FC9801 computer FDD replacement

clingel
Posts: 13
Joined: Wed Aug 15, 2012 2:21 am

Re: FC9801 computer FDD replacement

Post by clingel »

Jeff,
The hfe files and the imd file are zipped and located here: http://infoconix.com/programs/clingelfloppies.zip Also included is a text file with a summary.

The V2 software does indeed correctly show t0h0 as an FM track even though the status window during the dump indicated it was MFM.

When I run the IMDV viewer program it does not show anything past the first unformatted track in the imd dump. I verified the diskette by reading it using my software (it uses fdrawcmd). Summary is Total tracks 79 ( 0 to 78): Track0 h0 is FM with 26 128 byte sectors, track0 h1 to track 76 h2 are all MFM with 26 256 byte sectors. Track 77 h0 and h1 are unformatted, and the last track t78 h0 is formated MFM 26 256 byte sectors, while the t78 h1 is unformatted.

An img file created from the partial format hfe does not correctly show the data.

I also include an hfe for a diskette that has all 79 tracks formatted and the same data written into the appropriate sectors.

An img file created from the all formatted diskette does show the correct data ( it, of course, also shows the data from the unused tracks that were formatted ).

The diskette with the unformatted tracks works in the EDM machine. What is unknown is whether the diskette with all the tracks formatted will also work or not. I don't have immediate access to the machine, I have to send the diskette.

-Charles

Jeff
Site Admin
Posts: 8091
Joined: Fri Oct 20, 2006 12:12 am
Location: Paris
Contact:

Re: FC9801 computer FDD replacement

Post by Jeff »

clingel wrote:Jeff,
The hfe files and the imd file are zipped and located here: http://infoconix.com/programs/clingelfloppies.zip Also included is a text file with a summary.
Thanks. I am maybe wrong here, but i think that the machine will not read the tracks above 76. The others tracks seem to come from another live of the floppy disk :) .

Anyway the HFE file seems be ok and should be ok for the floppy emulation.

clingel
Posts: 13
Joined: Wed Aug 15, 2012 2:21 am

Re: FC9801 computer FDD replacement

Post by clingel »

Jeff,
Thanks for looking at this.

The machine definitely reads track 78, if it is not present then you get a message that the system disk is invalid. If the data in track 78 doesn't have the exact format it also shows the invalid message. It took us a while to figure this out when we first tried to make a diskette after the original diskette started getting occasional bad reads. We had an original backup copy(there were two with the machine) that we used to make a clone. Once we figured out the skipped track and the data in track 78 the cloned diskettes worked. By the way this machine is from the mid 1980's and is still running with the original floppies and computer - so we are getting a little nervous about it's reliability, mostly the moving parts - the floppies.

Regarding the hfe file with the partial formats ( the actual disk that runs the machine), when I use the Hcx software to create an img file the data in the last track is not correct. The machine won't run unless the last track contains the **** string. So I don't think it will work.

If the machine will run with a diskette with the formatted track 77 and 78 side 2 ( to be determined) then yes, I think it will work with the full format hfe file. It is possible that the machine checks for an unformatted track 77, however. That is what I need to determine.

-Charles

Jeff
Site Admin
Posts: 8091
Joined: Fri Oct 20, 2006 12:12 am
Location: Paris
Contact:

Re: FC9801 computer FDD replacement

Post by Jeff »

clingel wrote: Regarding the hfe file with the partial formats ( the actual disk that runs the machine), when I use the Hcx software to create an img file the data in the last track is not correct. The machine won't run unless the last track contains the **** string. So I don't think it will work.
For some reasons we got tracks 39 & 40 on the tracks 77 & 78. I will check if this can be a bug of the dumper.
Why the IMD file doesn't have the track 77 & 78 ? (can you activate the "full analyse" mode ?)

clingel
Posts: 13
Joined: Wed Aug 15, 2012 2:21 am

Re: FC9801 computer FDD replacement

Post by clingel »

I ran the full analysis. When the scan reaches an unformatted track it runs through a bunch of bit rates and then puts an 'N' at that position on the screen, so there are 'N's at track positions 77h0, 77h1, 78h1, 79h0, and 79h1 when it is done.

Here is the file: http://infoconix.com/programs/clingelfloppies2.zip

-Charles

Jeff
Site Admin
Posts: 8091
Joined: Fri Oct 20, 2006 12:12 am
Location: Paris
Contact:

Re: FC9801 computer FDD replacement

Post by Jeff »

clingel wrote:I ran the full analysis. When the scan reaches an unformatted track it runs through a bunch of bit rates and then puts an 'N' at that position on the screen, so there are 'N's at track positions 77h0, 77h1, 78h1, 79h0, and 79h1 when it is done.

Here is the file: http://infoconix.com/programs/clingelfloppies2.zip

-Charles
This file is perfect. Just drag&drop it on the hxc software and have a look into the Track Analyser. You will see this at the track 78 sector 1:
<<*JAPAX *********>>

And the track 77 is not formated.

clingel
Posts: 13
Joined: Wed Aug 15, 2012 2:21 am

Re: FC9801 computer FDD replacement

Post by clingel »

Yes! The img file output exactly matches the img file from my program. I'll contact my machinist friend and we'll start on the project. Thanks for working with us on this. Nicely done. I'll post progress as it happens - machinists aren't the fastest adopters of new technology so it will take a while.

