heads). Yes, I freed the heads, even though RD53s turn the disk if the pads have turned to goo and stuck the arm. The disk turns freely when I turn it by hand. and I have a copy of the fiche for the RD53, it is a bit hard to read but I will see if I can find where these components are. Thanks Rob

RD53, edges of swapping, Swapping the motor control board did indeed resolve the problem. So now I can look at the faulty board and see if one of the power transistors on there has failed. Thanks Rob

there I have been desoldering the power transistors and testing them with a DMM using the diode tester. There are eight in all, four test OK as two diodes. However, the other four are marked 8702 TIP125, have a Texas Instruments logo, and do not test as two diodes. Looking at the datasheet here http://www.datasheetcatalog.org/datasheet/stmicroelectronics/4128.pdf it would seem that it is not they are not simple transistors and so perhaps my measurements are showing them to be working OK. Here is what I measure with the diode tester: 1st - b-e 1.95V e-b 0.82V b-c OC c-b 0.66V c-e 0.58V e-c OC 2nd - b-e 1.91V e-b 0.83V b-c OC c-b 0.68V c-e 0.58V e-c OC 3rd - b-e 1.90V e-b 0.81V b-c OC c-b 0.67V c-e 0.59V e-c OC 4th - b-e 1.75V e-b 0.75V b-c OC c-b 0.56V c-e 0.58V e-c OC It looks like the 4th one is slightly different to the other three, but I have no idea if this is far enough out of tolerance as to cause the motor not to turn. Can anyone tell me if those values look OK, or what else to look at on the motor control board? Thanks Rob

to not. Holm, Thanks for the procedure for testing the Darlington transistors. One thing I wanted to check though. You say that the collector should be connected to the negative pole of the battery, I thought that, conventionally at least, collectors were connected to the more positive side and emitters to the negative? I have done a drawing to illustrate what I understood from your email (I used an LED symbol for the light bulb because I did not have a simple bulb symbol). You can view it here: https://skydrive.live.com/redir.aspx?cid=fc758a5a91b91301&resid=FC758A5… 1301!602&parid=FC758A5A91B91301!143&authkey=!AO-F3bI24frPXbg Let me know if that is correct. Thanks Rob

Rob Jarratt wrote: You have done the job pretty well in the meantime, so my information that you have the connected LED in the wrong direction on your circuit is proably a little late now. :-) And yes, as Peter already wrote the polarity of the battery is depending of what kind of Transistor you want to test. The TIP125 is PNP so my description was correct. Kind Regards, Holm -- Technik Service u. Handel Tiffe, www.tsht.de, Holm Tiffe, Freiberger Stra?e 42, 09600 Obersch?na, USt-Id: DE253710583 www.tsht.de, info at tsht.de, Fax +49 3731 74200, Mobil: 0172 8790 741

Rob Jarratt wrote: Yes, knew that already... Regards, Holm -- Technik Service u. Handel Tiffe, www.tsht.de, Holm Tiffe, Freiberger Stra?e 42, 09600 Obersch?na, USt-Id: DE253710583 www.tsht.de, info at tsht.de, Fax +49 3731 74200, Mobil: 0172 8790 741

to not. I have breadboarded the circuit above and tested the 4 Darlington Transistors. The one that was showing slightly different values to the rest with my diode test turned out to switch on the bulb regardless of whether the resistor was there or not. The others worked as they should. So I will now look to source a replacement. Thanks for the help with the test circuit! Regards Rob

FWIW, most ocmmon US (and Japanese) transistors are not hard to get here. Certianly 2N3904s and 2N3906s are very common (I tend to buy they 50 at a time..). I don't think you'll have any probkesm getting a TIP125 here eiter. If you can't find oen, let me know and I'll see if I can find a supplier. -tony

On 08/04/2012 21:34, Rob Jarratt wrote: No, a TIP125 is a TIP125. Any manufacturer's version will meet the same minimum specs, or else they'd call it something else. I'd guess the 8702 on the Texas ones is a date code (week 2 1987). -- Pete Peter Turnbull Network Manager University of York

On 2012 Apr 8, at 1:34 PM, Rob Jarratt wrote: Looks fine, the TIP125 is what matters. The 8702 on yours is most likely a date code. Many semiconductors and other components have a date code stamped on them in addition to the part number and manufacturer id/logo. One of the most common forms of date code (particularly for North American manufacturers) is yyww where yy=LSDs of year, ww=week of year, so 8702 would indicate the device was manufactured in the 2nd week of 1987. Look at the ICs in your drive, they may have similar codes for around the same period.

