>Date: Tue, 14 Jan 2003 11:39:20 -0500 >From: Dan C <[EMAIL PROTECTED]>
>At 07:28 PM -0600 01/13/2003, Jeff Walther wrote: >>Sure, bad RAM. Happens all the time. It used to be quite rare, but >>with the tight margins quality control seems to have suffered. I >>got seven out of eight bad 128 MB modules from Velocity Upgrades in >>one order. Three out of ten bad in another order. >7 out of 8 bad? 3 out of 10? That's totally unaccepable. I'd >switch suppliers PDQ! FWIW, over the past two years, I've made a >couple of bulk purchases from various places; resale to local >customers, excess via eBay and USENET. Before shipping 'em out, I >tested them in one system or another. Only four or five were bad - >total - in two years! Well, since Velocity Upgrades does not seem to be operating any more, I seem to have switched suppliers. :-) On a more serious note, the sequence of events went something like this... 128 MB DIMMs were approximately $130 each and had been or higher for a couple of years. Velocity showed up with DIMMs for $80 each. So I bought ten thinking this might be a short time promotion. Folks were always doing crazy below cost sales back then to "build our customer base". The first ten I bought had clearly identifiable Mitsubishi chips on board and worked great. The price dropped to $50 each. I bought ten more, again from Velocity (though other suppliers were priced similarly by this time) because I had a good first experience. This time three of ten tested bad. I sent them back. Two of the three tested bad. Return. One of two tested bad. Return. The final one was finally good. Getting RMAs from them was fairly painless and they turned the returns around quickly. Chips on this set were unidentifiable. I think they were from USA Tech or something like that. A friend in Britain wanted to get 8 modules and Velocity had the best price at about $30 each. I ordered them for him, because the logistics of ordering from Britain can be a pain and if you need to do returns.... Anyway, seven out of eight tested bad. Four of those seven would prevent the machine from booting at all. One only appeared as 64 MB no matter what gymnastics I went through. Etc. This may have been deliberate. When placing the order I had to phone in, because Velocity had redone their web site and their stupid Java programming wouldn't let one complete the order page. Furthermore, there was no longer *any* phone number of any kind on their page. I was only able to call them because I had an old invoice with a number on it. I berated them for their stupid decision to redo a web page that had been working perfectly well, into one that no longer worked with older browsers, didn't worked properly even with modern browsers and provided no way to contact them when their stupid web programmers proved their incompetency. So maybe they sent me seven bad modules on purpose. I was polite in my criticism, but I made it clear I thought they had let an idiot (the web programmer) take over the company. The fellow I talked to didn't seem to take it well. The replacements tested OK and I sent the modules on to Britain. After they arrived, one more failed, which may have been the remaining module that seemed okay orginally, but there's no way to know for certain. He mailed it back to me and I got it replaced and we haven't had any further problems. After that last episode, no I wouldn't have ordered from them again. I'm just glad that we tested all the RAM thoroughly using the RAM Sandwich method for at least 1500 iterations of RAMometer, because those lifetime warranties aren't worth much at this point. Now, all that said, I think there's way more bad RAM out there than folks realize. If you haven't tested it thoroughly using the RAM Sandwich method, you could easily have bad RAM causing you occasional problems and just not realize it. When I first started testing RAM, I thought 20 or 30 iterations of RAMometer was sufficient. And I thought, you just plug the RAM in any old way and test. Then I got inconsistent results and hang ups. After a bunch of experimentation, I realized you've got to have some known-good sticks at the top and bottom of your address space and the only DIMMs you're really testing are the ones in between those known-good sticks. Furthermore, I've seen DIMMs that fail consistently on the 1200 and something iteration. So you've got to go to some number larger than that to be sure your RAM is okay. I use 1500 at this time. Sure, a DIMM that doesn't fail until the 1200th iteration has one of teh rare flaws that doesn't occur often, but it's still going to bite you and cause a freeze, corruption or whatever from time to time. And most bad RAM does fail in the first 30 or fewer iterations of RAMometer. But some takes more. The really important bit is the RAM Sandwich configuration for testing. Because testing software doesn't test the portion of the DIMM in which the OS is residing. >But I've had to swap out dozens of returns from home-users... > >I don't care what joe schmoe pc "experts" say - ALL high-speed parts >*are* static sensative. They also tend to object to having finger >oils, chocolate, pencil eraser, coffee, kool-aid, and other >unidentifable substances schmered on the edge contacts... > >Please Please Please, everyone, remember to ground yourself and your >computer properly before and while touching any static-sensative >parts! Be especially careful if the air in your home/office is >particularly dry. I use good handling technique. Some portion of your users may not. Some of them may just be experiencing RAM that was bad in the portion of the DIMM you were unable to test properly. Jeff Walther -- PCI-PowerMacs is sponsored by <http://lowendmac.com/> and... Small Dog Electronics http://www.smalldog.com | Refurbished Drives | -- Sonnet & PowerLogix Upgrades - start at $169 | & CDRWs on Sale! | Support Low End Mac <http://lowendmac.com/lists/support.html> PCI-PowerMacs list info: <http://lowendmac.com/lists/pci-powermacs.shtml> --> AOL users, remove "mailto:"; Send list messages to: <mailto:[EMAIL PROTECTED]> To unsubscribe, email: <mailto:[EMAIL PROTECTED]> For digest mode, email: <mailto:[EMAIL PROTECTED]> Subscription questions: <mailto:[EMAIL PROTECTED]> Archive:<http://www.mail-archive.com/pci-powermacs%40mail.maclaunch.com/> Using a Mac? Free email & more at Applelinks! http://www.applelinks.com
