Although, it may seem that the heat while running can't effect a stored drive, heat is still a major contributor to such failures. Heat causes the seal on leads to fracture. While sitting on the shelf, moisture and other atmospheric contaminants get to the silicon inside. Most silicon has some layer of passivation but that can be compromised by heat as well. Most failures are still related to the bonding wires. Exposure to the atmosphere will oxidize aluminum bonding, causing failure. Dwight ________________________________ From: cctalk <cctalk-bounces at classiccmp.org> on behalf of Ian Finder via cctalk <cctalk at classiccmp.org> Sent: Wednesday, February 7, 2018 11:50:09 AM To: allison; General Discussion: On-Topic and Off-Topic Posts Subject: Re: Maxtor full-height 5.25" drives of death Yes, certainly heat is why these drives go onto a shelf working and come off of it broken with servo and head amplifier problems at an astonishing rate far higher than their contemporary brethren. It also helps answer my question of common failure modes- surely, it must be the refrigeration on the drives that fails while it is on the shelf. Perhaps the freon slowly leaks. Thank you for this insightful response.

On 2/8/18 12:05 PM, Ian via cctalk wrote: Little concrete analysis has actually been done on these drives that has been released outside the drive recovery shops. I would suspect the heads themselves before the preamps.

Al, thanks for the thoughtful response- I was genuinely unaware heads could simply fail outside of the obvious catastrophic modes. (I.E. head crashes) I don?t have a lot of background here- what goes wrong with them? I do wish that the cumulative knowledge developed by these data recovery outfits was just a smidge more public- and I know there are some outfits who claim to ?refurbish? old HDAs, although I have no idea what they actually do. I?d be very curious if you or any other list members have more insight into how these processes work. I would like to some day build an SMD and ESDI emulator, but until then there are a lot of misbehaved drives out there- and if there?s anything that can buy them time, even enough to get a good read, there?s goodness there. - I

Any idea if the drive orientation during storage might make any difference in the failure rate? Usual PCB side down, PCB side up, either side down, either end down? My guess it that it is best to store them in the usual PCB side down orientation, but that is just a guess.

On 02/09/2018 01:40 AM, Chuck Guzis via cctalk wrote: I've found flat pcb down seems better.? Those working are stored that way and those I've found for spares or parts are on their side currently.?? Most I've found have been fried already. Most common problem head crash, likely due to handling. Second up are servo issues maybe due to high temps (some were formerly used in PCs with very poor cooling) and the remainder had board level issues.? Of the board level problems the motor drive fried, and positioner driver dead.? One had a bad microprocessor n the positioner subsystem.? The remainder seem to have likely suffered some ESD or power supply hit.? Like I said earlier these drives used a lot of power and ran hot, the lonest lasting? and still working units were in BA123 which has better cooling.? I know time at DEC if the drive was in a MicroVAX2000 the drive life was about half and a drive in a ba23 was longer lasting.? Also those mounted on edge seemed a bit longer lasting.? The differential between the uVAX2J abd the BA123 life was about 2:1 (about 2 years 24/7 against better than 4), At the other extreme I have over a dozen Quantum D540s (RD52 31mb full height) that are really old and getting older and still work without issues.? Some of those have sat on the shelf for a decade and work fine when I need one. Allison

On 2/9/2018 12:25 PM, Chuck Guzis via cctalk wrote: A tangential question out of curiosity: who here has 5.25" MFM drives they're extremely surprised are still working, and which model(s)? I'll start. I've quite a few MFM drives, but the ones that really surprise me are (2) Seagate ST-506, and (2) IMI CM-5412. My surprise is b/c these are amongst the earliest of models. I've performed *some* work on all these drives, such as cleaning and caps reforming, but haven't yet discovered a way to lube motors & bearings. I think lube issues will spell their ultimate demise. - John

