Report on sequencer problems due to RAM

EJ2 · Post by **EJ2** » Wed Oct 17, 2007 2:38 pm

Greetings all,
A few weeks back I had reported on a problem with my sequencer acting up when loading a combi for recording. Well, it's fixed. Thanks to the genius of Daz (aka Darren Richards), my OASYS is back to good health.

Just a bit of background: Dan Phillips and Brandon Daniel, two super Korg R&D blokes, have been closely working with me to try to solve my persistent sequencer problems including a hard drive swap. Now I know how to really get inside an OASYS.

Hope I never have to do that again with these shakey arthritic fingers. Anyway, guys, I really appreciate all the extra effort and support you provided me.

Since the HD swap didn't lead to a cure, I decided to follow up on a suggestion that Daz provided near the beginning of my problems - check the ram sticks. A few days ago Brandon thought I should give this a shot too, after reporting that my hard drive worked perfectly at R&D. So, this morning, I opened my OASYS (again), removed the factory ram, moved the Kingston memory to the back slot, closed up and rebooted. Shazaam! The sequencer performed as it should!

I remember some discussion way back about trading the factory ram's position with new additional ram. So, I decided to leave the Kingston in the original factory location (closest to the rear) and to put the factory ram back where the Kingston was (closest to the keybed), closed up and rebooted. Shazaam!!! No more BS from the sequencer.

Daz, you're a genius! I owe you one. That's coming up, my friend.

So, here's crossing my fingers that this was the problem and that my OASYS stays healthy for a long time.

Thank you once again, Dan, Brandon, and Daz!

Cheers everyone,
Eric

PS: For those having difficulty loading combis into the sequencer, it's not the update 1.3.0 causing this. In all likelihood it may have something to do with your ram memory sticks as was the case with mine. Try Daz's suggestion and see if that improves matters.

Drew FM · Post by **Drew FM** » Wed Oct 17, 2007 3:20 pm

Drew FM wrote:Nigel,

I think you are correct. After a crash with high pitched squeal, I changed my memory around (put the Korg stick in the front slot, & the 1GB PC3200 that I bought in the back slot) & now it works fine. I do however suspect that if I do enough sampling to dip far enough into that front 1GB stick of Korg's, or whoever supplies them, I will probably run into trouble again. Seems to load and play the 503MB piano, and the EXs1 set at the same time ok so far. So the question is, is it the memory, or the memory translator on the motherboard itself? Or is it a setting in the bios causing these problems?

Andy

Mag66 wrote:Daz..

Two quick things.. One; My O isn't midi'd to anything so in my case, midi has nothing to do with my freezes... (I've been having so much fun with the O I haven't even connected it up to the rest of my system! ) and two, since I switched ram modules I have yet to have a freeze.. Today I have been working on the O for about 13 hours straight and not a single glitch.. In the past, I would have expected to have at least one if not two or three freezes during that time frame.

I'm not ready to say that the 'ram switch' fixed my issue for certain.. but so far.. so good.

Good luck!

Cheers..

Nigel

sebbytriton · Post by **sebbytriton** » Wed Oct 17, 2007 5:48 pm

RC-IA also got a similar problem I think.

Korg repair center constate also the memory slots was inverted and the second memory slot was fill with the first oen supplied with OASYS.

Strange ...........

RC-IA can you confirm this point ?

EJ2 · Post by **EJ2** » Wed Oct 17, 2007 5:52 pm

Hey, Drew. Thanks for quoting those. That should add to helping the fix.

That's quite interesting, Sebby. I haven't heard about that possibility, but it could explain some of the problems people are having. I don't know why that should be. Maybe Daz can shed some light on this.
Cheers,
Eric

Daz · Post by **Daz** » Wed Oct 17, 2007 5:53 pm

Very glad to hear that you're up and running again. As Drew very rightly alludes to, RAM problems are always going to be something computery folks will point the finger of blame at when intermittent problems occurs, because of how these things work. It's part of our 'lore' ... rather than genius

I personally felt it was more likely a RAM thing than a disk thing, because your Oasys often booted and operated correctly suggesting the disk and it's contents were in-tact. However it was when you started working with the sequencer that things would periodically go wrong. Imagine it this way; if everyday you travel to work on the same multi-lane highway and there is a pot hole in the second lane at a particular point between the junctions where you join and exit, you probably won't hit it every day because you might not be in that particular lane according to the prevailing traffic conditions. The same thing can happen with RAM, a couple of the memory cells die in a particular spot, but they are not always a location of memory that gets used even when you perform the same operation repeatedly, such as copying a specific combi to the sequencer, after powering up the Oasys. The memory manager doesn't always drive in the exact same lane to work everyday

Daz.

