[jan-jaap]

Jan-Jaap - Out of question (since I know you're an actual workshop with real equipment), has doing manual reflowing (with large area board preheating) on every component (slowing working your way around) on these huge boards ever yielded any fix?  I've seen you give the "wobbly board with broken joints" warning several times throughout the years, have any of them ever been brought back from the dead or dying?  It seems like large-graphics inventory is getting lower and lower (find another) that fix would become more necessary.

I have access to the lab at work for manual repairs, but we don't have a reflow oven big enough to pass a 9U VME card. I have friendly contact with the (former) owners of the facility where we have our products assembled, and they do have a reflow oven that should be capable. It's a big machine where you pass the items through on a conveyor belt. Of course they'd have to be slowly pre-heated etc. The people who ran this place were usually willing to think outside the box with oddball non-commercial jobs like this, but I do not know the current owners of the place... It's also a good hour and a half drive from where I live. So no, I never actually attempted this.

I must have somewhere between 5 and 10 dead Reality Engine RM boards. I also have some failed DG2 boards; they seem to mostly fail in the analogue parts, resulting in an unsteady picture on the monitor (like the monitor has trouble syncing). At some point I simply gave up buying them because they all arrived DOA even when the sellers claimed to have tested them before shipping. I got rid of my Onyx RE2. I have a working RE in my Crimson and I'm not touching it ;-)

Also is there any hope for a space on the edge of these large boards for some kind of reinforcement (metal frame or the like) to be added to working boards to rigidify them?

I don't think so, look for yourself. You need some space at the edges for the VME cage guide rails.


(picture from Gerhard Lenerz' SGIstuff)

The DG2 and RM4/RM5 PCBs are thinner than most other PCBs found in those systems (e.g. CPU boards, but also the GE8/GE10). So if you insert them in the system there's some sideways play of the PCB in the guide rails of the VME cage. It takes considerable force to insert one into the backplane; it almost makes you feel the PCB must look a bit like an accordion the moment you push it in. The same for the front interconnect. It takes force to push it in, and it must be pushed straight back into the system while engaging 2 ... 5 connectors simultaneously. Push it in crooked and you'll likely put strain on something not designed for it. Same for removing it. The force it takes is a good bit more than what it takes to move the entire system. So, you have two options: try to wiggle it free, or sit on your but in front of the system, holding it in place with your feet, and pull ...

[GanjaTron]

There is indeed too much slack in the RE boards, and I agree you need a lot of force to insert them (particularly the GE with its full-length, high density connector), as well as the front interconnect. The spare DG2 I fell back to has the same jitter in the analogue output that Jan-Jaap mentioned, which was surprisingly (still) absent from the one that's on the fritz now. And as for the RM4's, I've had 3 out of 5 go bad...

First time I tried inserting the Geometry Manager of my VGX I slammed the PowerSeries chassis against the wall -- the force I had to exert on the backplane was phenomenal. I don't think I've removed that board from the chassis since. But unlike RE, it's also been very stable.

So decades later, the perceived advantage of the big irons -- flexibility through a graphics subsystem that lives on separate boards -- is also its achilles' heel.

That said, are there any known instances of solder joints on the backplane going bad? I've had to access the rear of the chassis while refurbing the PSU just a year ago or so, and I'm really not looking forward to repeating that anytime soon...

[CB_HK]

That said, are there any known instances of solder joints on the backplane going bad? I've had to access the rear of the chassis while refurbing the PSU just a year ago or so, and I'm really not looking forward to repeating that anytime soon...

I pulled the backplane of my Crimson out when I was working on refurbing it a few years ago. I ended up recapping the backplane at the time since, like you, I didn't want to have to deal with it again and I was on a quest to "future" proof the system against any leaky electrolytics. I can't say that I saw any bad joints or areas that needed to be touched up. The soldering seemed to be quite good quality (as one would hope). I've had a lot of "fun" dealing with a faulty VGXT setup and repeated removal and installation of boards has yet to cause any visible damage--no bent pins on the backplane and no broken or cracked connectors.

I'm not advocating for beating up systems this old, but I can at least say that my experience has shown the backplane to be fairly robust and well built.

[jan-jaap]

I pulled the backplane of my Crimson out when I was working on refurbing it a few years ago. I ended up recapping the backplane at the time since, like you, I didn't want to have to deal with it again and I was on a quest to "future" proof the system against any leaky electrolytics.

Is that even possible with a normal soldering iron? Those are really really thick PCBs...

[GanjaTron]

I ended up recapping the backplane at the time since, like you, I didn't want to have to deal with it again and I was on a quest to "future" proof the system against any leaky electrolytics.

