What I am wondering about, though, is the extra current they draw while they are forming up while the power supply is running. The capacitor might survive it (not get so hot that it fails), but the things supply the higher than ordinary current to it might not. Killed a bridge rectifier on a PDP-12 that way. JRJ On 7/17/2015 1:31 PM, tony duell wrote:

Could you, please, explain why? And how often should this be done? Every week, every month, every year, or what? FWIW, the number PSU elecrtrolytics I have replaced can be counted on the fingers of one hand -- in unary. Well, perhaps both hands. But it's <1% of all the PSU electrolytic capacitors I own. Only 2 cases spring to mind : The PSU in my 11/44 had a high ESR capacitor on the +36V rail (all other caps in the machine were fine) I changed the 2 mains smoothing capacitors in my HP120 not because they were electrically defective (they tested fine) but because one was bulging a little on top and had it exploded it would have hit the neck of the CRT with all the problems that would be likely to cause. I do find this witch-hunt against capacitors to be curious, given how few I've found to have failed. I suspect a lot of it comes from audiophools who think this is the way to fix anything... -tony

some are good some are bad..... http://www.badcaps.net/ here is all the info you ever need on today's bad caps not so much on yesteryears bad caps On 7/17/2015 11:53 AM, Mouse wrote: -- The contents of this e-mail and any attachments are intended solely for the use of the named addressee(s) and may contain confidential and/or privileged information. Any unauthorized use, copying, disclosure, or distribution of the contents of this e-mail is strictly prohibited by the sender and may be unlawful. If you are not the intended recipient, please notify the sender immediately and delete this e-mail.

duell given marginally-rated it is a be and so When I repaired my VT100s I had to replace all the electrolytic caps on the monitor control board to cure the screen wobble. Before doing so I had reformed them all and I had tested them all for ESR and they had all tested fine so I was unable to determine which of them was the bad one. Perhaps there is other more professional test equipment I could use that would have helped, I don't know. I did keep all the original caps though (somewhere). Regards Rob

Thank you for the detailed information. I need to figure out how im going to get it out of the rack and moved to a place where i can test it over the next couple days where it will not be in the way. Ill find some way to do a dummy load and do an extended test to be sure the supply is working properly. All fingers crossed, god it better not have damaged any of the boards, i do not know where i would get replacements.It took years for me to get the machine, who knows how long it would take to find a specific board that is bad. I did buy an oscillicope and a logic analyzer well in advance in preparation for getting this machine, however short of pressing the power button no clue how to use them or basic troubleshooting procedures. Guess i just have to learn by doing.... On Fri, Jul 17, 2015 at 3:23 PM, tony duell <ard at p850ug1.demon.co.uk> wrote:

On 07/17/2015 11:17 PM, tony duell wrote: Many years ago, I managed to feed +12V into an Acorn Atom which was expecting +5V (the on-board regulators were bypassed to allow it to run Amazingly, it survived - although it makes me wonder if the experience dramatically reduced the lifespan of the TTL ICs though, even if they didn't immediately show any failure. cheers Jules

On Fri, 17 Jul 2015, tony duell wrote: It really depends. If I'm having to replace filter capacitors in a faulty PSU or monitor, even if I've isolated the problem to one or two capacitors, I would be a fool not to replace them all because the next part that fails will be one of those which I didn't replace (been there, done that, own the tee shirt). A standard 85C or 65C rated 20-30 year old aluminum electrolytic is simply past its useful service life. With modern electronics, there are high quality parts and there are really cheap parts. Even though the high quality parts are about the same price as the really cheap parts when purchased in small quantity (in the one-off to a few 1000 quantity), profit-driven consumer electronics manufacturers are still going to use the cheaper parts when they can save even $0.01 or $0.005 per component or even a few cents on the overall cost of the complete widget. Many of these manufacturers also design their widgets to last for the warranty term and no more. If the consumer gets 2-3 years or more out of said widget, those manufacturers consider it a loss in terms of potential profits. In these cases, wholesale replacement of really cheap capacitors with high quality versions from top tier manufacturers (Panasonic, Nichicon, Rubycon, etc) can be a /really/ good idea. That said, even the top-tier capacitor manufacturers have lower end lines of parts, so it pays to do your homework and choose a longer life (usually lower ESR, meaning less internal heating and thus longer service life) part when sourcing replacements. Even vintage Mallory and Sprague parts are not without their faults. I recently replaced a bunch of early 1980s era Mallory capacitors which tested good, but when desoldered from the board, had brown crusty stuff around the safety vent in their bottom rubber seals. They might have continued to work "ok" for another year or even 10 years, but the high quality replacement parts I put in will be good for at least another 20-30 years. (The replacement parts also worked better as the DC rails had lower ripple after they were installed). It really is a case by case basis, however in PSUs, CRT monitors, and similar where the components are exposed to higher temperatures, it really makes sense to replace 20-30 year old aluminum electrolytics wholesale. I guess another way to look at it is that an electronic device might still be functioning with old aluminum electrolytic capacitors, but is it still functioning as good as it was when it was new? A vintage device might have used really high quality aluminum electrolytics when it was made, but even high quality parts have a finite service life. What you have to decide is where to draw the when considering wholesale replacement of aluminum electrolytics, which with vintage electronics, I've found tends to be somewhere around the 20-30 year point.

