The [H] Retrocomputing Thread!

They're beasts. We been running them 24/7/365 since 2002/2003. In that time period, they've probably only been powered off 3 or 4 times at most, and only have reboots for required engineered software changes that need it. I think the last time some of them were rebooted was 3 or 4 years ago. They are the most stable computers I've ever used. (OpenVMS operating system)
 
And this is why I love the computers of yesterday--dead reliable stability. Today's systems stability would have been considered flaky garbage by yesterday's standards, except with servers--today's servers are quite stable. I know I only had to reboot my 2950 twice in a year--and that was when running it as a desktop, lol.
 
Thank you for posting these servers!!! I hate seeing equipment like this thrown away (there's a guy in CA on reddit homelabsales with some Dell 2850s and 1750s that's needing to just get rid of them and everyone there tells him to throw them away even though they look brand new aside from some dust). People don't realize how ungodly expensive this equipment was back in its day. Hell, even my DL380 g5 from just a decade ago msrp'd at over $7000! And this is a regular plain-jane server that wasn't top of the line or anything. That Dell 6600 had to cost north of $30k at the time. :eek: Lovely rack and setup! Thank you for sharing!
Thank you for your kind words!
You are right, those servers were insanely expensive, and the Intel Xeon MP 3.0GHz CPUs in them cost around $3600 each back in Q1 2004 - glad they aren't anywhere near that now. :eek:

Ouch, those poor Dell 2850s and 1750s deserve better, and new life can easily be breathed into them for a modest cost.
The PowerEdge 6600 was apparently used as a phone and voice system from the mid-2000s to up until a few years ago, and shockingly, there was almost no dust at all in the unit, and the fans were totally clean.

At the university I worked for back in the 2000s, we had a PowerEdge 6600 that hosted our website, help desk ticket system, print sharing, and file sharing over a SAN via 2Gbps fiber.
It was a joy to work on and maintain, so being able to find one in such mint condition, and being able to do real production work on it again over 15 years later, was a real treat.

Nice, those DL380 units were, and still are, tanks and have held up extremely well over the years - I'm glad you still have yours! (y)

Nice rack. Those were used for the Compaq Alpha line of server equipment. I have quite a few at work we just recently tossed, and still have several in service racked up with the matching Alpha DS10's, DS20E's, quorum disk arrays, RAID arrays, etc. Next year we are finally replacing our last production Alpha's. Luckily I grabbed some before scrapping this past go-around of upgrade replacements.
Thank you, I had bought it from said university nearly a decade ago, and ironically enough, it did have Compaq Alpha equipment within it - good memory!
Unfortunately, I wasn't able to purchase any of that equipment, but getting the 42U rack for $75 was a great deal.

Nice, I'm glad you were able to get some of that gear as well before it is phased out completely.
I'm impressed that all of that equipment has been in service this whole time, and is still running.

When most individuals think of Compaq, they always remember their desktops and laptops from the late 1990s and 2000s, which admittedly weren't that great.
However, their desktops and laptops from the 1980s were rock-solid (and are some of the first clone IBM PCs), and their server and enterprise equipment was absolutely amazing, both in reliability and longevity.

It will be sad to see the rest of those Alpha systems finally be retired, but perhaps you can get your hands on a few more of them before they reach their final destination. ;)
 
Thank you for your kind words!
You are right, those servers were insanely expensive, and the Intel Xeon MP 3.0GHz CPUs in them cost around $3600 each back in Q1 2004 - glad they aren't anywhere near that now. :eek:

Ouch, those poor Dell 2850s and 1750s deserve better, and new life can easily be breathed into them for a modest cost.
The PowerEdge 6600 was apparently used as a phone and voice system from the mid-2000s to up until a few years ago, and shockingly, there was almost no dust at all in the unit, and the fans were totally clean.

At the university I worked for back in the 2000s, we had a PowerEdge 6600 that hosted our website, help desk ticket system, print sharing, and file sharing over a SAN via 2Gbps fiber.
It was a joy to work on and maintain, so being able to find one in such mint condition, and being able to do real production work on it again over 15 years later, was a real treat.

Nice, those DL380 units were, and still are, tanks and have held up extremely well over the years - I'm glad you still have yours! (y)


Thank you, I had bought it from said university nearly a decade ago, and ironically enough, it did have Compaq Alpha equipment within it - good memory!
Unfortunately, I wasn't able to purchase any of that equipment, but getting the 42U rack for $75 was a great deal.

Nice, I'm glad you were able to get some of that gear as well before it is phased out completely.
I'm impressed that all of that equipment has been in service this whole time, and is still running.

When most individuals think of Compaq, they always remember their desktops and laptops from the late 1990s and 2000s, which admittedly weren't that great.
However, their desktops and laptops from the 1980s were rock-solid (and are some of the first clone IBM PCs), and their server and enterprise equipment was absolutely amazing, both in reliability and longevity.

It will be sad to see the rest of those Alpha systems finally be retired, but perhaps you can get your hands on a few more of them before they reach their final destination. ;)
I keep forgetting how expensive top of the line new processors cost...ah the great thing about picking up old stuff, lol.

