> From: Tothwolf
> The conductivity of tin/lead solder is simply higher than that of the
> wire and terminal alone
I think you meant 'resistance', not "conductivity", right?
> Can you tell what size the wire is?
Sorry, I'm not good at telling those larger sizes apart. I looked on the
insulation to see if it said, and all it gave was some manufacturer data, and
a 600V rating.
> Maybe they used too large of terminal for the wire and the crimped
> connection is bad?
I can verify that the opening in the terminal was 'full' of wire - i.e. wire
and terminal were size-matched.
I honestly don't think that it was the wire-terminal connection that caused
the problem. For one, the insulation on the wire right up at the terminal
shows no damage (it's still nice and flexible, I can put thumbnail prints in
it - unlike the insulation on the terminal, which was extremely brittle (I
wound up cutting the insulation off the terminal with a Dremel to check out
the crimp - replaced with heat-shrink), and if the wire-terminal connection
had been bad, I'd have expected the wire there to get pretty hot. Also, when I
pulled the terminal off the lug, melted solder from the terminal-PCB
connection had run down onto the lug, and you can see the shape of the
terminal in it. So either i) they soldered the lug onto the PCB after the
terminal had been placed on the lug (very unlikely), or the solder melted from
the overheat. (Some overheat!) Would that much heat from a poor wire-terminal
joint have made it to the PCB? Anyway, given both of these, I suspect it was
either the terminal-tab connection, or the tab-PCB connection.
>> Is that terminal a dual crimped type with a separate insulation crimp?
> If you look into the end of the connector, there would be a secondary
> crimp grabbing the wire's insulation.
I _think_ it's double crimped, but it's not 100.00% definite. Most of the other
terminals, all the smaller ones, are very definitely double crimped. The handful
of other large ones (from the full-wave bridge to the heat-sink/bus-bar) look
a lot like this one - strong indications of a double crimp, but not absolutely
positive.
> At the very least, the pc board mount tab should be resoldered.
Yep, did that - although I'm a tiny bit worried that I got a cold solder
joint on the pins, the solder didn't really flow up them the way it should.
There may be burned gunk (probably from the old rosin) on them; I may remove
the existing solder, and take corrective action.
(What I'm going to do first is turn the thing on with a bunch of Minimum Load
Modules in it, and run it for a while, and monitor the temperature of the
connection.)
Noel
PS: I just realized that perhaps the spade lug connector was under-specified,
and that's why it was getting too hot. So perhaps another alternative is like
the first (clean and put back together), only this time, solder the two
together; my intuition says that would increase the amount of amps it could
carry - or am I totally confused there?
Noel
First, thanks to everyone for the help; this is really a puzzler for me, and I
really appreciate the assistance. Anyway, on to substance (several replies all
packaged into one to minimize list traffic - and sorry the result is a bit of
a tome):
> From: Tothwolf
> What caused the excessive current draw though? Is there a marginal part
> on the pc board? You shouldn't be seeing that much current draw though
> a pc board.
I realized after I posted is that your first question is the $64K question -
and I don't know the answer. I suppose it could (broadly) be one of four
things:
- The previous owner tried to draw too much power (although the boards
that were plugged in when I bought it shouldn't have done that - I haven't
sat down to calculate the total draw, and compare it with what the PS can
supply, but the card cage was less than half full of stock items)
- There's a fault (e.g. partial short), or faulty component, somewhere
in the power supply
- The connector is under-specified for the current it could be/was asked
to carry
- There's a poor connection somewhere in the connector system, further
dividable into:
-- The wire<->terminal joint
-- The terminal<->lug joint
-- The lug<->PCB joint
I agree that without knowing the original cause, and correcting it, the
problem could recur - but I don't think I can (at this remove) work out which
one it was. So I can take some corrective action(s), but I don't think I can
be sure I got the orignal cause...
> The first question that comes to mind for me, is the terminal even
> original to the chassis or did someone replace it at some point in the
> field?
Well, I'd have said 'original' (it has the look/feel of original work), but
looking closely, I cannot be sure. It's the only connector of that type in the
PS - the other +5V lines are i) soldered to the tips of the rectifiers in the
full-wave bridge, and ii) use crimped bolt-on type terminals (i.e. not the
push-on tabs) for the rest. Those latter do have the same colour and general
look as this one, _but_ the crimping marks on their insulators are more
distinct. (Perhaps the heat removed them on this one?)