There is an 11/34 for sale on Craigslist in LA but, that's not why I'm posting. What's the Vadic device shown in one of the pictures ? Nothing is coming up in google http://losangeles.craigslist.org/wst/sys/2947330407.html

On 4/8/2012 5:01 PM, Dennis Boone wrote: That was quick. Makes sense.

On Sun, 8 Apr 2012, Dave McGuire wrote: That was my guess as well, with included modem (TXD, RXD, CTS, etc). Mike Loewen mloewen at cpumagic.scol.pa.us Old Technology http://sturgeon.css.psu.edu/~mloewen/Oldtech/

On 04/08/2012 08:45 PM, Adrian Stoness wrote: "F" for "Fixed". It's twice the capacity of a standard RK05, but the disk pack is permanently installed. I heard years ago that they did this due to the insertion/removal misalignment vs. the tighter tolerances required for the higher density. -Dave -- Dave McGuire, AK4HZ New Kensington, PA

Yes, pretty nice. If memory serves they show up as two separate RK05 drives to the controller. I have one of these, the only one I've ever seen in person. It's high on my get-it-running list. It's still down in FL but will be coming up in the next truck trip. (which I hope will be the last!) I do like RK05s. Actually I very much like most of DEC's big drives. -Dave On 04/08/2012 09:03 PM, Adrian Stoness wrote: -- Dave McGuire, AK4HZ New Kensington, PA

kool! the disk packs are indeed kinda neat now to just find some pdp8 rk05 packs and some cables and the temrinator thingy's for my controlor card set for the 8 On Sun, Apr 8, 2012 at 8:18 PM, Dave McGuire <mcguire at neurotica.com> wrote:

On Sun, Apr 8, 2012 at 9:18 PM, Dave McGuire <mcguire at neurotica.com> wrote: They do - that's why it has two unit numbers on it. I saw one in 1984. It was damaged beyond what appeared unreasonable to repair at the time, but today, I'd probably give it a shot. I did get the RK11-C and an RK05J. I haven't tried to restore the RK11-C but I've used the RK05J with an RKV11D. They are easier to repair than many other drives. I've found the lamp in the positioner a common point of failure - easy to spot, too - power up the drive and look through the door - if you don't see the light, the lamp is probably blown. Head crashes are not uncommon, but the heads are pretty robust. I haven't done it myself, but I've read tales from other members about vigorous cleaning of heads after crash events. I've also heard of blower motor insulation failures that have been known to be somewhat catastrophic, so if you are powering up a long-stored drive, inspect the motor and the air path. I'm sure there are other tips in the list archives (like always mounting a scratch monkey ;-). -ethan

is taking notes On Sun, Apr 8, 2012 at 8:57 PM, Dave McGuire <mcguire at neurotica.com> wrote:

Just looked at the pics, andlots of pack boxes in and on the cabnets. Looks like a BA11-FE expansion box in lower left cab. I've worked on numerous RK05-F's, abd they show up as 2 logical drives. you can convert to/ from an F. the inner parts are just the linear posistioner and the heads. the door can be rigged and any pack that matches your controller(12 or 16 sector) can be reformatted. It's pretty simple if you have the parts. Paul On Mon, Apr 9, 2012 at 2:18 AM, Paul Anderson <useddec at gmail.com> wrote:

On 4/8/2012 8:25 PM, Adrian Stoness wrote: Punch line of a BAD joke. Ben.

Duh! --- Enviado do meu Motorola PT550 Meu site: http://www.tabalabs.com.br ----- Original Message ----- From: "Adrian Stoness" <tdk.knight at gmail.com> To: "General Discussion: On-Topic and Off-Topic Posts" <cctalk at classiccmp.org> Sent: Sunday, April 08, 2012 11:57 PM Subject: Re: What is this Vadic thing ?