On 02/09/2018 01:05 PM, js--- via cctalk wrote: Working Drive some of which I bought new... These are all MFM: ?2 st506? 1 new in 1981, last tested 1 year ago and still has the base install of CP/M 2.2. ??? It was my first hard disk in my Northstar Horizon S100 crate and still visits there. ? ? Its hard to imagine when 5MB seemed like a huge unfillable space. 2 St412? both gotten used (one was RD51) both work but one has stiction so startup ?? the first time in a while requires manual twist of the drive(inertia). ?3 ST225 one was former RD31 all work though they have been formatted for RQDX2, ??? RQDX3 and WD1003 at different times. ? 1 ST250, out of a PC rescue back in the early 90s, still hasn't died. 14 Q540s some as old as 1985 all work no failures yet.? The first one I have was "tossed" ? ? ? literally as bad by field circus (dec internal in 1986) only to find the RQDX2 had ????? croaked.? By tossed I mean dropped from 3Ft into a steel trash can.? Reformatted ????? (high level using XXDP) for RQDX2 its served me since.? Yes, I pick trash. ?3? RD53 aka 1325s ll had spindown from head sticking to bumper, opened and fixed ??? with the oldest picked up used in 1991. ?1 CDC (wren) 3.5" MFM picked up used in '93 and paired with a Xybec controller in my ??? Micromint SBC180. SCSI group: In the SCSI class I have a large collection of PC and VAX drives from RZ22 though RZ56s and those that worked when I got them still do over 20 years later.? This is a mix of 3.5 and 5inch full height. Of those from PCs the Baracuda 1gb drives? have proven bullet proof.? I have 4 that came from system upgrades (2001) that have at least 2 years 24/7 when I got them. Note the replacement Maxtor drives (9gb) tended to fail after about 13 months (humm pattern, yes?). One noteable SCSI drive is a 3.5" Fugitsu SCSI 45MB given to me for a AMPROLB+ still running more than 20 years and used just yesterday to assemble some 8085 code. IDE: Large selection of pulls all tested good.? Noteables were the 420mb WD, 4.3GB WD and St3660A.? Oddballs like a 5.25 thin bigfoot (slowest and reputed to be the worst for failure) in linux 486box thats still running Slackware 1.2. In all cases the drive mentioned are not less than 15 years old most greater than 25 and a few pushing more than 35 years. The grand daddy of them all is the RL02 salvaged in 1983 and still running along with two packs from then.? Its life was 9 years of Monday through Friday power on and now intermittent about 5 times a month since 1994.? Still booting a PDP-11. Allison

On Fri, 9 Feb 2018, js--- via cctalk wrote: One factor may be that the earliest models were new attempts, and tended to be very carefully engineered, and often conservatively, due to lack of prior experience. On the later drives, they had "learned" and "realized" some of the corners that they could cut. "On analysis of failed drives even halfway through the life of the lubricants. Therefore, cheaper lubricants and smaller quantities with half the expected life would still be adequate to outlast, on average, the other components."? One of the first corners to cut, once analysis of failed units from other causes shows that it is outlasting the rest of the unit. In addition, of course, some lubricants can deteriorate, whereas many other components may have no significant impact from time. Lubricants can fail from contaminants that would not affect most other components.

/Here is my list. 6x RD54 (Maxtor XT2190) >2x OK, 2x Media Error, 1x Actuator Issue, 1x Head issue 3x RD53 (Micropolis 1325) >2x Actuator issue, 1x actuator issue followed by spinning issue (speed sensor?) 4x Seagate ST251 >4x OK 3x Seagete ST225 >3x OK 3x IBM Type 068 >3x Dead On the 3.5" side I have also many dead drives (<1GB capacity). Mostly sticky actuators and dead tantalum caps, but by now nothing I was not able to repair. Many Greetings Ulrich /

On 02/22/2018 07:45 AM, Ulrich Tagge via cctech wrote: 1325s with head stuck will normally spin down.? If the head does not position there is no servo so the motor control shuts the show down.? The fix is pop the cover and remove the offending goo that was the bead bumper stop. Allison

I'm definitely interested but didn't have anything to contribute other than that I have both XT2190s and XT1140s that are still working (at least for now). So yes, please share whatever you find. m ----- Original Message ----- From: "Ian via cctalk" <cctalk at classiccmp.org> To: "Tom Gardner" <t.gardner at computer.org>; "General Discussion: On-Topic and Off-Topic Posts" <cctalk at classiccmp.org> Sent: Thursday, February 08, 2018 3:05 PM Subject: Re: Maxtor full-height 5.25" drives of death On Feb 8, 2018, at 11:03, Tom Gardner via cctalk <cctalk at classiccmp.org> wrote: Good point. The lisp machines I?ve been using the drives with are way past their prime too- pushing 34+ years outside of their warranty. I think next time one fails, instead of trying to understand the problem and repair it so it can last another 30, I?ll just be thankful it lasted this long, gold scrap the boards, and build a gaming PC out of the case. It did, after all, have a good run. /s I will now try, likely in vain, to put this thread back on the rails- A colleague, offline, in the first useful and on-topic response to my query, tells me that there is an analog head preamplifier on the ribbon inside the HDA which is suspect- a drop out here can cause both the ?missing head? behavior, as well as the bad positioning behavior- if the head is on the servo track- common to these drives. This makes intuitive sense. As there are a number of machines that *require* the ESDI Maxtor models- IBM RT, early SGI, Symbolics- in the absence of an ESDI emulator- I think repairs are a good option. I?ll experiment on a 2190 MFM model I recently imaged that shows up with lots of errors on the same heads. It uses the same HDA as the ESDI models. I?ll report back here, though it seems no one on the list is interested in discussion beyond anecdotes when it comes to 80s hardware. My hope is that it simply needs the joints attaching it to the ribbon reflowed. - I

Indeed there are. I use the one designed by David Gesswein, it is great. However I think the OP was talking about ESDI as well. I don't know how ESDI differs from MFM. Regards Rob