exxos's Atari, Amiga & retro forum

stephen_usher wrote: 26 Feb 2025 21:52 Well, I was out all day today and, of course, all the post, including the chips arrived.

At least you know how to ensure things are delivered in future... ;)

JezC wrote: 26 Feb 2025 21:59

stephen_usher wrote: 26 Feb 2025 21:52 Well, I was out all day today and, of course, all the post, including the chips arrived.

At least you know how to ensure things are delivered in future... ;)

:lol:

OK. Lifting pins 2 and 6 (RX and TX) makes no difference. All the other pins other than 8, 9 and 10 are common between the two 6850s and buzz out. The only difference is that CS0, CS1 and CS2 are slightly different They all buzz out as well.

Keyboard & Midi have pin 8, CS0, in common. A2 is connected to CS2 of the keyboard and CS3 of the MIDI, Keyboard CS1 is Vcc and MiDI CS2 is grounded.



So, from what I can see nothing else in the circuit could possible be causing an issue, yet the MIDI 6850 is pulling the interrupt line low (pin 7 on both 6850s) constantly. If I lift that pin the system works.

I'm guessing that the CS pin combinations are used to select a mode, such as writing to registers, and something is not working when it's trying to access the MIDI 6850... but then why is it not affecting the Keyboard chip which uses all the same control lines (plus an extra one)?

Stumped. Anyone any ideas?

I guess that I'm going to have to get the 16 channel logic analyser out and hook it up to all the bus side inputs and outputs of U70. I'm guessing that a register isn't being set properly, causing the chip to raise an interrupt which is never being cleared by reading or writing to a register.

stephen_usher wrote: 26 Feb 2025 21:52 Well, I was out all day today and, of course, all the post, including the chips arrived.

Anyway, just been out to the outhouse and did a test with the 6850s... Well, well! That's interesting!

Install a working 6850 into the MIDI socket and the keyboard stops working. This suggests that something downstream from the 6850 is causing the chip to raise its interrupt line. I'm going to have to look at the schematic.

Oh, how strange. So it works fine with no MIDI 6850 installed?

Bit of a head scratcher there, be interesting to find the cause.

Yeah, it's definitely going to need a logic analyser clipped to the legs, plus a pull-up on pin 7 lifted as I need to see what this chip is doing rather than what the keyboard 6850 is doing, which shares the trace... and it's open collector.

Oh, and the second Sugon A9 solder station arrived today. I needn't have ordered the other one as the only difference between the 110V and 240V model is the kettle lead. And the difference between the A8 and the Pro version? The standard version has a sponge and the Pro has a brass brush for cleaning the solder.

I can just send it back but if anyone is interested in buying it for, say, £60, then I'll just pass it on.

In the meantime, I fitted a 74LS11 and changed the solder jumpers so I could fit an Exxos dual TOS board to get TOS 1.4 on the machine.

Couldn't get it to work...

Could see the output of the LS11 on the jumper but not on pin 20, just the non-jumpered /CE on there...

Eventually found that it was the ROM-2 sockets which had the 1MB /CE connected by trying to find pins which where connected to the jumper.

Obviously Rev B. boards are odd.

Logic analyzer all set up... run a test and...

THE FLIPPIN' LA IS GENERATING SPURIOUS INPUT!!!!! AAAAAARGH!!!!


/CS2 and Channel 15 are both connected to ground. I even had my oscilloscope connected to the Channel 15 clip and it was a nice, flat ground.

If you can't trust your test equipment then you're sunk.

stephen_usher wrote: 28 Feb 2025 11:48 Logic analyzer all set up... run a test and...

THE FLIPPIN' LA IS GENERATING SPURIOUS INPUT!!!!! AAAAAARGH!!!!


Screenshot 2025-02-28 11-39-52.jpg


/CS2 and Channel 15 are both connected to ground. I even had my oscilloscope connected to the Channel 15 clip and it was a nice, flat ground.

If you can't trust your test equipment then you're sunk.

Oh, that's almost as annoying as your postal service!!

ABRAIL25 is over, and so are the logistics. Phew!

Anyway, decided to replace the socket on my CPU switcher/relocator board in my accelerator STFM, which meant desoldering the connection to the motherboard. Knowing how fragile the traces are on the ST I tried desoldering on the top. Mostly went well other than one of the turned pins decided to break off. So, once the relocator board was out I had to clip around the damaged pin and desolder/remove that and replace it with a spare one.

With that done the removal of the old turned pin socket was done by removing the cross bracing and desoldering each side in turn until I could wiggle them out. Refitting a new dual-wipe socket was easy and the final reassembly onto the motherboard pins was OK too... All worked except that the TF536 was still flaky unless it was lifted in one position. The solder joints on the connector all looked OK but I reflowed and miraculously it started working perfectly. Ran for hours with GEMBENCH from ST RAM.

With that I decided to try the Pi Storm again... Well, that's a big fat nope. Strangely the ataritest --memory tests=r caused screen memory to change when it was doing its reads and the R/W line being high(ish) and noisy. That shouldn't even be able to happen. Anyway, I've given up on that again.