Re: DM44 - Droop in Volts reading or eventually lock down to reading 0033 (regardless of range)).

Fabio Trevisan
 

Hello David,

Thanks for your comment.

Although I sort of understood the mechanism how this filter works, I
couldn't find any theory of operation or reference to this topology.

The capacitors seemed original Tek parts, and there were no signs of
retrofit or re-soldering so, I think the failure mode shouldn't have been
temperature.

This takes me to another question... Apart from 2 x 820khoms in series (in
Volts mode, in Ohms mode, only 1 x 820k), I couldn't find anything else
protecting the input of the LD111, or even those 2 capacitors.
In the DM44's manual, Tektronix states that it's safe to connect up to
1200VDC (or Peak AC + DC) at any range...
Those capacitors are rated at 200V, and I don't think the input of the
LD111 can withstand even that... So, how come?

Rgrds,

Fabio











2017-07-04 14:49 GMT-03:00 David @DWH [TekScopes] <
TekScopes@...>:



On Tue, 4 Jul 2017 10:48:49 -0300, you wrote:

Yesterday, after I posted the question to the group, I went back to the
schematics, resolute to find some additional component or connection to
that node of the input of the A/D processor.
It happened that the answer was right there on Diagram <1>, the low pass
filter comprised of C3023 (22n), C3024 (33n), R3023 (2M2) and U3023D (1/4
LM324).

This low pass filter is engaged in all functions and in all ranges, EXCEPT
in OHMS *and* 20Meg range (which perfectly explains why, only when I set
that mode, I could see the right voltage).
This circuit implements a DC accurate low pass filter; the DC value
does not pass through the operational amplifier so offset including
the offset produced by the input bias current through the feedback
resistor is unimportant. Leakage however is critical as you
discovered.

Check out the Linear Technology LTC1062 and application notes 20 and
24 for modern examples of this:

<http://www.linear.com/product/LTC1062>
<http://www.linear.com/docs/4115>
<http://www.linear.com/docs/4119>

Upon returning home, I just lifted, one at a time, C3023 and C3024... and
VoilĂ ! C3024 (33n, PoliStyrene cap, 200V) was the one causing the droop.

I didn't have a PoliStyrene replacement, but did have a PP of exactly the
same value.
Polypropylene is a great replacement for polystyrene. Leakage is what
matters here although dielectric absorption will affect settling time.

Polystyrene capacitors fail at 90C where the polystyrene undergoes a
phase transition. I wonder if that happened here.

After replacing C3024 and confirming that the meter was stable again for
more than an hour (which never happened), I was able to remove the LF411
buffer from the A/D processor input, and it remained working fine and was
relieved to know that the LD111 was not faulty.
That is a great diagnostic use for an LF411. I am glad you found the
problem and it was easy to repair without dealing with the rare LD111.


[Non-text portions of this message have been removed]

Join TekScopes@groups.io to automatically receive all group messages.