Ummmm - his PDP-11/34 most certainly does use switching power regulators. ;) On 7/17/2015 4:06 PM, John Robertson wrote:

Oh, sorry, didn't realize they used switchers for the PDP-11s. However I was talking with a friend of mine last night about my error, and he told me that the switching supplies for the PDP-11s were very unreliable back in the day. He often had to troubleshoot the machines back then. A common failure was caused by static electric shock to the machine would blow the supply. NO carpets allowed!! John :-#)# On 07/17/2015 2:19 PM, Jay Jaeger wrote:

More myths and ledgends. As I remember it this all started with the arrival of FET's having a very high input impedance due to narrow gate areas. If you were daft enough not to have a path to earth and let a charge build up on the gate then you could exceed the breakdown voltage across the junction. I used to see the DEC field service statistics and PSU failures due to static build up were rare. Blown filter capacitors and burned up resistors were more the norm. TTL was was not imune to over voltage but not via static build up. Think about a CRT VDU. lots of high voltage and static but no charge up related failures. Rod Smallwood On 18/07/2015 20:31, tony duell wrote:

I assume this relates to my comments on static damage of PDP11 PSUs, I am not sure I would quite put it that way (not on a public list :-)) but (a) I have never seen a DEC PSU (in a PDP11 or elsewhere) damaged by static, certainly not to external connections and (b) there were some many PDP11 PSUs with totally different design topologies that it would appear to be impossible to generalise like that. -tony

Be careful, static daamge does not always show up at the time. You can damage an IC, have it work for some time afterwards and then fail. Yes. Static electricity is not magic. It follows well-known physical laws. In particular, Q=C*V (using the normal definitions). When you put a static charge on the human body, the voltage is high becuase the capacitance (to ground) is low. If you then touch something, effectively its capacitance is in parallel with yours. The charge redistributes as a result. If the touched thing has a low capacitance, like the input to a MOSFET or CMOS gate (which you want to be low capacitance in normal use so you can charge/discharge it quickly to switch said device) then the voltage remains high. Zap!. If you touch the output of a PSU with around 1millifarad of capacitance to ground then quite simply you are not carrying enough charge to put a damaging voltage on it. OK, the inductance of the circuit makes a bit of difference (it will slow down the charging of said capacitor so the voltage may have a small peak above the steady state) but it is not going to make that much difference. Put it this way, if you connect a 1000uF capacitor across something (as is done on the outputs of these PSUs) and can damage it by static from a human body I will be very surprised. Now let's look at the PDP11/34 PSU regulators. What is in there? Some bipolar transistors, including power types. And normal semiconductor diodes. Not things that are commonly thought of as very static sensitive. A 723 bipolar IC. Never heard of that being damaged by static either. And that's about it. As I have said before, I don't know about the later PSUs. It's possible the SMPSUs in say a PDP11/94 do have problems from static damage (although I seriously doubt it's caused by touching external connections). But not the 11/34 supply. Oh sure. No PSU is going to take a lightening strike on input or output. Even a strike nearby might cause a lot of damage. But I don't think that's what we are talking about. -tony

On Sun, 19 Jul 2015, dwight wrote: Oddly enough, I have a Sangean ATS-803A (Radioshack DX-440) receiver on my bench right now which has static electricity damage. They used two Sony 2SK152 JFETs (long since discontinued and virtually impossible to source) which can be damaged by static electricity merely by touching the telescoping antenna. I've replaced the two 2SK152 with Fairchild J113 JFETs and retrofitted a pair of inverse-paralleled 1N4148 diodes just before the antenna selection switch (between ground and the switched side of the external antenna jack) to help prevent future damage to the replacement JFETs. The receiver still isn't working quite right just yet though. It currently works fine on broadcast AM up to 1620KHz (internal loopstick antenna) and broadcast FM with the external telescoping antenna, but at 1621KHz and above in AM mode, the set is totally deaf. It is supposed to use the external telescoping antenna for AM operation above 1620KHz, but there is still a faulty component somewhere in that portion of the circuit.

