PlanetZ forums - scopeusers.com

Sooo...
Ive finally put my system together, but I cannot get the Creamware boards working reliably. 21 DSP SCOPE System ver 4.5, PCSystem as follows; PCI slots used in order from AGP slot:

XP Pro SP2
ASUS A8V with AMD64 3000+

ATi 9550 AGP video outputting at 1680x1050
PCI slot not used.
Pulsar2 SRB (IRQ 17)
Scope Plus (AES-EBU) (IRQ18)
Adaptec 2940U SCSI PCI card connected to K2500R/KDFX
Marvell 802.11g PCI card
________________________________________________________

Errors:

• >> PCI capacity limit reached. (Message repeats and will not be silenced)
  >> Instant crash on loading certain objects. For example I loaded UKnow007 and the screen went blank. No blue screen, no warning. Power stays on but all video output stops.

________________________________________________________

I hope someone can help. If this system is not workable I can sell it on ebay and buy something more appropriate.

If I am forced to change my motherboard is it possible to use scope on PCI-E dual graphics card systems (SLI/Crossfire)?

thanks in advance for any advice...

well, that chipset isn't the most brilliant in PC performance, but still your 'results' are a bit heavy.
your Scope cards might not be found in the proper sequence - it looks like the SRB is found first, and the bigger cards with the IO as second.
In the DSP meter box you'll find a 6 and a 15 DSP entry - which comes first ?
That one is the 'main' card

The 'optimal' sequence depends on the type of the 15 DSP board (i.e. if it's a first or a second generation card).

You can tell it by the IO extension - would it fit on the SRB as well ?
Then it's a 2nd generation card and you want that to be found first, as it's bigger and has the IO system.
Exchange the position of the boards if the 15 DSP card is listed below the 6 DSP card in the DSP meter. This could already make a big difference.

If the IO extender would not fit on the SRB then the Scope Pro has a slower IO system and you would want it to be the 2nd card if you want to take advantage of the lower latency the SRB (theoretically) offers.
Now since the SRB doesn't have a physically IO plate connected to it, there IS a performance penalty and it may not work reliably on some systems.
Your description reads a bit in that direction, as such massive symptoms are quite rare.
You may increase the ULLI latency and see if things improve (incase the Scope Pro is a 1st generation card)

In any case the WLAN and the SCSI card will most likely hog the PCI bus and are never a good idea in a Scope system.
WLAN is highly questionable anyway in a DAW, unless you ignore all safety aspects. A 'reliable' encryption is a tremendous processing load, anything else is simply cracked within minutes.
Mind you, people (breaking into networks) will not want to steal or compromise your private data - they will abuse your router and hence YOUR momentary IP adress for whatever criminal activities. It's YOUR IP that will be logged in such cases, and then go prove that it wasn't you who uploaded the files or wrote the published statement...

I'd rather get a Core Duo mobo as they are recently presented here.
Scope wants a good PCI performance and all I can tell is that a 1.6 CoreDuo outperforms my P4 2.4G by at least 30% .

cheers, Tom

I have removed the SCSI card now.

I also reversed the order of the creamware boards. However they still show up with the 6 dsp card first and the 15dsp card second. Do I have to re-install SCOPE before it understands a card change?? After changing the cards the OS prompted me to install both card drivers again and that went ok.

I am wondering if it is a RAM problem (not enough or unreliable memory)

Both cards use those little 1" long internal bus connectors for attaching to the three socket flat cable. (is this the IO extender?)

I still get the PCI capacity limit reached message. It is occurring randomly - it just happened and I had not even used scope. I just booted up and browsed the internet.

go to start/programs/accessories/system tools/system information/hardware resources/confilcts-sharing look to see if the creamware card shows up there. if so, what does it share with?

no, that cable is an internal connector for the DSPs between the boards called S/TDM - it is important that you don't have one end 'dangling', but use only the outer connectors (for 2 boards)
Otherwise the bus is not terminated and produces all kind of nnonsense.

the default installation procedure sets Scope to autostart, so even if you don't use it, it's still active - and that's why you still receive those error messages.
I once had a board with the (infamous) Via KT133 chipset which also gave constantly PCI capacity messages as soon as Scope was activated.

It is extremely unlikely that it is a Ram error (unless you have a specific reason to assume it's damaged), in particular with a VIA chipset and 2 PCI cards which are certainly not programmed in a cooperative way.
I would remove the 2 cards to verfify if they have something to do with it.

cheers, tom

OK --- S/TDM cable changed so there are no dangly ends.

There are no conflicts or sharing as per 'system information'

Boards are reversed so now Scope is on IRQ17 and Pulsar2 on 18.

