Oscillator coil with common wires?

Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
DUT = Device Under Test, and that is C2, but to make the sim work its internal resistance (ESR) is shown explicitly, and the 47µ value is just a random example.

In practice, you will connect the test terminals instead of the capacitor+resistor (in parallel with the two diodes).

BTW the diodes are just for protection, you can use 1N400x, 1N540x, or similar (or nothing if you like to live dangerously)
 
Last edited:
More streamlining:

The squarewave oscillator can be replaced by a sine one.

It spares one transistor, but more importantly, the cleaner waveform improves the performances (less sensitivity to the inductance of the test leads for example) and also allows some vector-sensitivity: the reactive part of the impedance will have less influence on the reading, improving the accuracy for low capacitance values.

It is not sufficient to make it a true ESR meter, but it is a step in the right direction.

Note that C1 and C4 need to be better than group II ceramic, like X7R and similar.

Plastic film, group I ceramics (not only COG), mica, etc. are suitable. Even mylar is OK
 

Attachments

  • EditedESR1.PNG
    EditedESR1.PNG
    51.2 KB · Views: 52
The 22nF are not critical: you can use ceramics or anything else. Mylar is perfect.

Only the 1.5nF and and 680p of the last version need to be a little better.

Only the two protection diodes have to be large ones.

D5/D6 should ideally be small schottky's: BAT42, 48, 81, 85 etc are suitable, but a small germanium type is also OK (even better): 1N34, OA90, AA119, etc.
1N4148's will work for test, and might even be satisfactory enough, but try to find the types indicated.

Note that the schematic I used is the simulation one. The real circuit certainly includes supply bypass caps, trimmers, etc: you should check the original project thread, but I will try to give some indications too, and I am in the process of simplifying the circuit again.
 
Further rationalization:

It is possible to eliminate one more transistor:

attachment.php


I made a quick and dirty breadboard sanity check:

attachment.php


attachment.php


It works quite well. I didn't have a 50µA meter, so I used a 25µA one with series and parallel resistors to emulate a 50µA/20K instrument, but I had to add a supplementary series resistance in the form of a 47K trimmer.

The movement I used was pulled from an old megger, and by an extraordinary coincidence, the scale happened to be a close match for the new use, with the megohms becoming ohms.

With the mods, the instrument works more or less like the original, but most of the performances are improved, the number of parts is significantly reduced and the current consumption is ~halved.
The temperature stability is somewhat degraded compared to the initial circuit, but for this type of instrument, it is relatively unimportant.

The values are optimized from the reality; the sim optimum is slightly different.
 

Attachments

  • EditedESR2.PNG
    EditedESR2.PNG
    49.8 KB · Views: 152
  • EditedESR3.jpg
    EditedESR3.jpg
    116.5 KB · Views: 208
  • EditedESR4.jpg
    EditedESR4.jpg
    120.8 KB · Views: 121
Phantom, don't hurry too much building your meter: I am going to use my breadboard prototype to test various components (including ordinary 1N4148's as diodes), and see where the calibration trimmer(s) need to be located.

I will also provide a really buildable schematic, not just a replica of the sim.

Please confirm that you are going to use a 50µA meter, because it will have an impact on component values.
 
Here is the final, tested version:

attachment.php


I have tested 1N4148's, and the circuit can be adjusted to accomodate them.
The only small inconvenient is that the left of the scale is more compressed, meaning resistances greater than 5~10 ohm become difficult to read, but that is not really problematic, because such a value already means that the cap has dried up.

I have tested ordinary ceramic caps: the 680p tolerates any type, and for the 1.5nF, only one posed problem: its losses were too high, meaning a reduced oscillator amplitude.
So, it is not too critical, but if you can find multilayers, it is preferable.
The 22nF is absolutely not critical.

To calibrate the meter you can use a new, good quality, large Ecap (1000µ or 2200µ) as a short-circuit reference, with known series resistors added.

First, you should leave the test terminals open, set R9 to its minimum resistance, and progressively increase the value until the meter deviation is reduced almost to zero.
Stop just before the actual zero.
Then, connect the reference cap directly (no series resistance) and adjust R1 for exactly the full scale deviation of the meter.

When that is done, your meter is ready to go, but if you want to have an idea of the scale, or write indications on the dial, you can connect various resistors in series with the reference and observe the deviation.

With 10 x1 ohm resistors connected in parallel, then in series, you can cover the whole 0.1 to 10 ohm range.

I think the original project has details on this procedure.

Note that the calibration/scale will be different if you change the diodes.

The circuit is designed for a 50µA meter having a total series resistance of 20K.
Other values would require modifications.
 

Attachments

  • EditedESR5.PNG
    EditedESR5.PNG
    48.2 KB · Views: 156

PRR

Member
Joined 2003
Paid Member
I'm not really sure how much µA my meter has....

200uA. We know this from the voltage, the (non-standard) resistance, the fact it is rated at 0VU while the scale goes to +3VU. Also it looks like they didn't use the last 10% of the range.

BUT this is probably an AC meter with internal rectifier. While the better plans above use a DC meter plus a rectifier, I'm not 100% sure it can be an exact swap. Maybe the ESR-meter gurus have insight.

Also: a working pro VU meter is worth much more than the type of $2 meter this project really deserves. (Sadly hard to find a $2 these days.)
 
I'm not really sure how much µA my meter has. I took it from an old professional audio console, it says on the front: 0VU = 1.228V = 7500Ω. @ 1000~. I really don't know if you can derive the Ampere value from this information.
And don't worry, I'm not sooo much in a hurry.
As PRR said, it is probably a 200µA AC, which is not suitable: I just tested a 200µA movement, but the circuit is incapable of supplying such a current, and it cannot be increased in a simple way, because of implications and constraints on other parts.
In addition, the internal rectifier could be selenium or similar, which is not suitable for 100kHz operation.

Depending on the internal construction, the actual, naked galvanometer might be more sensitive (because of scaling, half-wave rectification or similar cause).

If it is possible to dismantle it without risking damages (be super-careful!), you could see what is inside, and test the actual DC sensitivity.

Up to 100µA is manageable: I just did the test with a meter having 1K internal resistance and 6.8K series added.
It practically requires schottky or Ge diodes though: with ordinary 1N4148, the calibration remains possible, but the left of the scale becomes so compressed that it is not usable for more than 5 ohm.
Small Ge diodes are very common: all transistor radios used to have at least one.

There is an example of such a diode at the center of this picture:
 

Attachments

  • ESR1.jpg
    ESR1.jpg
    471 KB · Views: 46
The old Midas console had 3 of the original VU-meters and 4 of smalles ones that at some point were installed to preplace the original ones.
On the attached pictures you can see that the replacing ones have a rectifier of 4 1N4148 diodes installed on the back, which can easily be removed.
 

Attachments

  • MIDAS VU's front-min.jpg
    MIDAS VU's front-min.jpg
    396.5 KB · Views: 34
  • MIDAS VU's back-min.jpg
    MIDAS VU's back-min.jpg
    399 KB · Views: 38
You can check the sensitivity of the small one using a fresh 9V battery and a 180K resistor (in series).
You do not even need to remove the diodes.
If it results in a full scale deflection, it is a 50µA.
If it's half => 100µA
1/4 => 200µA

You can try the same manipulation on the larger one: depending on the internal circuit, you could have a deflection -or not-.
Try both polarity: if it has a halfwave rectifier, only one will work
 
Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.