ZX-WESPI - add a cheap wifi-enabled fileserver to your Zeddy!

[msknight]

OK - this is going to complicate things.

The first zeddy I worked on, had problems with the wespi in terms of even when I'd put in the jack, the zeddy would crash. An issue 3. I had to employ the separation of ear and composite to do simultanious loading from either wespi or tape.

However... I was recording the next video in my series... and I was filming installing a wespi into an issue 1 board.

Now... here's the thing. It was doing the sometimes boot to K thing, so I replaced the 1uF cap.

The wespi behaved not only perfectly, but I didn't even need to silence the thing in order to play from tape. Plugging in and out the ear jack didn't crash the system, and when I typed load "" with the jack in, the wespi didn't actually load its speed loader, but it loaded from tape OK.

I'm starting to wonder whether there might be some mileage in replacing that 1uF cap and seeing if the solution behaves better and not needing the switch?

[1024MAK]

If you are talking about C5, that's the capacitor used in the reset circuit for the Z80. As it charges via a 220kΩ resistor from the +5V line, it has a really, really easy life.

A slight loss of capacitance will not make any difference to the operation of the circuit. Only if the electrolytic capacitor becomes electrically leaky (becomes more like a resistor) or suffers a significant loss of capacitance will it cause problems. These faults would prevent the Z80 from resetting at all, or cause it to enter the special reset mode.

And once the Z80 is running normally, you could (if it was practical) remove it and the Z80 would continue running without even noticing.

So, I can't see how it would crash the system (assuming the leads don't become shorted out).

If a ZX81 (or ZX80 or clone) crashes when either the Ear or Mic sockets have a plug inserted or removed, then I would be looking for dry solder joints, cracked/broken PCB tracks, poor socket contacts and most likely, the contacts on the 3.5mm power socket not making good contact with the plug.

Mark

[msknight]

Well, I got really excited and instead of carrying on with putting it in a new case, etc. I decided to reinstall the UHF can, etc. and put it in a proper case... and.... after all that... it reverted to its previous behaviour.

Drat. Drat. Double Drat.... and TRIPPLE Drat.

[msknight]

I do have it caught working... or not working... whatever... on film, so once that's rendered I'll tell you where it is and see if you can see what I've done wrong... or right... er....

[msknight]

OK - i took it back out of the case and reverted it to how it was in the video ... and it's "working" again. Plug the jack in, and it loads from "tape" instead of the wespi. I now don't know what the heck is causing this.

Going to get a cup of tea and think.

Video is still rendering. Will be uploaded later.

[msknight]

Right - the video is up - https://youtu.be/Wzld5uvLBLA?feature=shared&t=3214

Can anyone shed some light on this?

[msknight]

OK - with the zeddy turned on, at pin 20 of the ULA I see what appears to be noise.
With the jack plugged in and playing through the TXDuino, I'm seeing a strong signal which appears nice and uniform.
With the jack unplugged, the wespi gives another slightly weaker signal.
With the jack plugged in, the load command given, I can see the blue LED on the wespi flashing, so it looks like it's transmitting, but on pin 20, I still see the same noise.

I'm still an oscilloscope newbie, but that appears to be 500mV per vertical division? I did check the probe and scope and both are 10x.

[msknight]

OK - measuring at the line in part of the ear jack socket...
1) Playing the wespi loader with no jack inserted into the socket. Nice, clear divisions.
2) Playing the wespi loader with the TZXDuino plugged in, the pulses from the wespi just become a jumble of small peaks which, it's no wonder the zeddy can't load it.

The question is... what's causing it? And if we can replicate this on demand, then it would negate the need for the switch.

[1024MAK]

As we were discussing how the reset system works, I thought you may like to watch this video. This NOP test board only has a 220kΩ resistor connecting the /RESET input (to the Z80) to the +5V rail. There is no capacitor connected to it. And, as I have not found my stash of PCB mounting tactile buttons, no switch or button either...

See also this video.

So I use a piece of thin solid core wire in place of the reset button

Mark

[msknight]

OK - so the way I look at this is... whatever is happening on this zeddy is an opportunity to enable the zeddy to bypass the wespi and take from tape without having to install a switch.

This system is currently mangling the wespi signal whenever the TZXDuino is hooked up, to the point where the zeddy can't see it and takes the TZXduino instead... as per the traces in the previous post. If the jack is in on the zeddy end, but not at the TZXDuino end, then the wespi signal is fine and the zeddy loads from that.

However, when I removed everything and installed it in a proper case, it reverted to normal behaviour.

I have tried measuring the resistance from the TZXDuino, which is about 1.7Mohm, and installing that over a jack... but that didn't change the wespi signal. So I'm not sure where to go next.