On Sun, 13 Dec 2015, Robert Jarratt wrote: Obviously zinc plating of the screw has come in reaction with something. It doesn't seem from the photo electrolyte could have reached there, however I'd double check that, in particular inspect the surface of the PCB throughout looking for any difference in appearance. This stuff can reach far, e.g. in one of my failed H7821 PSUs which was unfortunately stored at an angle to its proper operational position (or otherwise the upside down position of problematic capacitors have prevented the seal capacitor. And this screw is located in a bottom corner of the PSU when in its operational position so gravity will drive leaked electrolyte towards it even if capillary action does not. Other than that I have no idea. Lead or tin from the soldering alloy aren't as reactive as zinc is and therefore they don't show signs of corrosion so quickly as zinc does, although as I observed in the failed PSUs I have they eventually do too. As does copper from any exposed traces and can the solder mask or the epoxy substrate of the PCB itself too. Indeed, and good luck! NB the two brown capacitors seen in the middle top of your photo are among ones that leaked in my PSU; an SXF marking is barely recognisable on the edge of the lower one in your photo. After considerable effort earlier this year I succeeded with removing the intermediate heatsink blocking access to these caps, by desoldering the four rectifier dual-diodes holding the heatsink in place -- at the cost of losing a leg PCB. Fortunately MBR3045PT parts are still available and I was able to get a replacement, and the damage to the PCB is I believe not critical. What's important the larger rectifier dual-diodes (84CNQ045; no clear indication of the manufacturer) have survived intact, as these seem to have become unobtainium now. Apparently the last die foundry capable of making these parts has discontinued them earlier this year due to lack of customer interest, so the only source remaining might be part recovery from otherwise broken equipment. Overall I think I'll need better tools to be able to desolder such stuff in a more repeatable and less destructive way. The thick legs of the 84CNQ045 parts combined with the large volume and consequently thermal capacity of the intermediate heatsink seem to be able to take heat away virtually instantaneously. I'll appreciate your advice on choosing a good soldering/desoldering station, capable of handling such high-current (and consequently highly heat-conducting) parts. For the curious and possibly to provide some information on the parts involved I've documented progress with the disassembly of this module at: <ftp://ftp.linux-mips.org/pub/linux/mips/people/macro/private/h7874/> (large photos!). Removing the offending caps revealed C322 and C323 designation underneath; these are 330?F/25V parts. I decided not to move forward with installing replacements and reassembling the module, or proceeding with disassembling the other module without upgrading my soldering/desoldering tools first though. No idea. I've only recently started finding my way with this system; I'm mostly a software person. Admittedly quite a low-level one, but still on the software side. And it's been only earlier this year I discovered the issue with the dreaded SXF caps, even though I've been using various DEC hardware/PSUs for over 15 years now. Maciej

A bench test of the PSU with a small dummy load saw the PSU appear to work fine. It might help to load it more to see if it still operates correctly under load. The specs for the PSU are: 22 amperes at +12.1 Vdc 60 amperes at +5 Vdc 7 amperes at +3.4 Vdc 4 amperes at -12.1 Vdc Has anyone got any suggestions for how to create a higher load in a fairly simple manner? It would take a lot of resistors or vehicle light bulbs to create some of those loads, and something heftier than a breadboard to connect them together. Thanks Rob

Bulbs have a lower than normal resistance when cold so it is possible that you may be drawing more than 4A from the -12V line at the instant that you apply power. Bulbs are great for supplying a minimum load for a power supply that won't regulate without it but they are not so great for supplying a load near the maximum. In the latter case, the maximum load will be exceeded for a short time at startup. This is not a problem if the power supply is designed to cope with this (a power supply for motors for example) but computer power supplies will typically be designed to react quickly to overload conditions. Three or four 15 ohm resistors in parallel would probably make a better load for the -12V line. It seems hard to imagine what could normally draw something approaching 4A on the -12V line in a VAX 4000 though. If it was possible to get an ammeter in series with that line maybe you could get an idea of what the normal draw is on that line, if the power supply stays running long enough. This would give a better idea of what sort of dummy load is needed to simulate it. Regards, Peter Coghlan.

