exxos's Atari, Amiga & retro forum

I know that I suggested you look at /CE, however you haven't actually confirmed with the oscilloscope that all the clocks are good as I suggested you do before looking at /CE in my earlier post. You had previously asserted they were, but this was prior to your oscilloscope arriving, so you should check all of those. No clock == no go as @stephen_usher stated.

I'd also suggest you get some time with your oscilloscope on your other machine so that you can understand how to set it up to capture a TTL signal on address and data lines and also look at other TTL signals like /CE.

Using a working machine will give you a wealth of information on how things should look and will also give you the opportunity to get used to using the oscilloscope with something that is known good. This is how I learned to use an oscilloscope.

It's no good looking at something bad and stating it looks wrong/bad if you don't know what good looks like. Also, don't be in a hurry to fix it. I know that you are upset your machine is broken, and it was possibly something you did, but rushing can make things worse, so I urge you to be patient as it will help in the long run.

On the subject of /CE, @exxos is right, it doesn't necessarily mean that /CE is faulty. If the 74LS11 is receiving 3 inputs and there is a good output, then you're beyond that at the GLUE. You state the GLUE works in your other machine just fine, so then it is likely beyond the GLUE. I'd suggest you also follow his advice as he has helped me on more than one occasion.

The problem you have is that you have a complex fault, and no matter how many experienced people you get on here, trying to tell you what to do next is the difficult part.

Last but by no means least, the RESET line on the Shadow IC. I think your interpretation of it being active as it is high is incorrect. If you look at the image below, you can see that the line going into the pin shows /RST, meaning active low, so 0v.


Bear in mind these drawings/schematics were originally hand drawn, so there is always scope for error. In TTL reset is always low when active, at least I've never seen it work the other way around, at least not on any of the machines I've repaired.

exxos wrote: 06 Oct 2022 21:18 ...
This is why I mentioned earlier you need to test address zero.. Ideally by removing the CPU manually setting it because then if you are forcing address zero on the CPU, then CE will be low continuously. viewtopic.php?p=91608#p91608 IMHO You need to really do this first before anything else otherwise you will be chasing faults which may not actually exist!

I was really hoping not to come to this. Removing a 64-pin chip saving everything is not easy (not even for me who have 13 years of experience in soldering electronic boards, but zero experience in operating logic), unless I have a crucible at my disposal, what I do not have.
In that area it is not possible to use hot air to take out the CPU as there are plastic sockets and other plastic connectors that would be melted with the hot air technique. The only safe way is to sacrifice the CPU, cutting the pins one by one with a clipper, then subsequently heat each track with the soldering iron and remove the previously cut pin.
I have seen that you sell good quality low profile 64 pin sockets, however you have the 68HC000 CPU only in box. The 68HC000 CPU should differ from the normal MC68000 only in lower power consumption. I saw to my surprise that a MC 68HC000 CPU is not impossible to find, indeed, even the prices are affordable. the 68HC000P10 should support up to 10Mhz compared to the 68HC000P8 (Which supports up to 8Mhz) that mounts the Stacky, so it should be compatible, is that correct? My concern is that the P10 only works with a 10 Mhz clock (unfortunately there are no P8s on the market).

The link you sent me in your previous post I had printed on paper and studied carefully. I saw that there are two ways to test the zero address, the one without removing the CPU (which I am still studying how to proceed) and the one where the test is done by removing the CPU. Do you think I go directly to the removal of the CPU?
Thanks for your valuable advice

You can keep pressing reset to diagnose startup. But it's a lot harder when there is activity on the bus.

The speed of the CPU is not important as you're only using 8mhz. So P8,P10,P12 doesn't matter. Though note store CPUs are tested. You can of course buy wherever you like but you may buy a duff CPU.

Also I don't recommend using HC CPU on unmodified stock machines. They tend not to work well.

rubber_jonnie wrote: 07 Oct 2022 09:53 I know that I suggested you look at /CE, however you haven't actually confirmed with the oscilloscope that all the clocks are good as I suggested you do before looking at /CE in my earlier post. You had previously asserted they were, but this was prior to your oscilloscope arriving, so you should check all of those. No clock == no go as @stephen_usher stated.

