Previous Thread
Next Thread
Print Thread
Page 1 of 3 1 2 3
Bondwell BW-14 #116479 11/24/19 04:28 PM
Joined: Jan 2012
Posts: 1,028
rfka01 Offline OP
Very Senior Member
OP Offline
Very Senior Member
Joined: Jan 2012
Posts: 1,028
I received a Bondwell BW-14 from a friend and dumped its ROMs since it has a German keyboard. The character ROM indeed is different, I've integrated it into the driver and submitted a PR. The archive I've uploaded to the FTP also contains photos from my machine and some schematics I found online.

It would be great if one of the Devs could give the driver a look (it's marked as a Skeleton in the comments section but not in the driver description).
The PC I'm using for MAME has plenty of power (32G RAM, i7 4790k CPU).

Some observations:
- the ROM based parts (initial system check, reset) happen in "real time", the reset is almost too quick as on the real iron it shows two distinct pages of ASCII patterns
- there are two major delays during the boot, one after CP/M 3.0's "54K TPA" message (ca. 75sec), the other after the Digital Research copyright message on the next page (ca. 17sec.)
- the emulated BW-14 will boot from the software list boot disk but neither from the direct dump of the disk that came with my machine or from one I reconstructed using the software list boot disk and mine
- if I convert the softwarelist boot disk to .MFM using the HxC tools, the emulated BW-14 doesn't boot from it
- There's a 20s delay before the directory of a newly inserted disk in A: or B. is displayed, haltingly.
- formatting a disk in the emulated BW-14 works, but is very slow. When the real hardware formats the disks, then verifies it, verifying a track takes about 1sec, in the emulation it's ca. 17sec
- copying files (using PIP) shows the first filename, but then nothing more happens ... in the Debugger you can see that the machine is still alive, but doesn't proceed.




Last edited by rfka01; 11/24/19 04:31 PM.

NCR DMV- DEC Rainbow- Siemens PCD- ITT 3030-Oly People- Acorn A5000- Olivetti M20
Re: Bondwell BW-14 [Re: rfka01] #116482 11/25/19 01:53 AM
Joined: Jan 2012
Posts: 814
C
crazyc Online Content
Senior Member
Online Content
Senior Member
C
Joined: Jan 2012
Posts: 814
Quote
- if I convert the softwarelist boot disk to .MFM using the HxC tools, the emulated BW-14 doesn't boot from it


The rom appears to expect sectors to be numbered from 0 but hxc messes this up and renumbers numbers them from 1 instead.

Quote
- there are two major delays during the boot, one after CP/M 3.0's "54K TPA" message (ca. 75sec), the other after the Digital Research copyright message on the next page (ca. 17sec.)


It's reading the disk through this whole period so I don't know why it would be much slower than the real hardware if it is. It reads one sector at a time which is going to be rather slow.

Quote
- the ROM based parts (initial system check, reset) happen in "real time", the reset is almost too quick as on the real iron it shows two distinct pages of ASCII patterns


The ASCII pattern does appear in the emulation but it's so fast it can't really be seen.

Quote
- copying files (using PIP) shows the first filename, but then nothing more happens ... in the Debugger you can see that the machine is still alive, but doesn't proceed.


Trying to write to one of the cpm disks it overran the index while writing the last sector which confused the emulation so it gave up. The sectors are very closely packed on that disk but I don't know if the cpm floppy was specially written to pack extra data or our cqm image loader makes the intersector gap too large.

Re: Bondwell BW-14 [Re: rfka01] #116483 11/25/19 06:01 AM
Joined: Jan 2012
Posts: 1,028
rfka01 Offline OP
Very Senior Member
OP Offline
Very Senior Member
Joined: Jan 2012
Posts: 1,028
You can see my BW-14 booting CP/M 3.0 and loading WordStar here:

https://i.imgur.com/BGjskX4.mp4

The 75 and 17 sec delays in MAME seem a bit excessive smile

18 sectors per track is cramming things rather tightly. As for gap sizes, this is what IMD comes up with:

[Linked Image from i.imgur.com]

You can watch the output of the probe2.com utility here: https://i.imgur.com/SM5SaM0.mp4

The 22disk definition (which appears to work) is

Code
BEGIN BON2  Bondwell 14 - DSDD 48 tpi 5.25"
DENSITY MFM ,LOW
CYLINDERS 40
SIDES 2
SECTORS 18,256
SKEW 2
SIDE1 0 0,1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,16,17
SIDE2 1 0,1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,16,17
ORDER CYLINDERS
BSH 4 BLM 15 EXM 1 DSM 174 DRM 127 AL0 0C0H AL1 0 OFS 2
END


NCR DMV- DEC Rainbow- Siemens PCD- ITT 3030-Oly People- Acorn A5000- Olivetti M20
Re: Bondwell BW-14 [Re: rfka01] #116484 11/25/19 01:06 PM
Joined: Jan 2012
Posts: 814
C
crazyc Online Content
Senior Member
Online Content
Senior Member
C
Joined: Jan 2012
Posts: 814
Originally Posted by rfka01

SKEW 2


Interleave probably makes a big difference. The cqm images are loaded with the sectors in order. Since the machine reads one sector at a time it always misses the next sector and has to wait for a whole disk rotation to do the read.