-Charles

Jeff
Site Admin
Posts: 8091
Joined: Fri Oct 20, 2006 12:12 am
Location: Paris
Contact:

Re: FC9801 computer FDD replacement

Post by Jeff »

clingel wrote:Yes! The img file output exactly matches the img file from my program. I'll contact my machinist friend and we'll start on the project. Thanks for working with us on this. Nicely done. I'll post progress as it happens - machinists aren't the fastest adopters of new technology so it will take a while.

-Charles
Another point to check : Do you know if the floppy disk drive need to have the VFO circuit ? What is the original floppy disk drive model ?

clingel
Posts: 13
Joined: Wed Aug 15, 2012 2:21 am

Re: FC9801 computer FDD replacement

Post by clingel »

Our system is identical to the first set of posts in this thread. The floppies are NEC FD1135D and have the same extra white wires running to the inside of the drives themselves. How does a VCO/VFO affect all of this? Is this something the computer interface needs to unserialize the data stream? or maybe is it related to the fact that there are switches on the back of the housing to select different densities ( these switches are never changed, they are always in the 1Meg posiition)?

-Charles

Jeff
Site Admin
Posts: 8091
Joined: Fri Oct 20, 2006 12:12 am
Location: Paris
Contact:

Re: FC9801 computer FDD replacement

Post by Jeff »

clingel wrote:How does a VCO/VFO affect all of this? Is this something the computer interface needs to unserialize the data stream?
Exactly. The VFO circuit is commonly on the FDC board. But this not always the case with NEC / PC98 based system : For some reasons the VFO was placed into the floppy disk drive...

So you need the kitahei88 VFO board.

clingel
Posts: 13
Joined: Wed Aug 15, 2012 2:21 am

Re: FC9801 computer FDD replacement

Post by clingel »

Jeff,

On the V2 software track analyzer I notice that the for my 128 byte FM tracks the green graphic shows it as FM but the information in the status window shows it as MFM. I assume there is a bug in the status window report.

We only need to build the 500kbs separator - anyone have any ideas on now to obtain the SED9421 data separator chip? I don't see that the kitei88 is actually something you can buy but there is the schematic.

Jeff
Site Admin
Posts: 8091
Joined: Fri Oct 20, 2006 12:12 am
Location: Paris
Contact:

Re: FC9801 computer FDD replacement

Post by Jeff »

clingel wrote:Jeff,

On the V2 software track analyzer I notice that the for my 128 byte FM tracks the green graphic shows it as FM but the information in the status window shows it as MFM. I assume there is a bug in the status window report.

We only need to build the 500kbs separator - anyone have any ideas on now to obtain the SED9421 data separator chip? I don't see that the kitei88 is actually something you can buy but there is the schematic.
You should contact kitahei88 , he made a batch of these board some weeks ago.

clingel
Posts: 13
Joined: Wed Aug 15, 2012 2:21 am

Re: FC9801 computer FDD replacement

Post by clingel »

Jeff,

I guess I forgot to hit submit because the following post doesn't show up as far as I can tell, so this might be a repeat..

The data from the Hxc runs at exactly 500kbs controlled by a crystal, right? I mean there is no jitter like from a real floppy disk. I don't know if there are any spare I/O pins on the PIC processor and I don't know how tight your software loops are, but it would not seem very hard to generated a toggled bit that windows the data coming out. All that is needed is the square wave that windows the data pulses to simulate the data separator. The VCO in the VCO circuit really isn't doing anything since there is no speed variation for it to track.

Alternatively, a very simple data separator can be made from a crystal oscillator, a counter, and a few D flip flops. Both the Hxc and the counter are controlled by a crystal so the timing is more than close enough since there is no variation to track , it only has to get the phasing correct relative to the rd pulses from the Hxc which can be done by resetting the counter appropriately. For that matter it could probably be done with a single chip 8 Mhz AVRTiny.

-Charles

Jeff
Site Admin
Posts: 8091
Joined: Fri Oct 20, 2006 12:12 am
Location: Paris
Contact:

Re: FC9801 computer FDD replacement

Post by Jeff »

Yes but this is not so simple : The data separator need to be able to detect missing clock (in FM or MFM) and align data from the missing clocks.

clingel
Posts: 13
Joined: Wed Aug 15, 2012 2:21 am

Re: FC9801 computer FDD replacement

Post by clingel »

Check this app note from NEC

http://www.bitsavers.org/pdf/nec/datash ... _Mar79.pdf

On page 16 of the document ( 17 in the pdf list) it shows a data/clock sequence that is the original read data along with a clock centered around the data. The schematic to do this is on page 18 and since a VCO is not necessary you can construct the counter shift register set to produce the separated data and clock to feed a 765 controller.

I think you only need the data window clock and the original data stream. The counter is reset to start the counter whenever a rd data pulse is received, this synchs the clock , but it keeps running even if the pulse is missing, resynching at the next one that does arrive. As long as the original data is stable, the VCO is not needed to track speed changes and the clock will continue to bracket the data. The max time without a resynch is 2 clocks which will not affect the window because the Hxc source and the synch source are both crystal controlled.

-Charles

Post Reply