Mastech MS5308 LCR meter with ESR measurement - on discount at the moment

[jtronix]

Please upload video of in circuit measurement of capacitor and ESR.

[jwrtiger]

POWER SUPPLY vs. BATTERY
I finally picked up a large package of batteries , in order to do the test .
I powered up the meter , then went through the calibration procedure , then found a NOS 3 watt 100 ohm carbon resistor that was waiting to get put away .
It measured about 109.?? [ this was yesterday and did not write down the last digits ] and with the meter plugged in to the power supply , with batteries installed , I pulled the plug from the PS at the meter and no change to the reading , even at the the lowest number . , plugged back in PS and again no change .
Again thanks for the video . 

I was wondering if you would test a resistor 10 k-ohms or higher.  The reason is I think at low values there was not much change but at higher values the problem starts to appear.  I do hope you have a good external supply and again I am happy you found the video useful.   John

[KD0CAC John]

I'l check out with some other values and get here with the info .

[jwrtiger]

I did a follow up to the Mastech LCR Tester's external power module.  The video is on YouTube:

  http://youtu.be/hA_0GnNO_2A

John

[T4P]

HOLY CRAP IT ACTUALLY HAS A FUSE. 
Anyway, it lacks a inductor

[Shock]

Another great video John.  Good to know its in your power supply somewhere, better than a faulty meter.
Wonder if its a 120v 60hz problem (despite the rating) you checked the caps right?

Anyone else who has one of these confirmed the problem running on mains with the supplied power supply?

[SeanB]

It needs moving the positive lead direct onto the cap, along with placing a ferrite bead on the diode lead, along with a snubber over the diode. A 100n ceramic cap across the big secondary one as well will help, though for best noise removal you need a 3 pin power lead with the earth connected to the output negative, along with 2 1n class Y caps on the input side from the bridge + and - to earth. That will get rid of most of the noise.

Better though is to use this for another purpose ( run a light from it) and build a new unit with a mains transformer, bridge and voltage regulator on the output. As the meter uses less than 100mA it will be both cheap and small. You can repurpose a standard magnetic wall wart and place the regulator on it's output.

[rowifi]

I've just taken delivery of the MS5308, and checking some small ceramic caps ( 10uf Hicap )using the tweezers  notice that the capacitance measures quite different between the frequency ranges  ( 6uF ... 14uF ) - and not in a linear way.
I haven't had a chance to do any other checks yet but would be interested in any comments, and suggestions on best way to check calibration when I don't have a reference LCR meter.

Instructions are poor, but general unit and accessories appear nice.
Rob

[PA4TIM]

Capacitance does change with frequency and indeed not lineair
ESR too. It is at max at 0.00000000000001 Hz  and decreases rather fast until 10-20KHz, after that is goes slowly down and somewhere between 50 KHz and some MHz (depending the package ect)
It starts to rise again due to things like skin effect.

ESR is just one function. Not specified or a standard, everyone can call his cap low ESR. You need a datasheet to use this value so many digits is useless, if you really need that you need an vectoril impedance meter that makes freuency sweeps or read more about theory and think again if you need it. Also in tht case you never meaure in circuit, only if you have a high vslue you know for sure it is bad, a low value does not proof a thing. There can be others parallel ( coil, reistors, fuses, transformers, other caps ect). Important too is capacitance and for most DC leakage and DC bias.

Most datasheets give ESR for 1 KHz or lower and Impedance for 100 KHz. Many ESR or LCR meters do not measure ESR but impedance.
My home made ESR meter measures real ESR from 10 KHz to 100 KHz, but I had it working upto 300) it measures caps from 100 nF to over 15000 uF ( han no bigger ones) just 3 ICs, kelvin technique and a few dollars in cost. Checked it against two VNAs, a peak ESR meter and several HP and GR meters.

[KD0CAC John]

PA4TIM , could you give some more details on your home made meter ?
Thanks

[kripton2035]

PA4TIM , could you give some more details on your home made meter ?
Thanks

explained on his web site here : http://www.pa4tim.nl/?p=3775

[KD0CAC John]

Thanks , too bad the hole page does not get translated .
The column on the right is not seen by by my Google translate . 

