Re: ESR Meter - Roll your own - ESRrev0.JPG

[Ross Herbert]

Yup, it should be 2/4 terminal. Somebody makes some sorta-affordable
Kelvin clips. I'll try to find the source.

John

RG58U or similar small gauge coax makes a pretty good diy kelvin lead.

 

[Ross Herbert]

Yes.

Remembering that any shunting impedances can upset the veracity of any
particular in-circuit reading. When you don't get the result you are
hoping for you have to disconnect the cap under test anyway, just to
be sure.

 

[Winfield]

Mueller does:

http://www.muellerelectric.com/kelvin_clips.html

Note, the big "Check Inventory" buttons on that page
don't work, but the box on the right does.

May I point out both of those clips are horrible?
The first clip has far too weak a spring, not even
strong enough to hold against the pull of its cable,
and the second has ridges that are far too coarse
(and are misaligned), so that it fails to properly
contact small parts. A few years ago I had a Kelvin
cable with clips of a design that worked well, but
it has disappeared.

One solution is to find a good quality alligator
clip, one that works well in various circumstances
(happily one of the very common types works well),
and rebuild its hinge to insulate the two grippers.

 

[Joel Kolstad]

The first clip has far too weak a spring, not even
strong enough to hold against the pull of its cable,

One solution is to find a good quality alligator
clip, one that works well in various circumstances
(happily one of the very common types works well),
and rebuild its hinge to insulate the two grippers.

Hmm... wouldn't replacing the spring on the far-too-weak-spring clips be
easier?

I have one of these: http://www.m3electronix.com/lcr.html ...and it works
pretty well. It's not specifically designed for ESR, though -- it's a
general-purpose impedance meter at a half-dozen or so frequencies.

---Joel

 

[Baron]

Hi guys,

One solution is to find a good quality alligator
clip, one that works well in various circumstances
(happily one of the very common types works well),
and rebuild its hinge to insulate the two grippers.

Excuse me for butting in to this very interesting discussion!

I made some Kelvin clips from a pair of small (30mm) croc clips some
time ago. I filed off the teeth at the front and epoxied a strip of
gold plated pcb into the body at each side of the jaw. The pcb
strips came from an old, dead computer PCI card edge connector. I
found that the width of the gold plated bit a perfect fit inside the
jaw. I also used stereo headphone cable for the leads. Soldering
the screen to the body of the croc clip.

HTH.

 

[john jardine]

Hmm... wouldn't replacing the spring on the far-too-weak-spring clips be
easier?

I have one of these: http://www.m3electronix.com/lcr.html ...and it works
pretty well. It's not specifically designed for ESR, though -- it's a
general-purpose impedance meter at a half-dozen or so frequencies.

---Joel

Thanks for the mention. Nice bit of kit. I've just bought one!.
Past few months I've been drooling over a 500 euro, Chinese LCR but
couldn't honestly justify it work wise. This cheap unit fits the bill.
Next on the list is some kind of PC controllable network analyser

 

[Mark Zenier]

Since lots of single-chip uPs include decent adc's these days, why not
just digitize the voltage drop across the cap? One could then note the
cap charging slope and untangle it from the true esr. And measure
capacitance, for free.

That's what the Bob Parker meter is. A Z8, I think.

Mark Zenier [email protected]
Googleproofaddress(account:mzenier provider:eskimo domain:com)

 

[Mark Zenier]

10 ohms would be a practical upper limit for anything I
would need.


Up to this point, I would be happy with an analog movement
with a logarithmic scale.


Could we say 200 mV peak to peak, open circuit, from a 10
ohm source?

Hmm, just to keep someone from plagarizing it,

<ftp://ftp.eskimo.com/u/m/mzenier/esrmeter.gif>

A 30 kHz "ohmmeter" that I did 8-10 years ago.

Mark Zenier [email protected]
Googleproofaddress(account:mzenier provider:eskimo domain:com)

 

[Spehro Pefhany]

That's what the Bob Parker meter is. A Z8, I think.

Homemade slope ADC.

Hey, John, did you happen to look at applying the AD5933 in this kind
of application? It's not a super good fit perhaps, but a lot of stuff
in 16 pins for the dollars.


Best regards,
Spehro Pefhany

 

[John Larkin]

Homemade slope ADC.

Hey, John, did you happen to look at applying the AD5933 in this kind
of application? It's not a super good fit perhaps, but a lot of stuff
in 16 pins for the dollars.

That certainly gets close. The claimed impedance range is 100r to 10M,
so it would need some help to do esr. But a similar home-made thing
could be done with a cheap FPGA and a cheap ADC and DAC, and some
analog glue... synthesize a variable-freq sine wave and digitize the I
and Q parts of the component response.

A very sophisticated swept-frequency impedance meter could be built
for peanuts. It would just take a lot of engineering, and that mostly
firmware.

