Cartridge Port: Would a buffer board help with problems?

[stephen_usher]

The problems I had yesterday with the STE cartridge port (and previous problems on the TT) I was wondering if a solution to this would be a unit which sit in the cartridge port and did the following:

• Only switched on the +5V only when the voltage was stable (zenner diode?)
• Implemented buffering of all address and data lines
• Replicate the cartridge edge connector

The first would prevent any chips on-board coming up in a non-functional state due to under-voltage. Maybe add a status LED so show that the port is getting power.

The second would clean up the signals on all the lines.

I’m guessing that originally Atari only envisaged the port would be used for ROMs.

[exxos]

Putting a buffer there will load the bus anyway, so you end up with a buffered unstable bus. so its not a solution, the pull ups need replacing to fix the problem as I said in your other thread.

[stephen_usher]

So, would there be a way of doing this using an external board?

[exxos]

You could pull ups on the cartridge board. Though its nothing I have ever tried.

[frank.lukas]

I have a PCB Set for a romport buffer with 5 ALS645 and one ALS08. It is a old one from the c´t magazine ...

komplett_buffer.jpg (119.13 KiB) Viewed 1019 times

[stephen_usher]

Did it help with misbehaving peripherals, such as Spectre GCR and NetUSBee?

[frank.lukas]

I think not ...

For me it works fine with a Cubase Dongle and Logic Emagic Dongle and Netusbee. But with a Atari TT who the Netusbee are not working it helps nothing.

[stephen_usher]

I wonder what's causing the NetUSBee to fail on many TTs. Oh well.

[stephen_usher]

Well, I managed to get a 'scope onto some of the lines on the NetUSBee.

The data bus signals look nice and square. On the address bus, however, although many are nice and square others look like shark fins, just about reaching 5V by the end of the pulse. Those aren't going to work well! So, yes, it looks like a pull-up problem.

I bet your bottom dollar this is the reason why it (and the Spectre GCR) don't work on the TT too.

I'm wondering how feasible it would be to add pull-up resistors into the address lines on the cartridge. I've no way of making a PCB with a male edge connector and female edge connector.

P.S. I'm guessing that the old PSU had slightly higher voltage on the +5V rail which meant that the voltage rose more quickly, getting to the TTL high value fast enough for things to just about work.

[stephen_usher]

This is how shockingly bad the address bus is on the STE cartridge port... (Triggered on falling edge.)

IMG_1499.jpg (75.06 KiB) Viewed 913 times