I think that';s right. I have one. It's very much the same as a normal RK05J, but without the door,pack holder. The disk pack is fiotted when the drive is built and held down by a couple of tension srings. It still ahs only one disk and 2 heads, but has twice as many cylinders as a noraml RK05. The fact that the back is not normally removed emans that any slight run-out in the spindle is not important, since the disk will be in the same position for formatting, writing and reading. I beleive if you do remvoe the pack for any reason, you are supposed to do a low level format when you put it back in. Much of the internals are the smae as the RK05J. The heads are obviously different, you can recognise RK05F heds by the DEC convention that the high density version of somethign is blue (think of RL02 .vs. RL01 packs). The sockets on the end of the head cables are blue on RK05F heads. I assume the positioner graticule is differnet too, but I've never investigated that. IIRC the electorncis is the later RK05J stuff, but with link changes. It appears to the controller as 2 consecutive drives, each of the 'normal' RK05 geometry. -tony

I don;t rememebr that, but it;'s been some time since I worked on one. I don't recall there being any shims in my unit either. And I can't really see the need. Provided the disk hub doen't move relative to the spindle (they are matneitclaly locked together, like in a normal RK05), the tracks will stay in the smae places on the disk so there'll be no problems reading/wriging then. I wonder if you're thinking fo the trick of adding a small ship to the hole of a low-density alignment pack to let it be used in high density drives. I think that's in one of the RK05 manuals. There are certianly jumpers to set, to make it appear as 2 separate drives and to have twice as many cylinders as normal. I thought it was one cycle of the graticule signls for each cylinder in a normal RK05. In which case there mush be some changes for the RK05F. I really can't rememebr... I'll look at the prints sometime and check. -tony

Err, before long it'll be 'limited to only 2 houses' :-) Which HP99000 do you have? That is the perfect reason to get one. I suspect many of use knew little/nothing about resotring a particular machine before it ended up in our workshop. That certianly applies to me. -tony

On Apr 9, 2012, at 3:37 PM, Cameron Kaiser wrote: Try an 11/23. That's what I've got, and I run the backplane off a hacked-up AT power supply tied to the rails. One of these days, when I find one in good condition, it'll go in a BA11 or BA23 (though if I find a good BA23, I'll probably try to find something beefier to go in it). For now, though, it's pretty content in its bare 4x4 backplane (about the size of a VCR) with a decent fan blowing through it. Easy enough to contain for most people. - Dave

On Mon, Apr 9, 2012 at 11:33 PM, Cameron Kaiser <spectre at floodgap.com> wrote: It is not uncommon for the PDP-11/23 (or even PDP-11/03 for that matter) to be paired with a DLV11J quad serial card. One serial point is typically strapped as the console interface leaving three ports at your disposal. One thing that's easy to do is set up a DOS or Linux box as an emulated TU58 and boot RT-11 from that. Transfer rates are usually 38400 bps, and emulated tape skips the long seek times. One can also do this with a PDP-11/23+ board that happens to come with two ports (and boot ROMs) on the CPU board. In terms of traditional mass storage, an RLV11 was cheap back in the day ($50 to $100 when I was seeing them in the mid-1980s) but it needs a particular type of backplane and can only see, IIRC, 18-bits worth of memory. The single-board RLV12 works anywhere there's room for a quad card (i.e., nearly any one except for the dual-height mid-1970s backplane for the LSI-11/2) and sees all 22 bits in a full-blown machine. Either controller can address up to four RL01 (5MB) or RL02 (10MB) drives. More common was a floppy controller - an RXV11 that could only talk to an RX01 drive (or an RX02 set to work as an RX01), or an RXV21 that can talk to an RX02 in RX02 mode. Newer systems, mid-1980s and on, tended towards an RQDX1 or RQDX2 or RQDX3. There are a few differences between them, but they all talk to RX50 dual 5.25" drives and various models of ST506/ST412 hard drives ranging In terms of physical size, the emulated TU58 is only a serial cable to a laptop or desktop in the same room, the RQDXn interfaces to drives that fit in a shoebox, then the RX01/RX02 are microwave oven-sized, then the RL01/RL02 are as tall and wide and at least twice as deep as a floppy drive. There are other options, but these were quite common in their day and were typically larger still. In terms of size, I'd say an emulated TU58 has the smallest footprint, followed by an RXQD3 with a Teac FD55GFR strapped to be an RX33 (I think that's supported on the RXDX3 - it is on the MicroVAX 2000). -ethan

