Jeff wrote: ↑Tue Aug 14, 2018 10:56 am
A question to everyone present in this thread : Is the "Gotek" hardware is a must for you ?
I mean : Today there is definitively some more reliable and interesting solutions available. I am thinking about something designed around a Raspberry-Like solution.
BTW is there someone with a Raspberry Pi Zero (W) here ?
Jeff wrote: ↑Tue Aug 14, 2018 3:40 pm
Why low cost is important for you ? Do you plan to resell them ? Or plan to install thousand units somewhere ?
Neither. But, I do have 4 systems to update with support for hard-sector emulation. At 30 USD, the Gotek is a great deal. At 90 USD, the Lotharek would be prohibitively costly. I would like to spend no more then 50 USD on hardware if possible. Maybe it's not possible, but you did ask .
Jeff wrote: ↑Tue Aug 14, 2018 10:56 am
A question to everyone present in this thread : Is the "Gotek" hardware is a must for you ?
I mean : Today there is definitively some more reliable and interesting solutions available. I am thinking about something designed around a Raspberry-Like solution.
BTW is there someone with a Raspberry Pi Zero (W) here ?
The network support open the solution to tons of features. And the huge RAM quantities available eliminate the flash latency problem and make all image type support possible.
The in-middle solution is to emulate an USB stick with a raspberry-zero with wifi support. Already tried and this is working pretty well.
Hi Jeff
Did you try responding with a corrupt data a couple times? I'd think this would give you the time needed. You do need to keep the sector pulses going while doing other book keeping. Is that the problem?
Dwight
dwight wrote: ↑Fri Aug 24, 2018 11:11 pm
Hi Jeff
Did you try responding with a corrupt data a couple times? I'd think this would give you the time needed. You do need to keep the sector pulses going while doing other book keeping. Is that the problem?
Dwight
No i didn't. As i explained a specific firmware is needed. Currently the sector data flow is giving rythm to the index signal. To fake some sectors the firmware need to be modified : the index must give rythm to the sector data flow. This may sound easy to do but in fact all the pipeline need to be changed.
BTW Will all available software retry with an read error ? I ask because this is not always true on others systems.
Any access that goes through the bios will normally expect a possible error, at least on the first read.
I can't say that 100% of the software will do that.
I do hope you are stretching the timing between sector pulses. it isn't a lot of time but still some slack.
Dwight
I posted the maximum time you could delay a sector pulse. It is noticeably longer than a normal running floppy would be.
The hardware requires a continuous stream of sector pulses and index pulse. There is some slack on the timing but not a lot. That is why I asked if you were using timers to create the pulses.
Dwight
dwight wrote: ↑Tue Sep 18, 2018 9:12 pm
I posted the maximum time you could delay a sector pulse. It is noticeably longer than a normal running floppy would be.
The hardware requires a continuous stream of sector pulses and index pulse. There is some slack on the timing but not a lot. That is why I asked if you were using timers to create the pulses.
Dwight
Not currently : The index is recorded in the HFEv3 track stream. The current issue is that when the machine is stepping a bunch of tracks, some index may be delayed or skipped due to the flash access latency.
So as i explained i need to change the firmware design : the sector/index timer must be used as trigger to output the next sector.
I've been following this thread on and off over the months, but it's a long thread and hard to follow at times. I was just wondering if this is at a stage where it works reliably enough to try on my SOL-20?
namachari wrote: ↑Sat Dec 08, 2018 10:07 am
I've been following this thread on and off over the months, but it's a long thread and hard to follow at times. I was just wondering if this is at a stage where it works reliably enough to try on my SOL-20?
Thanks
Philip
Have already tried the current firmware with your SOL-20.
Hi
The Sol-20 didn't have an official controller. Most used a N* controller. Still, the software is different and may not be as ready to flag an error on a late sector pulse. Jef code works well for the data part. It is possible that the software written for the SOL-20 doesn't even pay attention to the sector error signal.
Dwight
Hi - a new guy here (actually a really old guy!) also looking for the best possible NorthStar disk data extraction for the HxC. May I ask the current status? Anything changed since this thread was last active? I just ordered a unit and am putting together an 'old' Pentium system which will hopefully have 3 1/2", 5 1/4", and 8" drives under MSDOS or FreeDos as well as Win98.
And, BTW, while many Sol-20 users did use a NorthStar controller, I would say that the Helios system with 8" PerSci drives would be the 'official' system though I am not sure what controller board was actually used.
rstek37 wrote: ↑Sun Jun 23, 2019 9:20 pm
Hi - a new guy here (actually a really old guy!) also looking for the best possible NorthStar disk data extraction for the HxC. May I ask the current status? Anything changed since this thread was last active? I just ordered a unit and am putting together an 'old' Pentium system which will hopefully have 3 1/2", 5 1/4", and 8" drives under MSDOS or FreeDos as well as Win98.
And, BTW, while many Sol-20 users did use a NorthStar controller, I would say that the Helios system with 8" PerSci drives would be the 'official' system though I am not sure what controller board was actually used.
Bob Stek
Save of Lost Sols
What do you want to do exactly ? Floppy Emulation or floppy disk read/dump ?
Dutchacorn wrote: ↑Fri May 18, 2018 1:55 pm
I have quite detailled documentation with my Vista disk drive. Does this image help? It is hard sectored on 10 sectors / track.
I dumped 5 off hard sector Vista floppies for the Exidy Sorcerer today using a Kryoflux.
As @Dutchacorn said above they are 10 sector with 512 bytes per sector, 40 track. Written in a single sided drive but with data on both sides and dual index holes. They seemed to read OK using a Mitsubishi M4853 drive although there is no Kryoflux profile to create an image and verify them.
I don't currently have a Vista controller but thought people might be interested in how close things are anyway even without explicit support.
I have access to another 15 discs.
HxC appears to be interpreting the geometry although the data itself is unsurprisingly not correct.