[time-nuts] 5370B Input Board Repair [WAS: 5370B Question / help needed]
th.allgeier at gmail.com
Tue Jul 18 04:41:25 EDT 2017
Hello All Again,
I’ve got my 5370B going now and in the process made a “discovery” which I thought might be worth sharing:
The A3 input board is a through-hole PCB with a few SMD capacitors and resistors on the reverse of the “switch area”. It turns out that on my 5370 (2410A00777) these components are not soldered, but fitted with conductive adhesive. I first thought it was solder with a black coating but under a microscope it is clear that it is not solder at all. Most probably it is a mixture of epoxy and silver particles, or a similar compound. So no going over joints with a fine iron…
Inspecting all this carefully under the microscope I discovered that the “joints” on 2 resistors (R23 and R56) had cracked. As you know this board gets heat from the hybrid amplifier IC’s and due to the way the board is mounted to the front panel I guess it sees thermal stressing when the instrument warms up and cools down. While this obviously lasts a long time it looks that on my unit the adhesive has eventually cracked in places. (Vigorous switch activation and pressing / pulling on the switch handles also won’t be helpful in this respect…) One of the resistors just fell off at the slightest touch with fine tweezers.
Anyhow, after removing the offending components, cleaning the pads of the adhesive, and soldering replacements in place, we have a perfect 99.9x ns with the 10 MHz on the commoned inputs. Happy days!
So if any of your 5370’s have the kind of intermittent fault I described (and one or two other people seem to have reported) or instability that seems to originate from the A3 board – check the joints around the SMD’s.
I wonder why / how it ended up having the adhesive instead of solder – were earlier / later instruments the same, or was this a build change introduced at a certain period?
Hope the above is of help to somebody else,
Date: Sat, 15 Jul 2017 21:54:47 +0100
From: Thomas Allgeier <th.allgeier at gmail.com>
To: Discussion of precise time and frequency measurement
<time-nuts at febo.com>
Subject: Re: [time-nuts] 5370B Question / help needed
<CAOT1cQLOjZ0=Xs_+nQ+MzAODboZG5ar0f8G_PydxJ=Oj6grB8g at mail.gmail.com>
Content-Type: text/plain; charset="UTF-8"
I'm coming back to this after a few days of prodding etc. and there has
been some progress. The fault has been tracked down to the A3 input board
and it points at a crack in the board or a joint. Here is the sketch:
After a good clean of the switches and re-seating all boards etc. the 5370
sprung back into life with almost-in-spec performance of 102.xx ns with
it's own 10 MHz. Trouble is it didn't last - after a while and triggered by
sliding some switches around it went back to the 14 ns. Fiddling with the
BNC's may have the same effect, i.e. there is a mechanical element to it
Turns out that running it with the front panel removed I can make it go
from 14 to 102 ns by slightly bending the A3 board, certainly while it is
cold. After a while this trick doesn't work anymore, my suspicion is
whatever crack/gap is causing the trouble has expanded too far to close it
tight. After cool-down we're back to square one.
I notice there are a few SMD components on that board, right in the middle
where it would bend most - basically on the reverse side of the switches.
Capacitors and resistors I expect. I wonder if it is worth going over the
solder joints of all these carefully.
What has stopped me so far are 2 questions:
Am I the only one with this observation or has anybody come across this
Secondly the solder on these SMD's is coated with a black substance.
Clearly this could be removed somehow but it is probably there for a
reason. Has anybody re-touched joints like these before?
On an instrument like this it is very much a case of "proceed with care"
and I'd hate to do more damage by rushing in.
On a related subject: last year there was a discussion over redesigning the
input board(s) for the 5370 / 5345. Did this get off the ground? If my A3
packs up completely...
Thanks in advance for any suggestions,
More information about the time-nuts