Here's a home friendly PDP once again located in LA. Seems to be PDP 11 season in Southern California for some reason. Just a bit too far away for me and a little pricy http://losangeles.craigslist.org/wst/sys/2950659229.html

On 2012 Apr 10, at 12:41 PM, mc68010 wrote: .. just be careful about what you're buying, the picture looks like to be a badly photoshopped version of this: http://en.wikipedia.org/wiki/File:DEC-PDT-11-150.jpg

I think this same guy had a MicroPDP 11/23 for sale last week. He used a stock photo for that too but, was pretty detailed in describing the machine when I contacted him about it. I got the feeling he did actually have a machine for sale but, just was lazy and used a stock photo. I never followed up so, I never saw the actual box or had any other dealings with him. On 4/10/2012 1:17 PM, Brent Hilpert wrote:

On 10/04/12 8:22 PM, Kirn Gill wrote: browsing of a certain internet cesspool imageboard, styled after a popular Japanese web-BBS, in my teenage years. They used Photoshop like it was going out of style. After a while, you learn how to discern the fakes.) i.e. you've seen a few 'shops in your time. --T

On Apr 10, 2012, at 12:33 AM, Cameron Kaiser wrote: Depends on what you want to do with it. I have a CQD-220 MSCP SCSI card (which I highly recommend) which will support SCSI disks on any operating system that supports MSCP (later versions of everything DEC, and I believe someone has working MSCP patched into 2.9BSD which is the last that'll run on a /23 due to the lack of split I/D). I also have a TSV-05 system (DEC Pertec controller and DEC-branded Cipher-F880 9-track tape drive, which is larger than the machine it's talking to). And I landed a DELQA (good Ethernet controller, ironically running on a 10MHz 68000 which is probably faster than the /23 it's running in), which will be useful once I get my act together to get something running with it. I'd say hunt for a good MSCP controller of some sort first; the SCSI cards are a great deal if you're into old Macs and have lots of older SCSI disks lying about. :-) It's not as simple to bootstrap something onto it, though, not nearly as much as the tape drive. I wrote a quick assembly bootstrap and corresponding Python script to load and execute programs over the serial port so I could e.g. write tapes out and boot from them, but MSCP is a much more complex protocol and I'm more interested in letting the OS handle that for the time being. I know lots of people like to do emulated TU58s (DECTapes) over the serial port. I haven't done that yet. As Ethan mentioned above, the 11/23+ has the serial ports built in - that's the one I have. Nice, all-in-one quad- size CPU board with serial and bootstrap... no fuss, no muss. :-) Of course, the built-in serial only goes up to 19200 baud, which means writing out a full 40MB 1600bpi tape takes... a while. - Dave

I think a related question is 'What will a PDP11/23 do that I can't do with my existing machines'. There are some pretty good PDP11 emulators, so I guess if you want to run the DEC OSes and related software you could do it one those rather than on the 'real' machine. OK, a PDP11/23 will let you talk to real Q-bus peripehrals. For mass storage, most people use a ST412-interfaced winchsseter drive on these machines. There are SCCI interface boards, but they're expensive. You can get SMD controlelrs, but I suspect that gets you into the 'too large' category fast. Waht this mahcine _won't_ do is have a blinkenlights panel. Nor will it really let you 'get inside' the CPU. The CPU is a couple of custom ICs, really. Pwersonally, I'd go for an 11/34 (with a keypad panel) or an 11/44 (no front panel, but an emulator in the CPU control processor ROM that lets you do frontpanel fucntions using a serial terminal). These machines have CPUs built from TTL, PROMs, PLAs, etc and you can spend many a long light with one and a logic analyser... -tony