Hard to get rid of 802.11g card, with no internet I am crippled and I dont have the cable to run to the router. I could buy a USB Wifi-dongle but not for a couple of days until I can get to the shops.

I am confused about what is meant by IO Extender. My Pulsar2 SRB board has a blank back plate and the Scope board has a multipin breakout to Midi, AES-EBU and balanced IOs.
Additionally the Pular2 has a multipin out next to the backplate, the same multipin out on the Scope board has a small PCB plugged into it.
Also on the Pulsar2 is a black plastic shrouded 10 pin out (looks like an IDE out but smaller). There is a 10 pin out on the Scope board too, but without the plastic shroud.

---Thanks everyone for such prompt and helpful replies.
---Sorry if I appear such a noob. Its 'cause I am!
---This is such a great community, Im very glad to have joined it finally!

araaran wrote:...Additionally the Pular2 has a multipin out next to the backplate, the same multipin out on the Scope board has a small PCB plugged into it...

well, that reads good and points to 2nd generation boards, makes things much easier to sort out.
I meant to remove the cards only temporarily, as the situation would change significantly if the cards have something to do with it - if it remains the same without the cards, then the source is somewhere else.

To be safe it would be best to uninstall them completely and remove them physically from the PC (maybe the install was just a short run ?)
Otherwise just disable them in the Windows device manager, but that may still leave some services and modifications active... it depends on the specific installation...

good luck and cheers, Tom

furthermore, when you have just 2 boards, remember you must use the connectors 1 and 3 from the stdm cable (i mean, do not use the center one, as it is dedicated to a 3 board system).

I would just add that, indeed, 2d generation boards (have a look on the white sticker -date, SN) are by far more efficient when it comes to run a host of "big" plugins (minimax, STW reverbs,etc..). And, yes, i would unistall all your boards and scope 4.5 and focus on a clean installation (your scope pro 1st, scope 4.5 softaware, then add the pulsar2 (restart the computer each time) ...

Like it is the case for other soundcards (uad, Pro Tools...), the choice of the motherboard really matters. Many of us have used this forum (i have alrady asked creamware staff too) to chose teh appropriate board. Hope you will fix your problem !

Jo

To change the sequence of the scope cards seen by your system open cset.ini and change tthe row on which cards are seen ,boardid numbers.
Regards,
Dimitrios

[board0]=1
[board1]=2
[boardid]=0

Thats an example of three cards where I changed the sequence.
Just change the 1,2,0 with whatever best until you get rid of pci overflow warnings.

as no one mentioned the samplerate setting, i do on what samplerate do you work? i'd recommend 44.1khz. If you have lots of ASIO connections, i.e. a big mixing project, and/or maybe one of the famous STW A/P100, you will never be able to work eighter on 96khz or sometimes 48khz, they are just too pci demanding...

and just in case you ask where the cset.ini is found: ../app/bin in your root Scope Folder, usually C:/Scope_'Version'

cheers
roman

i can run several P100 (full version) under 48khz without any problem but of course it depends on your motherboard and the way the pci slots are run...

Jo

but for shure not beside 24 ASIO2-24 channels

Thanks everyone.

Will do a complete de-install of all PCI boards and build system up board by board. Will report back in couple of days when Im off next.

I would run the system at 48kHz if I could. I predict my normal usage as follows:

• 16 inputs in use via adat but probably only 8-10 simultaneously.
  24 channel mixer, insert effects and master effects.
  2x Blue synth.
  a drum synthesiser plug in.
  Modular synth of Juno106 type complexity.

Sequencing and audio recording is done on another machine.

Of course I have no idea yet as to whether I will be able to do all this with 21 dsps... maybe I will need more?

Two further questions:
1 - Is there a thread that I have not found yet that has a list of user tested and known working motherboards that I could refer to when building a system

2 - Is there a known working PCI-E crossfire motherboard that I could use? [/list]

araaran wrote:...Of course I have no idea yet as to whether I will be able to do all this with 21 dsps... maybe I will need more? ...

I have a humble 12 DSPs on 3 Pulsar Ones and almost every synth and fx that's available...
of course I cannot run it all simultaneously, but imho reasonable - with almost twice the processing capacity you don't need to be that shy...

get at least some of the Wavelength freebies (if you don't want to spend additional cash) from the devices forum, though an investment in Adern, Zarg, SpaceF, STW and CWA devices (to name a few) is highly recommended (and a great value for money)

no experience with crossfire, but for a 'modern' system seriously check the Intel CoreDuo boards which Scope4live uses (there's a couple of threads mentioning them) - he uses them according to his nick ... LIVE
Don't spare at the mobo - spare a bit by choosing not the top of line CPU.
Pick a 1.8 instead of a 2.2 or a 2.4 instead of a 3.0 if money is a concern.

cheers, Tom

Hmm, to my knowledge I run both my Pulsar II cards hooked up with either connector 1 and 2, or 2 and 3 without any problems (not 1 and 3 as suggested - which is new to me). And I get some of the best scores on the Masterverb test here. :/

the s/tdm bus (cable) isn't related to the pci bus in any way, in fact in the MV test (which is more or less an empty project) there may not be much (any?) traffic

there IS electrical emission at the open cable end (detectable with an AM radio f.e. - though it may be difficult to distinguish from the mess of signals inside a PC), if it interferes with the operation of the device is pure chance.
Since the probability is rather high, it's better to avoid it.