Yes they certainly do. I hate people bashing older hardware that's perfectly working. To steal a line from the movie Seabiscuit, "You don't throw away a whole life just because it's banged up a little." I think the same applies to servers that are still ready to run, but maybe just not the bleeding edge. And there's nothing wrong with that in my book, especially when a 13 kid with an interest and aptitude for servers will be able to afford it and sink his/her teeth into it. That's much more value than the scrap metal anyday.

Super cool to know how much the 6600 could do back in its day. :)

I remebmer the compaqs of old and they were the first PC clones that gave IBM a run for their money. I believe they were also instrumental to establishing the EISA standard which brought 32-bit and still retained backward compatability--something that MCA, PCI, and PCIe have all failed to do even when they could have been designed to do so. HP I think has done a good job of incorporating compaq into HP and a lot of the HP lineup of today started out as compaq machines. :)

Those alphas will be just as awesome in their second life. I just hope that they get one!
 
Recapping a customers' Macintosh SE:

bcC1IoOl.jpg 9bqONX2l.jpg
DIaHpF1l.jpg XWiIptJl.jpg

Unfortunately recapping the analog board exposed more hidden faults. Now the screen has an even worse case of wobbling raster (it was already happening before, just not as bad):
I40ogYZl.jpg

This is after about 8 hours of tracing the circuits on the analog board. There were dozens of solder joints with hairline cracks from heat stress, which after fixing made the wobbling even worse. I suspect at this point there are components with internal damage, like arcing resistors. Or the TV deflection or quad input AND buffer is bad. Can't test those without damaging them unfortunately. It's definitely the analog board, because I can swap the one from my Mac SE FDHD in and everything just works:
VWlTp2Ql.jpg

I'm just going to recommend turning the junk analog board into a frisbee and get another used one that doesn't have so many problems.
 
Nice work! Keeping these older guys running is going to get ever more difficult as the years pass. :(
 
Not really sure what I did, but I got the analog board working better, but still not perfect. It's now back to the random glitchy raster, instead of a full screen spasm.

0YCy9All.jpg

I just changed the high frequency deflection capacitor back to the 3.9uF electrolytic that was in before and one more cap on the horizontal section. Also gave all three potentiometers a good clean and re-touched up some joints that still looked suspect.

There's still a bad component somewhere that has yet to be found.
 
No particular reason to suspect it. But for me, horizontal output transistors and the circuit that drives them have always been a source of grief.
 
No particular reason to suspect it. But for me, horizontal output transistors and the circuit that drives them have always been a source of grief.

You called it.

A0xpgxpl.jpg

I left the machine running and went to go out and get some food earlier. About 30 minutes after I left, got a frantic call from home that one of my computers was on fire. I returned home to a house full of burning electronics smoke and found the HOT had gone thermonuclear. I'll have to spend a bit of time testing the machine to make sure the HOT going bad didn't kill anything like the CRT or logic board, but the analog board is a write off, I wouldn't bother trying to repair it now that this happened. Getting another HOT would be a challenge anyway since they haven't been made in decades and only questionable NoS exists at high prices.

All of the random glitching must have been the HOT starting to fail and it just finally had enough and let go.
 
Dang, hope everything else was ok!
Excellent troubleshooting, I'm really impressed, and hopefully the board can be replaced without issue.
 
The machine definitely gave me the false sense that it was fine. I had it running for probably 8-10 hours yesterday and the raster glitching actually got better the longer the machine was running and there were no funny smells. The time from it running today to it failing was less than an hour, so I guess it had enough. I had actually looked over the analog board before I left and saw nothing amiss.
 
The thing is that there's really nothing around the HOT in this design besides a resistor:



It sits between two transformers, the flyback and a secondary isolation transformer. There's some stuff tapped off before the HOT, directly off the flyback and I tested all of these parts beforehand and everything was in spec. I have a second working analog board from my Mac SE FDHD which is the same design that I compared parts with. If it was another part that caused it to fail, I would have expected to see more burned components, like exploded capacitors or burned resistors from shorting out.

The analog board is not repairable, the HOT cooking off badly burned the board under it, making it conductive. It'd require major surgery with a drill press and/or dremel with a cutoff wheel to remove the burned bits just for starters. Who knows what it may have done to the rest of the board or even the flyback shorting directly to ground like it did.
 
Hmmmm.... No obvious freewheel diode.
When waterhammer hits the suddenly closed HOT collector, all that inductive inertia goes where?
Out the diode to Grids 2&4? Zorched is zorched. Like you said, the charcoal makes irreparable.
I just wanna know, cause I fought this battle half a dozen times, never to claim complete victory.

Or maybe CR3 is dumping the energy from primary side leakage inductance into C13 for re-use
next sweep. The leakage inductance meaning winds in free space as-if the core wasn't there, as
thats always in parallel with the core and other windings. Enough energy stays on the primary to
blow parts. Can't assume 100% makes a safe exit on the secondary side.

All complicated by energy in the deflection yoke.