Well, I'm far too young to really be expected to get into that level of arcane knowledge, but I'm working on a CPU emulator for a virtual ISA meant for use in a video game under development. It's extremely simple (16 opcodes - 15 basic ones, and a single "non-basic" one). It has 64k address space of 16-bit bytes (each address refers to 16 bits, and a separate set of 16 bits - 0x12345678 is 0x2 locations long - I believe that means that the byte is thus 16-bits long here) Addressing for instruction operands is extremely orthogonal. To the point that the stack manipulation functions PUSH and POP are operands and not their own instructions. To push an integer 1 on the stack, you do so: SET PUSH, 1 https://segin-utils.googlecode.com/svn/trunk/dcpu16/ contains the WIP source code for my emulator (and also serves as a "how not to write C code" guide, on purpose - I wanted to see how many tricks I could employ to keep code size down and execution speed high), as well as a copy of the CPU architecture spec. Those interested may want to browse the subversion history for the source to see the different approaches I've tried. For the lazy, use the web interface on the meta-project page: http://code.google.com/p/segin-utils/ under Source -> Browse But coding this emulator, thinking about how to structure the logic - yes, I know, probably too high level, but still, fetch the instruction, decode it, decode it's operands, fetch additional memory if operands are pointers or extend the instruction length, execute the operation, store the instruction, etc., while overly simplified, give me a fair idea what that billion-transistor slab of silicon in front of me (or in my pocket!) is doing when it's doing it's thing. I wrote a somewhat long memo detailing my thoughts on the architecture in general on the forum for the game, under the section for discussing the virtual architecture itself. Link: http://www.0x10cforum.com/forum/m/4932880/viewthread/2750850. Keep in mind that I might be talking out of my ass in huge swaths. Any and all commentary would be appreciated (and may even be used to edit and modify that post - I wish to inform, but I'm certain almost everyone on this list knows far more about real-world ISAs than I do; I was born 4 days before Windows 3.0 was released. Please forgive my mass ignorance.) (Note: I filter the list's messages into a separate IMAP folder that I check far too infrequently, so please CC me on replies, else it might be months, if ever, that I read your reply/) On Apr 10, 2012, at 3:14 PM, Tony Duell wrote:

YEs, and I'm probably too young/have the wrong qualifications/am over-educated to rewind motors, make HP9800 card reader rollers, repair TTL-based processors, align disk drives or do any of the other things I do routinely ;-) FWIW, I reparied my first TTL-based minicomputer (a Philips P850 when I was a first year undergraduate. Yo uare never too young to learn :-) I know it's easy to think that they stuff you enjoy is the stuff everyone should enjoy, and that it's totally wrong to assume that. But I really would urge anyone with a serious interest in how computers _really_ worked 30-40 years ago to get a TTL-based minicomputer (Philips P800, DEC PDP8/e or PDP8/a, etc, DEC PDP11 (one of the older Unibus ones, even an HP9800 desktop 'calculator') and spend amy days wit hthe scheamtics, the microcode listings and a logic analyser. Be warned it quickly becomes addictive 'Oh, I'll come to bed when I've understood just how this darn thing skips the next instruction/handles interrupts/...' -tony

--- On Sun, 4/8/12, Dave McGuire <mcguire at neurotica.com> wrote: And at last we know where this ended up: http://8bitaficionado.com/2011/07/25/auction-hunters-totally-miss-out-on-th… -Ian

You mean, Lord Vadic? http://www.bsd.org/decwars.html On Mon, Apr 9, 2012 at 9:42 PM, mc68010 <mc68010 at gmail.com> wrote: -- Yoyodyne Propulsion Systems: ?"The Future Begins Tomorrow" Visit us at: http://www.yoyodyne-propulsion.net -------- They view life as a bunch o' unconnected incidents 'n things. They don't realize that there's this, like, lattice o' coincidence that lays on top o' everything. Give you an example; show you what I mean: suppose you're thinkin' about a plate o' shrimp. Suddenly someone'll say, like, plate, or shrimp, or plate o' shrimp out of the blue, no explanation. No point in lookin' for one, either. It's all part of a cosmic unconciousness. -Miller, Repo Man

