Re: A stable and repeatable low value capacitance meter
thanks for all answers! I didn't know that the name of those clips was "kelvin".
BTW i checked digikey for them and was shocked by their prices! a set of 2 clips cabled to banana plugs was about 200€!!!
Claudio
|
Re: Open source impedance analyzer
Please put all your data.?
I have an HP 4284A Precision LCR meter. It hasn¡¯t been calibrated for about 4-5 years, but when it was calibrated it was done by Keysight rather than a Mickey Mouse lab. It came back from Keysight and hasn¡¯t really been moved much since it was calibrated. If you wanted something measured, I could do it with that if you wanted.?
The problem with the 4284A, and I suspect with your HP too, is that the uncertainty rises quite a bit when one is outside a modest impedance range. A 0.05% accuracy sounds good, but in practice then uncertainty can be a couple of hundred times higher.?
Hi all, It seems this audience might be interested in an impedance analyzer that I designed/built that works from DC-100kHz with 0.1% accuracy over much of the impedance range. The motivation for this was needing to get precise impedance measurements at arbitrary frequencies without spending a lot. I was also interested in learning more about impedance measurements and circuit design.? Some key details:
- Link:
- Auto-balancing bridge front-end circuit
- Data acquisition/digitization I use a NI-DAQ 6363. You could use other hardware.
- Software written in Labview
- TDMS file saving
- It can measure impedances from 10mOhms to 10MOhms
- Test at frequencies from DC-100kHz (up to 250 kHz still needs validation)
- Perform frequency sweeps on log or linear spacing
- If a battery-powered computer and DAQ is used, the entire measurement apparatus can be both portable and floating.
- BOM cost ~$50-$100 (depends which INAs, BNCs, etc. you use). Not including NI-DAQ.?
- Four-wire (Kelvin) measurement
- Impedance range selectable by jumpers
Board photo?  Example screen shot from GUI: Frequency sweep  Single frequency impedance statistics/time trace.  Accuracy data (1kHz, resistors) See ReadMe for more details.?  I'm avoiding putting all of the measurement data here to avoid clutter but happy to add more if there is interest. At higher frequencies (toward 100kHz) accuracy is closer to 1%, especially at loads much above or below the shunt resistor value. Also, very reactive loads (e.g. high Q capacitors) have closer to 1% accuracy. One interesting use case is looking at the impedance of a solenoid (~1 mm OD wire) wrapped on a copper tube (~50 mm OD, ~1.6mm wall), and compare it to a Litz wire air core toroid.? First the Litz wire:  As anticipated, Litz has constant resistance, and air core toroid has constant inductance.? Now you can see how profound of an effect the tube has increasing the real(Z) and decreasing L, as well as the skin depth increasing AC resistance.  My interpretation of this:
- Below 50 Hz or so, it is essentially a resistor, the inductive impedance is very small.?
- At 50 Hz - 1 kHz, the solenoid's impedance has an increasing real component and decreasing inductance (decreasing slope of imag.). This occurs as the inductor couples more strongly into the copper tube. The inductance decreases because the magnetic field in the center of the tube is decreasing as the wall effectively shields the interior, and inductance is a measure of magnetic field energy stored.?
- Between 1 kHz and 10 kHz the resistance nearly plateaus and the inductance starts rising proportionally to frequency. Presumably at this point, the skin depth of the copper tube has been met, so the magnetic fields are not significantly penetrating to the center of the tube.?
- Above 10 kHz the skin depth of the wire is being reached and the AC-resistance is increasing as the effective copper area for current to flow in decreases.
There still a lot of room to improve the system, but it is quite effective for day-to-day use in a lab where I do a lot of filter design/building. Going forward I'd like to test it on calibrated loads (R,L, and C), but I don't have access to a wide range of 0.1% components, so it is a bit of a challenge to calibrate the system. Currently I use an Agilent 4263B LCR meter at 1kHz, 10kHz, and 100kHz. -EM
-- Dr. David Kirkby, Kirkby Microwave Ltd, drkirkby@...Telephone 01621-680100./ +44 1621 680100 Registered in England & Wales, company number 08914892. Registered office: Stokes Hall Lodge, Burnham Rd, Althorne, Chelmsford, Essex, CM3 6DT, United Kingdom
|
Open source impedance analyzer
Hi all, It seems this audience might be interested in an impedance analyzer that I designed/built that works from DC-100kHz with 0.1% accuracy over much of the impedance range. The motivation for this was needing to get precise impedance measurements at arbitrary frequencies without spending a lot. I was also interested in learning more about impedance measurements and circuit design.? Some key details:
- Link:
- Auto-balancing bridge front-end circuit
- Data acquisition/digitization I use a NI-DAQ 6363. You could use other hardware.
