Fanuc System 9 Lost Memory.....

Good Afternoon,

7 days ago a technician came to me and explained a machine had alarm 904 or something and he had called Fanuc. Phone service instructed him to initialize the BMU, so he did. Obviously, now we do not have the "NC Software" or anything left in our machine. I joined the fight at this point.

Machine = Komatsu
Type = Mill
Control = Fanuc System 9

We do not have original tapes. We only have backup files for programs and params read off the serial port. I did Find a Greco Minifile plus, disks, and cables in the panel for our Fanuc System 9 machine. The machine is a 5 axis Mill. We invited a Fanuc service tech to join us and he stated without the original machine tapes there is nothing Fanuc can do for us.

The system boots into the screen which only allows 3 options:
I = initialize
L = load
G = boot

At this point the only thing to do was try to get the Minifile to save us. 4 days of struggles, tests, and scope measurements we now have the Minifile loading our backup files into the control. The control happily accepts the files, but when the load completes and we press "G" to boot, there is a scan of the memory and the message "incomplete" shows up.

I have a lot of questions, the main one being the files I have on disk... where did these files come from? According to the manual the only way to load the NC software is from tape. This means the only way to digitize the tapes would be to read the tape into a device like the minifile.

We have tried loading the files at different speeds and settings with no luck. We have tried sending the controls files we know are bad and the control recognizes the files are not what it is looking for. We are stuck, and hoping someone here has some advice. I understand being new here, I deserve none. My group has very little experience with older Fanuc controls, we are learning together.

One other thing. I have 2 bubble memory boards from a sister machine. I installed these 2 boards and the system will boot, but the sister machine only had 3 axis. The current machine has 5 axis, and I was not sure if it was possible to "install" new axis into the system.

Thank you for your time.

Is it possible that Komatsu can help? They might have the original machine files.

We reached out to Komatsu. A rep there contacted another in Japan and replied "There is no NC tape backup data for serial #". Cant say I blame them, keeping tape for 36 years would be extremely great customer service.

In that case, two thoughts come to mind: 1) a complete controls retrofit with modern hardware/software, or 2) boat anchor. If the machine were in my shop and mechanically sound, the former would be my choice.

Maybe someone here will have some old files that will work.

Best of luck.

Thank you Jim. I think we have a new machine being built as we type, but a few months away. I thought about a complete controls retrofit, but I figured the new machine would be done and ready by the time I am able to sort out and install the parts. As I am looking around and talking with different people different ideas are coming to my head. Tomorrow I will triple check all the hiding spots to ensure the tapes are not laying around here somewhere. I actually have one of the 4 tapes from the originals. The tape is 3 of 4 and is machine parameters.

I have loaded many system 9 system tapes and been personally involved in similar situations; unfortunately, at this point there is nothing more you can do; retrofit is the best solution,if the machine is worth it.

Memoryman, I have read your previous posts and learned much from you wisdom. Thank you for your response.

We had a breakthrough this morning. We had a sister machine, and I found the floppy disc containing the setup files for the sister machine. By looking at the files in notepad, I realized the NC OS files have some sort of header on them. The sister machine disc had 3 OS files and the broken machine only had 2. For fun I added the 3rd file to the broken machines 2 and loaded... boom the CNC will boot. There is still problems because I only have 3 axis but at least we understand a little bit more. Maybe what I am doing is crazy, I don't know. After 5 days straight of hitting walls Ive got tunnel vision and will try anything.

Great. I don't remember for sure, but it seems that the # of axes is a parameter, just as it is parameter 9000 in all Fanuc controls.

Quote:

---

Originally Posted by NickGutt

Memoryman, I have read your previous posts and learned much from you wisdom. Thank you for your response.

We had a breakthrough this morning. We had a sister machine, and I found the floppy disc containing the setup files for the sister machine. By looking at the files in notepad, I realized the NC OS files have some sort of header on them. The sister machine disc had 3 OS files and the broken machine only had 2. For fun I added the 3rd file to the broken machines 2 and loaded... boom the CNC will boot. There is still problems because I only have 3 axis but at least we understand a little bit more. Maybe what I am doing is crazy, I don't know. After 5 days straight of hitting walls Ive got tunnel vision and will try anything.

