From rick@dgii.com Sat Jan 10 15:53:46 PST 1998 Article: 61929 of comp.publish.cdrom.hardware Path: ix.netcom.com!newsfeed.direct.ca!news-peer.sprintlink.net!news-backup-east.sprintlink.net!news-in-east.sprintlink.net!news.sprintlink.net!Sprint!209.98.98.14!visi.com!news-out.visi.com!samba!ipcroe.dgii.com From: Rick Richardson Newsgroups: comp.publish.cdrom.hardware Subject: Re: Reasons For Philips Failure?? (long) Date: Sat, 10 Jan 1998 09:46:06 -0600 Organization: (missing) Lines: 299 Message-ID: <34B797BE.3F2595F1@dgii.com> References: <34AEFA79.3C3C@ghplus.infi.net> <34B74AEA.D4F3F27D@best.com> NNTP-Posting-Host: samba.dgii.com Mime-Version: 1.0 Content-Type: multipart/mixed; boundary="------------E886DE326ACD2494C47D10E" NNTP-Posting-Date: 10 Jan 1998 09:46:11 CST X-Mailer: Mozilla 3.04Gold (X11; U; Linux 2.0.32 i686) To: Andy Sparrow Xref: ix.netcom.com comp.publish.cdrom.hardware:61929 This is a multi-part message in MIME format. --------------E886DE326ACD2494C47D10E Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit Andy Sparrow wrote: > There's some conjecture BTW, that the "spring problem" affects mostly > Polish-built > drives, and the Dutch(?)/Belgian(?) manufactured units don't suffer from > it. > > My drive was built in Poland... > > Just before Christmas, the drive consistantly starting failing to write > more than > ~515Mb of data, giving mostly the infamous '50H, Write Append' error > (once > or twice it reported 'Buffer Underrun').. I checked and replaced the > SCSI > cabling and termination, heat build-up not a factor, tried every brand > of > media - this just resulted in a large pile of coasters. > > I was 200% convinced. Swapped the drive out without changing ANYTHING, > put a Ricoh MP6200 in there instead. Everything worked absolutely > perfectly, > burned 3 650Mb data CDs on the spot. 1000% convinced at this point... > > The Philips drive had successfully burned between 15-25 CDs. It then > burned > straight coasters 10 times in a row, all on attempts to burn large data > CDs. > Of course, if you can repair it yourself - all well and good. > If anyone can point me at the original, fully detailed text for the > repair procedure, I'd > be incredibly grateful - it's no longer on DejaNews it seems, and I > didn't save it when > I first looked at it... I couldn't locate the "original" spring fix on DejaNews either. but here is the one I wrote up a few days ago -Rick -- Rick Richardson Sr. Principal Engr. Digi Intl. Email: rick@dgii.com 11001 Bren Rd. East http://www.dgii.com/people/rick/ Minnetonka, MN 55343 Tel: (612) 912-3212 --------------E886DE326ACD2494C47D10E Content-Type: text/plain; charset=us-ascii; name="fix-hp4020i.txt" Content-Transfer-Encoding: 7bit Content-Disposition: inline; filename="fix-hp4020i.txt" HP 4020i / Philips CDD2000 Spring Fix for Write Append Errors ------------------------------------------------------------- My HP 4020i CD-R recorder (a.k.a. Philips CDD2000) recently (December 1997) started getting the infamous "Write Append" errors when trying to write CD-R's longer than about 500 MB's. I got a tip that this problem was due to a "spring" wearing out or of insufficient strength from David Neal on the cdwrite@pixar.com mailing list (Unix CD-R software mailing list). Armed with this clue, I searched DejaNews on the subject of CDD2000 & Spring. Here I found basically two theories for the problem - the spring theory and another one that said it is dust on the laser lens that needs to be blown off with compressed air (*not* wiped). Since my HP 4020i drive is over two years old, I figured I had nothing to lose by opening it up and seeing if I couldn't apply the "spring fix". Also, at the same time I blew off the laser lens with com- pressed air. After applying the "spring fix" and blowing the lens with air, the drive now works 100% again. I cannot say with confidence which of the these actually fixed the drive. I attach the approximate procedure I used for applying the "spring fix" below. I disclaim any responsibility for the correctness or in- correctness of this fix. Applying the fix will void your warranty. You should not attempt the fix unless you have the proper tools and ability. -Rick Richardson Tools Needed: - Torx T-10 screwdriver - Torx T-8 screwdriver (even smaller than above). I did not have this small a Torx screwdriver, so I used a regular slotted screwdriver with about a 1/8" blade from a jewelers screwdriver kit. I was able to wedge the slotted screwdriver into two of the Torx slots and get enough friction to remove the screws - *Very* small phillips screwdriver. Again, I did not have this screwdriver, but was able to use another of the slotted screwdrivers from a jewelers kit. - Two pair of small needle nosed pliers. - Pair of dikes (cutters) used for electronic work Parts Needed: - About 1.25" long, straight but springy wire, about .02" in diameter. I found a suitable wire in an old 5.25" floppy drive that was used to apply pressure to the spindle. I eyeballed its diameter as "a bit bigger than the current spring", which is rumored to be .012" in diameter. Skills Needed: - Good eye-hand coordination - Good eyesight - Steady hands - Patience The first step is to remove the case. Remove two screws with the T-10 screwdriver and four screws with the T-8 screwdriver. Remove metal case. Gently release the top circuit board from two plastic alignment posts -- this may require gradual rocking of the circuit board but do not stress the circuit board as it is very thin. The next step is to remove the main circuit board by disconnecting three ribbon cables and two sets of two pin connectors. The larger ribbon cables are released by moving two pieces of plastic at each end of the cable connector away from the connector body by about 1/8". The cables should then easily slide out of the connector. The smaller cable has a slightly different release mechanism, but again just move it about 1/8" away from the connector body. Slowly rock the two pin cables from their sockets. You should now be able to see 4 more T-10 screws. Remove these to free the drive mechanism from the other part of the metal case. The next step is to