- Software written in Labview
- TDMS file saving
- It can measure impedances from 10mOhms to 10MOhms
- Test at frequencies from DC-100kHz (up to 250 kHz still needs validation)
- Perform frequency sweeps on log or linear spacing
- If a battery-powered computer and DAQ is used, the entire measurement apparatus can be both portable and floating.
- BOM cost ~$50-$100 (depends which INAs, BNCs, etc. you use). Not including NI-DAQ.?
- Four-wire (Kelvin) measurement
- Impedance range selectable by jumpers
Board photo?  Example screen shot from GUI: Frequency sweep  Single frequency impedance statistics/time trace.  Accuracy data (1kHz, resistors) See ReadMe for more details.?  I'm avoiding putting all of the measurement data here to avoid clutter but happy to add more if there is interest. At higher frequencies (toward 100kHz) accuracy is closer to 1%, especially at loads much above or below the shunt resistor value. Also, very reactive loads (e.g. high Q capacitors) have closer to 1% accuracy. One interesting use case is looking at the impedance of a solenoid (~1 mm OD wire) wrapped on a copper tube (~50 mm OD, ~1.6mm wall), and compare it to a Litz wire air core toroid.? First the Litz wire:  As anticipated, Litz has constant resistance, and air core toroid has constant inductance.? Now you can see how profound of an effect the tube has increasing the real(Z) and decreasing L, as well as the skin depth increasing AC resistance.  My interpretation of this:
- Below 50 Hz or so, it is essentially a resistor, the inductive impedance is very small.?
- At 50 Hz - 1 kHz, the solenoid's impedance has an increasing real component and decreasing inductance (decreasing slope of imag.). This occurs as the inductor couples more strongly into the copper tube. The inductance decreases because the magnetic field in the center of the tube is decreasing as the wall effectively shields the interior, and inductance is a measure of magnetic field energy stored.?
- Between 1 kHz and 10 kHz the resistance nearly plateaus and the inductance starts rising proportionally to frequency. Presumably at this point, the skin depth of the copper tube has been met, so the magnetic fields are not significantly penetrating to the center of the tube.?
- Above 10 kHz the skin depth of the wire is being reached and the AC-resistance is increasing as the effective copper area for current to flow in decreases.
There still a lot of room to improve the system, but it is quite effective for day-to-day use in a lab where I do a lot of filter design/building. Going forward I'd like to test it on calibrated loads (R,L, and C), but I don't have access to a wide range of 0.1% components, so it is a bit of a challenge to calibrate the system. Currently I use an Agilent 4263B LCR meter at 1kHz, 10kHz, and 100kHz. -EM
|
Re: A stable and repeatable low value capacitance meter
Not finding an online source for
the "Hutale 4.2 LCR-DDS-F meter".
Where does one purchase it?
Ray, W4BYG
"...It appears to assume ideal
components - it does not give you series R or parallel R. The
test frequency depends upon the component, but as Stephen
Menasian pointed out, the test frequency is important. That's
not something I would personally build, even if I did not have
professional LCR meters. I think I would rather buy
and spend the time sticking it in my
own box. "
Dave
--
"If you want to build a strong house, I'll give you my engineer's number.
If you want to build a strong life, I'll introduce you to my carpenter."
Lebron and Heather Lackey
|
Re: A stable and repeatable low value capacitance meter
I have made Kelvin test clips using regular clip leads, just solder the wires as close to the tip as possible, or, at least the test current one. Preferably both.
?????????????????????????????????????????????????????????????????????? Mikek
|
Re: A stable and repeatable low value capacitance meter
On Mon, 5 Sept 2022 at 16:17, MAX < max@...> wrote: I made my own test clips.? You can see how I did it at, ?
? I hope posting this link is not against the rules.?
No, post any links you want, as long.
Dave (group owner)
|
Re: A stable and repeatable low value capacitance meter
I made my own test clips.? You can see how I did it at, ?
