I believe it about the thermal design, Chuck. A few years ago I was working with an RF power amplifier that dissipated almost 2 Watts. The package was 4 mm square, IIRC, so we thought it was a lot of heat in a small area and bought several types of heatsinks. Well, I had laid out the PCB such that it had a bunch, maybe 25, of thermal and RF ground vias directly under the package, and I had the fab shop plate the copper to shut the vias. Well, there was so much heat wicked out of the PA by the copper in the board that the difference in temperature at the package on the top side was essentially the same with or without a heatsink on the bottom side. I was surprised how well the PCB operated without any heatsink. Copper is your friend!
toggle quoted messageShow quoted text
------ Original Message ------
From: "Chuck Harris" <email@example.com>
Sent: 2/2/2021 1:37:47 PM
Subject: Re: [TekScopes] 2467 U800 heatsink
U800 is a strange and unique thing in the 2465 scopes.
It is fairly likely that early in the design, when the
IC packages were selected, so the main board could be
laid out, U800's designer wasn't very far along in his
design, and estimated U800 would dissipate a lot more
power than it ultimately did.
The U800 package is made to mount over a centrally
located heat sink that is sandwiched between the IC
and the circuit board. The IC's leads either passed
along side of the heat sink, or passed through holes
in the heat sink. The IC was made to be clamped to
the heat sink using mounting studs, or screws on either
end of the package.
In preproduction testing, they must have found out that the
finalized chip drew less power than expected by quite a lot,
and they decided to leave the heat sink off for the full
production life of the 2465 family. It was never added in
any form for any of the models, at any time.
Because the heat sink tab is the -5V power connection for the
IC, they had to do something to make the electrical connection.
If they had included the under chip heat sink, it would provide
the electrical path, but they didn't. Instead, they punted,
and just stacked a pair of star lock washers over the mounting
studs to lift the package up, clear the extra long leads, and
make the connection.
If the nuts that mount the package are over tightened,
or the star lock washers are omitted, the slightly longer
than usual DIP package leads will force the IC's lead frame
to crack the epoxy package, and allow moisture infiltration..
which is never a good thing. This is probably the cause of
the so called unreliability problems with U800.
The 2465 family is designed to operate at ambient temperatures
as high as 50C. U800 is designed to operate as it is applied
at that temperature as well. The fan must work, the scope
must be clean, and have room around it for air circulation, but
that was the design temperature. There are even holes in the
scope's case bottom right there to feed fresh air to U800.
I can see no reason to put a heat sink on U800, and I have
seen quite a lot of damage caused by poorly applied heat
sinks, like are shown in your picture. They break free,
and rattle around in the scope shorting this, and that.
I suspect that the installer of the heat sink on your scope
read some of the damage reports, and decided to mitigate the
problem on your scope... probably after the fact.
If it was necessary, tektronix would have made an ECO and a
kit available to install one.
Ondrej Pavelka wrote:
Heatsink is glued to the U800, the wire on the top is a secondary safety
measure. It doesn't really touch anything but the question is how much can
extra heatsink help.
I am also hoping for people like Chuck to give me the answer because there
seam to be many contradicting advices.
On Tue, 2 Feb 2021, 21:10 Jean-Paul, <firstname.lastname@example.org> wrote:
Hallo, where are you located?
The fine photos reveal the U800 is precarious as the bolts carrying the
wire lash up are at 87V and the heatsink is apt to fall out or move.
Suggest that you removed the HS carefully, see if the U800 is Tek original
or Maxim and followed other threads on U800. Very deep issues re heat,
heatsinks and failures.
I am sure that Chuck Harris or Dennison can give you better advice