We've been making our own DDS products lately, with just an FPGA and
some DACs. It ain't hard any more. Here's one,

http://www.highlandtechnology.com/DSS/V340DS.html

which uses a Spartan3 and four dual 125 MHz, 14-bit dacs. The waveform
tables are in ram, so later on we can do arbs, too. Even just a uP and
a dac can replace an analog oscillator these days, and do lots more
tricks.

John

 

[Winfield Hill]

Hmm... wouldn't replacing the spring on the far-too-weak-
spring clips be easier?

It's a bit on the weak side, but the real problem is that
it's got a smooth flat surface that can't grip and slides
right off. Making the spring stronger simply seems to
make it "spring" off. Note, this clip is used on one of
the very expensive HP Agilent four-coax probe sets.

I have one of these:http://www.m3electronix.com/lcr.html...and it works
pretty well. It's not specifically designed for ESR, though -- it's a
general-purpose impedance meter at a half-dozen or so frequencies.

Are you referring to their Kelvin clip, what's the link?

 

[Joel Kolstad]

Hi Win,

Are you referring to their Kelvin clip, what's the link?

Well, I meant the whole sheband -- the meter and the clips. Here's the page
with a picture of the clips on them: http://www.m3electronix.com/kitlcr.html

 

[John Larkin]

Done something similar some time ago, to save some flash space for bitsream
on a relatively small micro.

Check this thread on this newsgroup:

http://tinyurl.com/2xkt52

or search with google groups for "compressing Xilinx bitstreams, some test
data".

There are some measures from my real design, and a link to the source code.

Please share your results with the group!

OK, we have a company-standard rom-builder program that gobbles up
S28-format uP code and Xilinx .RBT files and builds binary rom images,
sort of a barbaric linker. I modified it to accept a /RLL switch, to
do essentially the same thing as your code (but recoded it in
PowerBasic!) It's byte oriented, and whenever I see a 0 byte, I output
it and count the total number of 0's in the run, then poke that count
into the next output byte. A single input 0 then becomes two bytes
(0,1) and a run of 100 zeroes becomes two bytes, (0,100). Nonzero
bytes just get shipped out unchanged.

I added 32 extra zero bits at the end (0,4) because the chips seem to
like that, and then (0,0) as the end-of-file marker, to tell the uP
config code when to quit. We also added a /FILL switch to load unused
image bytes with all 1's, to make eprom programming a bit faster.

A relatively simple VME interface chip, an XC3S200, crunches to 0.23
of the original size, over 4:1 compression.

A fairly hairy 8-channel massively pipelined DDS synthesizer in an
XC3S1500, using lots of resources, compresses to 0.43. If we
initialize the block rams, it's more like 0.46.

We could also RLL the 0xFF bytes, but that would help only marginally.
We have come up with some trickier dictionary-based algorithms, but
the RLL ratios are much better then we need at present, so we'll just
go with this.

Runtime unpacking and config bit-banging should be fast. When we see a
null byte, we can hop to a brute-force routine that reads the next
byte and bangs out up to 2040 zero bits as fast as it can wigwag the
CCLK port bit. It should be much faster than the old loop, which just
shifted all the data bits out.

Thanks,

John

 

[John Larkin]

[1] statistical analysis of some FPGA configuration patterns, leading
up to a fast, small compression/decompression algorithm. We need to
fit an application program and 6 megabits of Xilinx config stuff into
a 4 mbit Eprom.

I wrote the following two small C programs in 2001 for an Altera FPGA. You
can't get it much simpler, and from a whole bunch of config files it gets
about a 50% compression factor (plus or minus 15%).

You may need to set the 'most-common value' from 0x00 to 0xff - I don't have
a lot of Xilinx bitstreams here to check what's best.

You could actually check the 'golden' bitstream to count which byte value
occurs the most...

Good luck!



Ben

Actually, I think it was *your* code that I modeled my thing after.
Anyhow, thanks to everybody.

John

 

[ehsjr]

Hmm, just to keep someone from plagarizing it,

<ftp://ftp.eskimo.com/u/m/mzenier/esrmeter.gif>

?? Can't get there from here.

 

[MooseFET]

Hmm, just to keep someone from plagarizing it,

<ftp://ftp.eskimo.com/u/m/mzenier/esrmeter.gif>

Use one of the spare 4049s

POT
!\ 100K 0.1 1N914
U1-11 -----! >O---/\/---!!----+---->!---- Top of xformer
!/ !
---
^
!
GND

With the leads open, adjust the 100K for no deflection on the meter.

A 30 kHz "ohmmeter" that I did 8-10 years ago.

I've done a 16KHz 4 terminal ohmmeter. It has many more parts but it
also covers a much wider range.