? I hope posting this link is not against the rules.? ? 73 (Regards). ? Max K 4 O D S. ? I've Never Lost the Wonder. ?
toggle quoted message
Show quoted text
From: [email protected] [mailto:[email protected]] On Behalf Of Tom Lee
Sent: Monday, September 05, 2022 2:45 AM
To: [email protected]
Subject: Re: [Test Equipment Design & Construction] A stable and repeatable low value capacitance meter? When I googled, among the top several hits was this one:
There is a picture of the Kelvin clips in the doc.
--Cheers,
Tom
-- Prof. Thomas H. Lee Allen Ctr., Rm. 205 350 Jane Stanford Way Stanford University Stanford, CA 94305-4070 On 9/5/2022 00:05, i2NDT Claudio via groups.io wrote: hi Neil,
sorry...but what are those "Kelvin probes from Dr Le Hung"???
I "googled" but couldn't find any plausible answer!
best 73
de
i2NDT Claudio
?
|
Re: A stable and repeatable low value capacitance meter
On Mon, 5 Sept 2022 at 14:30, Jempi ( ON7MA ) < on7ma@...> wrote: Check this out
I will build it in a few weeks, as soon as the Nokia Display comes in
its also L and C meter.. and can work on higher frequencies
Nice jOB and project
JP
It appears to assume ideal components - it does not give you series R or parallel R. The test frequency depends upon the component, but as
Stephen Menasian pointed out, the test frequency is important. That's not something I would personally build, even if I did not have professional LCR meters. I think I would rather buy
and spend the time sticking it in my own box.
Dave
|
Re: A stable and repeatable low value capacitance meter
Check this out
I will build it in a few weeks, as soon as the Nokia Display comes in
its also L and C meter.. and can work on higher frequencies
Nice jOB and project
JP
|
Re: A stable and repeatable low value capacitance meter
For low capacitance values. I am happy with my vintage Tektronix 130. This was developed in the vacuum tube age and was probably originally intended for in-house measurments of CRT and vacuum tube interelectrode capacitances.
When mine developed power supply sickness, I replaced all the vacuum tube electronics with a solid state design. This instrument measures at a frequency in the 125-140 kHz range. With the solid state electronics, the frequency is stable to +/-0.1 Hz (the longest counter gate time I use is 10 S), corresponding to ~+/-0.002 pF.? I use 0.40 Vp-p voltage across the capacitor; this can probably be reduced to 0.20 Vp-p or less. Full scale ranges are 3 through 300? (either pF or uH). By measuring the raw RF frequency, the capacitance range can be extended beyond 1000 pF (inductance seems to fail with my oscillator circuit above ~600 uH; I could probably change a few component values to extend this limit).
It is somewhat tolerant of parallel resitance (75 k), but not as tolerant as the original Tektronix (20k). Warmup drift is <0.002 pF/min after 30 minutes. Total startup drift is <1 pF (first 15 minutes), which is easily zeroed out. It has a guard voltage output, which is very useful in getting rid of stray capacitance effects.
It does have a zeroing control. I consider this an advantage; it allows immediate use upon turn-on. Also, it allows capacitor matching to great accuracy (zero with one capacitor attached and measure the capacitance with the other to get the difference). Lead length is no problem if the instrument has a zeroing capability since one merely zeros out the lead capacitance.
At 0.4V p-p, it seems to work well for semiconductor junctions. I am still working on guarded fixtures for this have not yet done a complete evaluation.
As a final note, one should pay attention to the frequencies used to measure capacitances. A capacitor may measure quite differently at DC or audio frequencies from high kHz or MHz frequencies, where resistive effects may be less but inductive effects more.
Stephen Menasian
|
Re: A stable and repeatable low value capacitance meter
On Mon, 5 Sep 2022 at 09:42, Neil Smith G4DBN < neil@...> wrote:
Hi Claudio, it's the Hutale meter.? I
emailed Dr Le Hung directly to buy the meter.? I don't know if he
still has them for sale.
The video shows the use of the Kelvin
clips.
Neil The webpage says to contact him by email. I don¡¯t think it was any different - I don¡¯t recall a PayPal button for example.?