Yet your observation is pretty in line with experiences from (at least older) scsi devices, where the operation of the bus sometimes turned out to be black magic. Setups supposed to fail still worked, while others (carefully built according to documentation) just failed - but often could be made working again by re-arranging the sequence of devices...

You can in fact cutoff the dangling end with a sharp scissor right at the middle connector if that is an advantage in your specific setup. With a small screwdriver you can unlock the plug mechanism (to get rid of the cable end) and re-use the plug later (in case you get a third card).
'Officially' the connectors are 1-use-only, but I didn't have any problems with the afforementioned procedure. Application at your own risk of course

cheers, Tom

Nice info.

Well, I might try this sometime, but never change something that works of course .

No its sad to say - it is now *worse*

NOTE ON HARDWARE:

Scope is hardware version 1.01
Pulsar SRB is version 2.01

I removed ALL boards from the system and just installed AGP card and Scope card.

I de-installed Scope software and uninstalled drivers.

On re-installation of Scope software I did the following things:

• 1 - Change the mixer to 2448
  2 - Loaded ADAT Source blocks
  3 - Linked all inputs from ADAT Source1 to mixer inputs 1-4(stereo). No links for Source2
  4 - Changed Digital to MASTER, 48kHz Adat1.
  5 - Went to ULLI and changed latency from maximum to 2nd fastest: Resulted in message: This version of hardware can only support (13ms/44.1kHz). I changed to one down, system clicked through speakers as it reset. Then I noticed that from the lists of supported latencies, the hardware probably only supported the slowest mode, so I changed it to the slowest mode.
  6 - I went back to project page and attempted to load module SoundcardSource

At some point in this process I saved default and 'normal' projects. This order of things resulted in the system locking the video cutting out. Nothing but a hardware button reset will reboot.
Now on starting the system, boot up is normal until Scope loads into the tray then it locks in the same way as above.
System fine in Safe Mode, its definitely the Scope.

Only thing I noticed was that the system as I set it up was still running in 48kHz not 44.1, but it cannot be surely that the Scope board cant run in 48?

you made the scope card digital master? fine. but then the adat setting on the card is meaningless and you must set the converter box(es) connected via adat to SLAVE.

if you wish to slave the scope card to adat1, you must check the box for adat1 and then set scope to SLAVE. then set the AD/DA box to MASTER.

i hope you get my drift......

araaran wrote:No its sad to say - it is now *worse*
Scope is hardware version 1.01
Pulsar SRB is version 2.01 ...

sorry to read that, your latency figures verify that the Scope is a version one, with slower IO and less PCI capacity.

Unfortunately some modern chipsets don't support a generation one card as the SFP 'main' - card.
If you remove the piggy board from the Scope (the thing where the external connectors are mounted to the card) - does it look like it would fit on the SRB ?
I don't think so, but I've never seen such a card...
If it does fit, I would mount it on the SRB and check if the installation works with that card - as it's generation 2 it should work.
Leave ULLI latency at 13ms for the time being.
Does this work reliably ?

Then add the Scope, which will act like a DSP extender.
If that also works, reduce the ULLI latency and check if things remain stable.

If the IO plate doesn't fit the SRB, then it's a (slightly) different story.
The (basic) software should still be able to run even without the IO-plate (as that's what the XTC version does), but some devices (sorry I forgot which) require the IO plate.

As it seems your original setup was something like this (i.e. the SRB first card, but the IO on the 2nd) and it did not run stable or as intended, you might get along with 13 or 24 ms latency, but a first generation card will steal PCI performance in any case - of course not to the degree you've described above...

bottomline is that your setup depends on the SRB (possibly supplied with an IO extender) to be SFP's primary card, and there are ARE chipsets which will support an old Scope card as #2
You may even get along with routing the IO over the 2nd card, but obviously not with the VIA KT800 chipset of your mobo.
An old BX or Intel815 chipset would probably do the trick, but that's not what you were after...
In fact it's not very likely that any modern mobo will work (reliably) with low latency but IO from the 2nd card.

cheers, Tom