IET DE-5000 LCR Meter in UK

[Wytnucls]

All the meters based on the Cyrustek chipset seem to have decent specs. The more expensive ones have tighter tolerances and extra features like datalogging, 4 wire measurement and external power supply.

[BravoV]

All the meters based on the Cyrustek chipset seem to have decent specs. The more expensive ones have tighter tolerances and extra features like datalogging, 4 wire measurement and external power supply.

Is this one also using the Cyrustek chip ?

[Wytnucls]

if you are referring to the 50,000 count LCR meter Mike just acquired, the answer is no.
The test voltage on this one is quite high at 1.0V (0.6V on the Cyrustek ones), so probably not suitable for in circuit testing.

[mzzj]

I just ordered this http://www.reichelt.de/Induktivitaets-Kapazitaetsmesser/PEAKTECH-2170/3/index.html?;ACTION=3;LA=446;ARTICLE=113035;GROUPID=4060;artnr=PEAKTECH+2170;SID=13USNyKX8AAAIAAFVJPZYbc2f1faa3fc35d8d1c4071db95628b4c

Based on the same chipset as IET and has 4-wire measurement.
Can't comment yet about quality. Hopefully tomorrow I will get my hands on this meter.

+ has 4-wire measurement
+ 100 khz measurement like most? Cyrystek -based meters
+ EU- seller, no hassle with customs and reasonable postage fee of 19 euros.
+ Comes with kelvin clips

-Slowish measurement rate, I guess same applies to similar meters like IET-5000 as well.
-Crappy buttons requiring extra effort to press
-UI requires plentiful button presses to get series resistance.
-Build quality borderline acceptable.
-dual display somewhat useless with big electrolytic capacitors, 10000 µF capacitor has to be measured with 100hz test frequency and then you got only 0.1 Ohms resolution in dual-display ESR-reading. Useless.  Switching to series resistance mode allows measurement with higher frequency and better resolution.
-Kelvin clip wires are rather stiff, I am planning  to replace these with more flexible coaxial or twinaxial cable and Mueller BU-75 kelvin clips http://www.muellerelectric.com/kelvin_clips.html
- No AC-adapter included

Question for IET DE-5000 users: What sort of results you get from a 1uf or 2uF  polypropylene or polyester capacitors with different measurement frequencies? Preferably pulse rated polypropylene cap..

[PA4TIM]

What do you mean ?
I just grabbed a hand full of MKT, MKP, MMK ect from the caps drawer, measured them all in Cs, D and ESR ( not that that is an interesting number but its the buzz word today ;-) ) on 100, 120, 1K, 10K and 100K.
Results are like I would expect them. A ( very) small decrease in capacitance due to ESL and a rise in D with frequency. ESR high at 100Hz and decreasing with frequency.
I have not compared these to my vna or GR1620 but when I bought the meter I tested that with my test and standard caps and it is very accurate. Better as the specs tell.

[Wytnucls]

Here is the UNI-T version, also with Cyrustek chipset: UT-612

Sells for about 150$. 9V battery and doesn't have 4-wire measurement capability either. It does actually, but only up to the input jacks. Kelvin test leads won't work with this meter. It has data logging and external power supply through mini USB.
http://www.ebay.com/itm/UNIT-UT612-Multipurpose-LCR-Meter-Inductance-Capacitance-LCRDCRQDESR-/310546495592

[robrenz]

Here is a short video showing the DE-5000 using Rs on a 6000µF cap for ESR.  I am using 10kHz because it is the lowest freq. that gives 0.001 Ohm resolution because it settles faster than 1kHz.  Open and short calibration was done before starting the testing. The Micro ohm meter is calibrated and its worst case error at 0.010 Ohm is +/- 6 µOhm and at 0.020 Ohm is +/- 8 µOhm.  I am using pulse mode on the micro ohm meter because it uses 1amp drive on the 20 Milli Ohm scale and continuous mode would start to raise the copper wire temp and change the resistance. Pulse mode only gives a 166 Milli second pulse of current and latches the reading. This mode does add 2 µOhm to the above max errors but for this test it is irrelevant.  I have tested the DE-5000 Rs at single digit Milli Ohms  against the 510A and it is very good.  I would be very interested to see this test done on the Agilent 1733C (no Kelvin connections) but having 0.0001 Ohm resolution?