Re: Bondwell BW-14 [Re: rfka01] #116485 11/25/19 01:28 PM
Joined: Jan 2012
Posts: 1,028
rfka01 Offline OP
Very Senior Member
OP Offline
Very Senior Member
Joined: Jan 2012
Posts: 1,028
What's the interleave used to create MFM/HFE images from within MAME?

Last edited by rfka01; 11/25/19 01:29 PM.

NCR DMV- DEC Rainbow- Siemens PCD- ITT 3030-Oly People- Acorn A5000- Olivetti M20
Re: Bondwell BW-14 [Re: rfka01] #116486 11/25/19 02:27 PM
Joined: Jan 2012
Posts: 814
C
crazyc Online Content
Senior Member
Online Content
Senior Member
C
Joined: Jan 2012
Posts: 814
MAME just uses the sector order from the image. The sectors in the CQM images from the softlist have no interleave. The IMD images from the upload dir have the correct interleave and the first part of the boot is very quick. The second part is slow because it pauses for a few revolutions before reading each sector. The boot then hangs because it tries to write to the disk.

Re: Bondwell BW-14 [Re: rfka01] #116487 11/25/19 02:30 PM
Joined: Jan 2012
Posts: 1,028
rfka01 Offline OP
Very Senior Member
OP Offline
Very Senior Member
Joined: Jan 2012
Posts: 1,028
I mean, if I'm formatting an image from within the emulation (as I want to get a writable disk image) and chose MFM or HFE, what's determining the interleave? Is it hardcoded in MAME's image handling or does the FORMAT utility get a say?


NCR DMV- DEC Rainbow- Siemens PCD- ITT 3030-Oly People- Acorn A5000- Olivetti M20
Re: Bondwell BW-14 [Re: rfka01] #116488 11/25/19 02:35 PM
Joined: Jan 2012
Posts: 814
C
crazyc Online Content
Senior Member
Online Content
Senior Member
C
Joined: Jan 2012
Posts: 814
MAME will preserve the interleave if the image format supports it and the image save routine handles it properly. From a quick glance MFM format saves whole tracks at once so the interleave shouldn't be touched.

Re: Bondwell BW-14 [Re: rfka01] #116489 11/25/19 02:43 PM
Joined: Jan 2012
Posts: 1,028
rfka01 Offline OP
Very Senior Member
OP Offline
Very Senior Member
Joined: Jan 2012
Posts: 1,028
Thanks, I'll try to format a MFM image tonight and PIP the stuff from my German boot disk over ...


NCR DMV- DEC Rainbow- Siemens PCD- ITT 3030-Oly People- Acorn A5000- Olivetti M20
Re: Bondwell BW-14 [Re: rfka01] #116491 11/25/19 10:04 PM
Joined: Jan 2012
Posts: 1,028
rfka01 Offline OP
Very Senior Member
OP Offline
Very Senior Member
Joined: Jan 2012
Posts: 1,028
Here's the BW-14 booting the same disk (image) on real hardware and in emulation. I'm formatting a MFM disk image in the emulation and a physical floppy disk on the real BW-14. I've stopped recording the verification operation in MAME after three tracks smile

https://mega.nz/#!WFZnXYBY!edLcbH2dHHrFKfCpAzOyhDinaIH0nklf91b7ibI14M4

https://mega.nz/#!CMQRSIjR!twVYn_m3MVClTdErI0a0WQDKpmw1Z564whJH1h-L-Yo


NCR DMV- DEC Rainbow- Siemens PCD- ITT 3030-Oly People- Acorn A5000- Olivetti M20
Page 1 of 3 1 2 3

Who's Online Now
1 registered members (Doc Flareon), 148 guests, and 1 spider.
Key: Admin, Global Mod, Mod
ShoutChat Box
Comment Guidelines: Do post respectful and insightful comments. Don't flame, hate, spam.
Forum Statistics
Forums9
Topics8,722
Posts114,638
Members4,873
Most Online510
Aug 26th, 2019
Powered by UBB.threads™ PHP Forum Software 7.7.3