On 4/22/2011 11:40 PM, Adrian Stoness wrote: Oh, the Banana belt, compared to some places on this list. Nice sunny day with just a bit more snow to melt from Cold Lake here. Ask about Ethan's summer job some day. :) Ben.

On 4/23/11 12:57 PM, Curt @ Atari Museum wrote: Now, now...let's take it easy on the boy. At least for a day or two. ;) -Dave -- Dave McGuire Port Charlotte, FL

On 4/23/11 3:28 PM, Adrian Stoness wrote: I wouldn't if I were you. ;) You've come to the right place! Post when you're ready. -Dave -- Dave McGuire Port Charlotte, FL

On 4/23/11 11:58 PM, ben wrote: XX2247s aren't too hard to come by. -Dave -- Dave McGuire Port Charlotte, FL

At 10:54 PM -0600 4/23/11, ben wrote: Considering I use a key from a VAX to power up my 11/44 and my PDP-8's, I'm not sure DEC used more than one. Zane -- | Zane H. Healy | UNIX Systems Administrator | | healyzh at aracnet.com | OpenVMS Enthusiast | | | Photographer | +----------------------------------+----------------------------+ | My flickr Photostream | | http://www.flickr.com/photos/33848088 at N03/ |

The 11/05 and 11/10 used a flat padlock type key. Paul On Sun, Apr 24, 2011 at 12:59 AM, Ethan Dicks <ethan.dicks at gmail.com> wrote:

?

Gezundheit. On 4/24/11 10:43 AM, Alexandre Souza - Listas wrote: -- Dave McGuire Port Charlotte, FL

On Sun, 24 Apr 2011, Ethan Dicks wrote: If it hasn't already been done, it would make sense to post the depths of the cuts, so that anybody without one to copy can code-cut it. (I'm assuming that those who own them are no longer paranoid enough to want to keep it a "secret") -- Grumpy Ol' Fred cisin at xenosoft.com

On Sun, 24 Apr 2011, Patrick Finnegan wrote: What that URL says: Ethan said: Not really... the point of the XX2247 is that a full-service locksmith should be able to cross-reference that to a set of key depths, _presuming_ they can code-cut a 7-pin round key (not a universal expectation). My experience with code-cut keys is that a) they cut it, you pay, fit or no fit, and b) it's a few $$$ more than a copy. On Sun, 24 Apr 2011, Alexandre Souza - Listas wrote: "XX2247" is a "CODE NUMBER". Notice that it is SIX digits, two of which are 'X's. (which is NOT normally a cut number). The digits of "XX2247" are presumably COMPLETELY arbitrary, and serve to identify a given code, NOT give you the depths. (There ARE rare exceptions, where the code number and depths are algorithmically related, such as 1975 Honda door keys where the code number is just the list of depths in reverse order. I don't see an obvious algorithm. "XX" COULD be a "don't care, no pins in those slots", but we would still be missing one of the seven? cuts.) X X 2 2 4 7 is presumably the INDEX, not the VALUE (seven ints) stored in KEYS[XX2247] int KEYS[][7]; /* look up that key, get cut numbers */ /* apparently 7 spaces for this model lock */ float DEPTH[10]; /* look up each depth number to get distance */ /* # of different possible depth cuts may be */ /* other than 10, but is a constant */ /* for the particular model of lock */ A well equipped locksmith can look up "XX2247" in a "CODE BOOK", and receive a seven digit number, where each digit IS referring to a specific depth. Then, a quick glance at a "Depths and Spacings" chart will provide the actual depths of each cut number, although most code-cutting machines have databases (often in the form of dial cards) to look that up for you. If somebody MEASURES the cuts, or looks up "XX2247" in a CODE-BOOK and looks up the seven depth numbers in a Depths and Spacings chart, and then POSTS those depths, then the key can be cut manually (drill press recommended, if an appropriate keycutting machine is unavailable). -- Grumpy Ol' Fred cisin at xenosoft.com

On Sun, 24 Apr 2011, Alexandre Souza - Listas wrote: Somehow, I was stuck on the notion that this was a seven cut key. That sounds very reasonable and correct for a six cut key. Thanks for your patience with me, -- Grumpy Ol' Fred cisin at xenosoft.com

On 24/04/2011 23:27, Alexandre Souza - Listas wrote: Definitely not. DEC keys have seven cut positions, and on an XX2247, of which I have several originals in front og me right now, none of the positions are zero-depth cuts (one is very small, though). Similarly on a few other XX.... keys I have, there are 7 cut positions, and although on all of them one position is very small, it's not a zero cut, and it's in a different place on every different-numbered key. -- Pete Peter Turnbull Network Manager University of York

