CT63 only starts from 030 mode

[foft]

Ah yes in ct60conf.ctx/param.s in that change-set. As well as device id, it seems to save parameters at the end of the flash. At least I keep seeing addr+flash_size-param_size in that code. The blocks it needs to erase are a different size in that area. 64KB in the (usual) bottom boot device and 16KB in the top boot device. So it needs to send multiple block_erase commands probably (if it doesn't already, I only speed-scanned the asm...)

As for initializing without code changes: I was assuming it could be flashed via jtag. In fpgas this is often done by loading a temporary core to the fpga that flashes the chip. Another possibility is to use an eprom programmer before soldering it.

[foft]

On the other problem, cold reset

On the falcon schematic I can see a 555 reset circuit. I’ve never used a 555 though know they are very common.

So trigger goes active and then the pulse length is determined by the time to charge C1.

On the trigger input there is a bistable flip flop built from not gates, with the infamous C7 capacitor. So I push reset and while it’s pressed (gnd) the flip flop is outputting high. When I let go it’s low again.

Therefore the reset pulse is when I Iet go of the button, after c7 discharges and trigger is active (low). So C7 delays the reset pulse start after the button is released.

Now on power on what happens? No idea! I guess I’d better do some more reading… & maybe put the 555 on the scope so I can at least check what I wrote so far.

Though presumably the idea is that C7 is low for longer as it powers on while the psu stabilises. Though wouldn’t we want to delay the initial trigger low until it’s stable, rather than making the trigger longer?

[exxos]

Which fault exactly are you trying to fix ? If reset is switching, then why are you messing with it ?

[foft]

Which fault exactly are you trying to fix ? If reset is switching, then why are you messing with it ?

This thread started as CT63 only starts from 030 mode. So to get to 060 mode I had to:
i) Start up with 030 jumper
ii) Remove 030 jumper
iii) Hit reset

Since then I fixed my NVRAM battery (I'd replaced it was a near dead CR2032 by accident...). It was enough to run the clock, but not to reliably store settings.

Now it starts up in 060 mode, but I need to:
i) Turn on the Falcon
ii) Hit reset
iii) (Sometimes hit reset again)
iv) Boots with a while screen about CT60, often freezes
v) Hit reset
vi) It now works perfectly.

I read that increasing C7 should fix this, so I changed it to first 42uF then 100uF, with no change.

So now I'm trying to work out what is going on here. I thought a starting point would be to understand how the reset circuit works. It sounds like the problem is thought to be with ATX power rail stability timing, after telling the PSU to 'soft' turn on.

[exxos]

When it is not looting, is the floppy light flashing very fast ? If the system is running light will flash several times very fast, then stay off for a second and then flash several times again. If it is doing that then you have a RTC fault.

If pressing reset helps, then you either have a faulty reset switch. You already changed the capacitor.. So you could try adding something like 1K between pin 2 & 14 of U11.

[foft]

When it is not looting, is the floppy light flashing very fast ? If the system is running light will flash several times very fast, then stay off for a second and then flash several times again. If it is doing that then you have a RTC fault.

Nope, its not flashing.

If pressing reset helps, then you either have a faulty reset switch. You already changed the capacitor.. So you could try adding something like 1K between pin 2 & 14 of U11.

I'll give that a try tomorrow, thanks.

[exxos]

You could also try upping the value of C1. I think you shouldn't need higher than 100uf for C7 or C1. Normally 22uf-47uf seems fine on STEs at least.

[foft]

I've not tried larger C1 yet but I did try the pull-up. It didn't change much.

Though I put it on the scope and I'm a little surprised by what I see.
C1 -> sets minimum pulse width
Reset button directly seems to pull XRESET low
C7 doesn't do a whole lot, it discharges when I press reset and then slowly charges when I release it. However how long that takes to charge doesn't seem to change much.

Pink = XRESET (i.e. pin 2 of U11)
Cyan = C1
Yellow = C7

Now I thought looking at the circuit that C7 should DELAY the start of reset. Is that wrong?

ScreenImg.png

This is two pushes of reset, a long one then a short one.

Edit: So anyway I reckon I could increase C1, which would make the pulse wider and thus might fix it. But it doesn't seem the right fix.

[exxos]

Now I thought looking at the circuit that C7 should DELAY the start of reset. Is that wrong?

Only when you press the reset switch. C1 is when the machine powers up AFAIK. But i've not looked into the circuit past what I wrote in the fixes thread.

This is why C7 is a lower value than C1. Because machine is already powered up, so the reset does not need to be as long. So lower value of C7. C1 is higher to increase the delay longer because the machine power supply has to get up to voltage etc.

[foft]

Now I thought looking at the circuit that C7 should DELAY the start of reset. Is that wrong?

Only when you press the reset switch. C1 is when the machine powers up AFAIK. But i've not looked into the circuit past what I wrote in the fixes thread.

On power on it doesn't do much either... I wonder if D1 or the 555 are iffy. edit: D1 is good.

C1 does hold reset low for longer since it sets the min pulse width but...

ScreenImg(1).png