That might be a little different -- much more recent - presumably in the ear of flat screens and PCs where there have been times when manufacturers got fed bad capacitors for their boards - which then failed later. IBM replaced a whole series of motherboards in one organization that I worked at because of that (though those were not power supply capacitors.) On 7/17/2015 1:53 PM, Mouse wrote:

I tend to agree with your hunch. On 7/17/2015 1:55 PM, Todd Killingsworth wrote:

On 2015-Jul-17, at 11:42 AM, tony duell wrote: This is something Tony and I are quite in agreement on. Similar to Tony, (and as mentioned in discussion on this topic a couple of months ago): in the solid-state category, of the many pieces of 1960s & 70s and later equipment I have or have serviced, the vast majority are running with their original capacitors. If you're dealing with a 1936 or 1952 tube radio, a knee-jerk "replace the capacitors" is warranted. If you're dealing with a 1970s computer, it isn't (IMHO). Esp. when they're screw-terminal 'computer-grade' caps. My own perception of the concern is that it has been perpetuated over the years from the vacuum tube / antique radio arena. The issue of capacitors "drying out" dates from the days (1920s,early 30s) when electrolytics actually were filled with an active liquid which actually did dry up. "Dry electrolytics" were developed in the 1930s, and while early dry electrolytics also warrant replacement, the chemistry and techniques have seen a few improvements in the many intervening years, and solid-state equipment is not placing the same stresses on caps as tube equipment. In other arenas it's a real issue, in a modern arena it is largely lore. The point of electrolytic caps is to form an oxide to be the dielectric, formed (in part) out of the electrolyte, and while I'm no expert on the chemistry, I will point out the oxidised state is 'the' or 'a' low energy state, and hence relatively stable. Rust doesn't normally undo itself.

I was operating the panel when i first got it, now the numbers do not light up, panel is unresponsive, and run light stays lit.(just describing the behavior, i will not start it back up till I work on the power supply) On Fri, Jul 17, 2015 at 3:50 PM, devin davison <lyokoboy0 at gmail.com> wrote:

Replace - no, I don't agree - especially not for those of us who don't have the kind of budget that your organization has. In my experience, for equipment of this quality and vintage, 95% or more of the time an hour to a few hours of re-forming is all that is necessary - and as Tony has pointed out, even that is not often really necessary. JRJ On 7/17/2015 1:33 PM, Rich Alderson wrote:

Err, have you priced the screw-terminal 'computer grade' electolytic capacitors that were used in these PDP11 power supplies. They are not cheap, if you can find them at all. And of course NOS ones might be as good or bad as the one that's already there. Conversely when I had a major disaster in a DEC power regulator brick some years ago (blew almost all the transistors and the 723) the replacement parts were easy to get (exact replacements, not just equivalents) and were not expensive. [...] Odd... I know plenty of people who restore old valve radios and audio stuff and not one will blanket-replace all the aluminium electrolytics. There is a capacitor that I (and they) would check very carefully, but that's not an electrolytic. I refer of course to the coupling capacitor to output valve grid. In a lot of radios this is connected to the anode (plate) of the audio ampilfier triode so if it leaks it puts a +ve voltage on the output valve causing far too high an anode current there. But even then I (and everyone else I know) would test it, not just replace it. Some of those capacitors are very reliable and the replacements you get not any better. I probably would replace certain safety-related capacitors in live chassis sets, like ones that isolate external sockets, using class Y replacements. But that;s about it. Incidentally, do you shotgun-replace 7805s and other 3 terminal regulators? If not, why not? They can fail, and if they do they do a lot more damage than a failed capacitor. -tony

On Mon, 20 Jul 2015, Mark J. Blair wrote: Absolutely! I didn't mention those in my previous list since I was focused on aluminum electrolytics, but those yellow Rifa parts are an especially sore spot in older test gear. My own theory on these is that the swelling of the paper is what is causing them to crack. I'm not sure what chemical they treat the paper with, but it apparently doesn't hold up long term and/or this is just their failure mode as they age and wear out. While Rifa still makes these very same safety capacitors, I've been replacing them with MKP types from TDK which won't fail in the same way. Oddly enough, those failing class X and class Y Rifa parts I see seem to be early 1980s to mid 1990s vintage gear, which puts them in that same 20-30 year age I tend to use for replacing aluminum electrolytics. Another capacitor type which I replace on sight are any wax paper capacitors such as you would find in tube (valve) based equipment. The wax coated tubular types are easy to spot, but the epoxy covered parts (black beauty, bumble bee, etc.) fail the same way. After replacing 100s, if not 1000s of the things over the years, I've yet to find one, even NOS (which always seem to turn up in lots of parts from estates and such), which would pass a leak test. I replace wax paper types with polyester (mylar), polystyrene or ceramic discs, depending on how they are used in the circuit (note however that for wound foil types, modern replacement parts do not mark the outside foil, which needs to be at ground potential in many tube circuits, otherwise the circuit can pick up noise and hum).