On 04/09/2012 04:08 PM, Mr Ian Primus wrote: Well I sure am glad that got "discovered". I often wonder (and fume) about that machine. -Dave -- Dave McGuire, AK4HZ New Kensington, PA

googled "16 channel vadic" and this came up: http://www.arcelect.com/racal.htm Apparently some type of modem On Sun, Apr 8, 2012 at 7:50 PM, mc68010 <mc68010 at gmail.com> wrote: -- Joseph Giliberti Director of Buildings, Grounds and Facilities InfoAge Science / History Learning Center and Museum Get more information at http://www.infoage.org

On 2012 Apr 8, at 1:34 PM, Rob Jarratt wrote: BTW, the logo on the ebay one is STMicroelectronics (major Euro manufacturer, formerly SGS-Thomson), the 0001 is probably a date code, and STM has a manufacturing facility in Morocco. Datasheet from STM: http://www.st.com/internet/analog/product/65312.jsp

line be a either. Just in case anyone is interested, replacing the failed Darlington transistor fixed the problem. The RD53 now spins the disk again. Thanks to everyone who helped. Regards Rob

On Sun, 8 Apr 2012, Tony Duell wrote: Now that's very simple. Reichelt, TIP 125 STM, cost 34 Euro-cent. Was that hard? I think not. Christian

On Sun, 8 Apr 2012, David Riley wrote: I ordered 40 of them on Friday from Jameco for .35c each. g. -- Proud owner of F-15C 80-0007 http://www.f15sim.com - The only one of its kind. http://www.diy-cockpits.org/coll - Go Collimated or Go Home. Some people collect things for a hobby. Geeks collect hobbies. ScarletDME - The red hot Data Management Environment A Multi-Value database for the masses, not the classes. http://www.scarletdme.org - Get it _today_! Buying desktop hardware and installing a server OS doesn't make a server-class system any more than sitting in a puddle makes you a duck. [Cipher in a.s.r]

On Sun, 8 Apr 2012, ben wrote: Nonsense. They have stock of the .39c parts. Min qty is 1. The .35c part will be in stock in 3 days and the min. for THOSE is 10. g. -- Proud owner of F-15C 80-0007 http://www.f15sim.com - The only one of its kind. http://www.diy-cockpits.org/coll - Go Collimated or Go Home. Some people collect things for a hobby. Geeks collect hobbies. ScarletDME - The red hot Data Management Environment A Multi-Value database for the masses, not the classes. http://www.scarletdme.org - Get it _today_! Buying desktop hardware and installing a server OS doesn't make a server-class system any more than sitting in a puddle makes you a duck. [Cipher in a.s.r]

Er...it was a joke :o)

m?n 2012-04-09 klockan 02:09 -0300 skrev Alexandre Souza - Listas: It exists a rumour that Curt Nicolin once said: -- If left on their own and with to little to do, bureaucrats will start to invent new work (for themselves and everyone else.) -- The meaning is of course that the newly invented work wont be things useful to the organization. Bureaucrat = Transistor designer. Curt Nicolin was a swedish industrialist. Mainly he was part of the Wallenberg-sphere.

On Mon, Apr 9, 2012 at 12:18 PM, Henk Gooijen <henk.gooijen at hotmail.com> wrote: I do not have any pictures, but I should be shuffling storage this summer and I will be taking *lots* of pictures then. I got it as a discard from a company I later went to work for. They in turn got it from a university surplus auction. Neither they nor I have tried to power it up. I'm sure it needs a lot of cleaning and testing. I have experience with cleaning DEC backplanes (manila folders and isopropanol) and testing M-series modules. I do happen to have a set of the reproduction KM11s (http://www.shiresoft.com/products/km11/KM11%20Replica.html) that I'm sure will come in handy. Once I do have the time and space to restore it, I was planning on attaching it to my PDP-11/20 once I restore _it_. That one, as I've mentioned in the past, was literally pulled from a dumpster after it was parted out and gutted. I do not think I have 100% of the parts, but I do have three cabinets (CPU + two memory boxes), the DD11 backplanes, and a large box of loose modules. That will be more of a challenge to get running again. No less than two of the core modules were souvenired before I got my hands on the remainder, so I'll be lucky to get 16K working. I recognize that smell well. -ethan

