a friend that owns an Intel iMDX Series IV system.

We've never powered up the iMDX, and we had low faith that all of the multibus cards, power supplies, and the aforementioned HD would function, so the machine sat dormant for the last few months while I devised a plan to extract the data from the drive without relying on the original host machine.  Today, I'm pleased to share with you that I succeeded in that endeavor.

There were three key issues to solve:

• Power needs
• Drive controller emulation
• Data interpretation

Fortunately, the Priam and the iSBC 215 Winchester multibus card were both very well documented and the wonderful archivists of the internet had made digital copies available.  Documentation indicated that the Priam needs:

• 24V @ 4A (wow!)
• 5V @ 2A
• -5V @ 2A
• -12V @ 0.7A

I rigged up two bench supplies with current limiting for the +24V and +5V rails, and used Mean Well bricks (wired backwards) for -5V and -12V.

The trickier part was building a controller.  I decided to go with something Raspberry Pi Pico 2-based, since that board with its PIO functionality is now being used in lots of exciting projects to emulate retro storage controllers and devices.  Mostly what I needed was level shifting and to adapt the 50 pin ribbon cable to the pins on the Pi.  I designed a board in KiCad which I've coined "Hecuba" (queen of Troy and wife of Priam - I decided a little Greek Mythology was called for).  I sent my designs to JLCPCB and after about a week received 5 boards.  I soldered one up, uploaded my test firmware to the Pico, and wired everything.  Here's what the setup looks like.  Note that after I shot this video I ended up wiring the ground strap on the Priam to the star ground in my harness, and from there to the ground terminal of one of my bench supplies, just in case.

I'm a professional software engineer, but I have no real electronics background other than what I've gleaned from people more knowledgable than I am.  While admittedly the design is pretty basic, I was still thrilled to see it work on the first try!  This video captures the Priam roaring to life for the first time in several decades.

The subsequent several days were spent writing more firmware so that I could pull data off the Priam.  Note that all flux interpretation happens on the drive (the two massive PCBs sandwiching the platters), so my job was primarily to seek, rotate through each head, and read the full track.  I learned a lot about basic data recovery in the process, and ultimately figure out that most of the seemingly bad sectors were just missing their header, with the rest of the data intact.  The end result was a > 98% recovery of the files on the drive, which houses an iRMX volume.  This particular system was used for developing medical clinic software, which was built on top of iRMX 86.  Source code and compiler are intact, along with some real patient data which I'm in the process of scrubbing.  With the permission of the owner of the HD, I'll release the image and the extracted files.

I plan to open source the entire project (hardware, firmware, and Python data processing scripts) so that anyone with one of these old Priam drives can archive its contents.  Bear with me while I separate some of the details of this particular HD from my design and code files.  In the interim, if you own a Priam and want to give this a try, feel free to contact me and I can send you a board.  I have four extras I'm unlikely to ever need.

Once I've released the details for Hecuba v0.9 (the one I used for this project), I'm going to contemplate a re-spin so that the same board could emulate either the controller or the Priam drive itself, which would give real legs to any host system designed to interface with one of these drives.

Credit to David Gesswein whose MFM emulator inspired my work, and the owner of the hard drive, my friend and fellow club member Bryan Graves who compiled a massive amount of information about the drive, controller, and the iMDX in general.  Bryan's dedication to this project is what kept pushing me along, and I'm certain it would've sat on the shelf far longer without his enthusiasm and support.

Best,

Jeff

[al kossow]

I have a large number of priam 50 and 70 meg drives that I have never been able to get to spin up using the priam intellegent SMART-T controller

I would really be interested in getting one of your boards to see if I can get just the drive to spin up and try reading the data off of them

--
You received this message because you are subscribed to the Google Groups "intel-devsys" group.
To unsubscribe from this group and stop receiving emails from it, send an email to intel-devsys...@googlegroups.com.
To view this discussion visit https://groups.google.com/d/msgid/intel-devsys/1f80e488-8007-405e-95fa-c1009edb49d4n%40googlegroups.com.

[Jon Hales]

Hi Jeff

Thank you for providing documentation of your arrangements for recovering iRMX data from a Priam drive. That's really good news. I have an iSBC 215D (never used by me) that I think would support the Priam drive.