On Mon, 20 Jul 2015, Mark J. Blair wrote: I've often wondered why they even bother to put that polarity stripe on modern film parts when it doesn't actually indicate the outside foil terminal. Maybe this is something that has become lost knowledge to manufacturers over the years to the point where even Vishay/Sprague doesn't know what that black indicator stripe was actually used for?

On Mon, 20 Jul 2015, tony duell wrote: I've not replaced any in a PDP11 power supply, however I have replaced them in other equipment with equally large PSUs. In the odd case where a computer grade screw terminal capacitor is extremely expensive or completely unobtainable (those which I've purchased were under $20-30) I might be willing to leave an original part in place, *if* it can pass a leakage test. Testing each aluminum electrolytic and wax paper capacitor vs replacing them all doesn't seem to be the prevailing norm in the vintage radio communities today. Yes, the grid cap would /usually/ be a non-polarized wax paper type, which tend to be very unreliable. I've yet to find a wax paper type which will pass a leak test and those are also on my replace on sight list. Of course you wouldn't want to replace mica, ceramic, or plastic film parts without good reason, but if a set is going to be more than just a shelf queen, aluminum electrolytics and wax paper capacitors are a must-replace item. Carbon film resistors in this sort of equipment should also be tested, however I only replace those which are either bad or out of tolerance (some brands held up better than others). That's a good idea, however something to keep in mind is that class Y safety rated capacitors are not designed not to short (and not put say a floating chassis at mains potential). Safety rated capacitors are instead designed to blow clear while not catching on fire (or otherwise be self-extinguishing), should they short out across mains potential. This means should a class Y part short, a floating chassis could still be placed at mains potential, and that capacitor is not going to blow clear. That said, anything is going to be better than a wax paper capacitor, and a class Y safety rated part is also much safer in this sort of application than a ceramic disc. As a general rule, I do not make shotgun-repairs to electronic equipment. While I may replace certain failure-prone parts outright on sight, I still prefer to determine which part(s) are causing a malfunction. I consider replacing aluminum electrolytics to be preventive maintenance. One wouldn't drive a 20-50 year old car with original hoses, belts, and tires, and IMO it is just common sense to replace electronic components such as aluminum electrolytic capacitors which have extremely well documented life expectancies and failure rates. As far as shotgun-repairs go, one of my own pet peeves are those out there selling "cap kits" (usually really low quality [sometimes counterfeit] capacitors, too) to newbies which also include a bunch of semiconductors (diodes, voltage regulators, and transistors) on the theory that those parts fail because they run hot. I've gotten to the point where I will not even attempt to service a board which has been botched up by a fat fingered newbie who has attempted to install one of those kits. There is however one component besides certain capacitors which I absolutely will replace on sight, no exceptions, period, and those are selenium rectifiers. There is nothing good that can be said of selenium rectifiers, and it is absolutely trivial to solder in a silicon diode as a modern replacement.

On Tue, 21 Jul 2015, tony duell wrote: Huh? All devices with selenium rectifiers that I/we own are OK. And a selenium rectifier only fails if overloaded. See for example the power supply of the LGP-30: http://computermuseum.informatik.uni-stuttgart.de/dev_en/lgp30/lgp30.html Note: the height is about 60cm, power rating IIRC somewhere between 500 and 1000W! And all my radio/TV sets with selenium rectifiers are OK, too. I had to replace one in my TV as it had a too high resistance, but only because I didn't know at that time that I could re-fasten the nuts of the rectifier (selenium rectifier plates can suffer from contact resistances). Christian

