I found something that looks similar to the crazy 24Mhz rig I have ended up with:



The author's description states that it outputs the fundamental frequency of the crystal; I assume this means 8mhz for my crystal (ie, perfect). In order to implement it I only need a pair of caps; I have the resistors and there is a spare inverter on the LS93 for the buffering.

Any comments..? Can anyone explain the effect of the capacitor / resistors on the output frequency in plain English, please?

Web page is here http://www.gaby.de/z80/uexosc.htm

That's a good find. I can't explain it, sorry. The capacitor to ground looks somewhat like a low-pass filter, I think. (Because high frequencies see capacitors like shorts, so they see the capacitor as a short to ground, whereas low frequencies see capacitors like opens, so they don't see the connection to ground.) The resistors, as I understand it, are just biasing the inverters to their mid-points so they act as sensitive amplifiers.

1uF is probably too big and in any event an electrolytic is probably mostly
useless at those frequencies. I'd leave the 1uF in and add a .1 directly across chips pins. I generally try to avoid ceramic discs, because you
can't be sure what you're getting (too often they're just junk as capacitors
go).
The 1uF should help and shouldn't hurt, but I've heard tell of weird
resonances causing problems in similar circumstances, though I've
never observed that myself.

Ed's and bogax's explanations are probably as good as you'll get, as resonant oscillators involve some rather heavy math with the S-plane and so on, which is why it's often best to just use the complete crystal can oscillator modules. Otherwise, even with experimentation, you might come up with something that seems to work right but then goes nuts when you build it again with a slightly different layout or unforseen operating-condition changes come along, whether it means different temperature, different amount of stray capacitance or inductance somewhere, etc..

As for the decoupling capacitor, there should be one of .1 or .01uF from the +5V pin to the ground pin on the IC with the connections and leads absolutely as short as possible from the capacitor body to the IC pins. This is for reducing inductance, not resistance.

TTL oscillators never worked for me. I've always had much more success with CMOS CD4069 inverters. The CMOS oscillators have much higher values for the resistors in the oscillator circuit, hence my earlier comment about the values being too low. If they're correct, then I guess they're correct for TTL...


I agree. Do the 1/2 can can. It'll work. Easy to hook up. Less susceptible to bumps and bruises, but...

There are sawtooth and clipped sine and I'm sure some other type of wave generators out there. Obviously look for the square wave, and pay attention to the Voltage required. All kinds of ratings nowadays.

Most likely buying one will not only work for you right off the bat, but will probably be MUCH cheaper when you consider IC's, boards, wiring, etc.

BTW I like the close up pics . Good luck in your endeavor!

I think the picture of the rear of the board came out a bit blurred (crystal is in focus though). Taken on a Sony NEX-5 with the standard 18-55mm zoom lens which has an uncanny knack of getting the light just right when used indoors. I need to get a macro lens for it.

Have ordered a can oscillator, 4 legged, 8Mhz, runs off 5v. This is the lazy way out, sure, but I can't run the board at 3Mhz.. I may try using the spare counter too, depends on how much time I have.

Well at last, it is actually working.

Here's what I did to sort it out:
- fit a 100nf cap between the inverters in the clock circuit. The clock runs at 8Mhz just fine, with the 470R resistors. I haven't bothered with the 120pf earth cap.
- discovered ACIA ground pin was not connected, sorted that out
- discovered the max232 GND pin was not connected, fixed it

The serial circuit is left with the 22uF caps for now. I got the clock working and fitted the NOPgen and CPU. Was delighted to see (after a reset) that the address lines were all pulsing nicely. Then I fitted the rom, ram =, acia and max232 ICs. A small amount of debugging later, and I have a working Micro UK101. Huzzah! Will now post some pictures so you can all see my dodgy wiring.

Thanks to all for your help!

Fantastic! Glad you stuck with it!

Wow man! Freakin awesome!

Nice job

So there you go, one happy camper.



Now for some true retro computing (yeah, emulators, but they're not real).

I have to say I am amazed at how simple it actually is. To think I used to look at these boards with total amazement and awe when I was 15, and here I have just whacked one together (with a mallet, seemingly). Nothing beats the feeling. So I am wondering what to do next. I reckon I might fit a second serial port to it, and maybe a parallel port as well. It needs better I/O, and a persistent store of some sort (compact flash I think). Not sure I have the patience to write a DOS for it, might be able to use the UK101 original code (which has been disabled in this ROM image). What's interesting is that I could stick any old OS in there if I wanted, provided it honoured the existing memory mapping for the serial port. Even write my own. The BASIC implementation needs better editing, that's for sure, but it is impressively quick. The 0 to 10 count is faster on this than it is on the Spectrum and that is clocked much faster (3.5Mhz if memory serves).

I'll have to have a good think about this!

Thanks..

It's not as neat as GS's implementation, but this is the first time I've ever done something like this.

_________________
x86?  We ain't got no x86.  We don't NEED no stinking x86!

Having a little bit of a problem with loading and saving files. Well, save's easy, just capture text and type list. But trying to load it back in, it keeps getting confused, even with lots of spaces padding the text (before / after / blank lines). I think BASIC cannot keep up with it. Any suggestions?

Yeah, I know. I've been watching some contemporary TV - The Sweeney and The Professionals and I am into the late 70's groove, man. The only smoke coming of this little baby is the digital tyres burning rubber when it executes a program. It's fast alright, much quicker than I expected. It is quicker than BBC BASIC, I am sure. When emulating a BBC Model B at normal speed on a PC, the 1-10 count is slower. Not that all operations will be faster, but for now... whoosh!