If it was a poor wire<->terminal connection, could that really have melted
the solder on the lug<->PCB joint?
> Is that terminal a dual crimped type with a separate insulation crimp?
Sorry, how would I tell?
> The .250 tab type terminal on the board is likely made by either AMP or
> Molex and would probably be pretty easy to source as a replacement part.
I read this to mean you think I should replace the soldered on tab (presumably
as well as the crimped terminal on the wire)?
> Done correctly, using the right compression dies and tools, crimped
> connections are always far more secure and reliable both mechanically
> and electrically
Alas, while I have a fairly well-equipped shop, it doesn't run to the
right compression dies and tools...
> From: Chuck Guzis
> You may want to consider a standard PCB screw lug
Well, I'd have to modify the PCB, which I'm loathe to do. The current tab uses
two round pins to make the connection to the PCB, which I suppose I could cut
into a slot, but there are other components in the immediate vicinity (seems
odd for such a high-current connection, but...)
> From: Paul Koning
> Solder helps keep things mechanically sound, but for high current
> connections a well executed crimp is probably better.
Really? I thought that while AC was carried in the skin of a conductor, DC was
carried spread evenly throughout, so I assume(d) that the wider cross-section
of a soldered connection would allow it to carry more current.
Perhaps the industries that stopped using soldered connections in favour of
crimped did so because with hand-soldering, you can get variation (and also
of course bad solder joints), whereas with a crimp it's pretty uniform from
joint to joint?
> What I would do is pull the old lug off, clean off the wire, and crimp
> on a new lug
Ah, to make sure I understand, you're talking about replacing the terminal
attached to the wire, right? (I tend to think of the 'lug' as the
spade-shaped thing soldered onto the board.)
> with a good quality (commercial grade, not Radio Shack) ratchet crimp
> tool.
And if I don't have such a tool... ?
> If it's done right it should be on very tight; if you can pull it off
> without serious force it's either the wrong size lug or the wrong tool.
You're talking about the wire into the terminal here, right? Yes, I've
always made sure a crimped wire _cannot_ be pulled loose (whatever size
connector).
FWIW, the terminal onto the lug is a pretty tight fit, too - I can wiggle it
with my fingers, but to get it off I'm going to need some mechanical help.
Noel
I received this email from some TV people - maybe you can help out:
----- Forwarded Message -----
>From: Barbara Heller <hellerb6 at gmail.com>
>Sent: Monday, September 15, 2014 11:02 AM
>Subject: Computers for TV Show
>
>
>Hi Steve,
>
>I'm hoping you can help me. I'm working on season three of "The Americans" and we need three or four visual 1050 computers.
>
>They don't have to work but they do need to be in good shape.
>
>I'm hoping you happen to have them or that you can suggest a source!
>
>Many thanks and I look forward to hearing from you.
>
>Best,
>
>Barbara
>
>Barbara Heller
>Set Dec Shopper
>The Americans, Season 3
>TVM Productions
>210 Douglass St.
>Brooklyn, NY 11217
>c: 917-402-9884
>o: 718-650-6417
>f: 718-855-7898
>
>
>
>
>
I've been trying to get my IBM 3174 establishment controller (to which my IBM 3179 terminals are connected) communicating with the Hercules emulator over TCP/IP. Hercules is an IBM 370/390/z-Series emulator.
I wanted a more immersive feeling of actually using a mainframe than using TN3270 emulators. They just didn't CUT it (little pun there).
After much grumbling and pulling out what little hair I have left, I finally managed to get everything working as evidenced here: http://www.shiresoft.com/new-shop/Shiresoft/Mainframe_Fun.html
The real challenge was getting the 3174 configured as there are literally 100s of configuration questions that need to be answered correctly. The configuration manual for the 3174 is 800+ pages and the questions from the configuration utility are of the form Q110: ___. You *really* need to read and understand how the 3174 is set up in order to do this right (along with filling out what seemed like 100 worksheets).
