Hi All,

I have an issue with a spare RiscPC which exhibits a corrupted Pointer and
also a loud buzzing sound from the speaker on initial starting.

It did have a failed CMOS chip and this has been sorted by using a
CJEMicros CMOS and battery arrangement on the backplane.

It had been working Ok for many months since doing this fix.

It originally had a ARM 710 card in it and the issues are seen when a
StrongARM or Kinetic card is used.

With the 710 card it seems to behave itself and the usual memory fix, C32 removal, has been performed.

I have tried different memory boards from a working RiscPC and spares I
have.

If I put one memory board in slot 1 and have the StrongARM in then it
doesn't generate the buzzing sound but the pointer is corrupted on initial starting.

If the machine is left on for a while then the corruption goes away.

So is this something like board damage, though it appears OK around the
now removed battery, or is the board one of those that had issues with StrongARM's and required Acorn repairs over and above those noted in Engineering note 297?

There is mention of changing the alignment of two resistors on the
StrongARM board to make the data bus more stable but I am not inclined to
do that as it works OK in the other working system I have and my soldering skills have diminished with age and lack of practice :-).

Doug

--
Experience the future using ARM Technology - ARMBook,BeagleBoard -xM, PandaBoard,Raspberry Pi,iMX6/ARMX6,IGEPv5 & Titanium powered by RISC OS
5.28.

--- SoupGate-Win32 v1.05
* Origin: fsxNet Usenet Gateway (21:1/5)

Doug Webb <doug.j.webb@btinternet.com> wrote:

I have tried different memory boards from a working RiscPC and spares I
have.

If I put one memory board in slot 1 and have the StrongARM in then it
doesn't generate the buzzing sound but the pointer is corrupted on initial starting.

Have you tried swapping or removing the VRAM module, if any?

Theo

--- SoupGate-Win32 v1.05
* Origin: fsxNet Usenet Gateway (21:1/5)

In message <aPf*EPQzy@news.chiark.greenend.org.uk>
Theo <theom+news@chiark.greenend.org.uk> wrote:

Have you tried swapping or removing the VRAM module, if any?

Theo

Yes removed the VRAM and also tried a different hard drive as well as an outside chance as it also gets worse , with a StrongARM or Kinetic in, if
the disc is accessed.

There appears no issues if the 710 card is installed.

Doug

--
Experience the future using ARM Technology - ARMBook,BeagleBoard -xM, PandaBoard,Raspberry Pi,iMX6/ARMX6,IGEPv5 & Titanium powered by RISC OS
5.28.

--- SoupGate-Win32 v1.05
* Origin: fsxNet Usenet Gateway (21:1/5)

On 21/11/2021 12:11, Doug Webb wrote:

Hi All,

I have an issue with a spare RiscPC which exhibits a corrupted Pointer and also a loud buzzing sound from the speaker on initial starting.

It did have a failed CMOS chip and this has been sorted by using a
CJEMicros CMOS and battery arrangement on the backplane.

It had been working Ok for many months since doing this fix.

It originally had a ARM 710 card in it and the issues are seen when a StrongARM or Kinetic card is used.

With the 710 card it seems to behave itself and the usual memory fix, C32 removal, has been performed.

I have tried different memory boards from a working RiscPC and spares I
have.

If I put one memory board in slot 1 and have the StrongARM in then it
doesn't generate the buzzing sound but the pointer is corrupted on initial starting.

If the machine is left on for a while then the corruption goes away.

I think the VIDC might be failing. Does the pointer corruption occur in
any particular modes?

---druck

--- SoupGate-Win32 v1.05
* Origin: fsxNet Usenet Gateway (21:1/5)

In message <snfsk7$bsf$1@dont-email.me>
druck <news@druck.org.uk> wrote:

I think the VIDC might be failing. Does the pointer corruption occur in
any particular modes?

The issue occurs in any mode with a StrongARM or Kinetic installed but not
when a 710 card is installed.

Might have to tap out RP16 pack but that is for another day as my old test meter has given up the ghost as well.

The machine is just a spare to the Kinetic/UniPod/VPod enabled other one
so it isn't urgent to repair as it has been kept for its bits to replace
in the other one if that goes faulty.

--
Experience the future using ARM Technology - ARMBook,BeagleBoard -xM, PandaBoard,Raspberry Pi,iMX6/ARMX6,IGEPv5 & Titanium powered by RISC OS
5.28.