From: Patrick Finnegan: Sunday, April 24, 2011 3:29 PM Patrick is correct -- there appear to be 7 cut positions, one of which is actually uncut. Looking down the barrel of the key with the tab in the 12:00 position, and reading clockwise, I measure approximately 0.8", 0.0", 0.11", 0.44", 0.11", 0.8", and 0.11". Actually, rechecking this with the 10x magnifier and reticle (more accurate) gives right at 1mm, 2mm, and 2.6mm for the 3 depths of cut. (That's 39, 79, and 102 mils). Vince

From: Pete Turnbull: Monday, April 25, 2011 5:44 AM Yep. Though the later characterizations in mm and mils are correct. Rechecking the cut that I thought was zero depth with the magnifier, I do see a 10 mil nick there, so you are correct -- there are no zero depth cuts. (That 10 mil cut is nearly invisible in the corrosion of the key I was measuring. which still works :-).) Cool. I hadn't thought to measure those. I like the magnifier and reticle for that, as it makes it possible to sort of "visually smooth" the vagaries of the actual surfaces. I wondered about the tolerances, too. Seems like they can't be super tight, or the keys would be more difficult to use than they are. Would be nice if someone could look up somewhere what the spec's were/are. Vince

On 25/04/2011 15:22, Ethan Dicks wrote: Thanks for that one, Ethan; it's very useful -- I've bookmarked that! -- Pete Peter Turnbull Network Manager University of York

It's a very shallow cut on the key I've just looked at. But it is there. I would re-check that 0.44" one. For one thing no cut is anything like 0.4" deep. And for anther that cut is shallower than the 2 on either side of it, whcih you say are 0.11" -tony

I am pretty sure the old DEC frontpanel keyswitch has al l7 pins present. I can't remmebr if I've ever taken one apart (I've certainly done similar locks), but I possibly have, and I would have noticed 'misisng pins' 6 pins (spaced at 1/7th of a circle with a gap where the locator is) would be very unsual simply dcause these switches have their stable positions (off/on/paanel lock) at posuitons where the pins line up. That is, if the lock has 7 pins (spaced at 1/8th of a circle), the switch positions would be at 45 degree spacing, which is sensible. 5.142 degrees is not. [...] And to think I was going to hold the blank in a dividing head and guc the slots with an end mill.... -tony

I've come across 3 : The 'XX2247' tubular key A plastic tubular 'key' with no nothces, the lock switch has no pins. Any tubular key of the right size (including an XX2247) will operate that switch A small 'Yale type' key used in a wafer tumbler cylidner lock on the 11/05 and 11/10 -tony

You;re in the right place, then :-) Ask away :-). I know what _I_ would do, it's the same for jsut abotu any clasisc computer. My procedure is basically ; 1) Take apart anything that will sensibly come apart. No, I don't mean to deoslder all the ICs from the PCBs. But pull out all the boards, remove covers, things like that 2) A visal inspuction (and cleanup), looking for burnt/borken parts, cooling fans tht don't run freely, open-circuit light bulbs, blown fuses, etc. Anything taht can be easily checked, is... 3) Electricla safety check. Check the contiuuity of the earth (ground) wire, and the iunsulation resistnace between live parts and the case (at at least twice the mains voltage, I normally use 1000V). 4) Assemble the PSU section only, and if necessary connect it to dummy load resistors. Power up and check the output voltages. Don't go any further until they're cortrect 5) Assemble a 'mimimal system'. CPU + front panel + one memory board, something like that. Get that working (at least ot the extent of running a simple program entered on the panel). 6) Add the remaining parts one at a time and check. If you cna run the manufacturers diagnostics, do so. But be warned that in soem cases such diagnostics will tell you if soemthing is wrong, but may not be too useful in actually findign the fault. DEC ones are normally OK, though. -tony

