[time-nuts] HP 5370B low frequency modulation

Magnus Danielson magnus at rubidium.dyndns.org
Mon Jul 23 22:36:40 EDT 2007

From: "Didier Juges" <didier at cox.net>
Subject: Re: [time-nuts] HP 5370B low frequency modulation
Date: Mon, 23 Jul 2007 21:12:39 -0500
Message-ID: <000001c7cd98$1c8d3dd0$6601a8c0 at didierhp>

> Brooke,
> Thanks Bruce for the quick response, as always.
> In my business, we typically dissolve the gold by dipping the solder cup in
> a hot tin pot, which is replaced regularly, before soldering the wires. This
> has been blessed by the major defense contractors for military applications.
> Failure to do this typically results in weakening solder joints over time,
> and delayed failures, particularly when there is mechanical and/or thermal
> stress applied to the joint.

Gold will dissolve very quickly into the Sn/Pb solder.

If you have a too high gold content in the joint, it can form one of several
Au/Sn mixtures which is brittle and will crack easilly. We had trouble with
that under our BGAs so when the gold layer was too thick, the whole 400-500
ball BGA could crack loose of the board if you accidentilly hit the rather
flimsy cooling flange. On another board we glued a nut to the top of the BGA
and hanged 50-70 kg of it and while the board was clearly bent the BGA was
still firmly hooked to it. Now that's the difference between a few um of gold
too much or not.

There are other cases where the gold-plating is not an issue.

This has been covered in public sources, so a dig around would give some


More information about the time-nuts mailing list