--- SoupGate-Win32 v1.05
* Origin: fsxNet Usenet Gateway (21:1/5)

Doug Webb <doug.j.webb@btinternet.com> wrote:

In message <snfsk7$bsf$1@dont-email.me>
druck <news@druck.org.uk> wrote:

I think the VIDC might be failing. Does the pointer corruption occur in
any particular modes?

The issue occurs in any mode with a StrongARM or Kinetic installed but not when a 710 card is installed.

This:

I have an issue with a spare RiscPC which exhibits a corrupted Pointer and >> also a loud buzzing sound from the speaker on initial starting.

points to problems with Cursor and Sound DMA. In other words what's being fetched from memory for the pointer and the sound is corrupted in some way.

The sound has a separate DMA channel (Sndrq*, Sndak*) from VIDC into IOMD,
but the pointer doesn't. It's separate inside the VIDC but not separate
wiring on the motherboard. So a problem with the Snd wires could cause
glitchy sound, but there aren't any pointer-specific wires that I can see.
If it was all VIDC memory accesses, I would have expected screen corruption
as well (especially when no VRAM).

So either the pointer and cursor DMA parts of the VIDC are failing, or
there's something timing related that affects all memory access and it's
only noticeable during cursor/sound DMA.

The Kinetic doesn't use motherboard DRAM for most things, only for that
needed by DMA (the screen etc). But the SA110 card may exercise the DRAM
more than the 710 does.

A couple of other things to try (separately):

Try switching to a very conservative mode - perhaps mode 27 (16 colour VGA). That will drop the video bandwidth requirements and make demands on the
memory system 'easier'

Try *Cache Off Things will go slower, but that will push the DRAM harder.
I'm wondering if the behaviour changes.

I'd try these with one SIMM alone, and with one SIMM and one VRAM stick.
If you are able to check them a working RPC that will tell you if they are known-good.

Might have to tap out RP16 pack but that is for another day as my old test meter has given up the ghost as well.

I wouldn't be tinkering like that without a plausible reason why it might
help. The experiments above might provide information as to what's
happening.

Theo

--- SoupGate-Win32 v1.05
* Origin: fsxNet Usenet Gateway (21:1/5)

In message <dPf*2WUzy@news.chiark.greenend.org.uk>
Theo <theom+news@chiark.greenend.org.uk> wrote:

The sound has a separate DMA channel (Sndrq*, Sndak*) from VIDC into IOMD, but the pointer doesn't. It's separate inside the VIDC but not separate wiring on the motherboard. So a problem with the Snd wires could cause glitchy sound, but there aren't any pointer-specific wires that I can see.
If it was all VIDC memory accesses, I would have expected screen corruption as well (especially when no VRAM).

OK so if I put memory in to Slot 0 and start up I get the sound buzzing
and now I am getting screen corruption and no display.

This is with or without VRAM.

Previously if I put memory in Slot 0 I got the sound issue and also the
pointer corruption but otherwise the screen was OK.

Putting the memory in to Slot 1 just gives the pointer corruption.

Putting the same memory in to the other RiscPC shows no issues.

Swapping another VRAM in or memory , 32MB as opposed to 64Mb, gives no difference.

A couple of other things to try (separately):

Try switching to a very conservative mode - perhaps mode 27 (16 colour VGA). That will drop the video bandwidth requirements and make demands on the memory system 'easier'

Makes no difference pointer corruption when memory is in Slot 1.

Try *Cache Off Things will go slower, but that will push the DRAM harder. I'm wondering if the behaviour changes.

That seems to fix it though the machine is unusable as it is so slow.

So slowing down the machine with either a 710 or cache off stops the
pointer corruption and putting memory in to Slot 0 rather than slot 1
makes things worse.


--
Experience the future using ARM Technology - ARMBook,BeagleBoard -xM, PandaBoard,Raspberry Pi,iMX6/ARMX6,IGEPv5 & Titanium powered by RISC OS
5.28.

--- SoupGate-Win32 v1.05
* Origin: fsxNet Usenet Gateway (21:1/5)

In message <4c55748f59.dougjwebb@btinternet.com>
Doug Webb <doug.j.webb@btinternet.com> wrote:

Try *Cache Off Things will go slower, but that will push the DRAM harder. >> I'm wondering if the behaviour changes.