:-( As in, you don't have any such thing to put in series, or you don't see a good way to get it in series? I've got a couple of 0R1 resistors which work nicely for that in conjunction with a voltmeter that can handle millivolt measurements; you might be able to get something similar. Mine are 10W, and, while they are designed for bolting to a heatsink, they can probably handle at least 2 or 3 watts, which means some 5-6 amps, in free air. (Though that then drops about half a volt, which can be a problem when you can't put it on the input side of the regulation.) /~\ The ASCII Mouse \ / Ribbon Campaign X Against HTML mouse at rodents-montreal.org / \ Email! 7D C8 61 52 5D E7 2D 39 4E F1 31 3E E8 B3 27 4B

test full Rob, Seeing as it already seems to have over current sensing, I guess it has some small resistors in there so perhaps it is possible to measure the current. I notices one looks like this:- http://uk.farnell.com/welwyn/oar1-r020fi/resistor-1-0r020/dp/1200359 Dave

Something to watch out for with high current sense resistors: they often have four connections rather than two. This is a "Kelvin connection": two connections are for the current to be sensed, and the other two pick up the resulting voltage difference. The idea is that you get the voltage drop across the resistor itself, as opposed to the voltage drop across the resistor PLUS the solder joints. That makes a difference if you have a sense resistor of just a few milliohms. If you have a high current supply but it doesn't use Kelvin connected current sense, you're vulnerable to false shutdown if the solder joints are less than perfect. If so, switching to the Kelvin hookup, if you can see how, is a good answer; failing that, make sure the joints are in excellent condition. paul

It does look like a Kelvin hookup. Yes, usually they are two fat and two skinny connections, but so long as two connections carry the current while the other two do not, it's valid. paul

Robert Jarratt wrote: Just resolder the SMD Caps, they sometimes get a loose pad. You see if that happened while resoldering. 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

On 2015-Dec-21, at 1:14 AM, Robert Jarratt wrote: The lighter-beige ones, some of which are wider, are likely caps, for example the one beside the 34072 dual op amp looks to be a simple power decoupling cap (across the IC power pins). The reddish-brown units with 4-digit numbers are resistors: 3 sig figs & zeroes multiplier, e.g. 1002 = 10,000 ohms, 1330 = 133 ohms, 1181 = 1180 ohms. The unmarked(?) 'redder' ones I expect are resistors, for example the one above the "di" in "digital" looks like the feedback gain control resistor for the op amp, although in other places that would make 4 resistors in parallel which is unusual. Resistors can generally be identified with a multimeter, if you can come up with the same value when swapping the multimeter leads (polarity) and stepping through the multimeter ohm ranges you've probably got the resistance - making the multiple measurements to identify non-linear responses to eliminate conductivity readings from semiconductor junctions. There looks to be two very similar networks around the two op amps in the 34072. It wouldn't really be that difficult a board to RE, although it does look like some poking around with a multimeter for continuity will be necessary to track internal traces, but they look to be relatively limited in number.

