Cumulus Issues

[Xeron]

Could the cumulus sidestep any ROM issues, by mapping in its own ROM image read from the SD card when software tries to enable the ROM?

[NightBird]

Indeed it seems that mod. I wonder what "improved initialization on power up" means.
About your question, it is the cumulus which is powered from the Oric expansion bus. That is why the original PSU should be replaced by another able to give at least 1A.

Ok. So the Oric is powered throught the expansion bus... Like the Microdisc, and we know it is not a good thing (due to small section of the cable, measurements shows sometimes a lot of voltage up to 0.6V..., More, the regulator does not work in good conditions - 5V on output and nothing on input -...).

Cumana2 does not power Oric throught the expansion bus and use the original PSU to power the motherboard.
There are some docs in CEO-MAG about power issue. I think I will remove the 7905, and replace it by a BERG connector (like Cumulus) on the motherboard to use a PC standard PSU, and not power Oric through the bus (with an adaptor with 1 pin removed for example). so I will power Cumulus, Cumulusbus and Atmos with 3 Berg connectors. Il will test it before with real Microdisc (I am waiting for my 1st Cumulus unit)!

[metadata]

So the Oric is powered throught the expansion bus

no. the Cumulus is powered throught the expansion bus.

[Chema]

@peacer I don't think you have such resistor or something similar in cumulus. Even that logic could be embedded in the CPLD. I really think I cannot be of anymore help in this. The problem is beyond my knowledge and someone with good knowledge in electronics and Oric hardware is needed.

From my point of view there is a high probability than some capacitors in your Oric are failing and the clock signal, phi2, is not of good quality. Smoothed or disturbed clock signal may affect the cumulus, which is not working properly. In this case, the 'no operating system on disk' seems to indicate that some signal derived from the clock is occurring when it shouldn't, probably too late.

There is no way to corroborate this without an oscilloscope. Maybe locating the condensers that are more closely related to the clock signal and checking and replacing them is a good idea. As I said, some retro experts recommend to change all the condensers in old computers.

@xeron with the current cumulus, I am not sure. I in fact don't think so.

[Xeron]

an idea for cumulus 1.5 maybe...

[Chema]

Silly me. I just tried my cumulus and I also get the no operating system on disk as message. Then I select an image, reboot and I see the insert system disk for a second, and it starts loading the OS.

So back to the beginning. I still think the capacitors and phi signal is the culprit, but we cannot completely discard the idea of something else, such as the firmware, collaborating in the failure.

[peacer]

Even though we already checked all the capacitors located in the board maybe we shoudl check it again.

I tried to contact retromaster too. As you said , I'll wait hardware experts for other advices..

Thank you very much for all your efforts !

[Godzil]

One thing. You said you get a "no operating system on disk" message, not "insert system disk". That might give us a hint. When using a Microdisc if you get that message it means you have to adjust the variable resistor RV1, which does not exist in cumulus.

Have a look at this document http://www.defence-force.org/computing/ ... -11-14.pdf

Excerpt:

Si tout est correct, vous devriez obtenir ‡ la mise sous tension líun des messages suivants : ́ no operating system on disc a, ́ insert system disc a ou ́ RV1 adjustment required a en fonction du rÈglage de RV1. Le but est díobtenir le message ́ insert system disc a en tournant líajustable : tournez-le, appuyez sur le bouton Reset et observez le rÈsultat. Si le message níapparaÓt pas, recommencez líopÈration. Si la rÈsistance est trop faible, cíest ́ no operating system on disc a qui apparaÓt. Si elle est trop ÈlevÈe, cíest ́ RV1 adjustment required a qui síaffiche. Si vous níarrivez pas ‡ un rÈglage convenable, il vaut mieux observer les signaux suivants ‡ líoscilloscope : le signal díhorloge Phi2 prÈsent sur le TTL 221 et la sortie des deux monostables de ce mÍme circuit.

So it might indeed be a problem with your phi2 signal. I think it is getting mandatory to measure it with an oscilloscope.

Hum, I don't totaly agree with this excerpt, the real meaning of this message "No operating System On Disc" is when the MD is not able to found a valid bootsector on the floppy. So it may be an hint, but maybe not...

Could the cumulus sidestep any ROM issues, by mapping in its own ROM image read from the SD card when software tries to enable the ROM?

No the ROM cumulus is nearly the same as a real microdisc, the ROM selection is managed by the CPLD that behave like the original hardware, so the Cumulus firmware have nothing to do with ROM select.

[NightBird]

So the Oric is powered throught the expansion bus

no. the Cumulus is powered throught the expansion bus.

I misread . Thank you very much for your clarification! I now includes discussions on the power supply.
However I hope Cumulus/Cumulusbus have connectors BERG that would power them directly from a PC power supply type for example.
I know enough about Oric hardware powering problems. I'd love that power is not a source of problems for the reliability between Oric and Cumulus.
Best regards.

[Godzil]

The problem is:

http://retromaster.wordpress.com/2011/0 ... continues/

Metadata: you didn't do any changes on the hardware part of the Cumulus?

The solution may be to completely rework the least expensive part of the cumulus: the cumulubus, but it will need to add more component on it as most of important signals are in fact not buffered by it (all signal that come from the cumulus are not buffered, especially the Data bus, and /MAP

[metadata]

However I hope Cumulus/Cumulusbus have connectors BERG that would power them directly from a PC power supply type for example

No. I used a different connector. It is not pincompatible and the Cumulus use one pin for the reset. So no good idea to use a BERG connector.
And most of the PC power supplies only work correct if the 12V and 5V line is in use (there are a minimums round about a few watts). So don't use a PC power supply.

Metadata: you didn't do any changes on the hardware part of the Cumulus?

No.
I don't know if it will be better with more buffers. Maybe with very old/slow buffers.

@peacer: Did you try the shorter cable ends? Short both jumpers and connect the black end IDE to the CumuluBus and the grey one to the Cumulus.

[metadata]

just a thought:
could it be a problem with the cumulus when the clock in the oric is faster or slower?

[Godzil]

However I hope Cumulus/Cumulusbus have connectors BERG that would power them directly from a PC power supply type for example

No. I used a different connector. It is not pincompatible and the Cumulus use one pin for the reset. So no good idea to use a BERG connector.
And most of the PC power supplies only work correct if the 12V line is in use (there are a minimums round about a few watts). So don't use a PC power supply.

Metadata: you didn't do any changes on the hardware part of the Cumulus?

No.
I don't know if it will be better with more buffers. Maybe with very old/slow buffers.

@peacer: Did you try the shorter cable ends? Short both jumpers and connect the black end IDE to the CumuluBus and the grey one to the Cumulus.

Adding R on the lines, or maybe correctly sized capacitors to prevent to quick fall/rise signal could be a solution...

I think we may not need to change the cumulus itself, and only change the cumulubus.

(We may also add a buffer on the D line and /MAP signal (maybe on IOCTRL))

But at least making a Cumulusbusbus should be fairly simple to make with R or C on all the lines to prevent the ripple effect.

[peacer]

Yes. I tried to do jumper trick and connect cumulus as if it's secondary disk from the middle part of the cable. Nothing changed and exactly the same result occured.

[metadata]

But at least making a Cumulusbusbus should be fairly simple to make with R or C on all the lines to prevent the ripple effect.

i can try that.

On which line is on the original Microdisk the edititable resistor and what is the value of it?