I considered it too, but (a) was too lazy to place another order for parts, and (b) in my experience, big fat caps rarely fail. It's dem lil' guys ya gotta watch out for!
I keep an old tin full of dead caps I've pulled -- don't ask why. By now it reeks of electrolyte, and the insides have turned green from corrosion!

[weblacky]

I ended up recapping the backplane at the time since, like you, I didn't want to have to deal with it again and I was on a quest to "future" proof the system against any leaky electrolytics.

I considered it too, but (a) was too lazy to place another order for parts, and (b) in my experience, big fat caps rarely fail. It's dem lil' guys ya gotta watch out for!
I keep an old tin full of dead caps I've pulled -- don't ask why. By now it reeks of electrolyte, and the insides have turned green from corrosion!

Not to derail this thread but I felt the need to mention my experience and why I firmly DO NOT leave big caps alone during a rebuild. 

There is a bit of a fallacy in the statement people say. “Big caps rarely fail”, while it depends on the term “fail”, I find they absolutely do fail in a way many don’t consider, which leads to the rapid failure of downstream caps.  

RIPPLE. 

In most power applications, big caps are designed to store high voltage charge and absorb ripple, to keep downstream components from experiencing too much ripple themselves. 

I’ve seen two examples in power supplies (normal ATX PC) where the aging of the large caps allowed so much ripple through that it actually placed excessive strain on other caps but also caused total output ripple to be high enough to cause freezes, crashes, and blue screening of the system at high loads (like during OS updates or during OS installation). 

Just replacing the large cap(s) brought down the total output ripple to acceptable levels and the system has since ran without crashing, 8+ years old total, fixed 2 years ago). 

So this case taught me that a large caps, not performing as it should, will affect the whole. Just because the cap claims to still be the same farad reading and all that doesn’t mean it’s healthy. 

In short, I strongly advise the replacement of all large storage caps, for the above mentioned reason. Often they are the first level filter and largest filter in a power system. As they age the later caps have to take over handling the downstream ripple those old caps are letting through, which brings the whole design out of tolerance.  

My opinion and personal experience: please  replace All caps, not just the output caps. Because the output caps failing may have been accelerated by the filter cap(s) failing.

[CB_HK]

I pulled the backplane of my Crimson out when I was working on refurbing it a few years ago. I ended up recapping the backplane at the time since, like you, I didn't want to have to deal with it again and I was on a quest to "future" proof the system against any leaky electrolytics.

Is that even possible with a normal soldering iron? Those are really really thick PCBs...

It is with an adjustable iron, fat tip, and a lot of power! I also used a really nice Hakko de-soldering gun. It worked surprisingly well, though the whole process was a complete pain to do and I really don't look forward to ever having to do it again.

[GanjaTron]

My opinion and personal experience: please  replace All caps, not just the output caps. Because the output caps failing may have been accelerated by the filter cap(s) failing.

Your input is appreciated! I realise there are different degrees of failure -- from ripple to catastrophic, and that any "degraded-but-not-quite-knackered" component may upset a system that already operates within tight tolerances. Perhaps a dumb question, but what exactly are those big caps on the backplane for? As far as I can tell, they don't appear to act as bypass caps. Of course without a schematic, we may never know for sure...

[weblacky]

I don’t own any big iron but I’d speculate it’s for the same reason. Either filtering or large reservoir.

The only reason I know to use large caps is to “fill in” a gap of current (causing a local voltage line drop) when sudden current draw happens (think car stereo) or to absorb lots of noise and attempt to “clean up” the variance in the line. Really that’s the same thing.

I’d assume these machines have the same “sudden” power demands as a normal device. But they have much more distance to travel.

So I could easily see the caps being used for filtering, preserving line conditions, or as a stop gap from experiencing very large voltage swings (ohms law) as the device takes sudden current demands on the bus and it’s slots, which could cause (without caps) major differences in voltage values at each backplane connector based on each slot’s draw at any time.

But these would be conjecture. My main point being that whatever they are for, they are chosen for a reason (whatever that was) and often their ability to absorb is as important as their ability to suddenly deliver to smooth out voltage drops.

Fresh caps have the best charge/discharge rates. As ESR increases you get a resistance to those operations that’s both heat and improper timing to fill in the gaps in the signal line it’s on (charge/discharge was slowed, so now slightly out of phase with timing requirements to best fill dips in voltage in real time).

Depending on where you find them, it may just be for signaling voltages or maybe they are in the path of primary operating voltages (5v, 3.3v, etc) that get delivery to each backplane slot?

Either way, they try to bring calm to a turbulent stream.

[GanjaTron]

Either way, they try to bring calm to a turbulent stream.

In that case I need a few more caps in my life... 
Going thru a lot of ripple myself!