---

Sounds like you're making progress. Please keep us posted.

I figured I would outline what I have for materials here because I may have been confusing. We have 2 machines built 35 years ago, both Komatsu. One of them is no longer in service I will call that machine "sister". Sister had a Fanuc 6M and 3 axis. The machine still in service I will call "brother" and has 5 axis, and Fanuc system 9. We have a bunch of parts from sister and the actual boards are very similar. We have the following files created from original tapes:
Sister = 1-5
Brother = 1-4
Brother had 1 roll of original tape found buried in corner of electrical cabinet.
If we load files 1, 2, and 5 from sister we boot.
If we load files 1, 2 from brother and 5 from sister we boot with incorrect configuration.

I examined the tape we have and did some experiments online and think the tape we have never got digitized. I believe the tape we have is the missing piece. So Friday into today we began transcribing the tape by hand and will soon try to load into the machine.

We also purchased a refurb tape reader so we can load 1 and 2 digitally and the physical tape through the tape reader.

Memory man, I will check out the parameter 9000. Thank you for letting me know.

One other thing I will mention right now. I would love to buy a tulip memory board, but we are a sizable company with a large inventory of old fanuc parts. We found an SRAM BMU of 1M, and are loading to the SRAM now. I do not trust the bubble memory since the original faults were generated from the BMU board.

Make sure that that is a 1MBYTE board; afaik, Fanuc did not make a 1MBYTE BMU SRAM board, but it is likely only 128kBYTES, which may be too small for a system BMU.

Memoryman, very interesting. I made an assumption and thought the "1M" in the part # meant 1 megabyte because the original bubble memory also had "1M" in the name. I have attached a photo of the 2 boards I have installed in the machine at the moment. The 2nd slot is a 256 BMU and was always in there.

We finished transcribing the tape and error checking. We had a few errors installing the homemade file, but ultimately we were able to get the machine to boot from the hardware config shown in photo and now the CNC shows/moves all 5 axis.

Because we have the sister machine documents/files to compare to the differences between the 2 machines was approximately 500 bytes of data. Most certainly the tape labeled "system parameters" contains some sort of hardware specific setup. The other approx 12,000 bytes are identical to the sister machine.

We just got the CNC to boot about 30 minutes ago. We now need to load the programs and find out if making parts is possible or we have other issues to work through.

That is a common wrong assumption. The FS9 has 4 BMU slots; each can accommodate a BMU of up to 512kBytes; the first slot must be at least 128kBYTES and has the system loaded, with the remaining space used for part programs. However, the TOTAL part program memory cannot exceed 1.5 MBYTES combining all 4 BMUs.

Memoryman, thank you for explaining. We did run into issues yesterday, the machine will not load part programs. The machine states "00000 FREE".

We tried loading the exact same files into the original memory and the system fails to boot. Loading our files into the SRAM works, but loading the bubble memory does not.
The original configuration is:
Slot 1 = 1M BMU
Slot 2 = 256-I BMU

Today we will receive a reconditioned fanuc tape reader, so we no longer have to rely on a hand transcribed file. I suspect we still have some errors in our handmade creation.

I will also mention our inventory of memory below, all used:

The original boards are:
BMU 1M
BMU 256-1

Additional boards:
SRAM BMU 1M
BMU 1M
BMU 256-I
BMU 256

We have a lot of combinations to try, one thing I do not understand is the defect loop #s. What are they and specifically how to enter them accurately. Some of the BMUs have numbers which look like "***221***" out of the normal row of 3 digit numbers. I thought it was possible we were not initializing the memory correctly?

We received a fanuc tape reader today and now our machine is looking good to get back to production.

We are using the original BMU setup.

Our Fanuc 9 CNC OS is 3 tapes. 2 we loaded digitally via Greco Minifile and the 3rd through the tape reader. I am happy the machine is working, but left wondering a few things. We should be replacing the machine in a few months time so I am not sure doing anymore experiments is a good use of time. Thank you to anyone who has read or commented on the thread. Happy to share our experience here.

Thank you for sharing Nick.

The BMU boards are in BITS, so divide by 8 to get # of BYTES So your total BMU memory in BYTES is: 128k + 128k +32k + 32k; a single 512kBYTE board would give you more!