On Tue, 21 Jul 2015, tony duell wrote: IMO an alignment is simply part of the restoration process. When I service a set, I do so expecting that it is going to be used and thus needs to have an accurate dial vs just sitting on a shelf. Simply installing replacement aluminum electrolytics and wax-paper capacitors is not likely to affect alignment. It is extremely common however to find sets where someone else has previously mucked up the original alignment in an attempt to work around electrically leaky wax-paper capacitors which have caused the band the drift. Why is that inconsistent? If I test a carbon comp resistor and it measures within spec, there isn't much reason to replace it. Unlike an aluminum electrolytic capacitor, a carbon comp resistor is very stable chemically. Carbon comp resistors tend to drift due to absorption of moisture, and while it is possible to dry one out in a toaster oven at a controlled temperature, the resistor will again drift out again over time, so if one is out of spec, replacement is the best option. Except that the chassis in modern equipment is /expected/ to be connected to ground, unlike a floating or hot chassis in a vintage radio. Both class X and class Y can fail short. A class Y tends to have a thicker dielectric and/or a lower voltage rating, which means it is less likely to fail short, not that is cannot fail short. Possibly. Radio repair shops of the AA5 era also had a vested interest in turning a set around as quickly (and as cheaply) as possible, and a set back in again in the same year for another repair was also good for their business. Back then, consumers expected their radios to need "routine" service, so people were less likely to even question it. I've come to this conclusion based on the types and quality of radio shop repairs I've seen in these old sets. I have a radio in my to-do queue right now (an AA5) which was owned by my grandparents, where a shop needlessly hacked the leads off a Centralab hybrid module and replaced about half of its functionality with some really cheap wax-paper capacitors and a handful of resistors (after searching for a number of years, I actually managed to find a NOS module for it, so that part of the circuit will be restored to its original condition when I eventually get to that project). Cap kits or counterfeits? ;) Best way to avoid counterfeits...do not buy modern name brand aluminum electrolytic capacitors on eBay. When I've looked, it tended to be about 9 out of 10 capacitors were counterfeits when I searched for Nichicon or Rubycon parts. I wouldn't even really mind high quality cap kits for some projects, but those selling them usually make them from the absolute cheapest parts they can source, including the fake junk found on eBay, which includes counterfeit semiconductors (transistors and voltage regulators). They fail in the worst possible way, a rotten garlic smell that seems to completely permeate a room, which can take days, and sometimes weeks for the smell to completely clear. Not to mention how poisonous selenium itself is, which is best not released into an indoor environment. Since the selenium rectifier has basically zero reuse value, unless it is burned up, I usually just leave the old rectifier stack attached to the chassis and disconnect it.

Thank you all for the information. Ive been speaking with Paul Anderson on the phone, he went into great detail a general overview on how to get the machine apart and general beginner details. He referred me to the appropriate Documents on bitsavers, plan is to pick up some packs of paper and get them printed out tonight. On Sat, Jul 18, 2015 at 4:26 AM, Peter Coghlan <cctalk at beyondthepale.ie> wrote: On Sat, Jul 18, 2015 at 4:26 AM, Peter Coghlan <cctalk at beyondthepale.ie> wrote:

On Mon, 20 Jul 2015, Rich Alderson wrote: [...] Rich, Do you happen to have a list of whitepapers and/or links on hand? I too am getting tired of repeating the same thing over and over and compiling this sort of information in a single location might be helpful. Buried in my own archives I have quite a few papers and datasheets in pdf format, but they are all mixed in with everything else so finding them would be quite a challenge.

On Tue, 21 Jul 2015, Mike Stein wrote: I dunno about that. When I've done commercial boards such as industrial process controllers and CPUs for customers with nearly unlimited funds, I charged the customer based on an hourly rate. Since I use a vacuum desoldering tool, changing out 10-15 aluminum electrolytics on a board took me not much more time than 1-2. Most of the time spent on a board that comes out of the field is spent on cleaning, testing (before and after repairs) and prep, and it only takes a few seconds to pull the solder off of a couple of component leads. Replacing aged electrolytics wholesale on these types of boards also meant I didn't need to worry that the same board would be back on my bench again in the next 3-6 months. These days, I'm not taking on any new commercial work though, there was just too much demand due to all those shoddy far-east made capacitors, and it meant I pushed aside all my own projects. I guess from a business standpoint, if I had been trying to make extra money on boards repeat failing in the field and having to come back in for repairs over and over, changing out only 1 or 2 aluminum electrolytics would have made sense. That said, industrial process equipment tends to run 24/7 and is expected to be 100% reliable. If something shuts down, it tends to cost a heck of a lot of money, so I would no doubt have lost many customers. 100s? Where are you sourcing your components from? The typical board I rebuild has a component cost of about $20 or less. Smaller switchmode PSUs with a bunch of 10-18mm radials might be closer to $35-50. Larger PSUs /might/ cost closer to $100 if they have several large screw terminal capacitors in them. All things considered, that isn't very much money in today's dollars, and considering the full replacement cost of some of these boards (if they are even available), those preventive maintenance costs are an absolute bargain, /especially/ if you are doing the work yourself on your own time. How do you -know- they are "running perfectly"? Just because a widget itself is functioning, you have no way of knowing if that capacitor is working 100% properly /unless/ you actually remove it from circuit and run a full battery of tests on it. Simply measuring the capacitance with a DMM while a capacitor is in circuit isn't good enough. Given that a typical aluminum electrolytic capacitor costs anywhere from $0.12-$0.15 (4mm or 5mm diameter radials) to about $1.00 (12mm or 16mm diameter radial), it also doesn't make much sense to desolder a 20 year old part, spend at a minimum 5 or more minutes testing it, and then solder it back in. It it much more economical to pull the old part and install a new one and be done with it. (You also don't have to worry if the desoldering and resoldering process might have damaged the original parts end-seals.) That said, I personally pre-test new parts, in bulk, before I put them into my stock, so I know ahead of time that I'm installing known-good parts. On many occasions I've cut open old aluminum electrolytics, and the guts very much do deteriorate with age. In addition to corrosion of the foil (black spots and pitting) and foil to terminal junctions (corrosion), one thing I particularly noticed was the more operating hours an aluminum electrolytic capacitor had on it, the more its electrolyte and paper insulator tended to smell bad compared to an otherwise identical (same brand and series) part that had very low hours. These are all clear signs of deterioration. To those who advocate keeping old aluminum electrolytics in service, I have to also ask, would you also try to recondition 20 year old NiCd or SLA batteries and keep those in service too? The bottom line really is, if you want something to be as reliable as you can possibly make it, you replace old aluminum electrolytic capacitors which are outside of their expected service life. If you don't care if something fails over and over, or you actually like to have the same widget on your service bench year after year, or month after month, you just replace 1 or 2.