Or more generally 'make notes'. I never take anything apart without a notebook and pen to hand to record cable positions, board positions, which plug goes where, etc. It saves a lot of time later ;-) Some people like ot use a digital camera for this but I am not sure if it will show what needs to be recorded in all cases. So even if I had one, I'd want a notebook too. You may want to arbitrarily number the slots (say '1' at the bottom, it doesn't matter if you are consistent!) and then afix labels to the PCBs (e.g. tie-on tags with the string passing through the handle rivets) to indentify which board came from each slot. [...] Indeed. It is very useufl to have a configuration that you know has worked. If it doesn't work now, it means a component has certainly failed. And that's something you can trace. Rather than having a setup that maybe shouldn't work even if all the parts are good. Actaulyl, it's not an instruemtn I use all that often, but anyway... In some cases if the smothing capacitors still have some capacitance (albeit much less than they should do), you won't seen much, if any. ripple on no-load, but you will when you connect the dummy load (or the rest of the machine). In machines with swtiching regulators (and I have an idea that the 8/a is one of these), a dreid up or open-circuit output capacitor in the regualtor cicuit can put high voltage spikes (around 15V) on the output. These will hopefully trip the crowbar circuit, but it's something to watch for. If I did that, I'd have replaced every capactior in every machine I own, including the machine I am currently typing this on. I haven't. In fact many of my machines, including ones that are 40 years old, are still runnign with all their original capacitors. Yes, capacitors fo fail. I've replaced a number over the years. And they can fial from just old age if the electrolyte dries up. But in m experience it's not the problem that some people claim it is. -tony

If a power supply needs a dummy load to work properly, or to be run without damage, it needs that load even if you're not checking that output. So yes, you have to have load resistors on all the outputs. I think the 'B' outputs are ones from the optional battery backup unit. On most DEC macnhines if you don't have the battery abckup unit, the 'B' outputs are totally dead, and are strapped to the apporpriate 'normal' PSU rails on the backplane. I would check the schematics of the PSU, though. Right... 10K is far too large. You want to draw aa significant current from the PSU outputs, that's the point of the dummy load. I would say perhaps 1-2A (or maybe more) from the 5V output and at least 0.5A from the +15V and -15V outputs. Let's consdier the 5V output first. To draw 2A, you need a 2.5 Ohm (not k) resistor. R=V/I and all that :-). Also note that it will be dispating 5*2 = 10W as heat, so a small resisotr is not going to do it. One of the easiest things to sue for this are low voltage filament lamps 12V car bulbs (say a 12V 5W taillight bulb) is fine for a 12V supply. And probalby fine for a 15V supply too (yes, the bulb will have a short life at the higher voltage, but not so short as to worry you). For the 5V line, a 6V bulb is the best thing to use, many older cars had 6V electrics nad you may well be able to get a 6V 30W bulb (which will draw around 5A) from a vintage car parts company. I bought a couple of headlight bulbs (each with a apr of 30W or so filaments and a couple of stop/taillight bulbs (5W and 21W filamnets). That gives me a good selection of laod currents (and yes, I once parallel to two filaments of a headleam bubl to give a load drawing somewhat more than 10A to test a large SMPSU). In amny DEC machines, the fans run off the mains input to the power supply (possible via an autotransformer or equivelnet for 230V mains). They don't count as a PSU load at all. -tony

[Using car bulbs as dummy loads] Thank you. Yes, it;s often the practical details that are missing (I can think of several books that are seriosuly lacking in such information). I will make some suggestions... Car bulbs often have non-stnadard caps. The 'stop/tail' bulb has a B15-size (SBC) bayonet cap, but with offset pins so it can't be fitted the wrong way round in the car. So it won't fit a normal lampholder. Headlight bulbs are even worse. You won't find the holders in an elecrtronics catalogue, and if you ask to buy a spare light cluster at a car dealer, be sitting down when they quote the price :-). More seriously, I suspect you can get such things [ FWIW, when I replaced a headlamp unit on my father's car, I kept the wiring and connectors from the old one. I have no idea if I'll ever need them, or what for, but I don't throw things out if I can see a possible use for them :-)] Some headlamp bulbs (although I've not found a 6V one that this applies to) have terminals that will take 1/4" faston ('lucar spade') terminals, which fives an obvious way of connecting to them. Most of the time, though I cheat!. It's only a temporary job after all. The contacts on the base of the bulb are soft solder. The cap itself solders easitly too. I jsut tin the neds opf the wires and solder them to the bulb connections. This dependso n the machine, and in particualr how many I am likely to have to repair and how the PSU outputs are arranged. For example, if it's a separate SMPSU unit with a barrier strip or terminals for the outputs (as in a PERQ AGW3300), then I would just connec the lamp wires to the barrier strip. If it's a machine where there's a PSU regulator board and I am going to have several to repair (say an HP9826/HP9836, I own 3 such machine), I might make up a specail test unit to connect to the PSU board edge connector. The uint woudl cotnain dummy loads (possibly wirewound resistors, not lamps), monitor bulbs/LEDs and sockets to connect a voltmeter for testing. MAking such a unit is quite easy, but it can be a lot of work (in the case of the HP machines I just mentioned, designing it took about 10 minutes, soldering it up took about the same time, but drilling the box and making the spacers to fit the connectors at the right heights took a morning.) SOmetimes I just 'tack-solder' the load wires to points on the PSU board or backplane PCB. I don;t normally use crocodile clips, which have the amazing habit of springg off just when you don't want them to :-) My PDP8s and PDP11s have PSUs that are happy without dummy loads, you cna run them with no laod and just chekc the voltages. If I had one that needed a dummy load (or if I wanted to check on-load anyway), I'd probably make up a usnit ti the 15 pin plug wired to laod resistors/lamps. It's the sort of machine that is quite common, so it would be worth making up a test box. -tony