On 2015-Dec-21, at 3:10 AM, Robert Jarratt wrote: SMD markings and packagings are ambiguous, unclear, and all over the spectrum. Here's a good site for chasing SMD markings: http://www.s-manuals.com/smd REing boards like this is an iterative process with consideration of the device markings, pin use, measurements, what makes sense electronically, and design period. In the two cases you mention it looks like the H's are a slightly different font or size and spaced slightly differently, I suspect the id codes there are A4 and 2F. An initial guess for the "A4H" device that's right above the LM339 is an A4-code dual diode with only one diode used, as it looks (from the photo) like one pin is unconnected. http://www.s-manuals.com/smd/a4 A good guess for the "2FH" devices are 2F-code PNP transistors: http://www.s-manuals.com/smd/2f Similarly, the "1PG" or "1PC" device may be a 1P-code NPN transistor. Diodes and transistors are best double-checked by first characterising your multimeter with PN junctions in known transistors or diodes so you know what ohms-range and what reading to expect on your MM when you have a PN junction in forward-conduction. Look for the lowest ohms range that will give you a reading for forward conduction. Then use that knowledge to check the SMD transistors and diodes for sensibility. Some strategy: - Sort out all the network connections for the board, by visual observation and continuity measurement. - Identify the power supply pins to the board by chasing the power supply pins of the IC packages. This will include or begin with identifying the ground connection(s) to the board in part by chasing the ground from the main unit to the pins which this board plugs into. Is an internal layer on the board a ground plane? - From what I can discern it may be that the two ICs are supplied by +/- supplies, that is, the V- pin for the ICs may go to a negative supply rather than ground, so there may be ground(s), V- and V+ connections to the board, rather than just ground and V+. - I'd suggest drawing the two op amps one above the other with a positive bus horizontally at the top of the page. If there is a negative supply draw a negative bus horizontally at the bottom of the page. Draw in the networks around the op amps, it looks like they are similar and feed into the 3 terminal device located to the upper-right of the LM339, and thence into one of the comparators. That should take care of a lot of the board, then do the remaining 3 comparators. - within sensibility, try to draw everything so electron flow is 'up' the page, this means for example that PNP transistors will be drawn upside down, with the emitter at top-right, while NPNs will be conventional with emitter at lower-right. Everything should sit vertically in between the negative bus below and positive bus above or between grounds towards the bottom and the positive bus above. - within sensibility, try to keep signal flow left-to-right with only feedback paths going 'backwards' right-to-left.

Good news! It looks like I have managed to get this working again. I realised I had not checked the electrolytic capacitors on the main logic board, ones situated far from the main power output boards. I found one which had a higher ESR than two other identical ones, although the ESR was still theoretically OK according to the table on my meter it was at the limit, so I replaced it and the PSU now appears to work. It really goes to show, *all* electrolytic capacitors should be checked. Picture showing the one I replaced here: http://1drv.ms/1PlfKIV. Interestingly, I had had a similar failure in another of these PSUs, which I handed to a friend along with a VAX 4000-300 I gave him, in the hope that he would be able to fix it. As that PSU was already partially dismantled of doing the reverse engineering. I noticed that the same capacitor on that PSU also has a higher ESR, but again within range of what should be OK. So, looks like I might try this as a fix for the second bad PSU too, if he is OK with me trying that. Regards Rob

Tiffe In the light of what you say below, do you mean you only changed this one, or did you change all of the ones on the logic board? Does that include the really big ones too? Nothing at all was wrong with your idea. I had indeed followed your advice on the 12V board, on the basis that I could reach them and that it seemed to be the 12V side that was having the problem, so implying that other parts of the PSU were OK. I did not change any on the 5V side as that side seems to be OK and I don't really like the approach of pulling and soldering them them all desing Yes, and I may yet have to change those on the 5V side, because after posting the good news, I had one brief episode where the PSU switched itself off again after running for about a minute. On powering it up again it worked for a good hour without issue. So perhaps I should bite the bullet and replace those too, and also the other small electrolytics on the logic board too. However, I don't think I will replace the really big ones. Generally, I get the impression that these big ones suffer fewer problems, certainly their ESR is fine. I am also reluctant to change them because they are generally connected to large heat-sucking planes, making them hard for me to desolder without making a mess of the board and possibly damaging it too. Regards Rob

current. detecting random. I wonder if the voltage level detection thinks its out of spec, rather than the over current parts. These PSUs afre exceedingly complex... and I can't see the resistors themselves being the problem. Perhaps re-soldering them Dave G4UGM