I'm a long-time user of the Gesswein MFM Emulator and it's great to hear that this inspired your efforts. I'm sure David Gesswein would like to know about your new design. It's also great to hear that the Pi Pico was at the heart of your design.

While I don't have a Priam drive to recover, I would be very interested in the gerbers and BOM for your design - whenever you are ready to release the files.

I hope you'll receive responses from owners of Series IV machines who may have hints about how to restore your friend's machine to working condition.

Best regards

Jon

[Herbert Johnson]

So I have a Priam 7050-41, 70 MB apparently:

https://www.retrotechnology.com/aux/priam_drive.html

which has set idle for some time. I'm sensing that the drive won't spin
up without a controller? Jeff, what was your startup procedure? For
reference I found:

https://www.bitsavers.org/pdf/priam/Priam8_Maint_Jul82.pdf

DC power is a nontrivial consideration. You reported (for reference)
your 35MB drive power spec (also in the above reference) was:

> 24V @ 4A (wow!)
> 5V @ 2A
> -5V @ 2A
> -12V @ 0.7A
>
> I rigged up two bench supplies with current limiting for the +24V and +5V rails, and used Mean Well bricks (wired backwards) for -5V and -12V.

Presumably "backwards" means you used +5 and +12 supplies and reversed
ground and positive in use. 24V supplies I think were used in telecom so
should be obtainable as old-school. Or just build one with sufficient
amperage for power-up surges as I assume 24V drives the spindle and
maybe the head coil?

I read now from the above document, with PRIAM interface, power up
apparently requires a DRIVE SELECT signal (to ground) and then "issue a
Sequence Up command". Likewise a Sequence Down command to spin down.
Thus a need for a controller. Of course the spindle must be "unlocked"
and drive powered first.

Regards Herb

On 8/15/2026 2:24 PM, al kossow wrote:
> I have a large number of priam 50 and 70 meg drives that I have never
> been able to get to spin up using the priam intellegent SMART-T controller
>
> I would really be interested in getting one of your boards to see if I
> can get just the drive to spin up and try reading the data off of them

> On 8/15/26 11:08 AM, Jeff Flitton wrote:

>> I devised a plan to extract the data from the [Priam] without relying on

>> the original host machine. Today, I'm pleased to share with you that I
>> succeeded in that endeavor.

>> The trickier part was building a controller.  I decided to go with
>> something Raspberry Pi Pico 2-based,

>> I'm a professional software engineer, but I have no real electronics
>> background other than what I've gleaned from people more knowledgable
>> than I am.  While admittedly the design is pretty basic, I was still
>> thrilled to see it work on the first try! This video

>> <https://youtu.be/Kj5bFJDNB7E> captures the Priam roaring to life for

>> the first time in several decades.

>> The end result was a > 98% recovery of the files on the
>> drive, which houses an iRMX volume.

--
Herb Johnson, New Jersey USA
http://www.retrotechnology.com or .net
preserve and restore 1970's personal computing
email: hjohnson @ retrotechnology dot com
or try later at herbjohnson @ comcast dot net

[Bryan Graves]

Just wanted to add some context to our project.

[Jeff Flitton]

Hi Al, I'd be happy to furnish you with a board.  That's really cool that you have several recovery candidates, what a neat opportunity!  Can you confirm which model #s your drives have?

[Jeff Flitton]

Hi Jon, thanks for your interest!  I'll drop a link in this thread as soon as the repo is ready.  Bryan and I are excited to turn our attention to getting the full machine and its original terminal up and running, and we'll certainly post questions/and or notification of success or failure here!

To view this discussion visit https://groups.google.com/d/msgid/intel-devsys/CANPXPJk-X_91Em6jZa0Y6jh1vUu42ffK309m4i9maR02yoLHog%40mail.gmail.com.

[Jeff Flitton]

Hey all, just FYI I've finished documenting the hardware, firmware, and software I developed for this project.  It's all MIT licensed here.

Don't hesitate to reach out if you'd like to try archiving the contents of one of these drives, I'm happy to help!