Announcement

Collapse
No announcement yet.

Modded Promise ATA66 causing crackling noise

Collapse
X
 
  • Filter
  • Time
  • Show
Clear All
new posts

  • Modded Promise ATA66 causing crackling noise

    I just finished modding my Promise att66 controller card into a Fasttrak66 IDE Raid card. Everything seems to be working fine as most of the hard drive benchmarks I've ran have increased 75-100%. I'm using two 27.3GB 7200rpm WD drives btw.

    The problem is when I try to capture anything with my Matrox Marvel G200-TV PCI. When I capture anything, the sound crackles and seems to be distorted every few seconds. I've tried changing my SB Live PCI card out with a SB16 ISA card but this didnt help at all. Then I used the onbard audio of my K7M and it got better with less distortion and crackles.
    I removed the Fasttrak66 card and captured it straight to a single 27.3GB drive and everything was ok.

    I checked promise's site and they suggested that the problem was that the Fasttrak card was using too much of the PCI bandwidth along with my PCI Marvel card. I've tried it with nothing except the Marvel, Fasttrak, and a sound card and still get the distorted sound. I also tried using their suggestion to change the drives to a master/slave setup but that didn't help either.

    I was just wondering if anybody has had a similiar problem before or have any suggestions on solving this problem. I'm thinking that it's caused by Marvel being a PCI card instead of an AGP card. If it is, then there goes my thought of buying a RR-G to go with my G400 Max.

    Any help or input would be appreciated.

  • #2
    Yes, I had a problem with noise on my captured audio a while back, and it coincided with the addition of two more drives on my FastTrack. It drove me nuts, because I spent about a week uninstalling, reinstalling drives in different setups. The problem, however, turned out to be the stereo mini plug going into my sound card! Apparently, every time I breathed on it, it could go into a barely connected condition, and the fact that I was moving stuff around back there made it seem something else was wrong.

    Comment


    • #3
      I'm pretty sure that my problem isn't with the wiring as yours was. I removed the fasttrak66 card from the system and just hooked up one 27.3gb drive without changing any other connections or cards and was able to capture fine.
      I also tried to output straight from my vcr right into the sound card without going through the BOB. But that didn't help any.

      Thanks for your input though

      Comment


      • #4
        Hi !

        I have the same problem here with a Gigabyte 7IX board, K7-500MHz, SBLive 1024 and Marvel G400.

        Interesting is that I had the same problem with the unmodified Ultra66 controller. Here on my system the crackling noise becomes more worse the faster my drive/array is.

        I´m sure this is a PCI-timing problem caused by the VIA/AMD-mainboard-chipsets.

        Bye, Uwe.

        Comment


        • #5
          Hi,
          I just having the same problem as yours.
          My situtation are quiet similar to yours.
          My config :
          PIII 500e
          Marvel AGP G200
          Modified Promise U66 to Fastrak66
          20 G Maxtor Plus 40 x 2 (video capture)
          13 G IBM for O.S.

          VIA 693A mobo with onboard sound

          I got the same sound problem as yours.
          I 've even try installing another sound card.
          But the sound still distorted.

          However, I captured the video using the
          other IBM HD. The sound is 100% ok. While
          using those HD attached to F66. It is failed
          again. I do suspect the problem may be come
          out from the mod F66. May be the driver problem or hardware problem.

          James

          Comment


          • #6
            I had the same problem last year from January to june with amd K6-350, via chipset and FastTrack 33 and i'm disappointed to know that is present again with the new versions of chipset and fasttrack. I solved my problems when i changed mobo and cpu to Intel (PIII-450 and BX chipset).
            I wrote to Promise, they was so kind but they didn't solve my problem, i upgraded bios to FastTrack and mobo, changed the DMA drivers, but no way!
            In my opinion, the problem is the way the via chipset serve the DMA. You can try to use only one device that use DMA (eg disable the onboard ide controller and disconnect any SCSI controller), if it works is the same problem i got one year ago, you can also try to reduce the block size in the setup of FastTrack to 4K (but you reduce the performance too) or check if there are new drivers for DMA.
            Ciao
            Sebastiano

            Comment


            • #7
              Have you tried loading the UltraTune utility and moving the slider bar towards the "less" setting in the PCI bus utilization screen? I believe that is what Promise was alluding to when they suggested the PCI Bandwisth was being hogged by the FastTrak.
              Hope this helps,
              Perspective cannot be taught. It must be learned.

              Comment


              • #8
                Thanks for the reply everybody.
                Yes, I've tried to move the pci bus utilization bar done all the way but that still didn't help at all.
                It's kinda good news to hear that other people are having the same problems. I don't mean that in a mean way or anything. I thought it was just me that was having the problem but I guess other people with an AMD based motherboard are having the same problems.
                So I guess from what it sounds like, the only real way to get around this annoying problem is to go Intel. *sigh*
                If anybody has had a similar experience and found a solution to it, please let us know.

                Comment


                • #9

                  I brought this problem to the attention of a friend of mine who knows a whole lot more than I do about computers. He suggested it could be a latency problem. Try setting the PCI latency to 64 clocks. If the problems persist, continue to slowly reduce the latency.


