AADE L/C Meter drift..

[Robert Baer]

Insert battery, decide to test and find this one works up to at least
2.5uF.
Next day,

Power up
Wait for "ready measure u"
Select Cx
Press Zero (reads 0.0pF)
Let go and it reads -0.06pF (???)
Let it sit, see it drifts upwards;by day end reads 2.82pF.
Power down and re-try..now after pressing zero,it typically reads 2.58pF.
Press Zero (reads 0.0pF), Let go and it reads -0.06pF (???)
Drifts up as before.

WTF???

 

[Owen Roberts]

Mine drifts a bit more then 3 pf in a hour. Still best bang for the
buck I have found..
The repeatability is decent, in the short term, and that is what I
need.

Steve

 

[John Miles, KE5FX]

   Press Zero (reads 0.0pF), Let go and it reads -0.06pF (???)

   WTF???

If I had sold it to you, I'd apologize profusely, refund your money
immediately, and send you a prepaid shipping box with which to return
it. Then I'd sign you up for my competition's mailing list. :-|

-- john

 

[Robert Baer]

If I had sold it to you, I'd apologize profusely, refund your money
immediately, and send you a prepaid shipping box with which to return
it. Then I'd sign you up for my competition's mailing list. :-|

-- john

Usually it takes 2 zeroing attempts to get zero.
The drift, while apparently seen by others, is irksome and begs for
explanation as to why it exists.
To paraphrase the saying, just because 20 million Frenchmen see this
problem, does not make it right.
Hell, even the first Fairchild uA709 op amps did not drift like that..

 

[Jeroen Belleman]

Usually it takes 2 zeroing attempts to get zero.
The drift, while apparently seen by others, is irksome and begs for
explanation as to why it exists.
To paraphrase the saying, just because 20 million Frenchmen see this
problem, does not make it right.
Hell, even the first Fairchild uA709 op amps did not drift like that..

Apples and oranges. Do you realize how tiny 0.06pF is?

It's the capacitance of a fraction of a millimeter of
twisted pair wire. It's the capacitance of less than
one millimeter of (narrow!) PCB trace over a ground plane.
It's the capacitance between the open contacts of a small
low-capacitance reed relay. It's less than one tenth of
the gate capacitance of the smallest, lowest capacitance
discrete switching MOSFET I know of (SD214).

It's... well, I suppose you get the picture by now.

Jeroen Belleman

 

[Fred Bartoli]

Jeroen Belleman a écrit :

Apples and oranges. Do you realize how tiny 0.06pF is?

It's the capacitance of a fraction of a millimeter of
twisted pair wire. It's the capacitance of less than
one millimeter of (narrow!) PCB trace over a ground plane.
It's the capacitance between the open contacts of a small
low-capacitance reed relay. It's less than one tenth of
the gate capacitance of the smallest, lowest capacitance
discrete switching MOSFET I know of (SD214).

It's... well, I suppose you get the picture by now.

Yeah. You don't know enough MOSFETs

 

[Uwe Hercksen]

Apples and oranges. Do you realize how tiny 0.06pF is?

It's the capacitance of a fraction of a millimeter of
twisted pair wire. It's the capacitance of less than
one millimeter of (narrow!) PCB trace over a ground plane.
It's the capacitance between the open contacts of a small
low-capacitance reed relay. It's less than one tenth of
the gate capacitance of the smallest, lowest capacitance
discrete switching MOSFET I know of (SD214).

It's... well, I suppose you get the picture by now.

Hello,

very well said.

there is another prefix for so small values, 0.06 pF is 60 fF,
Femtofarad. The next one is aF or Attofarad. Both are so small that only
capacitances within an IC require these units.

Bye

 

[Jeroen Belleman]

Jeroen Belleman a écrit :

Yeah. You don't know enough MOSFETs

Please, enlighten me!

Jeroen Belleman

 

[Robert Baer]

There's no range switching in the AADE front-end. Full scale is 1 uF, so 1 pF is
1 part per million. 0.06 pf is 60 parts per billion. The inductor in the LC
oscillator, and the LM311, have got to have serious tempcos. 3 PPM drift in one
day isn't bad!

Incidentally, FR4 has a horrible capacitance TC, in the ballpark of 950 PPM/deg
C, so that must hurt, too. He has a bunch of parts and traces on the critical
nodes.

Check, spacing of red/input terminal is extremely close to a switch
(for starters).
Concerning range switching, the MCU does that by changing
frequency..which seems to put a hole in that analysis.

 

[Fred Bartoli]

Jeroen Belleman a écrit :

Please, enlighten me!

Jeroen Belleman

Sorry, I was just joking...

 

[Robert Baer]

There's no range switching in the AADE front-end. Full scale is 1 uF, so 1 pF is
1 part per million. 0.06 pf is 60 parts per billion. The inductor in the LC
oscillator, and the LM311, have got to have serious tempcos. 3 PPM drift in one
day isn't bad!

Incidentally, FR4 has a horrible capacitance TC, in the ballpark of 950 PPM/deg
C, so that must hurt, too. He has a bunch of parts and traces on the critical
nodes.

Put a scope and Eput meter on it, found it is just an oscillator at
least similar to -a- schematic; range switching is/in software for display.

Q: Obvious why frequency changes WRT capacitor, but why does the
amplitude run from around 500mV pk-pk (if i remember right) at lo
capacitance to around 8mV pk-pk (if i remember right)?
"schematic" indicates i was looking at the top of the parallel LC..

 

[Robert Baer]

The schematic is online on the AADE site. The uP measures the LC oscillator
frequency but can't change it. There is a CALIB relay that adds some test
capacitance, but it's additive, used at powerup. The "range switching" is
software, downstream of the frequency measurement.

Check.