SDHxC to emulate a IBM3740 Format?

[Frank]

Hi Jeff,
i have difficulties to emulate the original IBM3740 Format.

My Settings:
IBM FM / One Side / Auto Gap / 77 tracks
Bitrate : 500000
26 x 128Byte sectors
360RPM
Starting sector: 1

The Diskcontroller i use is a FD1771-01.
When i operate the CP/M system with real 8" drives (NEC FD1165), all is working properly.
An created empty floppy using the above parameters is not readable (sector not found).

I have no idea what is the problem.

Regards, Frank

[Jeff]

Can you try to dump the disk with imagedisk ? :
http://www.classiccmp.org/dunfield/img/index.htm

also : what is the system using this floppy disk drive ?

[Frank]

Hi Jeff,
the image, what i want to use is an Dave Dunfield *.imd image (use BB1.imd).
When i extract those images to a 8" disk, all is working properly.
The system that i have in use, is an Ferguson Big Board 1.

Frank

[Jeff]

Hi Jeff,
the image, what i want to use is an Dave Dunfield *.imd image (use BB1.imd).
When i extract those images to a 8" disk, all is working properly.
The system that i have in use, is an Ferguson Big Board 1.

Frank

If these IMD image work on a real 8" disk with your system, convert them with the hxc software should do the trick.
IMD files are directly supported by the hxc software.

What is the behaviour of the HxC with this board ? (Led / LCD behaviour).

[Frank]

Hi Jeff,
that is what i have done.
But it does't work with all the linked *.imd files.
The firmware is updated to the newest version, you provide on your site.
When i compare the data of the sectors with your new "Track Analyser", all the content of the sectors is OK.

Frank

[Frank]

Hi Jeff,
attached some pictures from the different timing of the real 8" drive and the SDHxC.
Both uses the same imd file as source.

[Frank]

And here the pictures from HxC.

[Jeff]

It seems that the VFO/data separator work in both case.
Regarding the data width difference , this probably because the HxC use a larger data pulse than the floppy disk drive.
Where have you took these measurement ? Right at the 1771 pins ? Can you take the VFO output clock too ?
what is the width of the floppy disk drive data pulses ?

[Jeff]

It seems that the VFO/data separator work in both case.
Regarding the data width difference , this probably because the HxC use a larger data pulse than the floppy disk drive.
Where have you took these measurement ? Right at the 1771 pins ? Can you take the VFO output clock too ?
what is the width of the floppy disk drive data pulses ?

Well, after a quick look at the 1771 / FD1165 datasheet and Ferguson Big Board schematic, i think that i have found the cause of the issue :

In external VFO mode the 1771 wait data and clock pulses from the VFO with a width under 350ns (150ns<>350ns).
The FD1165 generate data pulses with a width under 250ns (200ns +/- 50ns).
all ok... but:

The problem is that the Ferguson Big Board don't have the monostable circuit generaly used at the data input. So if the the data width pulses are wider than 350ns, the 1771 is not supposed to work correctly.

So the possibilites to get this work are :

- add a monostable circuit to reduce the data pulses under 350ns.
( this maybe done by a simple resistor/capacitor circuit :
http://www.premiumorange.com/daniel.rob ... stable.gif)

Or

- try to bypass Ferguson Big Board VFO and turn 1771 in internal VFO/data separator mode (XTDS = 1). It can support pulses up to 1000ns in this mode.

[Frank]

Hi Jeff,
thank you for your answer.
One of the very first tryouts was to use the internal data separator in the FD1771. This doesn't work.
I give the monostable multivibrator (LS123) a chance on the weekend, if there is time.
Is there a chance to implement a variable pulse-pause output parameter inside in the SDHxC?
The board i don't want to modifiy and leave it as it is.

Frank

[Jeff]

I give the monostable multivibrator (LS123) a chance on the weekend, if there is time.

Yes this problably the best choice. this circuit was commonly used to make this job on some FDC board. I will try to find an schematic.

Is there a chance to implement a variable pulse-pause output parameter inside in the SDHxC?

No, unfortunately this is not possible with the actual hardware. The USB version vhdl code maybe modified to reduce the pulse witdh.

[Frank]

Hi Jeff,
the shorting of the pulswide is working now over a 74LS123 up to about 450ns.
I am now able to read the contents from a "empty disk", created with the right parameters (all tracks, all sectors).
When i export the BB1.imd over "LOAD" and the export it to the SD-card, then i am only able to read track 0.
With the right parameters and "LOAD RAW" the BB1.imd and then export to SD-card, i am able to read all tracks with the right content.
Is there a way to convert the *.imd files to a *.raw file?

Frank

PS:The file BBCPM.IMD is not loadable in the software, why? Creating a real disk with IMD.COM has no problem with the BBCPM.IMD.

[Frank]

When i "reimport" the BB1_IMD.hfe in the new 2.xx HxC-Software and look on the datas in the sectors,
they are indentically in track 0 and in track 1. Also in track 2 and 3 and so on....

Frank

[Jeff]

Hi Jeff,
the shorting of the pulswide is working now over a 74LS123 up to about 450ns.
I am now able to read the contents from a "empty disk", created with the right parameters (all tracks, all sectors).
When i export the BB1.imd over "LOAD" and the export it to the SD-card, then i am only able to read track 0.
With the right parameters and "LOAD RAW" the BB1.imd and then export to SD-card, i am able to read all tracks with the right content.
Is there a way to convert the *.imd files to a *.raw file?

Frank

PS:The file BBCPM.IMD is not loadable in the software, why? Creating a real disk with IMD.COM has no problem with the BBCPM.IMD.

First it's good to see that you managed to get working the emulator.

PS:The file BBCPM.IMD is not loadable in the software, why? Creating a real disk with IMD.COM has no problem with the BBCPM.IMD.

Just change the file header to "IMD 1.07:".

Is there a way to convert the *.imd files to a *.raw file?

yes : export them the img file format.

[Jeff]

When i "reimport" the BB1_IMD.hfe in the new 2.xx HxC-Software and look on the datas in the sectors,
they are indentically in track 0 and in track 1. Also in track 2 and 3 and so on....

Frank

mhh i don't reproduce the issue here. Have you the last software ?
Do you have activated the double step mode ? You shouldn't do this for this system.