I wish I knew why the HP/Agilent/Keysight meters need one to enter the lead length. The 1 MHz 4284A will take cables up to 4 m in length (with option 006) and the 30 MHz 4285A up to 2 m. Given one performs an open, short, and optionally load at the end of the cables, I can¡¯t understand why the cable length is important. The dielectric is important too. I have a set of HP extension cables, and made my cables for the LCR meter the same electrical length, as I used a different dielectric, with a slightly different relative permittivity. I thought the electrical length was more likely to be important than the physical length.? -- Dr. David Kirkby, Kirkby Microwave Ltd, drkirkby@...Telephone 01621-680100./ +44 1621 680100 Registered in England & Wales, company number 08914892. Registered office: Stokes Hall Lodge, Burnham Rd, Althorne, Chelmsford, Essex, CM3 6DT, United Kingdom
|
Re: A stable and repeatable low value capacitance meter
On Sun, 4 Sep 2022 at 13:13, Neil Smith G4DBN < neil@...> wrote:
I've had one of those "Hutale" LCR
meters with Kelvin probes from Dr Le Hung for six years or so.? It
works well and gives sensible answers, but really needs to be
fitted into a box.? I broke one of the switches through
carelessness. It only goes up to 10 kHz sine wave for testing
capacitors, so the lead lengths are probably not critical.? I made
a four-wire jig for testing SMD components.
Neil
The latest version goes up to 50 kHz.?
I would suggest that if fitted in case, it would be sensible to fix 4 x female BNCs on the front, spaced 22 mm apart, in the order shown in the attachment, which is a photograph I just took from my Agilent 4285A LCR meter. Then it would fit any of the standard fixtures from HP/Agilent/Keysight or their clones. I attach a few pictures of such fixtures.?
Note the fixtures have male BNCs, with at least two having levers on them. Those BNC connectors can be found on eBay. You will probably need to take the levers off and put them in a slightly different position, because the levers need to be angled such that they are at the 1:30 position on a clock when mated.?
The red and black Kelvin clips can be found on eBay too, but I have found some need a bit of modification as poor manufacturing has resulted in the two halves of the clip being electrically joined. It may be necessary to take them apart to add an insulator.?
I have a couple of laboratory LCR meters (HP/Agilent 4284A and 4285A), but if I didn¡¯t, I would probably buy that one.?
The DE-5000 is a small LCR meter that was sold by IET Labs). There are many clones of that sold. I don¡¯t know if the quality of the clones varies. A friend has a DE-5000 clone. I loads him a few precision components?
1 H inductor? 100 mH inductor Various 0.005 or 0.01% resistors Variable 0-100 pF variable capacitor.
He was very impressed with his DE-5000.?
-- Dr. David Kirkby, Kirkby Microwave Ltd, drkirkby@...Telephone 01621-680100./ +44 1621 680100 Registered in England & Wales, company number 08914892. Registered office: Stokes Hall Lodge, Burnham Rd, Althorne, Chelmsford, Essex, CM3 6DT, United Kingdom
|
Re: A stable and repeatable low value capacitance meter
Hi Claudio, it's the Hutale meter.? I
emailed Dr Le Hung directly to buy the meter.? I don't know if he
still has them for sale.
The video shows the use of the Kelvin
clips.
Neil
On 05/09/2022 08:05, i2NDT Claudio via
groups.io wrote:
toggle quoted message
Show quoted text
hi Neil,
sorry...but what are those "Kelvin probes from Dr Le Hung"???
I "googled" but couldn't find
any plausible answer!
best 73
de
i2NDT Claudio
|
Re: A stable and repeatable low value capacitance meter
When I googled, among the top several hits was this one:
There is a picture of the Kelvin clips in the doc.
--Cheers,
Tom
--
Prof. Thomas H. Lee
Allen Ctr., Rm. 205
350 Jane Stanford Way
Stanford University
Stanford, CA 94305-4070
On 9/5/2022 00:05, i2NDT Claudio via
groups.io wrote:
toggle quoted message
Show quoted text
hi Neil,
sorry...but what are those "Kelvin probes from Dr Le Hung"???
I "googled" but couldn't find
any plausible answer!
best 73
de
i2NDT Claudio
|
Re: A stable and repeatable low value capacitance meter
hi Neil,
sorry...but what are those "Kelvin probes from Dr Le Hung"???
I "googled" but couldn't find any plausible answer!
best 73
de
i2NDT Claudio
|
Re: A stable and repeatable low value capacitance meter
A while ago I compared the performance of a number of units like that from several different sources.? Some of them were very good, some were really bad.? I tracked this down to (mostly) the quality of the reference inductor.? The good ones had inductances that remained fairly constant over the frequencies generated by the unit. The inductance of the bad ones varied significantly so the accuracy became worse as the calibration and? DUT test frequencies diverged (no compensation in the software as far as I could tell).? I suspect they were all "copies" of the same design, but some manufacturers saved a few pennies on the inductor.? Some of the worst were wound on toroids, but from measurement it looked as though the ferrites were designed for EMI suppression rather than inductor applications.