On 4/26/11 4:56 PM, Rob Jarratt wrote: Hey, that's the same monstrously complex power supply as is used in the 11/44. It is in fact a switcher. Tony can speak volumes about this particular power supply. I'm almost positive that it does, yes, at least on the +5V output. -Dave -- Dave McGuire Port Charlotte, FL

On 4/26/11 6:30 PM, Geoffrey Reed wrote: But they do! Their resistance goes WAY down if you give them a good jolt of high voltage first. -Dave -- Dave McGuire Port Charlotte, FL

A switcher simply won't run without a load. FYI -- Ian

On Tue, 26 Apr 2011 19:51:34 -0400, Dave McGuire wrote: There needs to be a small internal load than to maintain bias, there can not be 0 internal and 0 external load. something about efficency going to infinity. The other Bob

OK, it sounds to me that in your designs the difference between load and no-load conditions are sufficiently small that you are designing your control loop to no-load conditions and loading is not substantial enough to push your design out of that window. As I consider it, I guess I should not say that *any* switcher will fail to run without a load - I have a little wall-wart for my Android tablet that puts out 5V just fine (although I have to wonder if there isn't an internal load that consumes a small parasitic amount of energy). But for the designs we see in vintage systems - such as the H7100 - they require a significant load for the control loop to function. In such systems the load is an indispensable component in the control loop. -- Ian

On 4/26/11 9:41 PM, Bob Bradlee wrote: I promise you that I'm not trying to be either dense or argumentative here, but we're just talking about (assuming we're talking about the same thing) an electrically-controlled switch in series with a power input with a capacitor to ground on the other side, and a resistive voltage divider after the capacitor that generates a PWM control voltage to the oscillator that drives the switch. Nowhere in this arrangement is there any requirement for a load. Based on your comments, though, I'm starting to suspect that we're talking about two completely different circuit topologies. Why don't you (if you're so inclined and have a moment) describe in more detail the topology you're talking about, assuming I'm correct that it's something different. -Dave -- Dave McGuire Port Charlotte, FL

How do I put this... You've been thrown in the deep end :-) Am I correct that your 11/24 is in the 10.5" (6U) box? Boards going in vertically from the top? If so, it is the PSU I am thinking of. I came across it in the 11/44, and it's one of the most complciatred PSUs you are likely to come across. It is a swtich mode unit. Or rather it's several switch-mode PSUs. The basic design is as follows : Incoming mains is rectified ans charges a pair of coke-can size capactiros at the left hand side. This provides about 350V DC which powers 3 SMSPUS sections : 1) A small one, to provide power for the PSU control electronics. 2) The 'logic' supply. This is an SMPSU which provides the main +5V rial and also +/-15V rails for the RS232 ports, etc 3) The 'memory' supply. This is another SMPSU,. giving 36V. This 36V rail can be battery-backed by an optional unit which you probbly don't have. The 36V is brought down to +5V, +12V and -12V for the memory boards by switchign regulators (non-isolting ones, of course). it's also chopped by a full-H circuit to provide power for the 35V 70Hz cooling fans. This is one of the few PDP11 supplies where the fans do nto run off the mains. A word of warning. If you remvoe the top cover from the PSU you will see barrier terminal blocks at the tops of some of the PCBs. In particualr there ar 2-way ones with red and black wires on 2 of the PCBs. These wires carry the 350V DC input, which is not isolated from the mains. Touch those with the mains connected and you won't feel anything. Ever again. Seriously, it could well be fatal. That said, it's a more friendly SMPSU than most, in that much of the control circuity is isolated from the mains (so you can connect a 'scope or whatever to it without problems) and scehamtics are avaialble. I am pretty sure this is the same supply... Actually, I don't think it does. I don't remember needing one when I repaired the PSU in my 11/44. Certianly I've never met a DEC-designed supply that is damaged by running on no load (the output votlages may be low or missing, bit it won't damage any parts), so it's safe to try it. But a dummmy load is not a bad idea, particularly on the main 5V rail (ffro mthe logic supply, rated at 125A or something similar). I don't know if the 11/24 is similar, but on the 11/44, this output is a pair of studs/nuts that flexiprint tails from the backplane go onto. I think i'd remove all the logic boards form the machine (keeping a note of their positions!), conenct, say, 30W or more of 6V bulbs across the logic supply 5V output and power up. Check all the voltages. If any are out of spec, try adding a load to them beforediving into the PSU internals. -tony