remove the smaller circuit board on the laser transport assembly. Remove 4 phillips head screws. Pull the board up and lay it over - you will not be able to fully remove the board be- cause two wires are soldered to the motor. If you look at the transport now, you will see a worm gear which drives a regular gear which drives a rack gear. The rack gear is spring loaded to press up against the regular gear. According to the- ory, the spring which causes this pressure is worn out and/or not de- signed to be strong enough. The spring itself is a straight length of springy wire, about 1" in length and rumored to be about .012" in diameter. Since I don't own calipers, I couldn't verify this. The spring is held in place only by its springiness. Each end fits into a slot and the middle is bowed down an under a notch in the plas- tic rack gear. What you want to do is add a *second* spring wire. You should not need to remove the original spring wire. I found a suitable wire in a 5.25" floppy drive I had laying around. I removed the wire from the floppy drive and straightened out a couple of bends that were in it and ended up with a straight piece of springy wire about 1.25" long, which is longer than the spring in the CD-R drive. You want it longer for now because it is easier to install it that way. It will be cut to length later. Use the two pairs of needlenose pliers to straighten the spring wire. With your finger, rotate the worm gear on the shaft of the motor to move the transport carriage so that the center of the spring is under the center of the regular gear. You should be able to see the notches that hold both end of the spring now. Lay your new spring on top of these notches. Using the blade of a small screwdriver perpendicular to the length of the spring, press the spring down in the middle until it is underneath the slot in the black rack gear. This bows the spring about 1/8" in the middle. The spring should now be in place. Using your smallest dikes, cut the ends of the spring wire off so that they are the proper length. Wear eye protection when doing this, and if possible grasp the end being cut off with needlenose pliers so that the wire won't fly around the room or worse into the drive mechanism. With the drive still opened up, use a can of compressed air to blow off the dust on the laser lens. DO NOT touch the laser lens as it is magnetically floating in its holder. Now, reassemble the drive by reversing the disassembly instructions. -Rick APPENDIX A - A layman's analysis of the failure mode ---------------------------------------------------- As I said above, I cleaned the laser lens as well as adding a second spring. And I cannot vouch for what is the actual fix for the drive. But I couldn't see how a dirty lens would make the drive fail with Write Append errors only for large CD's (> 500MB's). Other failure modes may, in fact, be fixed by simply cleaning the lens. So, when I had the drive open and was looking at the spring mechanism I did some experimentation to determine how the spring failure mode could possibly occur. I cranked the laser transport all the way out to the edge of the disc by rotating the worm gear. The rack apparatus will look like this in ASCII: G @ XXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXX A# X C# #--------------X---------------# B X @ is the regular gear. X is the rack gear. A and C are the loose plastic supports on the transport that hold the rack gear. The "-" is the spring wire. B is the point where the spring wire slips on under- neath the plastic finger on the back of the rack gear. Point B is hard to draw in ASCII. Imagine in my picture that the spring wire is pushed down into your CRT and under XXXX exactly at point B above. The spring is then really bowed down into the CRT and also down towards the bottom of this post. It is thus applying two forces to the rack gear -- one that pushes the rack gear up and out of your CRT (and against points A and C) and one that is pushing the rack up towards the top of this posting and against gear G. Notice that the rack gear is cantilevered to the left of point A. This, more than any other, is probably the fatal design flaw, With the transport all the way out towards the edge of the disk, gear G falls at the end of the cantilevered portion. It can then exert downward force there, which will tend to want to make the rack move away from attachment point C. But attachment point C is nothing more than a groove. If there is insufficient force from the spring to hold the rack against the groove, the rack will "pop" out of the groove and away towards the top of this posting. Thus, gear G looses good contact with the rack. I tried this several times and was surprised at how easily I was able to make the rack "pop". Yet, it maintains sufficient contact with the gear to return back into place when you crank the transport back to home, thus covering up the problem by making it no longer a "hard" failure. -Rick APPENDIX B - Source for Springy Wire ------------------------------------ From: Date: Fri, 9 Jan 1998 08:23:50 -0600 (CST) Cc: Rick Richardson Subject: Re: HP4020i/CDD2000 spring fix Hello; Last evening my son and I also successfully repaired two HP4020i which had the spring problem. In our case we used the .02 inches in diameter wire normally used to cleaning the welding/cutting tips of my cutting torch. The wire has the proper tension and 'spring' to it. Once the new 'wire/spring' was in place and the drives reassembled they were back working. We also applied some teflon lubricant to the guide rod, and the oppo- site surfaces. Terry L. Ridder Blue Danube Software (Blaue Donau Software) "We do not write software, we compose it." APPENDIX Z - Miscellaneous -------------------------- CD Burning Software: cdrecord 1.6a7 and xcdroast on RedHat Linux 5.0 Linux kernel 2.0.32 CD Burning Hardware: HP 4020i, BusLogic 948 SCSI, IBM 9GB UltraStor Pentium Pro/233 (overclocked from 200) Document location: http://www.dgii.com/people/rick/fix-hp4020i.txt Primary Author: Rick Richardson http://www.dgii.com/people/rick/ Last Update: Sat Jan 10 09:19:21 CST 1998 --------------E886DE326ACD2494C47D10E--