Yes...but ammeters (for nontrivial currents) are generally just voltmeters across current-sense resistors, and it's the current-sense resistor that needs to carry the current and thus requires attention. /~\ The ASCII Mouse \ / Ribbon Campaign X Against HTML mouse at rodents-montreal.org / \ Email! 7D C8 61 52 5D E7 2D 39 4E F1 31 3E E8 B3 27 4B

it is sure finding and I was just reading the BA440 enclosure manual, and realised that the PSU will shut down the system if it thinks the fan has failed. When I have the PSU on the bench I don't have the fan connected of course. This would explain the shut downs on the bench even when under low loads, it could be just that it is detecting fan failure. At this point I realise I could be barking up completely the wrong tree in suspecting faulty overcurrent detection. This PSU is really hard to diagnose, because it is so complex, especially because it has a pretty complex interface to the backplane, not just power, but also, seemingly, a fair number of control signals. I can't really tell why it is shutting down, which makes it difficult to know where to begin once the basics have been checked. Regards Rob

I Well, I have the exception that proves the rule. My H7864 PSU (from an rtVAX 1000, same PSU as MicroVAX II) had a switching transistor and resistor fail, no cap failures.... :-) Regards Rob

I to locations Tonight I replaced the leaking capacitors on the +12V/-12V board, and also replaced the two similar ones that looked and measured fine. I put the PSU back together, put in a couple of less important boards and drives, and the machine now seems to power on OK. I measured the ripple using the power connector for the front panel and that looks OK too. Unfortunately though, that connector only sends out +12V, +5V and -12V. It does not have output for +3.3V, so I have not, so far, been able to check the ripple for this. It is a bit awkward to test on the bench with a dummy load. As I did not replace the capacitors on the +5V/3.3V board, because the underside is very hard to access, and I am reluctant to pull them off and solder from above, I would like to be sure there is no ripple on the 3.3V supply. So, I was wondering if anyone has any neat tricks for probing the +3.3V supply with the PSU installed in the machine? Next job is to replace the capacitors on the little DSSI terminator, they are easy to change. Thanks Rob

Robert Jarratt wrote: ...sounds very similar to my experience with that PSU. The cure seems to be simple run that machine then mine has stopped powering off now it seems. I don't even know if the power off is because of an detected overload at all, I had problems to even power on the machine the first time before changing the caps. It would start only the approx 5 time if I switched it on and I think that those problems are related. I've installed VMS 7.3 and some additional packages and played for some days with that machine and it is running stable now if I power it on. The only thing that I still have to complain about is that the NiCad Pack is empty again a little to fast for my taste, the Battery was a new one.. 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

Robert Jarratt wrote: ..sounds at least interresting. Maybe you have a bad Tantal Capacitor somewhere on the boards.. Try to pull boards until the current decreases... Whatever you want.. I'm curious what you will find out. The machine is build to contain those NC Packs, and yes I know that gases But the machine had some years running time on it and the corrosion effects are pretty small, so I think with the fitted new battery it will last some additional and I could handle this. 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

On Thu, Dec 10, 2015 at 11:42 PM, Holm Tiffe <holm at freibergnet.de> wrote: I took a look at the DSSI & SCSI termination power supply board in my VAX 4000 BA440 chassis. It's the small board in the storage area of the backplane mounted in place by a single screw. There are two identical circuits based on the LT-1086 with the adjustable output voltage set at 5.30V [1.25V * (1 + 392/121)]. Each circuit has a 100uF 35V 105C nichicon cap between the input and ground and the output and ground. Both circuits on my board still output 5.3V from a 12V input, but I think the caps are starting to leak electrolyte on the PCB and I should replace them all (4 total) before things get ugly. Each circuit has three diodes, an SB520 and a 1N4735A for which I can find datasheets, and a third one which looks like a General Instrument MP654 for which I can't find a datasheet and I'm not sure if that is the correct part number. I can't quite read the marking all the way around the body of the diode. If someone else takes a look at their BA440 DSSI & SCSI termination power supply board let me know if you can make out what that part number is, or if MP654 is a GI part for which you can find a datasheet. -Glen

Robert Jarratt wrote: All of the Nichicons in my PSU and that Terminator Power Source looked ok, but that was all...totally dead from the ESR side of view. Don't bother with what you measure, replace them all if you can. The leaking Caps in my PSU where from another Manufacturer (don't remember wich one) and had a black color. 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