The kelvin clips used on the Micro ohm meter and the DE-5000 are my design and construction as shown in this thread. They do not come with the DE-5000.  A Pomona pincer is used as the starting point but they are modified to provide a true kelvin connection.

[PA4TIM]

I not yet watched your video, but I will. I tested the DE-5000 tonight on some coils.
I made a 7 uH coil with a DC resistance of 0.6993 Ohm measured with kelvin clips on a 7.5 digit prema.
It was 0.70764 Ohm at 1KHz on a GR1608
It was 0.700 Ohm at 10KHz and 0.708 at 100 KHz on a Boonton 63H
So all very close, the deviations are because of temp and presure ect, i did not wam everything up for a few hours, just a quick test.

Then i measured it with the IET at 100, 1, 10 and 100KHz and at 10 and 100KHz with the boonton ( value between brackets) the Boontol is restored, repaired and calibrated by me using GR inductors, ESI resistors and a vna and more stuff

100Hz : 0.70 Ohm, 9 uH
1 KHz : 0.707 Ohm, 9.1 uH ( 0.70764 Ohm on the GR1608, that did not manage to measure inductance because the Q scale starts to high for this poor coil)
10 KHz : 0.699 Ohm, 9.15 uH ( 0.700 Ohm and 9.01 uH)
100 KHz: 0.656 Ohm, 9.011 uH ( 0.708 Ohm and 9.005 uH)

So you see they are all rather close. By the way, the IET is not specially calibrated before this meaurement. I used the standard delivered croc adapter that I calibrated the lat time I used it. If I had the beautiful clips Robrenz made I think they would be even closer.

So if you want a real good meter, the IET is the one. That it uses the ( or a ? ) chipset that is used in other meters too, does not make those other meters as good as the IET, or even good meters. If you manage to screw up something as simple as the PSU i do not believe they do much of a job on the hard part ;-) To be as accurate as the IET the rest of the design must be good too. It is not for nothing the others talk about 1% while the manufacturer of the chipset, as does IET, talks about 0.3 %. And that is not just a marketing number.

[Wytnucls]

@ Robrenz
Measuring resistance in Rs mode at 10 kHz has a resolution of 0.001 and a stated accuracy of 1%+/-3d on the DE-5000, so your measurements could have been 4 counts out either way. Just because yours is spot on in this case, doesnt mean it will be next time around or that the next meter will have the same accuracy. You just cannot objectively trust the meter readings down to a one milliOhm accuracy.

I'm not convinced that you're actually measuring ESR in that 10 kHz Rs mode. It is certainly not the way the manufacturer calculates ESR for capacitors.

I think that the stated accuracy of 1%+/-3d is also a bit optimistic, when the manufacturer (Cyrustek) publishes 1%+/-5d.

Cyrustek garantees accuracy if proper ratio resistor calibration and open/short calibration has been performed. So it looks like all the hard work is done by the chipset and all meters using the Cyrustek set-up should have about the same accuracy if calibrated properly..

