Hello!

I am writing this message here into this forum in hope that someone has some worthy ideas for me.

After a longer delay my Case showed up here some days ago it's a Yeong Yang A206 with 200 Watts PSU. I started to assemble the Pegasos II inside it. Mounted the DVDBurner and decided to go out to buy the remaining parts like RAM, Graphiccard and Harddisk. But then I said to myself that it would be better to test the Motherboard with what is there so I pulled out some Micron 128mb modules and mounted them on the mobo + placed a Matrox G400 AGP Graphiccard inside it and turned the machine on. Nothing much happened the Monitor was showing NO SIGNAL. The idea by using the Matrox card came because I talked with a few people who have such a card running in the Pegasos together with Linux so I assumed to show some ordinary OF booting text this should be enough.

But no luck, I then pulled the Graphiccard out and put a Radeon 9200 SE inside it (XpertVision 64mb) the same that someone else is running successfully in his Pegasos 2 as well but again same issue.

Someone suggested me to get a debug serial cable (nullmodem cable) and I connected that one between the Pegasos 2 and my normal ordinary XP 2600 and started capturing logouts.

Well after some time and several conversation with people such as Laire and some other experts I was suggested to ramtest only. There is NO need to connect a Graphiccard to get into the OF (via debug mode) and there is NO need to attach a Keyboard too. The normal procedure would be that the init process of the Pegasos II starts realizing the RAM, fails on the Keyboard, lineary fill the Memory, go into stage 2 and initialize OF which then seriously searches for all Hardware components found on the BUS.

So far so good.

I didn't have much success with the RAM modules so I started to get some different RAM modules amongst them an 256mb Infineon module and an 512mb Lei module. Same problems while the init procedure detects the RAM in the CSx slots and adds it's total sum it 99% of the time fails in lineary filling them. No matter what I stuff in there they usually fail.

I only made it 4-5 times into the OF sometimes by pressing 'q' in the IKARUS entry or sometimes with luck it went smoothly through into OF but this is really rarely happening. And when it happens turning Pegasos 2 off and on again will hang again in the IKARUS prompt with not doing anything. Often it sits in the IKARUS after linear filling failures and sometimes it shows corrupt memlist entries. Well I could say easily it's a RAM problem with one Module but I tested 3 different Brand of RAM modules and all behave similary. When I listen to people who own a Pegasos 2 they usually go into a store buy some ordinary RAM plug it in and continue installing MorphOS. Here I fail with the early init process.

I made various logs about all this which you can find on my Homepage:

Galaxy's Pegasos 2 Debug LOGs

You can depack that archive by using bunzip2 and then tar. Please start reading with 'peg-kaputt.txt' and then dive under into the logs itself.

What can I exclude (only to avoid these questions to show up)

- The CPU card is sitting perfectly otherwise the init code wouldn't get executed nor the OF code.
- The PSU works perfectly it's a 200 Watts P4 suited NEW PSU and some other Pegasos 2 users have the same one (in same Cases).
- There are NO wires on the Motherboard it's a plain connected Motherboard with no IDE, Floppy cable connected, from the Frontpanel there is only the Power wire connected, no other PCI cards or AGP cards were connected.
- No this is no GFX card problem since I don't come into the ability to pass the RAM problem that I encountered.
- I used 3 different RAM modules from different brandings.
- No I never saw the OF or Ambient screen, I only saw the OF output in the debug mode on my normal Linux machine (the XP 2600)
- The RAM I tested is NOT broken since I use them in 3 other computers here in this house.
- The Radeon 9200 as well as the tested Matrox card are AGP cards that have the 'universal' snotches. These cards being universal means they DO run in 3.3v mode as well as in 1.5v mode and you can be 100% sure of this. The snotches tell you this. Here is the Radeon Card's homepage:

Radeon 9200SE XpertVision

But again I don't think this to be a gfx card problem since I never made it so far.

I am a bit worried and don't know exactly whether this could be a faulty Motherboard or RAM problem but again if 1 RAM module fails then it's understandable or better explainable, but if 2 or even 3 different RAM brands fail then it starts worrying.

Any ideas and suggestions are welcome but please serious feedback. I think if you start reading my logs and the histroy in 'peg-kaputt.txt' to get an understanding. So far I circled the problems between the Mobo AND the RAM and exclude everything else.