Hang on to your good unit!
PeterS???????? G8EZE
-----Original Message-----
From: Hans Wagemans <on4cdu@...>
Sent: Sun, 4 Sep 2022 11:57
I use this LC meter and am very satisfied with it:
LC meters on Ebay that look the same have a much worse specification.
73, Hans, ON4CDU
_._,_._,_
|
Re: A stable and repeatable low value capacitance meter
I've had one of those "Hutale" LCR
meters with Kelvin probes from Dr Le Hung for six years or so.? It
works well and gives sensible answers, but really needs to be
fitted into a box.? I broke one of the switches through
carelessness. It only goes up to 10 kHz sine wave for testing
capacitors, so the lead lengths are probably not critical.? I made
a four-wire jig for testing SMD components.
Neil
On 04/09/2022 12:59, Dr. David Kirkby,
Kirkby Microwave Ltd wrote:
toggle quoted message
Show quoted text
On Sun, 4 Sep 2022 at 11:57, Hans Wagemans < on4cdu@...> wrote:
I use this LC meter and am very satisfied with it:
LC meters on Ebay that look the same have a much worse specification.
73, Hans, ON4CDU
I don¡¯t have one, but this LCR meter from
Hungary quite impressed me
Note it is 4-terminal, and works on the
principle of an autobalancing bridge, like at least some of
the HP/Agilent/Keysight ones do. My Agilent 4284A (0.05%
basic accuracy) works on that principle.
|
Re: A stable and repeatable low value capacitance meter
On Sun, 4 Sep 2022 at 11:57, Hans Wagemans < on4cdu@...> wrote:
I use this LC meter and am very satisfied with it:
LC meters on Ebay that look the same have a much worse specification.
73, Hans, ON4CDU
I don¡¯t have one, but this LCR meter from Hungary quite impressed me
Note it is 4-terminal, and works on the principle of an autobalancing bridge, like at least some of the HP/Agilent/Keysight ones do. My Agilent 4284A (0.05% basic accuracy) works on that principle.
On my HP 4284A one has to enter the length of the test leads (either 0 m, 1 m, 2 m or 4 m). There¡¯s no facility to enter the test lead length on that meter.?
I don¡¯t know why HP consider the test lead length important, but they obviously do. On the 4284A, option 006, which adds support for 4 m test leads, was an extra cost option. It¡¯s a software only option, but the instrument needs recalibration - I assume to check the 4 m results meet the specification. My meter didn¡¯t originally have that option, so I asked Keysight the cost of upgrading. They said when I sent it for calibration they would add it free of charge. The maximum frequency of the 4284A is 1 MHz (wavelength of 300 m), so even 4 m test leads are very small compared to the wavelength.?
Obviously 2 different imp¨¦dances would present the same impedance at the end of test leads, if those leads were of a different length. But I would have thought that as long as the 4 test leads were of the same length, the 4-wire Kelvin connection would have compensated fully for the length.?
Dave?
-- Dr. David Kirkby, Kirkby Microwave Ltd, drkirkby@...Telephone 01621-680100./ +44 1621 680100 Registered in England & Wales, company number 08914892. Registered office: Stokes Hall Lodge, Burnham Rd, Althorne, Chelmsford, Essex, CM3 6DT, United Kingdom
|
Re: A stable and repeatable low value capacitance meter
I use this LC meter and am very satisfied with it:
LC meters on Ebay that look the same have a much worse specification.
73, Hans, ON4CDU
|
Re: A stable and repeatable low value capacitance meter
John K,
I appended the circuit to show just how simple the meter is even with an added display.? The basic measurement code is in the original article. There are dozens if not hundreds of examples of LCD/OLED interface code for Arduino, not least on the Arduino Forum. ?
I'm afraid this is an old project for me now, I came across the files and thought the principles might be of interest to this group.
PeterS???????? G8EZE
-----Original Message-----
From: John Kolb <jlkolb@...>
Sent: Sat, 3 Sep 2022 22:14
So a meter that doesn't drift and has fast cycle times is desirable, but
one tied to a computer for readout is not useful and my brain is far too
old to learn how to modify Arduino code to add a LCD display to make a
portable unit. If someone does, please publish.
John
|
Dr. David Kirkby, Kirkby Microwave Ltd
Ray, W4BYG
Jempi ( ON7MA )