On 4/27/11 5:31 PM, Adrian Stoness wrote: Mmm, nice HP scope, 1707A maybe? And some of the little modular HP counter/DMM units on top of it? What models? What's the large unit on the shelf above the scope? It looks extremely GenRad-ish, or possibly early Boonton, but I don't recognize the model. And what's the walkie-talkie-looking radio, ham HT or something else? Someone stashed a tire under your electronics workbench. ;) -Dave -- Dave McGuire Port Charlotte, FL

peeps keep telling me this On Thu, Apr 28, 2011 at 1:07 PM, Tony Duell <ard at p850ug1.demon.co.uk> wrote:

On 05/02/2011 05:42 PM, Curt @ Atari Museum wrote: Curt, If you don't find one let me know. I believe I have spare one complete in the box. Allison

On 4/27/2011 5:31 PM, Adrian Stoness wrote: Uhhh, okay. You're definitely not the most clean or the most organized, though, huh? :) Keith

It's a lot tidier than my workbench, I can assure you ... Remeber 'An empty bench is a sign of an empty mind' :-) -tony

On 4/28/11 3:49 PM, Sridhar Ayengar wrote: Hey, I'm pretty sure "Indian style" refers to different Indians. Are you sure it's legal for you to sit that way? ;) -Dave -- Dave McGuire Port Charlotte, FL

On 4/28/11 4:30 PM, Teo Zenios wrote: As it should be. -Dave -- Dave McGuire Port Charlotte, FL

On 30 Apr 2011 at 1:07, Alexandre Souza - Listas wrote: Yes, or sometimes "Asian Indian". For the local variety, often "Native American". It can get to be quite confusing, occasionally requiring some adjectival help. Throw into that mess, the people from, say, the Phillipines, who are "East Indians". Another old attempt at disambiguation was "red indian" for our pre- Columbian inhabitants, but that's largely passed from current usage-- some felt that the term had a negative tone to it. But that's because some stupid European didn't know where he was and got everything confused. When I was young, I wondered why any 19th century European would write a novel about Hoosiers... --Chuck

lot cleaner right now then in that shot On Thu, Apr 28, 2011 at 2:17 PM, Keith M <keithvz at verizon.net> wrote:

Who cares?

a20a1 87 accord motor block with million miles on it sitting weighting for new rings as it smokes the tire is off the same car handy talker is just a old ge also got a bunch of motorolas something machean the black thing on the shelf? thats some sorta low frequency genirator from the 40's my dad built this work bench i just taken it over since he left all his stuff my understanding about the motorola spool thing is it was ment for old school computer wire wraping i know i have tools for that stuff somewhere. theres also a really early ibm laptop sitting on the bench On Fri, Apr 29, 2011 at 10:45 PM, Alexandre Souza - Listas < pu1bzz.listas at gmail.com> wrote:

You clearly like living dangerously... HAve you done any electrical tests at all? At the very least you need to check the power supply voltages, without power the machine is not going to do a lot. They, I guess, you have the fun job or debugging the logic They normalyl need that. The manuals are not hard to get off various web sites, I would recoemnd reading them before you start (I did my firstASR33 with no docuamtation at all, I did all the adjustments by thinking what they should do, and it worked second time... But then I had a mis-spent childhood dismantlign and repairing all sorts of things). You want to get the parts book too. It has exploded diagrams of the complete machine, very useful when you want to know where 'that washer' should go... It's been some years since I worked on an ASR33, but it'll all come back to me very quickly... -tony

