long term reliablity computer boards

I got curious when I should have been doing some productive work and found some USB to ISA adapters out there. It presumably requires some really oddball drivers, but they seemed to feel it would even work in DOS, as long as you were running it out of Windows.

Per other suggestions, it may be better to find another way to drive the machines, even if it means doing a ton of work on software.

Those are different than the low ESR capacitors used in switchmode power supplies. They use different construction, and some are four separate electrolytics in parallel to reduce the ESL, as well. The higher frequency used in switchmode power supplies puts a lot of ripple current through the electrolytics. When they start to fail, they heat up and can vent. That is why the tops of the aluminum cans are scored. If they can't vent, they explode.

You need a higher than operating voltage to reform an electrolytic, so if you suspect they are failing on a motherboard it is better to replace them. The average cost is 10 to 20 per motherboard if you buy in small quantities.

That is all that most are warranted for, and rarely ar the run at full temperature, so you extend the useful life. Motherboards use 105 C rated electrolytics, but the boards are never run at that temperature, the 85 degree rated electrolytics drop like flies.

Use a grounded vacuum desoldering iron to remove the bad caps, and make sure to clean off any electrolyte that leaked. Inspect the traces near the bad caps for erosion of the copper before soldering in the new caps.

If the motherboard was built with cheap electrolytics, you may not get a year out of it. A lot of Chinese capacitors were made with a low grade electrolyte that causes the foil to be destroyed. It wasn't much better than the salt water used in early homemade electrolytic caps in the early days of radio. There were dire warnings not to use any electrolytics back then, because they would develop high leakage currents and burn up the power transformer.

I have had new computers die in less than a year, due to bad electrolytics. One started randomly rebooting, and the other refused to turn on. Not all systems are the same, used the same, or built with the same grade of parts. Look at the motherboard and power supply in one of those servers and compare it to a cheaper desktop computer. You'll find a better grade of capacitors, and more attention to air flow in the design. You're comparing apples to crab apples.

The original XT motherboards are damn near impossible to kill but there isn't much there to generate heat, either.

Good for you. Now try running those servers in a 120 degree room with no air circulating and see how long they last. Have you never heard of derating components to extend the operating life, or the term MTBF? There is a reason most server rooms are cold. A/C is a lot cheaper than new servers. The same goes for a radio or TV control room. They aren't 60 to 68 degrees to make the operator comfortable.

Try that near a large industrial park sometime, in a building with over 300 SMPS power pieces of equipment. The harmonics and noise on the neutral will cause it to overheat. Decades ago, it was common to use two gauges smaller for the neutral on three phase. When they started having electrical fires n office buildings, they discovered that the harmonics was the cause, and require new systems to have the neutral larger than the three phases.

The whole industrial park and a small subdivion was on the same small substation. The spikes and surges caused the hundreds of UPS to complain constantly.

Unfortunately, at the current required by the CPU, they have no other choice than to put the caps next to the CPU. if they were even two or three inches away, the voltage drop would be excessive. The CPU can consume over 100 watts at 1.8 volts. That is over 50 amps, so the I/R losses have to be kept to a minimum. The older Pentium chips consume 50 watts at 2.8 - 3.3 volts. which is still a lot of current to push through 1 oz. foil on an interior layer.

An ESR meter will let you test them on the board. I use the original Dick Smith Electronics. A newer version is out. This website has the details on both versions:

Depending on the board, sometimes you just shotgun them when they start to act up, and replace all of the critical capacitors. The ones near the CPU socket are part of the switchmode power supply for the CPU, and are the most likely to fail. That supply drops the 5 volt or 3.3 volt supply to the voltage needed by the CPU and is critical to the motherboard's operation. Erratic operation of some daughterboards and servo controllers may be from the same cause. The top two failures in modern electronics is bad caps, and cracked solder joints. My solder work is in orbit, aboard the ISS where I tested and calibrated boards hot a communications system, and I spent four years doing Surface mount rework, most of it under a stereo microscope.

If I can finish the repairs to my shop building I would be able to do the work for the people on the group.

Yeah, I used to have the same problems when I was turning mine off and on daily.

Why, if you are not going to provide adaquate cooling, you deserve what ever you get.

You're not listening. Again, if you don't isolate the computers from the electric motors, you deserve what ever you get.

I'm printing this out and putting inside the computers. Would you mind emailing me (karltownsendembarqmail.com) your contact information to put in there? It looks like I should wait for trouble, may be a while till I need your help.

Karl

I have only seen clone XT motherboards go bad, and I rebuild lots of old XTs when Ihad a used computer business. Even the IBM AT mother boards had a higer failure rate than the clone XT boards whe tey went to the first custom chips, to eliminate the 'glue logic'.

Yawn. Always have to play the dumbass, don't you?

They weren't air conditioned back in the vacuum tube days, but the equipment was designed to work in higher temperatures. It was no fun sitting in a control room with the transmitter a few feet from your side in a room that was well over 100 degrees. Some of them had no herat, either.

