Re: Are all PS/2 to USB splitter adapters bad for old PS/2 keyboards?

On Thu, 7/9/2026 2:00 AM, Ant wrote:
> Hello,
> 
> Are all PS/2 to USB splitter adapters bad for old PS/2 keyboards? I have 
> two different brands PS/2 to USB splitter adapters for my old PS/2 
> keyboards & mouses for two different old home built PCs (updated Windows 
> 10 Pro. and Linux/Debian trixie stable [even in its text mode]). I 
> didn't want to buy new mouses and keyboards since I have many old ones 
> to use. I'm a fast typist on clicky keyboards, but I noticed results are 
> not fast and good. When I hold a letter/number down and then press a 
> Shift or Caps Lock key to show its opposite, the repeating results just 
> stop showing (nothing show anymore). This is annoying when gaming like 
> when I use W to move forward and want to use shift key. When I type 
> fast, I get incorrect, missing, or duplicate characters.
> 
> It's very annoying! PS/2 mouses are fine though! Are splitters not good 
> and I have to get USB PC104 keyboards? :( Thank you for reading and 
> hopefully answering soon. :)
> 

   computer --- USB ----- PS/2 keyboard       # Something in the path does 8042 emulation ?
                    ----- PS/2 mouse

There might not be many implementations of this silicon.
It would appear as a composite device with two HIDs.

There was one chip to do this, with mask ROM on the top layer
of the chip. The first batch of chips were "defective" in that
the device stopped responding after ten minutes or so.

The next batch (with new mask ROM on the top layer), presumably
fixed this issue.

Versions were made, for PCI bus cards, consisting of a USB host
chip plus the DIP packaged chip that does the conversion. By the time
those were coming out, they would have the newer mask ROM on them.

It's about as spooky an implementation as a KVM. Whatever is in the
mask ROM, determines the response. If they had used an EEPROM inside
the device, so that it could be flash upgraded, then it would
likely have been a champ. If they had put an I2C chip interface
on it, for an external EEPROM to be connected, there would similarly
be a means to update how it runs inside.

Just for contrast, I have a Griffin chip with Apple legacy AppleDesktopBus(ADB)
interface on one end and USB on the other end, for a USB host computer.
This allows legacy Apple keyboards to be used on a PC. The "Power" button on
the keyboard, serves no function on a PC, but everything else works. And
that device seems to have more on-the-ball than the PC-ecosystem chip.

Summary: Just buy a USB keyboard. With the various clicky keys.
         There is never going to be a perfect adapter for the job.
         And the reason, is the inferior (cheap) silicon implementation.
         It should have had an EEPROM on it, to "rescue" the badness of the design.
         There is a processor inside the chip. And the mask firmware is how it does
         whatever it does for 8042? emulation.

   Paul