On Mon, Apr 9, 2012 at 3:20 PM, David Riley <fraveydank at gmail.com> wrote: Unless something bad just happened... then it smells like burning. (thankfully, I've not had that experience with an RK05, but there was this one RK07 with a defective cap in the main PSU that took out the rectifier diodes - it was like a smoke bomb went off in the cabinet). -ethan

On 04/09/2012 03:20 PM, David Riley wrote: Nice. :) -- Dave McGuire, AK4HZ New Kensington, PA

IIRC, the TIP125 is an NPN darlington, basically 2 transsitors with one driving the other to get a higher gain )at hte expense of some other parameters). You won't see the normal 2-diode behaviour with those. It doens;'t look wildly out to me. I assuem you don;'t ahve a transistor curve tracer (it's not a common piece of test gear, most of the time hobyists get them becuase somebody gives one away....). You might try a simple swtichign trst using this sort of circuit : 1k +---\/\/------+------------------------o +ve | | \ | \ (X) 12V, 2W (or so)_ o | bulb | ----- 12V PSU | | c | +----------|b | Device under test | e | ----- | +------------------------o -vve You can build ti o na solderless breadboard, there's nothing critical. For the 12V pSU, use a bench supply if you have one, or the 12V output of a PC PSU 9which you know to be working!). The switch could even be a couple of wirses to toch together, the lamp should light only when the switch is closed. -tony

On 2012 Apr 8, at 1:17 PM, Mouse wrote: As shown in the datasheet that somebody provided earlier in the thread, it's not a pure darlington pair - it has built-in pull-down resistors across the internal BE junctions and a reverse-voltage- clamp diode across CE.

It's been about 10 years since I looked at one, and then I didn't need to repair anything in that area. So I am not very familiar with it to put it mildly. There are normally 2 sections of such drives that use power transistors. One is the spindle motor driuer, the other is the power amplifier to drive the voice coil positioner. The latter is likely to be a full-H circuit, using 4 transistors (2 NPN and 2 PNP most of the time). Most spindle motors are 3 phase, so expect 3 identical transistors to drive the windings. There could eb anotehr transitor for speed control, or braking, or... I don;t know how hard it would eb to trace the connections from the positioner conenctor or the motor connecotr back to the transistors, but uou might discover which are which that way. Well, it might tell you whether the fault is in the electronics (repairable) or in the motor itself (much harder). My worry is that if there's a shorted power transisotr, or a fault that causes oen ot be turned hard on, it could damage the motor windings. Or if there's shorted motor winding, or a fault with the hall effect swtiches that cuases one transistor to be on all the time, then it might damage the tranasistors. But I think that;s unlikely, and it should eb OK to swap the boards. Just don't blame me if it isn't :-) I just hope it's not a problem in the motor. With a very few drives -- the Micropolis 1200 series 98" winchesters), it's possible to dismantle the mtoor without having to pull the HDA apart, but even then you need special tools. With most drives it's impossible to repair winding or hall effect switch trouble without taking the heads and platters out. Good luck! -tony

On Fri, 6 Apr 2012, Tony Duell wrote: I've disassembled probably a dozen failed HDAs by now to see how they're made, but the direction the platters are supposed to turn isn't obvious to me... should it be 'into' the ends of the arms that carry the heads or 'away' from them? I always thought it should be 'away', so that the pressure of the circulating air on the arms will lift the heads off the platter, but on one HDA I took apart the heads tore off of the arms when I gently turned the platter in the 'away' direction... Alexey

I assume you swapped the boards indivdually, so you know the fault is o nthe motor driver PCB, not the main logic board. When I ahve some spare time I'll dig out an RD53 and see if I can roughly figure out what is going on. -tony