I have one of those usb-to-true-LPT adapters. Works quite well, but needs Windows drivers. I doubt it would work in DOS. I haven't thought of using it for Covox, but may take a stab at it perhaps next week. Sorry, I'll be at work all week-end and a few days at the beginning of the week before getting any time off.

stamasd wrote:

I have one of those usb-to-true-LPT adapters. Works quite well, but needs Windows drivers.

Which one, which chipset does it have ?

Nostalgia! I did a writeup on xkcd forums nearly nine years ago:
http://forums.xkcd.com/viewtopic.php?f=36&t=2 … =944602#p944602

Stiletto wrote:

Nostalgia! I did a writeup on xkcd forums nearly nine years ago:
http://forums.xkcd.com/viewtopic.php?f=36&t=2 … =944602#p944602

Great overview !
The MCS7715 is still sold, and a new chip AX78140 was recently introduced.
http://asix.com.tw/products.php?op=ProductLis … =74&PSeries=109
-->will check these out!

I think this, along with a timing-accurate USB serial port, could be very interesting for legacy (industrial) applications.
There are two problems with modern hardware:
1) As you say, most LPT/COM ports are designed as buffered devices, and as such, have poor timing accuracy.
2) Running a modern multitasking OS adds more inaccuracy with timing.

So there are a lot of legacy hardware devices that need to be interfaced through 'real' ports, the non-buffered ones, and need to run under real DOS, not in some kind of DOS-emulation environment under a modern OS.

I used to work at a company that specialized in offshore surveys, and designed their own hardware and software for sensors and positioning etc. We had designed our own device, which was a microcontroller with 8 COM ports, that could be connected to an external time source (a pulse-per-second device), such as a GPS receiver (all GPS satellites have atomic clocks in them, so they are highly accurate time sources).
The device connected to a PC via ethernet.

For sending data to a device, you could either just send it in buffers, which would be output at the correct baud rate by the microcontroller. Or you could use the 'send on time' functionality, where you'd send a buffer of data ahead of time, together with a timestamp when the microcontroller should send it, using its internal clock, calibrated to the PPS signal.

For incoming data, the device would timestamp every single byte received, and send them back over ethernet. This allowed super-accurate logging of data, completely independent of what the PC was doing.

Just random background info, but perhaps it gives you some ideas.

dreamblaster wrote:

stamasd wrote:

I have one of those usb-to-true-LPT adapters. Works quite well, but needs Windows drivers.

Which one, which chipset does it have ?

TBH I don't know. I got it about 2 years ago off ebay- it's a Chinese clone, but works well with Willem's programmer (also a Chinese clone) using the USB2LPT driver. I haven't opened it up, but may have a look inside in the next day or two.

dreamblaster wrote:

2/ Convert my old parallel port BUNG gameboy cart programmer to an usb device
3/ Reviving a friend's old dos based parallel port chip programmer, so he can use it on his laptop (which has no parallel port).

I've got a Willem programmer, and there are two problems with USB to parallel adapters:

• Most adapters are not bidirectional. I don't remember if willem needs bidirectional support, but some devices (scanners) will fail.
• Although Willem programmer software is windows based, the LPT addresses are hard-coded. Some PCI to parallel cards use alternative port addresses, so the software must be patched (there is a method involving an alternative portio.dll). I guess that USB to parallel converters will fail, too.

Zup wrote:

Some PCI to parallel cards use alternative port addresses, so the software must be patched (there is a method involving an alternative portio.dll). I guess that USB to parallel converters will fail, too.

What is that portio.dll, is it a specific patch for the willem programmer ?

It's a DLL that allows (or emulates) direct access to hardware (something forbidden on Windows 2000, XP and later). Willem uses it to access LPT port directly.

There is an alternate (or maybe a wrapper) DLL that "translates" between addresses, so a write to port 0x378 may be redirected to 0x9800 (an address that one of those PCI cards can get).

Note that this alternate DLL only works on programs that use portio.dll. Dosbox MB6 (and Daum) uses porttalk instead (that dll seems to be able to redirect ports) but I don't know what DLL uses Dosbox SVN (if it has LPT passthrough).

Also, I don't know if those USB converters are presented as LPT (device) or are presented as a hardware address. Maybe you won't be able to use them to drive that EPROM programmer.

Sorry, but I was wrong.

The DLL used by willem programmer is io.dll.I don't know why I had that "portio.dll" thing stuck on my mind.

Yes I had to use a similar method to get the Willem working with the USB/LPT adapter in Win7. But it proved to be troublesome so eventually I delegated one of my vintage laptops with Win98 and a real LPT port to eprom-burning duty. That's why I haven't been really using the USB/LPT adapter for a while. To be honest I don't even know where it is now. 🙁

the CY7C68013A-56 EZ-USB dev board arrived... )
looks good.

The Willem should work on linux without any problems.
Maybe give it a try 😀

i also read the original site above some time ago.

I did passtrough of PCI LPT Card to KVM for a Friend.
This worked for me and he can use his Devices under VirtualMachine on Linux.
Keeping some old Medical Equipment alive 😀

Hi
check this thread for the board based on these experiments : USBLPT :
USBLPT : a 'real' parallel port on USB