Do you have documentation for the scope? The HP stuff is usually good about "first time user" guidance. Agilent (the test equipment spin-off from HP) has it online: Google hp1703a manual for the link (it's a weird dynamic link or I'd just include it here). I have two scopes of this line, I can't recall exactly which ones but I think the 1701 and 1702. It's a very good scope, but there are a lot of knobs on the front. :-) Fortunately most things require using only a few of them. You'll want to look at volts/div, which is the input sensitivity and controls the vertical aspect of the display; input coupling (AC/DC/GND), which either allows or removes the DC component of the input signal (or grounds out the input to the channel); time/div, which is the sweep speed and controls the horizontal aspect of the display; and triggering, which controls whether the sweep actually... sweeps! Set it to 'auto' for most purposes, but if you read nothing else from the manual, read about triggering. It gives you a lot of control over what you see on the screen. What I would recommend is using the scope to look at all the voltages from the power supply. Probably the easiest way to do this is to set the input coupling to AC and (with a 10:1 probe, which is what most scopes come with) set the input sensitivity to 0.1v/div, and set the time/div to maybe 10mS or so. This means that each vertical division on the screen represents one volt. If you see one or more volts of waveform with this setup, you have a problem. A ripple of 1v on a 15v line isn't great but isn't a deal-breaker; on a 5v line, it represents a 20% variance over time, which is bad. (Actually, you really don't want more than a token ripple on the 5v rail - even a half-volt is 10% and can really confuse the logic.) Ideally, you would see a flat line; if you switch the input coupling to DC, the trace should displace 1v/div (e.g. the 5v rail would displace five divisions from a no-input trace). To reiterate: with AC coupling, if you're looking at a pure DC input there will be a flat line at zero volts. I know it doesn't need to be said, but I'll say it anyway: be sure you've hooked the ground lead from the probe to a good voltage ground! On the stuff you're working with here, the chassis of the power supply is a ground reference. Cheers -- Ian

"Just keep swimming, just keep swimming...." :-) The 10006D is named as the stock probe for the HP 1703A - do get the manual, it's a lot of help. -- Ian ________________________________________ From: cctalk-bounces at classiccmp.org [cctalk-bounces at classiccmp.org] On Behalf Of Adrian Stoness [tdk.knight at gmail.com] Sent: Wednesday, May 04, 2011 2:57 PM To: General Discussion: On-Topic and Off-Topic Posts Subject: Re: Scoping the supply lines (was RE: new here) i duno what probe to use i got a pile of them hp 10005b 10006b 10006d p6021 ac current prop tektronix 010-0237-03 feel like i've gone over my head

http://img708.imageshack.us/img708/1060/probes.jpg

Before you use the 'scope to check for supply line ripple, I would check the DC votlage of all the suppl lines using a multimeter. I asusme you have a digital multimeter rather than an analogue one, but I have no idea which one you hve and whether it has automatic or manual range selection. But the procedure is muc hte same in all cases. If it's an autoranging meter, switch it to the DC votlage (DCV) setting. If it's a manual range meter, switch it to a DC voltage range highr than the rail you awnat ot check, but not too high. Most meters have ranges 2V, 20V, 200V. The 20V ranges is the one you want for most computer PSU voltage tests. Connct the black lead of the meter to the computer 0V (logic ground) rail. In most DEC machbines this is connected to the metal casing. Touch the red probe onto the power rail you want to test with the machine turned on. Finding the power rails is the major problem, actually. You might use the pwoer pins of an IC on one of the boards. Or a pin on the connector between the power suypply and the backplane. Read the votlage on the meter :-). Now for 'scopes. I don;t know your 'scope (I prefer Tektronix...), but all 'scopes have simialr features. I would try to get the manual for the 'scope you have, it may well be on the Agilent website (they are putting up scans of a lot of their older instrument manuals). However, most test instument manuals assume you already know the basic principles of using such an instrument amd just explain the features of that particular one. The older non-digital CRT-based 'scope is a fairly simple instrument, and modern digital 'scopes have generally beed designed to work in much the same wayy for the user. The basics are : The electron beam in the CRT makes a bright dot on the screen. There are obvious controls for controlling the brightness and shapness (focus and astigmatism) of the spot. The input signal moves the spot vertically. There will be an intenral amplifier which will have a gain control calrated in 'volts per cm' or something similar (sometimes 'volters per division where a 'division' is one the squares on the graticule on the front of the CRT). If you wet that to 2V/cm and the spot moves by 2.4cm high on the screen, then that means the input is 4.8V Horizotnally, the spot is moved by a 'ramp' voltage from an internal timebase circuit. It moves from left to right at a constant speed and then quickly flies back to the left and repeats. The speed is controlled by a control calibrated in 'ms/cm' or something similar. Thus the 'scope ploits and displays a graph of the input voltage against time. You can m,easure the time between parts of the waveform (say time between peaks) or the votlage between parts of the waveform There are 2 other fauters that need mentioning. The first is the tirgger sicut. The purpose of that is to make the timebase wait for a particular input voltage (and the right direction of change) of the input signal. The idea is that assumign the input signal is repetitive, the timebase will start on a corresponding point of the wavefor each time, so you will get a steady trace on the screen that you can easily view ans makemeasueemtns from (as an aside, the purpose of storage 'scopes, whether analoge or digital, is to be able to hadnle non-repetitive waveforms. You effectively record the waveform for one sweep of the timebace and display it for evermore). The oterh feature is 'AC coupling' Most 'scopes have a swixch in thge vertical section for htis. It has the same effect as connecting a capacitor in series iwth the input signal (in fact in most caes that's what it really does). It removees the steady DC component of the inptu signal and only lets the AC part through to the amplifier. For example, suppose you have 100V supply rail with 0.1V ripple on it, and you want to look at the ripple. To display that supply rail, you probably have to set the Y amplifier to 50V/cm, which means the ripples is 0.02mm high. You can't see that. If you switch to AC coupling, the steady 100V is ignored, so you can turn up the gain of the Y amplifer, perhaps to 0.1V/cm at which point the ripple is easy to see. A word about probes. Most times you use a proper 'scope probe. This has a gorund lead which is actually conencted to mains earth throguh the 'scope so you can't connect it to just any point in the circuit. However, connecting it to the logic ground rail is fine (and what youshould do). Do not however assupe that the conenction through the mains earth is enough (and that you cna ignore the ground lead), it may be OK for low frequencies, but at high frequences the inducatance of that path through the mains earth wires is enough to mess up the display. Most 'scope probles sau '*10' or somethign similar on them. This means they attenuate the signal by a factor of 10. If you set the 'scope to 0.5V/cm when using such a probe the real sensitivity is 5V/cm. There are good reason ofr having this attenuation in the probe, mainly reducing the effectif ogf the capacitance of the cable to the 'scope on the circuit under test. Will that get you started? -tony

