[time-nuts] Cautionary tale #4711

[Magnus Danielson]

Fellow time-nuts,

I am sure several of you knows from your own experience, that there are 
times when the equipment in the lab should have been protected from you, 
as you posed the largest danger in the house. Well, I must confess that 
I did it again. Open confession should be the punishment I need in 
addition to the failures in the lab.

My lab has been in somewhat of a sleeping beuty sleep for some time. 
Sub-system after sub-system has failed as is to be expected of an old 
castle (Girl Genious reference). Recent events have resulted in an 
upswing of activity, but the combination of tiredness and still running 
the anti-virus program on the protein robot occasionally seen in the lab 
with associated fever prooved to be less than ideal with combination of 
the wish to get things started again. Running out of cables results in 
stranger and stranger optimations. Instead of using the old PSU for 
feeding 5V to the GPS antenna and the associated GPS distribution boxes 
(a HP 58517A and a GPS Network L1/L2 passive with DC-bypass) I used 
another PSU of more general kind, as I could use the cables otherwise 
occupied for other things... all nice and dandy... I have been tracking 
down problems relating to my GPS system, the RG-58 (yes, yes... I 
know... cheap unstable cable but it was at home when I needed it) cable 
to my L1/L2 antenna kept shorting out in the BNC. As I want N-connector 
regardless, I cut that connector off and transplanted the N-connetor 
over from another cable... still a short, but I was able to find it and 
secure it. Fine, now the cable was doing nice and the powered Z-12 liked 
it. However, whenever I used the distribution amp it didn't work. I 
checked the GPS Network L1/L2 passive one but it now had something like 
470 kOhm of resistance between the ports supposed to be DC thru. Load 
that with about 630 Ohm of antenna amplifier and it is no wonder the 5 
VDC isn't doing much good. Luckilly I bought a little stock of thees, 
pulling another one out and it was less than an ohm between those 
points. Fine, so I shift that in. Problem 2 solved. But as I want more 
L1 C/A receivers hooked up I continue to hook in the 58517A and still no 
luck on the thunderbolt while the Z12 happily locks. Then, bypassing the 
58517A proves that it too has failed, but this time it fails to pass L1 
signal but happily feeds the 5 VDC thru so it powers the antenna. Thus, 
problem 3 identified.

When testing all this, I have pulled the old hairy mess down and hooked 
things up again. I am back with my trusted PSU and such. However, the 
temporary PSU left a little hint on what went bad. That is a general 
PSU, it is an old Oltronic which is nice lab gear for its time. The 
voltage dial goes to a tally and has a nice locking-ring attached to it, 
which is nice mechanical lock to keep the intended voltage from 
accidental change. However, being a general PSU it has a voltage range 
slide-button. This puts it in either a lower range, higher current or 
higher range lower current. It was in the propper lower range.

Now go back to problem number 1... shortage in the connector. In certain 
bending forms, that BNC shorted. When feeding from a Thunderbolt, no 
worries... it current limits. The trusted PSU could have a current 
limit, but I haven't checked. The Oltronics also has one... but at a 
rather high current. And I recall hearing and seeing the relay and 
indication as it goes into current limit (constant current) rather than 
constant voltage.

That then burned the active 58517A and the passive L1/L2 splitter. It 
didn't burn the antenna... but I think I was saved by the short... but 
that broke the RF choke coil in the passive L1/L2 splitter. When opening 
up that one the odour finaly escaped a day after the event. No grey 
smoke, but the smell. After that the antenna was safe... but I think 
that maybe after that the range button may have shifted... and burned 
the 58517A...  So, now you know my bad habbits in my own lab.

The L1/L2 splitter seems repairable, I now have the set of SMD inductors 
to just shift in and see if it works well, the electrolytic cap seemed 
just fine, so a short seems reasonably the reason to caused the fault there.

The 58517A may just have a broken voltage regulator, and that is what I 
hope for.

In the meanwhile, two passive splitters and a pair of LNAs let me have a 
Z-12 and Thunderbolt lockup and run. The trusted (sufficiently weak?) 
PSU is being used again and it looks nice.

I hope you crinch in disgust of my bad lab habbits, as I expose them in 
the open like this, but it is the therapy session I obviously badly 
needed. To compensate for that, I reacted to somebody elses bad habits 
of leaving 20 GHz and 50 GHz scope inputs opened instead of terminated. 
I was called over to look at that scope since it didn't work and showed 
an error message. I quickly terminated all unused inputs properly.

Sloppyness in the lab comes at a price. Added confusion and broken gear. 
Don't do it.

I think others can share similar experiences when they think back a 
little...

The tidying up in the lab seems more and more sensible.

Cheers,
Magnus - should protect his gear from himself