[PA4TIM]

In what would you translate it, it is in English, and i think that is your native tongue ;-)
But that is my explanation about ESR, not about the meter.
http://www.pa4tim.nl/?p=1728
This is about meters. On the bottom there is a handdrawn schematic. I use a standard voltage panel meter for read out.
There is a new version based on this one on its way. It is probably gonna be a kit in the near future, I'm working with some friends on that. ( i have not much experience in pcb drawing so they do that part) and it must be affortable and allthough components are not expensive, all together including display, cabinet ect will make it more expensive like i want. It is not to make a profit, i do not care for that, but it i do not want it to cost me money snd I'm no company so I can not buy big numbers cheap. So the form of the kit will be a questionmark.

The changes are small. First i changed the oscillator and flipflop idea because that performance turned out to be component dependend. Then the inverters used to split the signal had to go ( 74LS14 and 04 ( old stock) are tried worked on my proto very welk but on the pcb proto we found out new ones were not consistant in dutycycle, so to much product deppendacy too,) so i thought of a better solution. The oscillator frequency is now doubled. This is no problem at all. Just a bigger potentiometer. Mine runs 20-200KHz.
Then I use a 4013 divider to divide the frequency by 2. Very simple, and i have two nice opposite phase signals and a perfect 50 % dutycycle.
This setup i made about a month ago and all tests are good now.
Now we must design a better pcb, the first one was not up to my wishes and standards. ( I'm an analog guy, the ocb artist is nost times doing digital stuff) To much crosstalk and leakage, to thin traces ect. But the schenatic as it is on the link, will work. But not optimal, but if you find a Parker or Peak meter good, and use it at 100 KHz only this will be more then good enough. With the changes it will be a lot better.

[KD0CAC John]

The way the site shows up on my computer the column / box on the right is in , I guess your language .
Inside the top of the box on the right is the zoeken / search , then below that is the categorieen / category and everything else in that box .
It would seem that my automatic search option does not see the box I am referring to  ?

[The Electrician]

Capacitance does change with frequency and indeed not lineair
ESR too. It is at max at 0.00000000000001 Hz  and decreases rather fast until 10-20KHz, after that is goes slowly down and somewhere between 50 KHz and some MHz (depending the package ect)
It starts to rise again due to things like skin effect.

I don't find the described behavior of capacitor ESR to be common, especially with respect to aluminum electrolytics.  It does seem to be more common with non-electrolytics, but even then it's mostly older capacitors with dielectrics that aren't as good as the modern materials.

For example, attached are two images from an impedance analyzer showing the impedance and ESR vs. frequency.  The frequency is swept from 50Hz to 5 MHz.  The vertical scale is logarithmic with 100 ohms at the top and 1 milliohm at the bottom.  In these plots the impedance (Z) is the green curve and the ESR is the yellow curve.

The first image shows an old cap with unknown dielectric; it might even be paper.  The second image shows a modern MKP capacitor made by WIMA.

The ESR is always less than the impedance, and just touches the impedance curve at the capacitor's series resonance frequency (SRF).  For a non-electrolytic capacitor like these two, the losses are low enough that the slope of the impedance curve increases substantially as the SRF is approached, with almost a sharp point at the bottom of the curve.

Notice that the old cap ESR decreases by about 3 orders of magnitude from 50Hz to the SRF.  The ESR of a non-electrolytic capacitor at frequencies below the SRF is typically dominated by the dielectric losses, and with modern dielectrics the variation is not so great compared to the older, inferior, dielectrics.

The ESR for the MKP capacitor is so low (3 milliohms at the minimum) that it is difficult for the analyzer to measure accurately.

I find that aluminum electrolytics typically don't have much increase in ESR as the frequency decreases to very low values because the non-frequency-dependent losses in the electrolyte, etc., are so great that they dominate.  I'm going to start a new thread in the Projects, Designs and Technical stuff forum and post a number of plots of capacitor parameters vs. frequency.

[PA4TIM]

I have done a lot of VNA  meaurements on capacitors. But also using RF-IV like you do ( but I can do phase too) i can sweep from 1KHz to 1500 MHz. Also have several bridges including the famous GR1620.

It is long told the ESR is at it lowest point at SRF, but like your plot of a new cap hows, it is not.
Indeed the ESR can never be higher as Z and is equal to Z at resonance, but that does not have to be the lowest value.

About decreasing frequency and increasing ESR, your plots show what I mean, it is not lineair, it falls off quick from allmost DC to about 20 Khz and then it goes more slow.