check thinking mis-spent to I hope to get an ASR33 one day. When I do, the plan is to get it connected somehow to SIMH running TOPS-20 on the PDP10 emulation, and probably also to my PDP11, when I get that working.... Regards Rob

across it Yes, this is an exact description of what I have. basic at I saw those capacitors and realized I need to steer well clear of them after power has been applied. and a of there ar the Thanks for the warning. I find PSUs worrying at the best of times, even before your warning this one already scares me rather more, which is why I am trying to do things more correctly than usual. control whatever repaired is Sounds like I could give it a quick whirl with everything disconnected just to see if there is anything that fails even at that first stage. Then I can proceed to adding a load and checking voltages, I can't do that yet because the bulbs I ordered have not arrived yet. (ffro the that I found three types of connector from the PSU to the rest of the machine. There are six 15-pin connectors (5 rows of 3 pins), with only three connectors actually connected. Then there are two connectors I don't know how to describe but are each a single row of about 10 or so pins, the connector is very like the ones you find on a MicroVAX II power supply if you have seen one of those.. Finally there were two wires with spade connectors on them which connected to a couple of tabs on the PSU. of spec,

without I have just this minute attached a dummy load to one of the 5V outputs and powered it on. Using a multimeter I measured 5.4V on the 5V outputs and 15.8V on the 15V outputs. I ran it with just the fans running (forgot to disconnect them) for about 5 minutes and then switched it off. Are those voltages within tolerance do you think? They seem close enough to me to warrant putting some boards back, but I would like a second opinion before risking the boards. I am not sure what you mean by barrier strips, you mentioned them before but I could not see what you are referring to. a lot circuitry is machine. Indeed that is what I did. wired ask wasn;t

I loaded one of the +5V outputs on one of the backplane connectors and tested every 5V output I could find on the backplane connectors, they were all the same voltage. I used a 6V headlamp bulb on the 5W part, so drew about 1A. I could try again to make doubly sure and perhaps load one of the other 5V outputs. Do you know off hand which output is the one for the memory so I can load that too. I have two 6V bulbs available to me now so I can load logic and memory if I know which is which. Would you recommend me using the 21W filament instead? I am not sure how much current the PSU will supply, but guess 5A should not strain it. so no to the strup fix that site In that case I think I do know what you mean, it is just that there are no red wires, just black ones.

some difference P2 Just a bit more info. I loaded the PSU a bit more. I used two 6V/21W filaments on the 5V output and now I measure 5.4V. I used a 12V/21W filament on the 12V output and got 12.8V (forgot to check this before) and one 12V/21W on the 15V which now gives me about 16V. When I used two bulbs on the 15V output it went down to about 10.5V.