One military radio & TV station I worked at had a steam plant on the other end of the building, but all the station got was whatever heat was left in the return lines, and that went to the offices. Some days it was below zero, and you sat next to the transmitter whenever you could, becasue it was closer to 32 degrees.

You have no reading comprehension. "the hundreds of UPS" weren't there for people to trip over.

You're the one with the comprehension problem. This branch of the thread was clearly about leaving computers on 24-7, as opposed to switching them off daily. Who cares if you didn't have UPSs 30 years ago, or didn't properly cool your computer equipment. Try to keep up or abstain from calling someone a dumb ass, while you drift from one topic to the next.

I've had one 40MHZ ISA dos box running almost continuously since about

Karl Townsend wrote:

Sent. :)

On Wed, 29 Apr 2009 12:17:06 -0500, Pete C. wrote: ...

Besides the above items and the disk, power supply, and fan issues others have mentioned, on-board batteries may be an issue. For long-term storage, I suggest removing batteries from the motherboard, and using new batteries when bringing up a stored system. Write down bios settings before taking out the batteries, and perhaps use a save-and-restore-settings program. If the bios itself is in flash memory, make sure you have copies of the bios and bios-loader on durable storage media.

FWIW, reforming is a technique that's performed to aged electrolytic capacitors to help reduce the amount of leakage caused by age (leakage in terms of internal current leakage, as related to insulation breakdown).

Reforming doesn't improve or change any other electrolytic capacitor test parmeters other than internal leakage.

A lot of folks that restore vintage electronic equipment will slowly bring up the line input voltage with a variac, the first time they power up the equipment, to reduce the possibility of damage caused by excess leakage in the capacitors. High current leakage paths in capacitors will often destroy other ciruit components, compounding the restoration costs and effort.

The capacitors in vintage gear are often the size of the paper tubes in the center of a roll of toilet paper, not typically anything to be found in modern low voltage circuits today.

The reforming process for (aged old stock) out of circuit electrolytic capacitors is to apply a very low voltage from a low current source while monitoring the current with a meter. It's very important that the current is limited to a low value, even a current limited power supply of only 10mA will be adequate.

The current limited applied voltage is gradually increased when it's determined that the current has decreased to near zero mA from the previous voltage setting. No electrolytic capacitors have zero leakage at their working voltage, but nearly zero current should be seen at low applied voltages (uA if the display is sensitive enough to indicate below mA current levels).

When a capacitor's current leakage doesn't achieve near zero current before reaching the rated working voltage, it's discarded because reforming won't correct the excessive current leakge condition present in the capacitor.

The applied voltage is only increased to the capacitor's rated working voltage. If the level of internal leakage is determined to be within acceptable limits, the capacitor can be considered safe to be installed for use in a circuit application appropriate for the capacitor's rated working voltage.

Reforming can take up lots of time, and it's not a practical procedure for most repair technicians.

All of Sencore's test equipment that will test components for leakage are clearly labeled: Do not hold the capacitor (device being tested) in your hand while testing.

Who can find a repair technician? 99% of consumer grade crud it's cheaper to toss than to repair - unfortunately.

By the time the machine is out of production, technology has moved on to the next level (three times) and you have to repair it to archive the materials, you can't get the parts. Or the schematics and manuals.

Example: NASA with all the high-resolution video of the 1960's and

Is that anything like "Do not look direcly into the laser with your one remaining eye..." (Pictogram: White cane, red tip.)

Needs a nice gruesome two part pictogram of an explosion, and a stump with a spurting artery, hand laying on table...

-->--

BTW, I have several hundred spare keyboards, most are AT or PS/2

Solid state hard drive prices are dropping rapidly. They are plug in replacements for mechanical hard drives. Standard power connectors & EIDE cables.

I have 360 K, & 1.2 M 5.25", and some 1.44 drives, along with older CD-ROM drives to install your software. I have at least a half dozen 100 MB ZIP drives, and some oddball streaming tape drives.

You can always partition a larger hard drive, within reason.

There are plenty of them online, but all I've seen bad were fans, shorted semiconductors & bad electrolytics. The datasheet of the switching regulator usually gives enough data to figure out the individual design.

I have a pile of working XT, AT & mini tower power sullies.

I need a couple USB keyboards if you have any. The wireless keyboard that went to my Viao flatscreen got screwed by the grandkids and I am using a waterproof industrial keyboard....that might be great for a machine tool....but sucks for an 80 words per minute typist.

Gunner

"Somewhere a True Believer is training to kill you. He is training with minimum food or water,in austere conditions, day and night. The only thing clean on him is his weapon. He doesn't worry about what workout to do--- his rucksack weighs what it weighs, and he runs until the enemy stops chasing him. The True Believer doesn't care 'how hard it is'; he knows he either wins or he dies. He doesn't go home at 1700; he is home. He knows only the 'Cause.' Now, who wants to quit?"

NCOIC of the Special Forces Assessment and Selection Course in a welcome speech to new SF candidates