My Klyball 600D build

[ronin47]

Yeah, that would be it.

On the 600D, you have to hold <Break> down for bit for it to work, possibly two seconds or so.

Great, thanks.

Yes, I'd noticed it took a long time to work when I was testing with the oscilloscope earlier. Got to hit the sack now, midnight here in the UK, will test some more tomorrow.

Appreciate the pointers, thanks.

[ronin47]

You can check if the CPU is actually executing the ROM code properly by hitting <Break>C<return><return> and then SAVE<return>, looking at the signal on the TxData output of the ACIA as you hit that last return. If things are actually running correctly, you should see almost one full second of activity on that line as the ACIA outputs the message <cr><lf>OK<cr><lf> at 300 bps, with 10 NUL characters inserted after each <cr>.

I tried this and didn't see any activity, so not sure if it's running or not...

At present I have removed the ROMs all bar the CHARGEN ROM, but only get intermittent blips of stuff like this:

PXL_20240127_150301174.jpg (2.64 MiB) Viewed 1554 times

And as of yesterday, for some reason, despite seeing a 4Mhz clock going into the LS93 (Like waveform 1 from the service manual), I'm only getting 500Khz coming out. I'm going to go over the PCB again looking for solder bridges, but any ideas at all would be welcome at this point.

Thanks

[bxdanny]

Hmm. As you know, I tend to focus on the software side of things, and I was simply thinking that the 6850 ACIA is the one place, other than the video screen, where software on the 600 board is intended to produce output. The one I used to have was a 600B, which by default had the 6850 wired only to the cassette interface. But the 600D comes with the RS-232 circuitry installed, and I think a jumper may need to be connected to one of the board's Molex connectors to ensure that the 6850 is getting a "Clear to Send" signal (a logic 0 on pin 24, the /CTS input), without which it will not transmit. Also, make sure you pause a second or so between the two <return>s after that initial C for Cold start, to give the memory test time to determine how much RAM there is. That's about all I can say about that. But it is still certainly true that the video screen would not be BLANK if the video circuitry was working, unless all the video RAM was filled with $20 and/or $60.

[Klyball]

You Should see 500k on pin 12 but 1mhz on pin 11going to the 6502 , i leaning towards a video problem
check for clock pin 2 U42, what are you using in U24,U25 and file are you using for the character rom

[ronin47]

You Should see 500k on pin 12 but 1mhz on pin 11going to the 6502 , i leaning towards a video problem
check for clock pin 2 U42, what are you using in U24,U25 and file are you using for the character rom

I was working through the clocks today and I'm currently getting just 500KHz on pin 11 and 250Khz on pin 12 of the LS93. I have tried another IC but it's the same.

It looks like there's no output on pin 6 of U79 either, the LS86, even though pin 5 has 4Mhz.

I have tried swapping U58 and U29, to no avail, and I'm waiting on replacement parts for the LS86, U79.

U24 and 25 are MC8T28P's as specified in the BoM.

Once I get the replacements for U79 I'll go back and check the clocks again, but wondering if the part I have is marginal even though it tests OK.

One question, what should I see at pin 1 of U29, the LS93? Should it be 4Mhz or 8 Mhz, as currently it's getting 4Mhz but the outputs are divided by 8 to deliver 500Khz on pin 11 of U29 and not the 1Mhz as expected.

[ronin47]

You Should see 500k on pin 12 but 1mhz on pin 11going to the 6502 , i leaning towards a video problem
check for clock pin 2 U42, what are you using in U24,U25 and file are you using for the character rom

BTW I'm using the file from here for the character ROM: https://github.com/jefftranter/6502/tre ... board/roms

The ROM ICs are 28C16's, pin 2 U42 has 2Mhz on it.

[jbtech]

It looks like there's no output on pin 6 of U79 either, the LS86, even though pin 5 has 4Mhz.

I have tried swapping U58 and U29, to no avail, and I'm waiting on replacement parts for the LS86, U79.

U24 and 25 are MC8T28P's as specified in the BoM.

Once I get the replacements for U79 I'll go back and check the clocks again, but wondering if the part I have is marginal even though it tests OK.

One question, what should I see at pin 1 of U29, the LS93? Should it be 4Mhz or 8 Mhz, as currently it's getting 4Mhz but the outputs are divided by 8 to deliver 500Khz on pin 11 of U29 and not the 1Mhz as expected.

Hi, yes that seems to be the issue. The first stage of U79 is being used as a buffer and should have 4MHz on pin 6.

The second stage of U79 takes two 4MHz signals which are partially out of phase (biphase) to create a 8MHz signal.