EJ2 · Post by **EJ2** » Wed Oct 17, 2007 5:56 pm

Thanks for sharing that, Daz. Interesting analogy. Now, why would simply switching the ram positions correct the problem? I hope this holds true for my O, but so far so good.
Cheers,
Eric

RC-IA · Post by **RC-IA** » Wed Oct 17, 2007 6:14 pm

i don't know exactly! they changed the dimm i bought a year ago telling that it was not compatible (but it works well before i have the update problem)......

Daz · Post by **Daz** » Wed Oct 17, 2007 6:16 pm

EJ2 wrote:Now, why would simply switching the ram positions correct the problem?

It may not be correcting the problem, per se. You may have simply the moved the problem somewhere else, somewhere you're less likely to run into it. It's like you've shifted the pot hole into the high-occupancy vehicle lane that you rarely use because you mostly drive to work alone

You may find that a new problem manifests itself when you load a lot of samples in RAM or do something else that forces the Oasys to use that specific memory location.

Daz.

sebbytriton · Post by **sebbytriton** » Wed Oct 17, 2007 7:50 pm

Daz, you exactly point the problem where it is.
I don't know Hard disk is a problem, there is a problem base on memory for sure !

Hope Korg could tell us more !

Daz · Post by **Daz** » Wed Oct 17, 2007 7:56 pm

When Nige (aka mag66) and I were having a problems a great tip he shared was to create a bootable CD running memtest86 to thoroughly test the Oasys RAM.

http://www.memtest.org/

You can download an ISO disk image from the memtest site :

http://www.memtest.org/#downiso

... and just boot the Oasys using that CD much like doing an OS update. It won't do anything nasty to the Oasys, it just tests the RAM. IIRC The display is a little squiffy for some reason, but you can pretty much make out what is going on.

Daz.

Daz · Post by **Daz** » Wed Oct 17, 2007 8:01 pm

c.f. http://www.korgforums.com/forum/phpBB2/ ... est#122836

EJ2 · Post by **EJ2** » Wed Oct 17, 2007 10:06 pm

Thanks again, Andy, Daz, Nigel. s**t, now I have to look our for pot holes while driving my OASYS...

Cheers,
Eric

Drew FM · Post by **Drew FM** » Thu Oct 18, 2007 4:46 am

sebbytriton · Post by **sebbytriton** » Thu Oct 18, 2007 5:28 am

Daz wrote:When Nige (aka mag66) and I were having a problems a great tip he shared was to create a bootable CD running memtest86 to thoroughly test the Oasys RAM.

http://www.memtest.org/

You can download an ISO disk image from the memtest site :

http://www.memtest.org/#downiso

... and just boot the Oasys using that CD much like doing an OS update. It won't do anything nasty to the Oasys, it just tests the RAM. IIRC The display is a little squiffy for some reason, but you can pretty much make out what is going on.

Daz.

It seems for you and Nige, memetest didn't find problem.
Nige/you replaced or inverted memory bank ?

Mag66 · Post by **Mag66** » Thu Oct 18, 2007 1:59 pm

Yes, I inverted mine and there was a small improvement over time for me at least.

Im not sure I actually posted this next bit here.. but I did find some random errors via Memtest, in the upper end of the Factory 'stick' that could have given issues when that particular part of the mapped memory was being used... ie, under heavy memory usage when, if I understand correctly, that upper end memory was being used by the OS which would be infrequently. I took it upon myself to replace the factory stick with another stick I had lying around here that matched my additional stick exactly, and since then I have noticed a marked improvement in stability. Maybe using matched sticks had something to do with it too?

I think that action along with the then latest update (1.2.2 that fixed an issue with the Program select screen) resolved almost all the issues I was having on a daily basis... I still think there is something buried deep in the O's code that can cause the very rare 'hang' during heavy memory usage (or it could be that it's another mem issue that only rears it's ugly head during such heavy mem usage) but for the most part, compared to how my O was prior to switching memory sticks over (and then replacing the factory stick) and the v1.2.2 update, my O is as stable as I would hope it would be, all things considered.

Cheers..

Nigel