SE30Analog_CleanupAttempt3.png


Cleaned up best as I could without hallucinatin' details I can't accurately read.
Lots of copy/paste legible over not-so-legible. Hopefully didn't change nothin.
Two caps on the -5 look upside-down to me, but I left them exactly as found.
Perhaps they are unpolarized...

Who spells Anode with extra O or D? If thats ANDode, I wanna see XORode.
Not my typo, just poking fun at what seems all too legible original nonsense.

I may have misread the HOT as BU486 and cleaned around it as-if so.
But that part doesn't seem to have existed. BU406 with a slashed zero
probably makes better sense at that location. I've botched this drawing
by shaping that misread to look even more what it probably wasn't.
https://www.onsemi.com/pub/Collateral/BU406-D.PDF

Weirdo 58V capacitors are probably 50V with a slashed zero too.
We see legible number 8's at least twice near the deflection yoke.
Don't look anything like I mistook for 8's earlier, had to be zeros.

Then again 8 in 68K looks much like slashed zeros but standard
value 68K seems much more likely than 60K.

Some of these problems are fixed now, others pending a bill of materials.
 
Last edited:
The HOT is definitely a BU406, a long obsolete and hard to get part. Nobody needs deflection driver transistors anymore.

Reading the schematic has definitely been a headache, the overly bold font makes it hard to differentiate between 6, 8 and 0. I had to go between both analog boards and the schematic to make sense of some parts. Off the top of my head, the 68k resistors are 68k. The analog board from my Mac SE FDHD is a bit different than the one that went on fire, some of the diodes are much beefier, like CR2 is literally three times the size.

There is only one capacitor on the analog board over 50v, it's C19 which is 10uF at 160v. The rest are 50v or less. The one "gotcha" is C15, which is the horizontal deflection capacitor. It's an electrolytic bipolar 3.9uF 35v high frequency rated. These are not available anywhere, the only substitute is a 250v high frequency polypropylene film capacitor. https://www.mouser.com/ProductDetail/667-ECW-F2395JA/ Fitment is obviously a problem, so you have to bend the legs and get creative.

Installing a regular bipolar capacitor in place of C15 will cause very bad things to happen. If the original electrolytic capacitor isn't faulty, I'd say leave it alone.
 
Last edited:
Mouser must have just gotten those in stock because they definitely weren't available just a few weeks ago.
 
I found this system at the local thrift store tonight and I think it's a 386, simply from the back panel, but I can't figure out how to open it!

PXL_20210211_004225161.jpg


I have that front cover. I removed it to see if I could catch a glimpse of what's keeping me from getting inside.

PXL_20210211_004234100.jpg


PXL_20210211_004247593.jpg


Here it looks like someone tried to pry it out. Makes sense, as I would suspect that when removing the three rear screws, that the case would slide back. It doesn't. It actually has metal portions holding it on. It tells me it has to slide forward, but HOW? I tried to pull the front panel off and that thing isn't budging. Even tried my little spudger and not a chance.

Starfalcon thoughts?
 
Does the front panel pop off?
Some of those older 90s desktop cases had front panels that could pop off and reveal more screws or inserts, which when removed allows the surrounding frame covering to be removed or pushed forward.
 
Looks like an upside down baby AT case, what does the bottom look like. Also is that a switch on the front next to the floppy drive?
 
Does the front panel pop off?
Some of those older 90s desktop cases had front panels that could pop off and reveal more screws or inserts, which when removed allows the surrounding frame covering to be removed or pushed forward.
I cannot get the front to budge at all, even when using a spudger to get between the case and front panel.
 
Looks like an upside down baby AT case, what does the bottom look like. Also is that a switch on the front next to the floppy drive?
Nothing underneath but feet.

Edit: that black thing next to the floppy drive is a latch for a door. The case originally had a door to close over the front.
 
So what goodies did you find in your computer treasure vault?

Nothing too fancy, although everything is CLEAN.

Sound card

View attachment 328115

Trident video card with all slots full. Probably still 1mb at best. I haven't checked the info on the chips yet and likely will just boot it to see those deets.

View attachment 328116

View attachment 328118

All ram slots are full!

Cpu is an AMD 386. I can't tell much more as there is a sticker over it from the system builder.
Nice work!!! (y)

Thanks!

Well hopefully you didnt find an exploded and leaking varta battery.....pretty common on that era of boards.

Battery looks good to me!

View attachment 328123

Over all I was VERY impressed with how clean the case was inside.

I haven't pulled the IDE or LPT or second IDE card yet. My experience in this generation is minimal, but I haven't a card like the secondary IDE one before. It has the Creative CDROM connected to it and the card output on the back is composite sound.
 
Weird sound card too, never seen that brand before. Was wondering what it was as I had not seen many with the spin knob besides adlib cards.
 
Looks like battery is starting to leak though, see some green fluff on the bottom along with that brown goo.

I am working my way down farther into the cards. One is screwed in properly but definitely not seated nicely.
 
Ouch, looks like someone tried to make it fit. that back bracket is bent really badly
 
Back
Top