A little personal info. I am educated electronics engineer thus you can rely that the Motherboard + CPU + RAM and everything around it is perfectly assembled, with a lot of care and following the instructions about 'electrostatics' things...

Feedback welcome.

This might be foolish but do you have keyboard and mouse plugged in? Try that

magnetic

> This might be foolish but do you have keyboard and mouse plugged in? Try that

Well I already excluded this as factor but YES I also tried this with keyboard and mouse applied (as you probably read in the 'peg-kaputt.txt').

Although Keyboard and Mouse are NOT required from what I was told by Laire.

If a Keyboard/Mouse and Graphiccard is not found then it skips over instead showing the OF on the Graphiccard it then shows the OF in the debug window.

The Pegasos 2 works in two steps as far I realized the first step is the early boot sequence which primitive initializes the RAM and checks for some simple other things.

Exactly here I fail 99% of the time. You see that the RAM is checked and no errors were found but then fails the linear fill one step later. And this is indeed quite paradox to me.

The 2nd step is initializing the OF which then checks for all types of Hardware correctly. Brand, Size, etc. If nothing else is found then it prompts the OF stuff otherwise boots into the OS (booting into the OS can be seen as 3rd Step, e.g. booting Linux or MorphOS).

So you see I fail even before reaching the 2nd Step e.g. loading and initializing the OF.

I like to add that I can exclude the 'setenv' issue as well since I never ever had the chance to setenv anything :) Even if I would have abused the 'sentenv' before flashing the new OF then it would still not affect my problem since OF is the 2nd Step and I hung in the 1st Step.

The IKARUS prompt is also the prompt you receive when you by mistake made some sentenv changes before re-flashing the update. For this bPlan has post these few cryptic lines that you need to cut&paste into the Terminal to get the Pegasos 2 running again. It's exactly that prompt.

I am now the 4th Pegasos 2 owner that encountered similar problems, one has the same problem like I have (as I was told). Targhan has some 00000003 Problem with 1st Step too (as he was told it's a faulty Initcode or something), then I with the problem as described and someone else whom I was told about but never spoken with. Maybe there are more non-public reports. Well I don't want to scare anyone nor do I want to make much noise about this but I think we could discuss this in a fair way to figure some solutions that even other people can benefit from.

I also do recommend everyone who plan to buy a Pegasos 2 and plan to buy other Hardware as well.

>>> DO NOT FORGET the NULLMODEM CABLE. <<<

It's a total lifesaver and you may need it later on to sent debug messages (backtraces) from MorphOS to the developers so they can fix the code and offer you an even more stable MorphOS with each incremental version.

Quote:

Sorry, I have no idea concerning your problem, since indeed it seems unlikely to be the RAM after three different modules, including an Infineon.

But in general that what I quoted is no guarantee, as I experienced:

My first buy was a module from Xtreme (DDR 266, 512 MB). With this one, the Pegasos II booted into the OF and MorphOS, but memtest failed, sound was distorted and format&verify used to fail while testing full write. On a PC, the module was okay, though, since the computer shop tested the RAM with a special program when I returned to exchange it against another brand.

With the second module, from TwinMOS (also 512 MB, but a DDR 400 one), everything runs fine.

So I advise future Pegasos buyers to either buy their RAM also from their Amiga dealer together with the board or to check with their local shop if at all and for how long they exchange RAM modules in case of incompatibility.

Hi,

Since I don't do much unpacking on MOS, what prog I can use to unpack
your log? Or can you put a lha file there? Normally I use my XP Laptop
and then I can unpack it right away, but I want to stay on my Peg.
She's got to learn.

Coder

@Coder

bzip2 -d <filename>
tar -x <filename> or untar <filename>

Hi Hooligan,

Thnx!

Coder

Hello,

I would like to add that I slightly altered my previous two replies e.g. removing some minor errors and rewrote some sentences to make it more clear what I wanted to say. So maybe you like to read them over again but the context itself didn't changed much. Furthermore I wanted to thank everyone who spent the time replying to me.