Do you seriously replace both headlight bulbs when one fails? I know of nobody who does that. Generally you carry a spare bulb kit and a screwdriver and if a bulb fails, pull over and change it. Why not change all other bulbs on the car at the same time? Or change them all once a year as part of the service [1]? What about bulbs in your house? [1] OK, this is done for things like lighthouse bulbs which are replaced after so many hours of operation, but that is a rather different case. -tony

ARGH!!! But presumably you carry a jack and tools to change a wheel. Can you not just remove the wheel on the correct side to reach the hatch (not that I want to work on a car not supported on proper axle stands...) Having had a number of bulbs that failed shortly (but not very shortly) after installation (nothing to do with headlamps, and not quartz-halogen bulbs so it was not contamination of the envelope that was the problem) I wonder if necessarily changing a good bulb is a good idea... Why not just change them both as part of the 12000 mile service (or whatever)? I am sure I once heard of a regulation that car bulbs (at least the legally required ones) had to be changeable at the roadside with a minimum of tools. Looks like that is universally ignored.... -tony

On 7/22/15 7:43 AM, Tothwolf wrote: Then why are tail light bulbs sold in pairs? I just had one go, and replaced both sides.

On 07/22/2015 04:39 AM, tony duell wrote: Automotive bulbs (at least those used for external illumination and indication are used in pairs. Most incandescent bulbs fail with age (manufacturing defects are an exception) and their light output decreases--the filament evaporates and darkens the envelope. It's logical to anticipate that if an old lamp fails, its partner is not going to be long in following it. The effort to replace two bulbs is not much different than the effort involved in replacing one. In my home, if an old fluorescent tube in a multi-lamp fixture darkens and fails, I replace all lamps in the fixture. This also applies to multi-bulb incandescent fixtures. Again, balanced with the bother of hauling out a ladder and, in some cases, disassembling the fixture, it just makes sense. Maybe I'm the only one who does this (and perhaps I'm the only one who coats incandescent lamp bases with a thin layer of dielectric grease), but it works for me. Non-solid electrolytic capacitors do have a rated service life and so merit replacement after a certain age (depending on service conditions). --Chuck

On Tue, 21 Jul 2015, Mike Stein wrote: Like anyone else, I have to buy the parts I use for my own equipment. If someone else wants to volunteer to buy them for me, I'm certainly not going to argue though ;) Without an actual list of components required and without knowing which vendors you are getting your price quotes from, I have no way to verify if your $120 total is representative of the norm. If your Cromemco system is still functioning to your satisfaction, and you have zero interest in replacing aluminum electrolytic capacitors as part of preventative maintenance, why are you even bothering to price them? I also seem to remember saying earlier in the thread: "In the odd case where a computer grade screw terminal capacitor is extremely expensive or completely unobtainable (those which I've purchased were under $20-30) I might be willing to leave an original part in place, *if* it can pass a leakage test." Well, if you want to bring those up and expand on the list of "bad caps" I mentioned, early SMD solid tantalums seem to be quite problematic in terms of spontaneously shorting out and going up in flames, even when operated at half their rated voltage (as specified by the capacitor manufacturers). I can't say I've seen a higher failure rate with newer SMD tantalums than say modern SMD multilayer ceramics, however after having to scrape the remains of many charred SMD tantalums off of (unobtainium) boards undergoing repair, I can't say I really trust them. YMMV. I've previously done a great deal of commercial work (not now though), however I still do the very same work on my own equipment. I currently have somewhere north of 300 projects in my to-do queue (everything from modern stuff made a few years ago to test equipment and radios from the 1950s and earlier) which I've already purchased and kitted up parts for (of which I'd say about 2/3 are aluminum electrolytic capacitors). I know /exactly/ what *I* spent on my parts (I have it all organized in spreadsheets, just like I did for commercial projects), and my own parts costs do not at all seem to match up with what you are describing. I can also state from experience that the majority of capacitor failures (wear out; change in capacitance and increasing electrical leakage at working voltage) do not exhibit visual signs of failure or impending failure. The main exception are some of those really low quality far-east parts made in the last decade or so which manufactures use in consumer grade electronics. You didn't answer the question. How do you know those aluminum electrolytic capacitors are functioning just as good as they did when they were new? Unless you've tested them out of circuit, you simply cannot make that assertion. Just like the NiCd and SLA batteries I mentioned, aluminum electrolytic capacitors by their very electrochemical nature degrade as they age and as they are used. You cannot claim that a 20-30 year aluminum electrolytic capacitor is going to function as good as it did when it was new. If the designer of an electronic widget designed in an adequate margin (which as I mentioned elsewhere in the thread, is a problem with modern made electronics), a 20-30 year old capacitor may still be functioning today. As it ages, how much longer that widget continues to function properly and at 100% reliably is going to be wholly dependent on how large of a margin the widget's original designer planned in. As I also previously mentioned in the thread, my own experiences with repairing and servicing equipment (of all types), both for myself and others, seems to point to about a 20-30 year mark as being the far end of the bathtub curve for aluminum electrolytic capacitors. At 20-30 years, their failure rates seem to increase dramatically, and if increasing failure rates are a concern, then you replace those capacitors. [...and if you don't care, then why are you even reading this thread?] Other components such as resistors and semiconductors seem in general to be much, much more reliable, and do not follow the same curve as electrochemical components such as aluminum electrolytic capacitors.