I don't know the 11/24 that well (at least not this version, mine is in the 5/25" box with the more standard SMPSU). In the 11/44, there's a main 5V supply for the CPU logic, the Unibus cards, etc and a separate +5VB line (I think that's what it's called) for the memory boards only. I cna take a look at the prints on bitsaers and see what goes where. The main 5V output of this supply is rated to supply up to 125A (!). I think I'd load it a little more heavily than that. I cna't remmber what the memeory supply can provide, but I'll bet it's more than 5A. The CPU supply is 125A I think (maybe 120A, it's a lot anyhow). These supplys will shut down without damage if you try to draw too much current from them. So try whatever bulbs you like. -tony

Oh, okay.... we'll let him slide on the first round of beers ;-) Dave McGuire wrote:

At 9:47 AM -0700 4/23/11, Ian King wrote: I have to agree on both counts! I'm trying to clear out a lot of stuff, but wow... On an unrelated note, I was had my C64 playing SID's while cleaning out in the garage earlier. First time in months I've had a Classic powered up. Of course it was playing SID files located on a SD flash card. :-) Zane -- | Zane H. Healy | UNIX Systems Administrator | | healyzh at aracnet.com | OpenVMS Enthusiast | | | Photographer | +----------------------------------+----------------------------+ | My flickr Photostream | | http://www.flickr.com/photos/33848088 at N03/ |

On 4/24/11 10:04 AM, Tony Duell wrote: Three 8/e systems (the "big one" has an RX01, TU56, 2xRK05, and a hacked HP optical PTR), an 8/m, and an SBC6120 here. I'm STILL looking for an 8/a and STILL can't seem to get my hands on one. I have the same problem with the Heath H-89...despite the fact that they're not particularly rare or sought-after, and opportunities keep popping up, they keep slipping through my fingers...sold or given to someone else at the last minute, owner disappeared, etc etc. Very frustrating. -Dave -- Dave McGuire Port Charlotte, FL

On 4/24/11 10:51 AM, Tony Duell wrote: Crap, that sucks! If you find an 8/a and I find an HP1000, we've gotta trade! ;) -Dave -- Dave McGuire Port Charlotte, FL

On 4/24/11 3:24 PM, Tony Duell wrote: *pout* -- Dave McGuire Port Charlotte, FL

At 5:44 PM -0500 4/24/11, Adrian Stoness wrote: Now that's far more interesting that some crusty old computer! The best audio is still tube based. Preferably fed by a nice turntable. Zane -- | Zane H. Healy | UNIX Systems Administrator | | healyzh at aracnet.com | OpenVMS Enthusiast | | | Photographer | +----------------------------------+----------------------------+ | My flickr Photostream | | http://www.flickr.com/photos/33848088 at N03/ |

Any interesting computer tubes? -- Will

Many of us here wilbe happy to explain how to use your 'scope. I asusmne you know the basics -- a 'scope displays a graph of the input voltage against time. You can think of it like this (and it's fairly accurate for older, non-digital 'scopes). The spot of light on the screen is produced by an electron beam hitting said screen. The beam is deflected vertically by the input signal, there wille ne a gain contrl calibrated in 'volts/cm'. If you have a vertyical defleciton of 2.5cm and a setting of 2V/cm, then tha's a total fo 5V. The beam is moved horizotally by a ramp voltage from the 'tiomebase' circuit'. This has a control calibtated in time/cm (so perhaps 20us/cm). So if you see that 2 points of interest on the waveform are 3.4cm apart and the timebase is running at 2ms/cm, then that's 6.8ms between them. There's a lot more to it, of course. Some 'scopes have all sorts of interesting extra facilites whcih are very useful but which may confuse you when starting out. Post the make/model of the 'scope and it's likely somebody hnows it :-). I don't know what your electronic background is like, but I am still going to suggest you look at a book called 'The Art of Electronics' by Horrowitz and Hill. It's not cheap, but it's very well explained and it covers most (but not all) areas of electronics. -tony

And as you may have guessed one of mine too. So much that I have both 1st nad 2nd editons and both of the student manuals. The latter are excellent too, but remembr they are written for a university course, so if you want to repeat the experient,s you need to have some reasonable test gear (bench PSU, 'scope, etc) which many beginners won't have at the start. Gettign bag to 'TAUE', it's one of the few books that start from the real basics like 'what is a resistor' and go on to microprocessors, etc. It's not a book you will 'grow out of'. And as I said, I found it very clearly written. Yes, there are some things in it that I would probably disagree with in the sense that I would design <foo> a different way. But I don't think there's anything there that made me exclaim 'They're suggesting WHAT?'. To be honest, if you read and understnad it, you may well think up other ways to do things. That's good. But you won't be learning stuff you have to unlearn later. My origianl comment could be taken to mean 'I don't know how much electronics you already know, but no matter what level you're at, read TAOE' :-) -tony