Here again a correct list of RAM modules that I tried with my Pegasos 2:

Micron: mt8vddt1664ag-265a1
2x128mb, single sided

Infineon: hyb25d128800at9207by
1x256mb, double sided, cm266/256/3rd, 16x8 pc266

Lei: 88lshdl0-1pd6
1x512mb, double sided, 64x64, pc2700, cl2.5

A Little note about RAM Testers, there are various ways to test RAM modules one of them is via Software such as Memtest86 which you can find here. The procedure of testing RAM is usually writing a byte, reading the byte and comparing the write- vs. readvalue whether they are identical. There are tons of algorythms outside such as generating a md5sum over the entire Memorypool and then re-calculate the md5sum various times to see whether they match or filling the ram with static bytesequences and much much more. Another way of testing RAM is the Hardware way, some bigger stores have little black boxes where they pull the RAM module in and have it do various checks. You then receive a certificate of the checks that has been done which is some sort of proof that the RAM is ok. But after all it's software that checks the modules.

So I seriously wonder why the init sequence of the Pegasos 2 detects the RAM in the CSx slots, it also sums the amount of maximum memory correctly, it does check the RAM from startaddress to endaddress and usually returns 0 failures but then is incapable doing the linear fill.

I fear that during the delivery of the Pegasos 2 that there may be some transport damages e.g. like throwing the box around in their stations (UPS) and that there may be hair cracks in the MOBO or some other damaged analog circuit or cold soldering points or whatever. I'm not inside the Hardware, just speculating here.

Hi,

How about Kingston memory?

Coder

Quote:
Yes I would be willing to buy Kingston RAM modules as well as I could buy Corsair or Samsung RAM but honestly I am not on a circus move, going from store to store only to find some modules that the Pegasos 2 may like. Until now I haven't heard from anyone having problems with RAM + new Northbridge, people even went into store and bought any ordinary regulary RAM modules and pulled them in (as it has to be) - not to mention that not every shop allows to return electronic semiconductor material due their nature.

I also believe that the INFINEON RAM that I tested as one of the two others is quite reliable.

Some people even told me that they bought their components before the Pegasos 2 arrived, they assembled everything and turned the system on and that's all about it.

If I would go out buying Kingston RAM now and it fails again then the next would tell me, 'hey Kingston is crap please get Samsung' then the next tells me 'hey get TwinMOS' etc.

Due to the request of some people I tried a different PSU to power the Pegasos 2 up. Just for the protocol - I just tried a 433 Watts PSU the problem still remains.

Hi oGALAXYo

Dont you have a local Pegasos reseller somewhere near you, you could try to ask ?

Probably he easely can see what the fault is.

Regards
Gunne

i dont understand your problem since there are a list about DDR ram that rtuns perfectly on pegasos...

I know from some people that have pegasos II, if the program or installer says that the ram isnt compatible, isnt compatible...

You must get a memory that are on the list for not have problems.

I really dont understand your minds, do you will use Pegasos II for linux or MorphOS?

I know some people that dont have knowedgements of electronics and not computer experts, no programming... and have Pegasos II running perfectly...


GET DDR RAM LISTED AS COMPATIBLE!! THATS ALL!!!

Hi again oGALAXYo

> No I don't have a local Pegasos reseller here.

Thats not good

Where do you live ? in Germany somewhere ?

> I doubt they have the technical understanding and capabilities...

Hmmm...

I believe all resellers have very good knowledge about the Pegasos, and could help you quite fast to come to the conclusion what the fault is, and will probably also offer you a solution.

But of course, its your decision...

Quote:
Hello Timofonic,

With my bPlan as well as Genesi Manuals there is NO sign of Compatible RAMs, they only describe to get some DDR RAMs 184pins and stuff them in. To quote some lines:

;---- quote
128 Mb - single sided
256 Mb - single sided, double sided (2x 128)
512 Mb - single sided, double sided (2x 256)
1024 Mb - single sided, double sided (2x 512)

Note: Only use Modules of same Manufacturer and Type, Max. size 16 or 18 Ram units per Module.
;---- quote

No Manufacturer or Type were namely mentioned. The mentioned RAM modules here on MorphZone (RAM that doesn't work) was made and necessary for the Pegasos 1 System which had issues with the ArticiaS for losing data during transfer so they needed a special version of Buffered RAM modules.

This IS NOT required for the new Northbridge and as you can read in the new started Pegasos II Memory compatibility list every RAM people bought worked. And I do claim that INFINEON is a quite good brand besides Kingston, Corsair, TwinMOS etc. Three types of RAM that fails is curious and NOT ok and I don't intend buying 20 types of them only to get 1 working.