This thread has gone on for a while and I think we all get the points here, but one other consideration - how will removing and replacing a component damage the board? Damage the board and it's game over. One should always take the overall board's ability to handle replacement. With the board in mind, I avoid any part replacements and try to keep them to what is proven necessary only. On Wed, Jul 22, 2015 at 8:18 AM, tony duell <ard at p850ug1.demon.co.uk> wrote: -- Bill vintagecomputer.net

I am reminded of something that was written either in the manual for my 'scope or in Tekscope (I forget which). Namely that 'The best tube tester is the circuit that uses the tube'. Tektronix deprecated the use of tube/valve testers for finding faults in their instruments It applies to capacitors too. If the _circuit_ (power supply or whatever) works correcty the the capacitor is good enough for that circuit. Whether or not it meets some published specs or not. I don't remove parts unless they have something to do with the problem I am solving. If the power rails are the right voltage with sufficiently low ripple then I look elsewhere for probkems. You claim that electrolytics deteriorate with time whether used or not. How do you know the ones you install haven't deteriorated since you tested them? Firstly the paper is not the insulator. After all, it is soaked in electrolyte. And 'smell bad'??? OK, so some chemical change has taken place, but how do you know it is detrimental to the performance of the capacitor? -tony

On Mon, Jul 20, 2015 at 4:46 PM, Rich Alderson <RichA at livingcomputermuseum.org> wrote: That's very interesting. I haven't seen those white papers, but the "no matter what" must in fact depend on something, since on the PDP-1 Restoration Project we found that most of the 40 year old aluminum electrolytic capacitors still met their original specifications, including capacitance within rated tolerance. Of the few electrolytic capacitors that had failed, the problem was a catastrophic failure, not the capacitance being outside the rated tolerance. In the PDP-1, we preferred to keep the original components as much as possible. Had there been a capacitor, the failure which would have caused extensive damage to other components, we would have given serious consideration to replacing it. However, that was not the case for any of the capacitors in the PDP-1. Had our analysis indicated any expected benefit to replacing all of the electrolytic capacitors, we would have done so, and bagged and tagged the originals similar to what we did with failed components, so that they could be replaced if it ever was desired to return the artifact to its pre-restoration condition. I'm not recommending against LCM's policy, but I also wouldn't necessarily encourage anyone to adopt it, nor to adopt the practices of the CHM PDP-1 Restoration Project, without studying the issue. Eric

On 7/21/2015 9:04 AM, Lyle Bickley wrote: The whole problem with the caps is the water between foil. As modern caps use more and more tricks to improve the surface area,the water margin gets thinner and thinner.At least with GOOD vacuum tube equipment you could replace caps. Ben.

