The twins paradox

[RobS]

Here's a photo of two capacitors that I removed from two identical eight inch diskette drive power supplies out of a couple of Honeywell Level 6 minicomputers. I was modifying the PSUs to use on my Honeywell 200 project and didn't need these components, but I noticed something about them. They were made by General Electric and Plessey and both have the part number 70930238-013 but one is marked 12500UF and the other 125000MFD, which is ten times the capacity where I come from. Now is this a relativistic effect where one has travelled at great speed, maybe from a galaxy far far away where the manufacturing technology is much more advanced or can we just not trust manufacturers' own labelling? When working on vintage equipment without the benefit of a schematic one has to have some faith in manufacturers' markings, but this has got me a little paranoid. Perhaps a batch were incorrectly marked and Honeywell got them on the cheap or didn't care as they went by the part numbers. Has anyone else encountered this problem? I haven't bothered to check their actual capacities but the truth is out there.

 

[RobS]

Whoops, sorry, not Plessey, Mallory. I didn't have the things themselves to hand when I wrote that. Okay, so anyone can make a mistake, but I don't build mission-critical technology, although my wife would prefer me not to burn the house down. No doubt some people do and have with wrongly labelled components.

 

[Chuck(G)]

Not surprising, really--and either might have worked just as well in a power supply as a filter cap. The blue one's labeling is consistent--a 1253U is definitely a 125K uF cap. Maybe it's a replacement and someone in purchasing simply added one too many zeros to their RFQ.

I replaced all of the AC star/run caps in my heat pump this past summer (a good idea after about 10 years--it'll add life to your compressor). In every case, I had to fashion a new bracket as the replacement caps (same ratings) were far to small to fit the original brackets. Time marches on...

 

[RobS]

Run that by me again. What does 12500UF mean in more common units? Does capital U equate to ten times lower case u then? As it's not that long since I moved from using smaller capacitances rated at HT voltages for valves (tubes to you) I'm not used to all the zeroes on ones used for low voltage technology, so maybe I've never seen this terminology before. Actually it is quite a while since I was working with 400 Volt PSUs but it seems like only yesterday. Cold cathode logic was great fun and very pretty but the tubes tended to stop working in the dark if they didn't have primer discharges. Fancy having a computer that's scared of the dark. Here in England work is being done to restore a very old complete cold cathode computer that uses Dekatrons by the way. They're scouring the planet looking for spares.

Back to the point. I've been working with capacitors in the Honeywell 200 PSUs, which are much larger apparently if these really are 125000 uF. I should have put a ruler in my photo. These capacitors are only 3 inches long. The 90000 uF 20 Volt capacitors in the H200 PSU are huge in comparison, twice as long and twice the diameter and also made by Mallory, but there can only be maybe ten years difference in their dates of manufacture, so electrolytic technology must have bounded forward in between. I know that one has to consider ripple current as well but the current ratings of both PSUs are pretty similar. Perhaps I should check their actual values to resolve this in my mind, otherwise one day I'll get them out and start wondering again. My test equipment thinks these big caps are canned black holes so I just need a resistor and a stopwatch...

 

[Chuck(G)]

We commonly use "u" for Greek "mu" (μ) because of its resemblance to the the Greek letter signifying "1 millionth". No other reason. In the old days, "M" would be used as in "MFD", but the problem is that "M" can also mean 1/1,000 or even 1,000,000 if you're restricting yourself to majescules. Note also, the phonetic abbreviation "MFD" for "microfarad". Old-style stuff still in use today.

Things have changed a lot over the last 30 or so years. You used to see "MMFD" for "pF"; nF was completely unknown and a Farad was an inconceivably large unit (I recall seeing a rack of capacitors used to fire a ruby laser back in the 60s--I was told that it was basically a 1 F capacitor). Now you can get multi-farad surface-mount capacitors.

It's always possible that the blue cap marking is a mistake--a check with a meter should turn that up. If that's so, it's incomprehensible why Honeywell would have allowed those to be installed in the field uncorrected. You'd think that at least a paper sticker correcting the value would have been used.

 

[RobS]

Having checked, the bottom line is that both components are actually 20,000uF, consistent with the specified 12,500uF + 75% tolerance, so Mallory got their labelling wrong. Let's hope that sort of thing hasn't happened too often.

I remember once visiting a chiropractor whose surgery was in his country cottage outside Tunbridge Wells. Wedged into a back room he had an old full-sized clinical x-ray machine which had a mammoth container of capacitors taking up a significant amount of the available space. You don't expect to find equipment like that installed in a cottage in the English coutryside.

Regarding surface-mount capacitors, I found some 1930s-40s style "surface-mount" examples in my spares. In those days "wireless sets" were constructed breadboard style with hand-tightened knurled nuts as connectors instead of solder and the connecting wires had a square cross-section to fit tightly under the nuts. I recollect a reader of an electronics magazine writing to complain that he could no longer find a supplier of square wire. So in a way nothing changes and yet it does as people complain about the demise of square wire, punched cards, floppy disks ...

 

[Chuck(G)]

In old WWII-era military equipment you'd also find "bathtub" capacitors mounted to the chassis. Very often they were multiple-section.

Also, high-voltage mica caps similar to what you have pictured above often also came with molded mounting holes.

Of course, by "surface mount", I meant a component that was held to a PCB using solder only.

 

[RobS]

No need to go back to WWII for bathtubs. Here's one I found in 1970s Honeywell computer equipment.
You're right about multiple sections. When I was a kid I dismantled a very old radio and found that all the foil/paper capacitors in it were in a single bathtub, sealed in with pitch. It was very economical packaging.

Of course I was joking about surface mount devices. It's just that as my eyesight gets worse they keep making the components smaller, which seems extremely unfair. That's a good reason for me to concentrate on rebuilding 1960s technology. At least I can see it clearly, well most of it.

 

[Chuck(G)]

My eyesight's terrible and always has been so. Even when I read the newspaper, I'm using one of these:

I've several of the things as well as a stereo zoom inspection microscope. My SMT skills are as good as anyone's. My hands, however, are less steady than they once were...

 

[rorypoole]

We commonly use "u" for Greek "mu" (μ) because of its resemblance to the the Greek letter signifying "1 millionth". No other reason. In the old days, "M" would be used as in "MFD", but the problem is that "M" can also mean 1/1,000 or even 1,000,000 if you're restricting yourself to majescules. Note also, the phonetic abbreviation "MFD" for "microfarad". Old-style stuff still in use today.

Things have changed a lot over the last 30 or so years. You used to see "MMFD" for "pF"; nF was completely unknown and a Farad was an inconceivably large unit (I recall seeing a rack of capacitors used to fire a ruby laser back in the 60s--I was told that it was basically a 1 F capacitor). Now you can get multi-farad surface-mount capacitors.

It's always possible that the blue cap marking is a mistake--a check with a meter should turn that up. If that's so, it's incomprehensible why Honeywell would have allowed those to be installed in the field uncorrected. You'd think that at least a paper sticker correcting the value would have been used.

maybe it had a paper sticker, and it fell off?

 

[RobS]

maybe it had a paper sticker, and it fell off?

No chance. The capacitor and the rest of the PSU were inside a completely enclosing steel box with only grilled vents for the fan and air intake. I was probably the first person to open the box since it was manufactured. The box was probably needed to shield the adjacent floppy disk drives from the EM field leaking out of the constant voltage transformer.

 

[Chuck(G)]

So some guy in Purchasing probably got a heckuva deal on the mislabeled jobs and it was determined that the mis-marking didn't matter to the vendor. Rather than scrap the whole run of caps, it's better to find a customer for them, even if you take a shellacking on the price.