Calibrated good fixture

Calibrated but bad fixture for the same cap

BUT, do you use a real calibrated fixture for these sweeps ?
And has your calibrationkit known parameters ?
I use a calibration that involves open, short, load and a series of knoen capacitors so allmost all strayes are compensated for. The VNA then uses 12 term error correction. I use special low Z fixtures for ESR and high Z fixtures for things like measuring fF ( and customtraces to transform S21 into R + jX.

What instrument you use, the plots look very good, similar to what I get ( how the Rs and Z behaves)

Problem with real good ( mica, MKP ect) caps is to find the SRF because, deembeded and clibrated right it can be several hundred MHz. Then calculate ESL from that and make a corrected jX sweep. If calibration goes wrong you are only looking at the SRF of the setup.

[The Electrician]

I have done a lot of VNA  meaurements on capacitors. But also using RF-IV like you do ( but I can do phase too) i can sweep from 1KHz to 1500 MHz. Also have several bridges including the famous GR1620.

I started another thread in what seems to be a more appropriate forum:

https://www.eevblog.com/forum/projects/capacitor-measurements-on-an-impedance-analyzer/

The analyzer I used to produce the plots can do phase, and every other parameter you can imagine.  See the first attached image for a screen shot of the available parameters.  I mainly want to show the behavior of electrolytics and there doesn't seem to be much point in attempting to measure parameters into the GHz range.  It's really hard to fixture electrolytics to a GHz VNA.

It is long told the ESR is at it lowest point at SRF, but like your plot of a new cap hows, it is not.
Indeed the ESR can never be higher as Z and is equal to Z at resonance, but that does not have to be the lowest value.

About decreasing frequency and increasing ESR, your plots show what I mean, it is not lineair, it falls off quick from allmost DC to about 20 Khz and then it goes more slow.

Your VNA plots are using a linear frequency axis.  This is typical of VNAs, although some can be set to display on a logarithmic frequency axis.  Most plots of ESR and impedance that can be found on manufacturer's web sites have a logarithmic frequency axis.  I find that it's difficult to draw conclusions about how fast ESR is falling off at low frequencies when the frequency axis is linear.  The second and third attached images show a sweep of an electrolytic with both a log sweep and a linear sweep.  You can see that it's difficult to see in much detail what's happening at the lowest frequencies.  Can you change your plots to have a logarithmic frequency axis?

BUT, do you use a real calibrated fixture for these sweeps ?
And has your calibrationkit known parameters ?
I use a calibration that involves open, short, load and a series of knoen capacitors so allmost all strayes are compensated for. The VNA then uses 12 term error correction. I use special low Z fixtures for ESR and high Z fixtures for things like measuring fF ( and customtraces to transform S21 into R + jX.

In the other thread I give a link to the impedance analyzer I'm using.  These relatively low frequency analyzers (5 MHz highest frequency, compared to several GHz typically for a VNA) only need to calibrate with an open and a short.  The fixture is connected to the front panel's 4 terminals and the short and open are applied to the fixture's measuring points; it is this style of fixture: http://www.home.agilent.com/en/pd-1000000479%3Aepsg%3Apro-pn-16047E/test-fixture-40-hz-to-110-mhz?nid=-34051.536880747.00&cc=US&lc=eng

The open is just the absence of anything in the fixture, and the short is a small shorting bar that comes with the fixture.

What instrument you use, the plots look very good, similar to what I get ( how the Rs and Z behaves)

Problem with real good ( mica, MKP ect) caps is to find the SRF because, deembeded and clibrated right it can be several hundred MHz. Then calculate ESL from that and make a corrected jX sweep. If calibration goes wrong you are only looking at the SRF of the setup.

As I said, I'm mainly concerned with electrolytics, and even for the non-electrolytics that I may show, the frequency range I'm using is reasonably similar to what can be measured by the low-cost ESR and LCR meters that are often discussed on the forum.

[PA4TIM]

My VNAs are able to make plots from 1KHz but using some special settings it can go down to  10 Hz. Several of my bridges do 10Hz to 100 KHz too. On VNA does 100 KHz to 110 MHz, the RF-IV does 1 KHz to 100 MHz. My home made ESR meter does 10 to 100 KHz.