 

[Mark Zenier]

Use one of the spare 4049s

POT
!\ 100K 0.1 1N914
U1-11 -----! >O---/\/---!!----+---->!---- Top of xformer
!/ !
---
^
!
GND

With the leads open, adjust the 100K for no deflection on the meter.

And add an emitter resistor on the voltage regulator transistor so
there's a path for the current to take.

It's not a bug, it's a feature ;-). As it is now, there's not enough
current through the junkbox LED (voltage reference) to light it up,
so having the meter off zero is the Power On indicator.

As I remember, it draws about 6 mA. I need to go back and read Thompson's
threads about CMOS oscillators and overloading the input diodes and see
if it could get a little lower. Or use a CMOS 555.

This was done in response to downloading the Creative ESR meter schematic
and seeing that it drew 100 mA or something stupid like that. IMHO,
it's comparable in performance (0-30 ohm), but a whole lot less parts
and power consumption.

Mark Zenier [email protected]
Googleproofaddress(account:mzenier provider:eskimo domain:com)

 

[Chris Jones]

On Sat, 14 Jul 2007 14:06:40 -0500, Spehro Pefhany

On Sat, 14 Jul 2007 10:07:11 -0700, the renowned John Larkin

Fred Bloggs wrote:
John Larkin wrote:
Fred Bloggs wrote:
John Larkin wrote:
Fred Bloggs wrote:
John Larkin wrote:

That's about right, but it's still going to be simpler than
most of the goofy and truly terrible "esr" meters we've seen
here lately. You may as well measure leakage and dielectric
absorption while you're at it; code is cheap.

Granted they are amateurish junk, but your designs and ideas
are so *ugly* they are borderline grotesque. When you have
something artful to suggest that will be the day....

You don't approve of my products? Show us some of yours.

As they say, I don't have to be a chef to know the food stinks...

Show us something you've designed.

You show us an esr meter that doesn't require a VMX crate.

I suggest we have an esr-meter design contest. However, to
make it usenet friendly, and accessible to the readers, all
entries must be made in ASCII drawings, annotated with text.

Why not make it a group project, for real? I'd be willing to do the
first cut at schematic and algorithms. What we really need is someone
to hack the real code; I hate to program, and I'm going to do it most
of the weekend likely [1], so I'm not going to volunteer for that
part!

I'll volunteer, on provision that you, and anyone else contemplating a
serious contribution, read a few things so as to understand the
problem domain (testing, and in particular in-circuit testing of
e-caps) better.

It looks logical that it would be your task to do that research and
post a concise summary of requirements.

John

Well, the problem is that there's a dichotomy (or maybe a trichotomy)
in the wish lists.

First we have the service tech's unit:

- range up to 10 or 100 ohms with 0.01 resolution
- measurement frequency the equivalent of 100kHz
- no need to distinguish between impedance and ESR
- accuracy is of little importance, 10% is more than good enough,
we are looking for order of magnitude changes
- must be able to measure in-circuit, and preferably without
concern about polarity, so voltage should be low (say < 200mV) and
current should be 50-100mA RMS maximum.
- two wire probe is probably all that's practical, we have to beat
the convenience of tacking a known-good cap across the part and
trying it out
- it's reasonable to expect the meter to survive connection to a large
capacitor charged to ~400VDC without complaint
- should be simple/cheap/reliable/rugged and easy to use
- a "Pike" model is possible that would analyze the cap and deliver
a go/no-go indication (or possibly two for low-Z and regular caps)

That list of requirements is satisfied nicely by Bob Parker's ESR meter.
There isn't much point re-inventing that one so people might as well
concentrate on the fancier list of specs below.

- 0.001 ohm resolution
- reasonable accuracy (say 1% or better)
- must measure actual ESR
- able to measure relatively low value ceramic caps as well
as e-caps
- Kelvin measurement
- variable frequency measurement (maybe 100Hz to 1MHz)
- milliohmmeter for inductors
- no need for in-circuit but should not be damaged by cap
charged to ~50V
- reasonable cost


and hey, if we're dreaming

- complex and real impedance of inductors, resistors and caps over
that range
- ditto with DC bias on capacitors 0-10V and current on inductors
0-2A

One other feature:
Should work even if one side of the capacitor is already grounded. E.g.
should be able to measure the input impedance of a pin of a complicated IC
with respect to ground, even when the IC is soldered down to a board, and
with the IC ground connected to the ground of all sorts of test equipment,
power supplies etc. Lots of fancy impedance analysers assume that
components are always floating, which makes them useless in a lot of
situations.

Chris

 

[SensizAsLa.Net]

have built about a dozen Dick Smith ESR meter kits and found the
instructions to
be quite good but I could see how an inexperience kit builder could be
somewhat
overwhelmed with having to make some of the decisions... generally one
would
expect that if you have purchased an ESR meter kit that you would
have ...
http://www.sensizasla.net http://www.sensizasla.net/forum