On Tue, 21 Jul 2015, Lyle Bickley wrote: How often is CHM's PDP-1 powered up and operated? If LCM's computers are going to be powered up and used routinely, it actually makes a lot of sense to go to the trouble to replace really old aluminum electrolytics, even if they seem to test good, since doing so is going to increase the reliability of frequently operated equipment. This is the reason why /I/ replace aluminum electrolytics when I'm making major repairs or fully reconditioning electronic equipment...I want said electronic widget to be as reliable as possible because it is never good when something breaks down while you are using it, especially with gear which needs to run 24/7/365 for years and years at a time. One example I can give are some Pentium P55C architecture (Socket 7) systems which I've been running with minimal downtime for ~15 years. The original power supplies with their original (and relatively low quality) capacitors lasted about 15 to 17 years (I think the manufacture date code stamped on the oldest one was 1998) before the systems began to develop stability issues, requiring me to rebuild the power supplies with new capacitors. I fully expect that the replacements would last even longer than 20 years, however I rather doubt I'll be running those computers by then.

On Wed, 22 Jul 2015, Chuck Guzis wrote: I believe there are a few webpages out there written by people who have tried it. From what I remember reading about them years ago, they had no success when they tried to use them as replacements in switch mode power supplies (no surprise, since the solid polymer parts they attempted to use had way to low of ripple current rating for that application) but had better results with certain PC motherboards. I use solid polymers as replacements in some applications, and as they continue to decrease in cost, I've been considering using them more for replacement of aging SMD aluminum electrolytics. One application where I particularly like solid polymers is for replacement of the vcore regulator filter capacitors on Pentium 4 industrial single board computers (yes, the P4 is still /widely/ used and extremely common in that market, although it is slowly being replaced by the Core Duo). The original aluminum electrolytics in that application are usually 6.3V rated parts while the solid polymer replacements are 2.5V or 4V (vcore is under 2V). In addition to long term stability, another major benefit to solid polymers is that unlike aluminum electrolytics and solid tantalums, solid aluminum polymers they can be used at their full rated voltage with no ill effects. The only real downside that I know of for a solid polymer is that they have an incredibly low ESR (less than 0.01 ohm), which can actually upset older circuit designs which were not designed for capacitors with such a low ESR.

Did the team ever consider some sort of active monitoring of the power, beyond whatever DEC implemented? The big, chunky power supplies from the 1960s and 70s have the advantage that you can very often make completely reversible changes to the circuitry quite easily, due to all the screw terminals and the like - adding current and voltage sense points, crowbars, and other gizmos that just might save a transformer or two. And with today's technology, those gizmos could be very small and tucked away. -- Will

On 7/23/2015 12:06 AM, William Donzelli wrote: I think that is their choice not ours. Ben.

That seems odd, considering the lengths you guys took on the caps, and the whole museum mentality of keeping things "safe" for the artifacts. Microcontrollers and such probably are too complex - simple over-voltage* and over-current circuits to trip the EPO (or whatever DEC called it when the line was cut) would suffice, and give some of the more delicate components a fighting chance of survival. * I once talked to a power supply guy about over-voltage sensors and the like, and he said that over-voltage conditions (like a power supply freaking out and giving TTL 8 Volts or something) due to component failures are exceedingly rare. When it does happen, it is almost always due to human error - mis-installing sense lines, cranking a trimmer too far, setting line voltage improperly, and so forth. -- Will

I wrote: On Thu, Jul 23, 2015 at 12:00 PM, William Donzelli <wdonzelli at gmail.com> wrote: There's no artifact safety issue for the PDP-1 power supplies. They use a ferroresonant transformer, rectifiers, and filter capacitors. If any of those fail, the machine won't work properly, but it won't be damaged. If there had been voltage regulators, the failure of which could have resulted in serious overvoltage, we probably would have added crowbar circuits. In the Type 30 display, it's possible, though rather unlikely, for a failure in the deflection power supply to blow the deflection drive transistors. Adding a microcontroller isn't likely to avoid that. A crowbar circuit might be useful, though the damage happens so quickly that the deflection transistors might still fail.

On Mon, 20 Jul 2015, Rich Alderson wrote: Please excuse me, but this is utter nonsense. Most electrolytics in our machines are 30 years and older, and they just work. Those caps that I checked (mostly large filter/smoothing caps), e.g. those from the LGP-30 (nearly 60 years old) or Mincal 523 (44 years old), are just fine. Smaller ones don't even have to be bothered with. OTOH foil caps from the 50s/60s (e.g. ERO/EROFOL/EROID/Wima) tend to lose a bit of their isolation and become resistive (several MOhm). This can be a problem with AC coupling in tube circuits. Also problematic are more modern foil caps in line filters (e.g. X/Y caps), or even oil filled MP caps in power supply (magnetic constanters, filters) or motor applications (phase shifters). But admittedly I don't know what crappy electrolytics you have encountered in your "industry grade" machines. Or are we talking of modern machines (<30 years) ? Statistics... don't believe any statistic that you haven't faked yourself. Honestly, IMO this doesn't really qualify you as expert in capacitors. I think those statistics are based on running the caps 24/7 at their nominal ratings, but surley they don't apply to moderate museum usage. Christian