Hi rubber_jonnie, in one of the posts I wrote previously, I wrote "the second test I did was to confirm the presence of the correct clock signal frequency for each component. I confirm that the signal is present and the frequency is correct ". Let's say it's one of the first things I did when I got the oscilloscope, as it's very easy to do. I also wrote that all the components are powered correctly by + 5V, and I was able to verify the +12 Volt and -12 Volt needed for the RS232, only from one component, the 1488, U11), others I have not found.

rubber_jonnie wrote: 07 Oct 2022 09:53 I'd also suggest you get some time with your oscilloscope on your other machine so that you can understand how to set it up to capture a TTL signal on address and data lines and also look at other TTL signals like /CE.

Using a working machine will give you a wealth of information on how things should look and will also give you the opportunity to get used to using the oscilloscope with something that is known good. This is how I learned to use an oscilloscope.

It's no good looking at something bad and stating it looks wrong/bad if you don't know what good looks like.

Yes, in fact I have both an Atari 520 ST and a Mega ST on hand, which I certainly find the latter closer in terms of components to the STacy. For example, here's how the address and /CE signals look perky on my Mega ST.

rubber_jonnie wrote: 07 Oct 2022 09:53 Also, don't be in a hurry to fix it. I know that you are upset your machine is broken, and it was possibly something you did, but rushing can make things worse, so I urge you to be patient as it will help in the long run.

Thanks for the encouragement

rubber_jonnie wrote: 07 Oct 2022 09:53 On the subject of /CE, @exxos is right, it doesn't necessarily mean that /CE is faulty. If the 74LS11 is receiving 3 inputs and there is a good output, then you're beyond that at the GLUE. You state the GLUE works in your other machine just fine, so then it is likely beyond the GLUE. I'd suggest you also follow his advice as he has helped me on more than one occasion.

I think the next step I will take will be to remove the CPU to test address 0 (unless I can somehow do this without taking it off).

rubber_jonnie wrote: 07 Oct 2022 09:53 The problem you have is that you have a complex fault, and no matter how many experienced people you get on here, trying to tell you what to do next is the difficult part.

Last but by no means least, the RESET line on the Shadow IC. I think your interpretation of it being active as it is high is incorrect. If you look at the image below, you can see that the line going into the pin shows /RST, meaning active low, so 0v.

Bear in mind these drawings/schematics were originally hand drawn, so there is always scope for error. In TTL reset is always low when active, at least I've never seen it work the other way around, at least not on any of the machines I've repaired.

Yes, it is true, in the logic diagram the RESET has the line above.

exxos wrote: 08 Oct 2022 17:33 Also I don't recommend using HC CPU on unmodified stock machines. They tend not to work well.

The Stacy has a HC CPU from the factory. I've used both HC and non with no problems.

stephen_usher wrote: 06 Oct 2022 22:08 I take it that you've checked the clock on the CPU? No clock == no go.

Yes, I have tested all the clocks to my knowledge based on the component pinout I found in the documentation. All components receive (or output) the correct clock signal.

Is there any activity on any of the output pins of the CPU at all?

stephen_usher wrote: 09 Oct 2022 13:44 Is there any activity on any of the output pins of the CPU at all?

No activity on addresses and data. Always a constant +5 V both for the addresses and for the data on the STacy. While taking measurements on another Atari (Mega ST), the addresses and data are working and I see the signal changing (similar just like the last two photos I posted). At the weekend I should meet a guy who lives not far from me and he gives me an MC68000 CPU. So I'll take out the old one (I'll have to break all the pins to safeguard the motherboard), to put a socket in it, then test address 0 as suggested by exxos.
Perhaps the CPU itself may also be the problem.

Out of curiosity, what is pin 17 (/HALT) doing? If its low, it can either be the CPU or something else on the board pulling it down as its bi-directional.

Also, what are the address and data pins doing when you power on? Do you see any signal before they go high and stay there?

rubber_jonnie wrote: 12 Oct 2022 12:40 Out of curiosity, what is pin 17 (/HALT) doing? If its low, it can either be the CPU or something else on the board pulling it down as its bi-directional.

Also, what are the address and data pins doing when you power on? Do you see any signal before they go high and stay there?

I took some pictures of pin 17 / HALT of both my STacy comparing with pin 17 / HALT of my Mega ST. I think the Single SEQ mode should correspond to the first instant where the signal change is detected, so I have also reported the phases in which the signal is displayed over time in the two computers. I did the same thing in one of the registers, in this case D4 of pin 1. consider that on mega ST all data and address registers, during operation there is a certain activity (as in the photo), while on STacy the signal for all registers (data and addresses) the signal is static, in some registers stopped at 2v, in others stopped at 5v