• Please review our updated Terms and Rules here

ISA USB board

I'm not using image file's, I'm using FDISK to Partition and Format the drive's as normal, Works fine with original Driver V1.9A, How exactly are you creating the Image's ?
I create an image file, use dosbox or VirtualBox to partition and format it and then I burn the image file to the USB stick ( since I don't have any 512M sticks).

Attached are speed test results. One is my MFM drive for comparison. One is the D2 driver and the other is the D4 driver. (D4 is faster than the D2)

MFM:
IMG_20220517_140834__01__01_800x600.jpg

D2:
IMG_20220517_141017__01_800x600.jpg


D4:
index.php


Original 1.9A:
IMG_20220517_145137__01_800x600.jpg
 

Attachments

  • IMG_20220517_141752__01_800x600.jpg
    IMG_20220517_141752__01_800x600.jpg
    123.3 KB · Views: 151
Last edited:
I create an image file, use dosbox or VirtualBox to partition and format it and then I burn the image file to the USB stick ( since I don't have any 512M sticks).
I remembered i had PCem on a laptop so used that, I created 2 image files one of 512M and the other 128M and used a 1gig flash drive, With the patched D4 driver i still get the same "CH375 Error Boot Sector", I switch back to the original V1.9A driver and both work fine.

I'm thinking it's my Lo-tech variant of the USB controller with the patched driver, It works fine with the original V1.9a driver, I can live with the slower W/R and messing with Image files is a pain in the ass and time consuming, I'd much rather use FDISK / Partition and format the drive and know it will work.
 
I remembered i had PCem on a laptop so used that, I created 2 image files one of 512M and the other 128M and used a 1gig flash drive, With the patched D4 driver i still get the same "CH375 Error Boot Sector", I switch back to the original V1.9A driver and both work fine.

I'm thinking it's my Lo-tech variant of the USB controller with the patched driver, It works fine with the original V1.9a driver, I can live with the slower W/R and messing with Image files is a pain in the ass and time consuming, I'd much rather use FDISK / Partition and format the drive and know it will work.
I used what's easiest for me.. (and i can dump the same images to multiple drives).. and because the docs said it was limited to 512M (and since these are all 8088 machines w/ DOS 6.22). I'm on a Mac not windows, so I have to use a VM for the fdisk/format, for ease. When i first tried this, I didn't know what would work and not work since the prior threads were all old and had minimal success. I'd rather have the speed .. the "messing with images' wasn't a pain nor time consuming.. it took me 5 minutes, and i have a template image now that has DOS on it, that i simply write to any usb stick and move on. (I have no need for > 512M on an 8088 machine, which is all I use these USB cards for).. When I looked at the Lo-tech cards eons ago.. and with nothing available and the bios having been 'alpha'.. that's why I went the aliexpress route. To each their own I guess. I'll be continuing to help test whatever mods/patches come about (including any possible patches to the BIOS since I like being able to boot from the USB stick(s)) instead of just treating it as a dump drive. These cards are cheaper and I can use any old usb stick.. whereas things like XT-IDE/CF need (getting harder to find) CF cards and cost twice as much (even to build) as this card. Something to note.. _unless you fix the geometry yourself_ it's wrong inside the driver and/or the bios, and you have to patch that yourself, it's locked to different geometry specs (which is probably what the 1.9A is using).
 
I create an image file, use dosbox or VirtualBox to partition and format it and then I burn the image file to the USB stick ( since I don't have any 512M sticks).

Attached are speed test results. One is my MFM drive for comparison. One is the D2 driver and the other is the D4 driver. (D4 is faster than the D2)
Yes, In D2 I optimized only the Read, In D4 I improved again the Read and Optimized the Write.

D2: Replace the Read sub function by direct I/O Read
D4: Replaced the Write sub function by Direct Write and Unrolled the Read Loop (Perform 2 Read per Loop and use STOSB Instead of MOD ES[DI],AL INC DI)

There is not sufficient space in the code to do the same for the Write.

Your Disktest software is interresting, where can I get it ?
 
I used what's easiest for me.. (and i can dump the same images to multiple drives).. and because the docs said it was limited to 512M (and since these are all 8088 machines w/ DOS 6.22). I'm on a Mac not windows, so I have to use a VM for the fdisk/format, for ease. When i first tried this, I didn't know what would work and not work since the prior threads were all old and had minimal success. I'd rather have the speed .. the "messing with images' wasn't a pain nor time consuming.. it took me 5 minutes, and i have a template image now that has DOS on it, that i simply write to any usb stick and move on. (I have no need for > 512M on an 8088 machine, which is all I use these USB cards for).. When I looked at the Lo-tech cards eons ago.. and with nothing available and the bios having been 'alpha'.. that's why I went the aliexpress route. To each their own I guess. I'll be continuing to help test whatever mods/patches come about (including any possible patches to the BIOS since I like being able to boot from the USB stick(s)) instead of just treating it as a dump drive. These cards are cheaper and I can use any old usb stick.. whereas things like XT-IDE/CF need (getting harder to find) CF cards and cost twice as much (even to build) as this card. Something to note.. _unless you fix the geometry yourself_ it's wrong inside the driver and/or the bios, and you have to patch that yourself, it's locked to different geometry specs (which is probably what the 1.9A is using).
Hi,