That is then divided by 2, 4 and 8 in the '93 to give the 1MHz clock on pin 11 and further divided by 2 to give 500kHz on pin 12.

With only one input the second stage of U79 will output 4MHz causing everything after that to be halved.

Rough sketch follows - wouldn't hurt to check around U79 for shorted tracks before replacing it. Regards, John.

IMG_1778.JPG (4.24 MiB) Viewed 1238 times

[ronin47]

It looks like there's no output on pin 6 of U79 either, the LS86, even though pin 5 has 4Mhz.

I have tried swapping U58 and U29, to no avail, and I'm waiting on replacement parts for the LS86, U79.

U24 and 25 are MC8T28P's as specified in the BoM.

Once I get the replacements for U79 I'll go back and check the clocks again, but wondering if the part I have is marginal even though it tests OK.

One question, what should I see at pin 1 of U29, the LS93? Should it be 4Mhz or 8 Mhz, as currently it's getting 4Mhz but the outputs are divided by 8 to deliver 500Khz on pin 11 of U29 and not the 1Mhz as expected.

Hi, yes that seems to be the issue. The first stage of U79 is being used as a buffer and should have 4MHz on pin 6.

The second stage of U79 takes two 4MHz signals which are partially out of phase (biphase) to create a 8MHz signal.

That is then divided by 2, 4 and 8 in the '93 to give the 1MHz clock on pin 11 and further divided by 2 to give 500kHz on pin 12.

With only one input the second stage of U79 will output 4MHz causing everything after that to be halved.

Rough sketch follows - wouldn't hurt to check around U79 for shorted tracks before replacing it. Regards, John.
IMG_1778.JPG

That is super helpful John, thank you. What I was sort of expecting, but it's nice to have the verification

I swapped out U58 this morning, I had a 7400 in there as per the BoM, but dropped in a 74S00 instead and started to get 8Mhz on pin 8 of U79, however the clock is really unstable, flashing up and down in frequency. I have changed C57 and C55, though not the crystal, though scoping that does seem to show it running consistently at 4Mhz. I'm certainly getting consistent values at U58, ad though the voltages are a bit low, what's coming out of pin 8 is good, and what's coming out of U29 pin is good voltage wise, even if the frequency out of pin 8 is only 4Mhz.

But U79 is weird. Sometimes I get 8Mhz at pin 8 as described bit sometime 4Mhz and sometimes its flapping between those values, and that's then reflected up the line.

I have new LS86s coming, but worried about the overall instabilities I'm seeing in the clock circuit.

Sometimes, and I'm assuming it's because the clock gets somewhere near where it should be I get a sort of video signal, but until the clock is stable, which it isn't ATM, all bets are off.

[bxdanny]

I swapped out U58 this morning, I had a 7400 in there as per the BoM, but dropped in a 74S00 instead and started to get 8Mhz on pin 8 of U79, however the clock is really unstable, flashing up and down in frequency.

So maybe you should go back to a 74LS00?

Perhaps I should stay out of this, but I'd hate to think that my earlier comments made you switch away from using a 74LS00 there, and that turned out to be a mistake which is now causing problems.

But I'm certainly confused as to why the circuit would be designed so that it tries to derive an 8 MHz signal from a 4 MHz crystal. Wouldn't it have been easier to just use an 8 MHz crystal and divide it down as needed? Maybe OSI still had a large stock of those 3.932160 MHz crystals, and simply wanted to keep making use of them?

[ronin47]

I swapped out U58 this morning, I had a 7400 in there as per the BoM, but dropped in a 74S00 instead and started to get 8Mhz on pin 8 of U79, however the clock is really unstable, flashing up and down in frequency.

So maybe you should go back to a 74LS00?

Perhaps I should stay out of this, but I'd hate to think that my earlier comments made you switch away from using a 74LS00 there, and that turned out to be a mistake which is now causing problems.

But I'm certainly confused as to why the circuit would be designed so that it tries to derive an 8 MHz signal from a 4 MHz crystal. Wouldn't it have been easier to just use an 8 MHz crystal and divide it down as needed? Maybe OSI still had a large stock of those 3.932160 MHz crystals, and simply wanted to keep making use of them?

Don't worry, I was happy to try anything at this point. It seems rather strange that you would derive 8Mhz from 4Mhz, just seems like unecessary complications.

Funny you should say though, but in the UK-101, the UK specific copy of this machine, they do use an 8Mhz crystal and it's all very much simpler.

I'm talking to somebody over on another forum and they are suggesting I replace the xtal with an 8Mhz one and cut out U79, and I am considering that and have ordered the Xtal.

I will go back to trying the LS00.