If you spent some times in actually reading what I wrote then you figured out that the RAM is detected and checked to have 0 errors it only fails one step further with sequential fill.

And it has nothing to do with Linux, this is quite inflamatory and pure bullshit. Please if you have some serious comments to do then please add them but don't reply with inflamatory crap in mind. I really had three bad last days and I feel not very good with this myself. Coming here and writing about these problems was some sort of 'last try' to get information about what's going on and I have no interest to continue talking with you. But thanks for your interest anyways.

Quote:

@Timofonic

Well sir, it really doesn't help much getting more ram chips if YOU HAVE GOT A FAULTY hardware

NOTHING IS PERFECT (including manuals).

Why not stop saying excuses and get a ram knowed for runs perfectly on Pegasos II? There are few that people use it without problems, there are others that has problems...
If the problem are on the Pegasos II board, then send to DCE and that they will repair your board... dont need crying more here...

STOP SAYING EXCUSES FOR NOT START WITH ATLANTIS, PLEASE BE LESS AS "PROMISES INC" AND TRY BE MORE BETTER PERSON...

All people has problems, you arent only that has it, then not say stupid excuses...

Could be very good if you had Pegasos II running and start making real programming on atlantis for MorphOS, and make it a good program, but i dont think that you make it, i think that you will never make atlantis for MorphOS...

If you are an "autenthic programmer" and you will make that you was said, then you must ignore that people says about you and make the things, if not, say bye forever... real programmers dont have weak mind...

When you has atlantis for MorphOS and be a good program, i will change that i think about you to better...

(and yes, i have a bad english, not all world are learned english perfectly...)

@Timofonic

Of course he will start porting Atlantis when hardware problems are solved. Thats why he has gotten a Pegasos. Chill a bit now.. relax and have a brewster.

@oGALAXYo: the problem of ScummVM isn't relationated about i said.
You always says all time the same, about people that ignores me, that about #### ( you was writed with 1337 letters ) and naming to Trollofonic, Troll-o-phonic or something... You has a lot of arguments for try to shut up me...

We are waiting you atlantis running on MorphOS correctly...

Another question, how do you can know a lot about programming, electronics... and not working? LOL

@Timofonic

Please give oGALAXYo some room to work please. You have to understand and be patient. He is already helping out in other subtle ways, even if his Pegasos isn't working yet. I'm asking you kindly to refrain from haggling the man.

@oGALAXYo
I realize the frustration you must feel, but please do not resort to name-calling. If you feel that someone is trolling you, don't bite on the bait.

I'm not trying to get on to anyone, so please don't take offence. I'm not trying to chase anyone off either; in fact, I'm glad you guys are both here posting and helping other users. Just try to calm down, okay?

Hello,

I probably found the problem of the Pegasos II acting that strange. I'm still not 100% sure whether this *IS* the cause but after taking the Pegasos II out of the case again and having a closer look at the bottom of the Board I detected that soldering lead and soldering lead splints have caused a mechanical contact in the RAM module pins of the backside of the Pegasos II Motherboard (the side that usually looks at the metal surface of the case).

Then I started looking even closer and saw that there is another 'bridge' on the Diskdrive connector. Two pins got contacts with soldering lead that also crosses a bunch of routes from other lines. This is probably not wanted and a production manufacturing issue.

This also explains why I sometimes reach the OF and often not, it's probably when replacing the RAM Modules that the mechanical force that one causes when pluggin the RAM Modules in will temporarely put the Board into a condition where it works. But well I don't know exactly - thus those who created it will be the better people to talk about it.

I only wanted to let you know so we can end this Thread since I was now clearly instructed to have this Board sent to bPlan for a closer review.

when things seem to not be going right, step back and say:

"A subtle thought that is in error may yet give rise to fruitful inquiry that can establish truths of great value."
Isaac Asimov

logic is good. and causes much fewer black eyes!

I had a bit similar problem with my Pegasos 1, only left (or was it right) audio channel was working. Most people "knew" it was becouse of AHI or PegasosMixer settings, but it couldn't be AHI becouse CDDA playback also played only one channel...

Anyway, I was offered possibility to send my board for replacement, but when I took it off, I noticed that one pin of audio connector wasn't properly soldered, so I took my soldering iron and fixed it myself (Well, otherwise I would have had to send my Pegasos away for a while )