I used a 1Gb DOS 6.2 partition without problem, it seems there is no partition size limit.
The limit exist with the BIOS.
I still did not finish to decompile the partition detection code. (If you have infos about the geometry problems.... because I had issue on a 512Mb DOS 3.31 image but the 1Gb DOS 6.2 is Ok)
 
Last edited:
Your test are interesting, because the raw read/Write speed is x3 and as the access speed is improved as well, during my test (Copy a big file) I calculate an increase by 4.
It is normal, as the total copy time is Nb of acces x Access time + Read/Write time

Also, take into account that the IRQ are disabled during the Read/Write, so the calculation are correct if the tool perform the mesurement reading the Timer registers, and not the Timer IRQ Counter.

And what is the machine you are using ?
 
Yes, In D2 I optimized only the Read, In D4 I improved again the Read and Optimized the Write.

D2: Replace the Read sub function by direct I/O Read
D4: Replaced the Write sub function by Direct Write and Unrolled the Read Loop (Perform 2 Read per Loop and use STOSB Instead of MOD ES[DI],AL INC DI)

There is not sufficient space in the code to do the same for the Write.

Your Disktest software is interresting, where can I get it ?
It's linked prior in the thread.. it's made by pearce_jj: https://www.lo-tech.co.uk/wiki/DOS_Disk_Tester you have to be on the disk you're testing it seems, so I have it on the C drive and switch to D and just run it as say: C:disktest
 
Hi,

I used a 1Gb DOS 6.2 partition without problem, it seems there is no partition size limit.
The limit exist with the BIOS.
I still did not finish to decompile the partition detection code. (If you have infos about the geometry problems.... because I had issue on a 512Mb DOS 3.31 image but the 1Gb DOS 6.2 is Ok)
the geometry bits are mentioned earlier in this thread (and the other couple threads about it).. but they were mainly in the optional bios so you can boot from this (which I like to do, but it's easy to tweak)

And what is the machine you are using ?
a PC's Limited Turbo XT (8088-2 Clone) and an IBM 5150, though the tests were done on the PC's Limited box.

There is not sufficient space in the code to do the same for the Write.
Not enough space in the code? Aren't you disassembling and reassembling?
 
Last edited:
I used a 1Gb DOS 6.2 partition without problem, it seems there is no partition size limit.
The limit exist with the BIOS.
I still did not finish to decompile the partition detection code. (If you have infos about the geometry problems.... because I had issue on a 512Mb DOS 3.31 image but the 1Gb DOS 6.2 is Ok)
I have used a 1Gb, 2Gb DOS 6.2 partition with no problems with the original V1.9A driver, Also have used a 4Gb Flash drive with 2Gb partitions with no problems.

Are you using Image files ?, Or just FDISK / Partition and format the drives ?
 
I have used a 1Gb, 2Gb DOS 6.2 partition with no problems with the original V1.9A driver, Also have used a 4Gb Flash drive with 2Gb partitions with no problems.

Are you using Image files ?, Or just FDISK / Partition and format the drives ?
I wrote an existing compact Flash image file directly to the USB Key
 
Hi,

I disasembled the driver completely, so I will be able to change everything I want.
 
Oh don't keep me in suspenders, Have you still got to use Image's ?
It is working as before, but faster.
We can do anything we want with a little time, like mount all the partitions.

I can't find anymore the XT IDE BIOS source code, to update it....
 
It is working as before, but faster.
We can do anything we want with a little time, like mount all the partitions.

I can't find anymore the XT IDE BIOS source code, to update it....
What do you need that source code for? It's still out there but that's not the BIOS that was being talked about. There's a bios specifically for this card.
 
It is working as before, but faster.
We can do anything we want with a little time, like mount all the partitions.
Can you post a pic of your ISA - USB adapter ?

@sorphin Can you also post a pic of your ISA - USB adapter ?

If you had CD's with your adapters does it contain the schematic's of your adapters and do they match the adapter's you have ?.

I have been looking around and there seem's to be 2 or 3 different variations of these Chinese ISA - USB adapters going around.
 
What do you need that source code for? It's still out there but that's not the BIOS that was being talked about. There's a bios specifically for this card.
The BIOS for the CH375 Use a unique disk geometry and XT IDE can adapt to any, so analyze the XT IDE code is needed to understand how they do. Then, rewrite/Adapt the existing BIOS to have it standalone or integrate it to the XT IDE BIOS
 
Back
Top