                  Comment


                  • #10
                    This is set to 64 for me.
                    I think it is because my sblive is on the same IRQ.
                    I heard that if I move to a different PCI slot, the IRQ will change, but I guess thats not it.

                    Comment


                    • #11
                      SBLive on the same IRQ as a Fasttrak of any type = bigtimetrouble.

                      Dr. Mordrid

                      Comment


                      • #12
                        This thread puzzles me.

                        My setup:

                        K7-500 (prod. week 37)
                        FIC SD11 Mobo
                        3x128 MB SDRAM CAS 2
                        Marvel G200 AGP 8+8MB
                        SBLive Value (the very first edition)
                        All SCSI setup (2940UW + 2904CD)
                        Ultra66 cont. (yet to be modified - no cash = no drives )
                        10/100 MBit network

                        All five PCI slots loaded = bad
                        AMD/VIA chipset = bad
                        SBLive = bad
                        Promise = bad

                        Yet I have never ever experienced trouble with the sound, regardless the OS. I never loose frames with MJPEG ccapture (it can occur when drive is heavily fragmented)

                        In the old days (one year ago) I had an K6-2 350@385 on Ali chipset. Still there was no trouble with that setup.

                        But don't get the idea that everything in my system is flawless, cauz it's not. I have other probs and some I even gave up on fixing.

                        Anyways, I only ment to say that the HW above who other people is fighting with is not the HW I fight with. It can work together, apparently.

                        Ghydda - having all PCI items on IRQ 9, including my Marvel AGP (w2k does that to you).
                        As I always say: You can get more with a kind word and a 2-by-4 than you can with just a kind word.
                        My beloved Parhelia was twotiming with Dan Wood - now she's gone forever and all I got is this lousy T-shirt
                        |Stolen Rig|RetroGames Rig|Workstation Rig|Server Rig|

                        Comment


                        • #13
                          Perhaps this crackling noise is attributed to ground loops? If I recall correctly, the resistor is tied at one end to "ground" via the FT bracket, which in turn should be connected to digital ground. If the screws holding your MOBO's and Cards aren't securely tightened, that could cause problems.

                          Just something to consider. Hopefully I didn't cause more confusion.

                          ------------------
                          K62-300, 96MegRAM, FastTrack IDE RAID (14GIG), Marvel G200, TRV-310 (real soon now)



                          [This message has been edited by JerryH (edited 24 June 2000).]

                          Comment


                          • #14
                            Ghydda;

                            These things are not *always* bad, but the majority of times this problem arises one or more of these combinations is involved.

                            Ex: there is a LOT of variability in the quality of VIA chipped mainboards, even more so than with Intel chipped boards, because of their price point. This causes a lot of inconsistancy in how they can handle heavy PCI bus loads, thereby affecting devices like the SBLive, video editing cards and the Fasttraks.

                            That the SBLive is itself a major bus hog that needlessly hits the sytem timer thousands of times a second also plays into it.

                            Dr. Mordrid

                            Comment


                            • #15
                              Hey Doc. M

                              I finaly found the cash to buy the bleeding drives (2x30GB Maxtor 40+).

                              And I must say that I'm more puzzled than ever. After modifying the Promise, I hooked the drives up, created a RAID0 array, bootet into Win2k and partitioned/formated the new drive (NTFS-partition).

                              Then I began capturing to see if I now would get the infamous crackling noise. And I did. So now I have joined the club. Hip-hurray...

                              Turning down the utilization slider in the Promise conf. program helped somewhat on the prob but it is still there.

                              Then I decided to replace my SBLive with my old and trusty SB32. Booted into Win2k, loaded drives -> all fine. Began capture to IDE-RAID. Still crackling noise. Again the slider can inc/dec the amount of crackling noise.

                              When ever capturing to SCSI I get perf. sound quality, regardless PCI/ISA soundcard.

                              I have not tried anything in Win98 yet.

                              And on a sidenote: Performance is NOT what it should be with my IDE-RAID. Something is terribly wrong. The best benchmark i can get is 30-something MB pr sec. And this is very inconsistently. When moving data back and forth between my SCSI setup and the Promise setup, transferrate is useally between 5-10 MB pr. sec. NOT right, right?!!? Total CPU utilization is lov, 15% when doing this.

                              I think the problem in my particular case lies with the Promise cont. Regardless of the situations (games/rendering/capturing/etc) speed is not up to par whenever the Promise is involved.

                              I hope you could have some additional info to throw this way. Things I should try out. I just would like to point out that Win98 is not the most stable thing to run on my machine (configuration problems/lockups), so any test would be grately appriciated if they could be done from Win2k. However, if Win98 is required, then I will comply.

                              Hope to hear something soon.

                              Ghydda

                              ------------------
                              2+2=5 - but only for extremly large values of 2.


                              correcting a slight typing error

                              [This message has been edited by Ghydda (edited 06 July 2000).]
                              As I always say: You can get more with a kind word and a 2-by-4 than you can with just a kind word.
                              My beloved Parhelia was twotiming with Dan Wood - now she's gone forever and all I got is this lousy T-shirt
                              |Stolen Rig|RetroGames Rig|Workstation Rig|Server Rig|

                              Comment

                              Working...
                              X