If I use a vna for electrolitics i most times sweep 1 KHz to 100 KHz, or 1 to 1000 KHz or whatever I want.
I also use a special fixture that sets 50 Ohm in series with the DUT for measuring ESR ( but not using a RF-IV testhead)

It can do log, lineair , custom, smith, polar and radar plots. But that does not matter, optical it looks different but its all the same. It is al based on the waveparameters and then converted to scattering parameters and thnigs like Rs, phase, Xc and all the other things you like ( from RL to group delay)

The magnitude at frequency X does not change wen changing the axis, only the position on the X-ax. If you want to look at the non-linear decrease of ESR vs Frequency it makes only sense to use a linear sweep.

But i have not invented this and It is not really worh the discussion, you have the gear to test it yourself, and you can read it in every good book about capacitors ( or general radio experimeters, app notes from lcr\bridge\vna\Impedance meter manufacturers, literature from cap manufactures or books like analog SEEcrets)

This is not to offend you, just a few little minor points. I Think your comments in discusions about cas are very good. And it is nice to know I'm not the only one mad enough to do a lot of capacitor resource.

[mrubbert]

MS5308 died directly when connected power adapter, total dead and wont fire up with battery either. Brand new 30min test.

Did ofcourse check the chines poweradaptor DC 12v fore Ac-rippel and more. It have that 100kHz nosie shown in youtube review.

Just wonder if it happend to other owners ?

[jwrtiger]

Sorry to hear about the failure to your LCR Tester.  I am still playing around with that power supply for fun and I have pretty much created a schematic diagram of the circuit.  The output after the transformer is just a half-wave rectified, capacitor filtered circuit.  It is still working when I connect it to my LCR Tester but of course the error is still there in the resistance modes.  I tried an $8.00 (USD) linear power supply from Jameco  and initially it seems to work... need to check it out completely.  I still plan to make my own supply in the next few weeks.

[mrubbert]

And iam sorry x 10, i start to like the meter very much and have around 8 devices to check dissipation factor. Also NOT funny to get email answer from seller " check battery " . In the end i think i will a get new unit. So its warranty issue.

Heard from other person that had similar problem: when push "startbutton" 12v will go to voltage regulators and out to the board, the small IC got 5v and the big IC and around it only gets 1.2v. Something is pulling it down. And he thinks its regulator curcuit or IC,  also he dident try cut of delivery to IC.

People shoud be careful with putting on Powesupply when meter is on battery !
Specially with that shitty original build PSU. No warnings in manual either.

The killing thing must be more like bad revision on PCB, sensitive to transients or something, how shoud i know.
 Come on mastech this can be a winner.

[alm]

I got my hands on one of these a while ago. This post is based on my recollection.

It had the same problem of inaccurate resistance readings when on AC power as reported by John. Replacing the power supply with a 12 V linear wall wart fixed the problem. One of the on-board voltage regulators, either the 5 V or 3.3 V regulator, were specified for up to 8 V input, 10 V absolute max. There were a few diodes in series with the power input, but the voltage at the regulator was above 11 V. I installed an 7810 (with filter caps) between the DC power input and main PCB. This appeared to work fine, apart from the battery level indicator showing only two bars when on AC power.

[mrubbert]

Sounds like your recollection is spot on.

The other guy told me he got 12v in to the regulator, Pin1-1.2v pin2-12v pin3-12v on his meter. He didnt se any 3.3v anywere. Great Mastech !

Just wonder how it can survive battery 12.5v, maybe the diodes in series you talk about is zener that keep the voltage down.

Many thanks.

PS.
The PSU will hang on chrismastree next year and its not importent, i did or do like the meter it self.

[Smokey]

So what is the consensus at this point about this meter?  Messing up the voltage regulators is pretty amateurish.  I had very little faith in Mastech as a brand already and it doesn't sound like it's getting any better.  But on the other hand, this MS5308 seems to be the goto meter that people recommend over the only slightly cheaper models using this same seemingly capable chipset.  At 200USD this really isn't a new toy impulse buy anymore where I don't care if it doesn't work as expected. 

Is this MS5308 still the best in this price range and feature set?  What is the next step up, and not just build quality but also feature set?

[kripton2035]

actually the top in its class has been discussed in eev forum - it seems to be the IET DE-5000 - but arround $350
the probes accessory are important and the low end meters have crappy probes.
you need 4 wires probes to make real measurments on an lcr meter
they all use the same chip, I even found a "copy" here in france for 120 EUR...
http://www.ietlabs.com/de5000-lcr-meter.html