That seems to fix it though the machine is unusable as it is so slow.

So slowing down the machine with either a 710 or cache off stops the
pointer corruption and putting memory in to Slot 0 rather than slot 1
makes things worse.

Forgot to say also no difference if I have a PCCard installed or not the
RISCOS pointer remains corrupted.

But there is one strange thing in that with a PCCard installed and booting
in to Windows 98 in full screen/single tasking the Windows pointer is not corrupted in


--
Experience the future using ARM Technology - ARMBook,BeagleBoard -xM, PandaBoard,Raspberry Pi,iMX6/ARMX6,IGEPv5 & Titanium powered by RISC OS
5.28.

--- SoupGate-Win32 v1.05
* Origin: fsxNet Usenet Gateway (21:1/5)

In message <4c55748f59.dougjwebb@btinternet.com>
Doug Webb <doug.j.webb@btinternet.com> wrote:

So slowing down the machine with either a 710 or cache off stops the
pointer corruption and putting memory in to Slot 0 rather than slot 1
makes things worse.

OK one further point is that with the 710 card in and the memory in Slot 0
I still get the loud distorted sound from the speaker so I think this
issue i.e what ever is causing the distorted sound/screething noise is
what is causing corruption on the system but with a slower setup it isn't
so apparent.


--
Experience the future using ARM Technology - ARMBook,BeagleBoard -xM, PandaBoard,Raspberry Pi,iMX6/ARMX6,IGEPv5 & Titanium powered by RISC OS
5.28.

--- SoupGate-Win32 v1.05
* Origin: fsxNet Usenet Gateway (21:1/5)

In message <598f93124eSpambin@argonet.co.uk>
Stuart <Spambin@argonet.co.uk> wrote:

Do you have an oscilloscope Doug, can you have a look at what's coming out
of the power supply?

Hi Stuart,

No I don't have a oscilloscope.

Need to fix my multimeter and find my logic probe as well :-)

I may have a spare PSU somewhere if I could only remember where I put it
for safe keeping as I don't want to rip apart the working RiscPC I have.

--
Experience the future using ARM Technology - ARMBook,BeagleBoard -xM, PandaBoard,Raspberry Pi,iMX6/ARMX6,IGEPv5 & Titanium powered by RISC OS
5.28.

--- SoupGate-Win32 v1.05
* Origin: fsxNet Usenet Gateway (21:1/5)

In article <e966868f59.dougjwebb@btinternet.com>,
Doug Webb <doug.j.webb@btinternet.com> wrote:

OK one further point is that with the 710 card in and the memory in Slot
0 I still get the loud distorted sound from the speaker so I think this
issue i.e what ever is causing the distorted sound/screething noise is
what is causing corruption on the system but with a slower setup it
isn't so apparent.

I wonder if there is a de-coupling cap somewhere on it's way out?

Do you have an oscilloscope Doug, can you have a look at what's coming out
of the power supply?

--
Stuart Winsor

Tools With A Mission
sending tools across the world
http://www.twam.co.uk/

--- SoupGate-Win32 v1.05
* Origin: fsxNet Usenet Gateway (21:1/5)

Doug Webb <doug.j.webb@btinternet.com> wrote:

Forgot to say also no difference if I have a PCCard installed or not the RISCOS pointer remains corrupted.

But there is one strange thing in that with a PCCard installed and booting
in to Windows 98 in full screen/single tasking the Windows pointer is not corrupted in

VIDC provides a pointer that's drawn by the hardware.

Windows doesn't use the hardware pointer, its rendering is entirely
software, which is why that works - just like any other drawing on the
screen. As I suggested, it seems to be affecting the pointer and sound but
not regular drawing.

That it works with the cache off does suggest it's a general memory problem, rather than something specific to just cursor/sound DMA, and it's just being tickled by cursor/sound DMA transfers.

As Stuart says I think this is oscilloscope time...

You might also ask on the Stardot forums, where they have probably the most experience repairing Acorn hardware.

Theo

--- SoupGate-Win32 v1.05
* Origin: fsxNet Usenet Gateway (21:1/5)

In message <aPf*SJZzy@news.chiark.greenend.org.uk>
Theo <theom+news@chiark.greenend.org.uk> wrote:

As Stuart says I think this is oscilloscope time...

