Turning on STacy only 2 times out of 10 does it start.

[stephen_usher]

The problem, as I said previously, is probably that the 5V rail is taking longer to get to 5V than the reset circuit is holding the CPU in reset.

Turning the voltage up to 5.8V will probably cause the voltage to rise more quickly and beat the reset circuit timeout. However, 5.8V *WILL* damage your machine. Look at the capacitor in the reset circuit.

With respect to the oscilloscope, the readings are showing the degree of noise on the rails. overshooting and undershooting is reasonably normal and is caused by the switching of the components. However, the base voltage of 5.2V is still rather high. It should never be above 5V and the max shouldn't really be more than about 5.1V. Modern oscilloscopes can also give you, via the measurement option, the max, min, average and peak-to-peak information. The average should never be more than 5.1V for TTL chips otherwise they'll be damaged over time.

Forget about the 12V for the time being. That's only used for the serial port and powering the hard drive and maybe the floppy motors.

[rubber_jonnie]

I have to agree with @stephen_usher regarding setting your 5v rail to 5.8v.

Look at the datasheets for 74LS logic ICs and they are on the whole expecting a max VCC of 5.25v, so you are certainly risking damage to your machine. The 6v you mention is really too high.

Take a close look at the reset circuit, but I'd also suggest hooking up the machine to 5v on a bench PSU if you have one. 5v is all that's needed to boot, plus you can current limit the PSU whilst you are troubleshooting, whilst knowing that you are working with correct values and can look for the real problem safely.

You may also have a PSU issue and that it is incapable of delivering the correct power at 5v, in which case you need to look closely at the PSU as well as the reset circuit.

[stephen_usher]

Look at the datasheets for 74LS logic ICs and they are on the whole expecting a max VCC of 5.25v, so you are certainly risking damage to your machine. The 6v you mention is really too high.

Indeed. And CMOS has a lower high voltage tolerance too, and the STacy, being a low power device designed to run on batteries, will most likely be CMOS rather than TTL.

[rubber_jonnie]

Indeed. And CMOS has a lower high voltage tolerance too, and the STacy, being a low power device designed to run on batteries, will most likely be CMOS rather than TTL.

Yes, very good point.

[Roberto]

Sounds like the reset circuit is taking the machine out of reset too early, before the 5V rail has stabilised. Without the other peripherals the PSU is bringing the voltage up just about enough in time for the CPU to come out of reset.

Check the capacitor in the reset circuit and, I think it's add a bigger capacity one to slow it down. @exxos would know better about that.

I only saw your comment yesterday and I can confirm that it is exactly as you say. I tried both with the normal power supply and with the professional bench one and the behavior does not change. I did a test and it gave me confirmation of what you say, that is: I brought the trimmer value of the DC/DC module to 5.6v and then I started the STacy with scsi controller, floppy and other things. So right after it started I dropped the voltage down to 5V. In this condition even if I do a reset the system always starts. I then searched the schematics for the reset circuit (which should be near the on button) but I don't know which capacitor is the one you're talking about. I took a picture of what I found. The one circled in red is a 100uF, in your opinion is that what I have to change with a higher value? Consider that I changed all the electrolytic capacitors of the motherboard too and what you see is new.

0F2672D6-FAB3-4794-A786-F398B125D756.jpeg

792FA2AC-AADC-4F5F-8C7C-F1609A66D831.jpeg

7ECBD7F2-D8E6-4796-ADD4-79F3259A22B7.jpeg

[stephen_usher]

Apparently the TL7705A time control capacitor (CT) is C3, which the specs say is 10uF. The larger the capacitance then the longer the delay.

You could try replacing C3 with a new 10uF capacitor, or maybe a 15uF if you want to delay 1.5 times as long (from what I can see). etc.

See TL7705A data sheet: https://www.ti.com/lit/gpn/tl7705a

[DoG]

Here is the reset circuit for reference:

Reset STacy smaller.png

[Roberto]

Apparently the TL7705A time control capacitor (CT) is C3, which the specs say is 10uF. The larger the capacitance then the longer the delay.

You could try replacing C3 with a new 10uF capacitor, or maybe a 15uF if you want to delay 1.5 times as long (from what I can see). etc.

See TL7705A data sheet: https://www.ti.com/lit/gpn/tl7705a

Today I had in my hands both the original 10uF capacitor and a 47uF one (unfortunately I only had this one available at the time, closer to the 10uF value), both at 16V of course. I tried both but I didn't notice any delay compared to the tests, and therefore I didn't get any improvement by setting the DC/DC module to 5V. Maybe I need to use a 15uF one to have the desired effect? I noticed from the @DoG diagram that there is also another capacitor involved, namely the 0.1 uF C1 and C2 (it seems to me not polarized). Maybe I need to change those too? All the non-polarized ones have not been changed, only the polarized electrolytics have been replaced.

[Roberto]

I have some news… I feel that I am getting closer to the real problem. I've done several tests using different capacitors, of different capacities at the spot indicated by @stephen_usher and noticed (especially using a 100uf capacitor) that there is a good turn-on delay but that doesn't seem to be the cause of the problem. rather I did a new test, i.e. to connect a device rather stingy with electricity, i.e. a real hard disk (surely working), and I noticed that from an initial setting of 5.1v (set using the adjustment trimmer) the hd is slow to spin and I have a voltage drop in the DC/DC module which reaches 4.5 volts. Even if (when it's on) I try to adjust it to 5v by turning up the trimmer, I see a stability problem. Do any of you know which component of the DC/DC module is responsible for voltage stability?

I also did the tests using a good professional bench power supply and the results didn't change.

[Roberto]

the test I did in the previous post is not reliable, in the sense that I realized that the STacy hard disk is not the same as a mega ste; the hard disk of the atari mega ste consumes much more current than that of the STacy. So I confirm that the problem is only with the ignition.