The big breakthrough came when I figured out that the 3174 was complaining that I didn't have enough memory for the configuration I was specifying. It seemed odd, since I had 4MB of memory...the max on most 3174s (some could go as high as 6MB). After finally cutting the configuration down to the absolute bare minimum I was finally able to get the "Connection Menu" to show up which allowed me to select from the configured hosts.
For those paying attention, the 399 on the 3174 status display is indicating that the SDLC link is not connected. I don't particularly care, so long as the ethernet and TCP/IP work. ;-)
If there are others on this list that are trying to do the same thing and haven't made progress, let me know...the details are still fresh in my head. ;-)
TTFN - Guy
> From: Sean Caron
> So you've got a spade lug terminating a line from a power supply, onto
> what I assume is a Q-bus backplane.
Not quite. The wire is from the main transformer (I'm pretty sure it's the
+5V center-tap), to the power supply board.
> Honestly it looks in pretty good shape to me .. you are seeing some
> discoloration on the insulating outer plastic on the spade lug.
Ah, there's also crusty black gunk on the spade lug, where the connector
slides onto it; and on the PCB around the outer perimeter of the solder bead
at the base of the lug. See the picture again:
http://ana-3.lcs.mit.edu/~jnc/jpg/PDP11s/SigmaBurnedConnector.jpg
Maybe this is just a sign of a connector running hot, but that sure looks
like it was running a _lot_ hotter than such a connector should.
> I don't see any corrosion in there, or major discoloration on the PCB
> which would indicate very high heat at that point
There is a brown spot on the PCB, but I can't tell if it's surface deposit, or
goes deeper; from the other side, I do see a small area of slight
discoloration.
And I've just noticed that the solder job on the second of the two through-
board pins is crappy - maybe that's the source of the problem?
> If .. you have sufficient slack in the wire to work with
I don't, alas (see previous message).
> If you are really concerned about more power being drawn through
> whatever rail that is, than it is really rated for, you should really
> be focused on the supply wire and not the lug!
Right, but the wire looks fine - the signs of over-heating (and therefore the
voltage drop) is only right at that connector.
> Is it not running? If that is the worst you can find in the PSU, I'd
> bet it would run fine if you fired it up.
Yes, that's the only thing I saw.
The seller had powered it on, and the 'lights came on', so I'm pretty sure it
is working, but...
> heat the surface not the solder
I've found that when working with large things and a small iron, you may not
get good enough heat transfer with the bare iron to heat the item up enough
(no matter how hard you press). So sometimes I use a small bead of solder to
get better heat transfer from the iron to the piece. But yeah, the piece needs
to get hot enough to melt the solder itself.
Ah, well, off to try a few things. Thanks everyone!
Noel
A few weeks back somebody said that if I ever got my hand-drawn
reverse-engineered DEC rainbow PSU schematic scanned they would like a
copy. I've now managed to do this,. but I've lost their details.
If you get in touch mwith me (private email) I will see what can be done.
-tony
On Fri, Sep 12, 2014 at 10:58 AM, Paul Koning <paulkoning at comcast.net> wrote:
> That reminds me of an article, perhaps in the late 1960s, in QST by John Troster, w6ISQ, author of all manner of joke articles. The title was something along the lines of ?Murphy?s laws of electronics?.
> One of them: ?A transistor, protected by a fast acting fuse, will protect the fuse by blowing first?.
> Another one: ?A dropped tool will land where it does the most expensive damage ? this is know as the Law of Selective Gravitation?. (That came with a cartoon showing a hammer nudged off a shelf, on a direct path to an expensive vacuum tube in an open chassis on the bench.)
Amplifiers oscillate, and oscillators don't.
(Software equivalent: Constants aren't and variables won't.)
On Fri, Sep 12, 2014 at 4:52 PM, Tony Duell <ard at p850ug1.demon.co.uk> wrote:
> My point is that this does not go far enough,. If you protect a transsito
> with an expensive FF fuse, then not only will the transistor fail first,
> it will do so i na wat to blow the fuse. It will not protect the fuse.
> That owuld be contrary to Muprhy's Law...
Not necessarily. There's a meta-law "If Murphy's Law can go wrong, it will."
That has also been stated as the Ultimate Law: "If something could
have gone wrong, but didn't, it would have ultimately been better if
it had."