Yep, looks like that way though perhaps some simple continuity testing
around RP16 and associated elements may be a first option once I get my multimeter fixed..not a good week for my old hardware :-)

As to an oscilloscope then I don't think I'm not going to purchase one for something that was there fore its spares.

You might also ask on the Stardot forums, where they have probably the most experience repairing Acorn hardware.

Yes seen some similar issues raised there so that may be an option if the continuity testing fails.

--
Experience the future using ARM Technology - ARMBook,BeagleBoard -xM, PandaBoard,Raspberry Pi,iMX6/ARMX6,IGEPv5 & Titanium powered by RISC OS
5.28.

--- SoupGate-Win32 v1.05
* Origin: fsxNet Usenet Gateway (21:1/5)

In article <4382948f59.dougjwebb@btinternet.com>,
Doug Webb <doug.j.webb@btinternet.com> wrote:

Hi Stuart,

No I don't have a oscilloscope.

Need to fix my multimeter and find my logic probe as well :-)

I may have a spare PSU somewhere if I could only remember where I put it
for safe keeping as I don't want to rip apart the working RiscPC I have.

Fair enough. One day, we'll have a face to face MUG meeting and maybe we
can find time to look at it. I do also have some spare PSUs in the loft
but I'm about to start work as Santa again, this time in "The Toy Box" in Ironbridge, and I won't have time to dig them out and test till January.

--
Stuart Winsor

Tools With A Mission
sending tools across the world
http://www.twam.co.uk/

--- SoupGate-Win32 v1.05
* Origin: fsxNet Usenet Gateway (21:1/5)

Fair enough. One day, we'll have a face to face MUG meeting and maybe we
can find time to look at it. I do also have some spare PSUs in the loft
but I'm about to start work as Santa again, this time in "The Toy Box" in Ironbridge, and I won't have time to dig them out and test till January.

Hi Stuart,

Firstly good luck with your Santa stint.

I have done some more testing today , now I have repaired the multimeter,
and tapped out continuity from IC33 /RP16 to the VIDC and also checked
RP11, 13 & 6 as well and they are all fine.

I get +5 to IC33, 26, 30 & 22 as well.

All those are by Memory slot 0 which when it has memory in gives a
distorted pointer and the loud screeching sound.

We plan to do a hardware repair video/simple steps presentation as one of
the virtual rooms for the MUG summer show next year so if we can sort
something out in the spring to catch up in person as a group then that
would be good if you can assist in testing the RiscPC.

Take care.

Doug

--
Experience the future using ARM Technology - ARMBook,BeagleBoard -xM, PandaBoard,Raspberry Pi,iMX6/ARMX6,IGEPv5 & Titanium powered by RISC OS
5.28.

--- SoupGate-Win32 v1.05
* Origin: fsxNet Usenet Gateway (21:1/5)

In article <598f93124eSpambin@argonet.co.uk>, Stuart <URL:mailto:Spambin@argonet.co.uk> wrote:

In article <e966868f59.dougjwebb@btinternet.com>,
Doug Webb <doug.j.webb@btinternet.com> wrote:

OK one further point is that with the 710 card in and the memory in Slot
0 I still get the loud distorted sound from the speaker so I think this issue i.e what ever is causing the distorted sound/screething noise is
what is causing corruption on the system but with a slower setup it
isn't so apparent.

I wonder if there is a de-coupling cap somewhere on it's way out?

or a dodgy track/connection that acts as a capacitor/resister which a
StrongARM will be pushing harder!

I'm not sure a meter will find such a cause.

Stardot users are very good at helping on this sort of problem!

Do you have an oscilloscope Doug, can you have a look at what's coming out
of the power supply?

Chris Evans

--

****** IGEPv5: The fastest RISC OS computer so far! *******
------------ http://www.cjemicros.co.uk/igepv5 ------------
CJE Micro's 'Raspberry Pi & RISC OS Specialists'
Tel: +44 (0)1903 523222
chris@cjemicros.co.uk http://www.cjemicros.co.uk/
Unit 16 Arunside Ind. Est., Fort Road, Littlehampton, W.Sussex BN17 7QU

Don't let the urgent things in life, crowd out the important things!

--- SoupGate-Win32 v1.05
* Origin: fsxNet Usenet Gateway (21:1/5)