[robrenz]

@ Wytnucls

I have done enough repeated testing of the DE-5000 on Rs to say that MY meter is much more accurate than the stated Rs accuracy. It is true that I may have the only one that happened to fall in the center of the bell curve.  But it looks like PA4TIM also happened to get one that is exceptionally accurate. The fact that IET Labs will sell you a calibration certificate implies it definitely meets its stated specs. The video showed the 5000 exactly matching the 10 and 20 milliOhm values of the test wire with no bobble at 1 milliOhm resolution. It also was dead stable with no bobble when the Rs of the capacitor was included. and it repeated exactly on the second pass. The accuracy of the 59 miiliOhms Rs of the capacitor is a separate issue which I have no means of checking but PA4TIM's readings show that it is very accurate also .  The 510 micro Ohm meter is substantially more accurate than the 3458A at 20 milliOhms and below. At three digits more resolution than the 1milliOhm and a worst case error of 0.04% at 20 milliOhm and 0.06% at 10 MilliOhm  I don't think I am stretching things.
I have caracterized the 5000 starting at zero and doing 1 milliohm steps thru 10 milliOhms. It is spot on if you feed it exact whole milliOhm values. Its max error is one milliOhm on any reading if you take 250µOhm steps from zero to 10 MilliOhms.  The 510A does offset compensated ohms measurements so thermal emfs of the measurements are eliminated. And since Rs is using an AC excitation it also eliminates any thermal emf issues.  My readings are not 4 counts out either way and this performance has been checked multiple times over the 2 years that I have had the 5000.

My point?  This is an exceptional piece of equipment that has conservative specifications.  If you ordered the Cal with the 5000 from IET and it failed some ranges you would send it back. There is no ability to adjust/calibrate the 5000 that I am aware of and I don't think IET is going to risk their reputation with shaky specs.

ESR vs Rs  If you look at the ESR readings of the 5000 they are just lower resolution and sometimes frequency limited versions of Rs.  Rs allow you to use any frequency you want on any capacitance.

My hunch is the chipset is available in various accuracy grades and DER is using premium chips for the IET branded DE-5000

[Wytnucls]

There is a one-turn trimpot on the PCB next to the input ADC (VR2). No sign of VR1 anywhere, which could have been removed in later revisions.
The CEM version has 2 multi-turn trimpots, with one next to each ADC.
I don't think that there are several version of the Cyrustek LCR chipset. The Q probably stands for QFP-100L and the S for SSOP-48L.
The market is just too small for that and Cyrustek is usually very good in listing all the versions of their chips on their site. Also, the difference in quoted accuracy is only a couple of counts on all ranges, making it even more unlikely.
There is always the possibility of IET custom chips, but that would push up the unit price to prohibitive levels.
The fact that the chips were epoxied over also means that they are not custom made, but rather freely available in quantity from the manufacturer.
Perhaps IET spends more time with each device, tweaking them to the highest accuracy possible.
It would be interesting to compare the 5000 with other devices using the same chipset, to see if there is really a noticeable difference in performance.

[PA4TIM]

It is realy measuring ESR in Rs mode, in all frequencies. Rs = ESR, besides that, I checked readings agains the D of my GR1620 ( and tested it for other ranges against my standards) and it is realy ESR, not Z like on most ESR meters.

Problem is the test done by Robrenz does not prove it measures ESR. ( but it does !) But it proves the accuracy, and I do not think our meters are the only two, a compare would be nice but difficult, i o not think someone will buy two of them and I'm not Dave, the manufactures do not send me standard caps and coils or instruments as a test ;-)

A Z meauring ESR meter will also show an acurate increase because they operate at 100kHz and Xc is neglectble there for larger capacitors, ( datasheets give Z at 100 kHz, the times I saw ESR it was at 1 kHz together with D, what is stated always at 1 kHz or 120Hz) that is why many peope have so much faith in their ESR meters. But that test is not the way. Measuring DF or phase angle and transformimg that to ESR is the right way. The IET DOES measure real ESR in all C ranges and in Rs.

A very old test for LCR meters is to add 1 K in series with the capacitor, the meter should still give the same capacity. Many meters fail this test and give the capacitance based on the |Z| meaured.

If accuracy is based on the right R ratio calibration and Open/short calibration. I think there could be the huge difference between IET and other chipset users. The R-ratio calibration is done by the manufacturer and I think the resistors under the analog front end take care of that. And using cheap non precision chinese resistors or high quality precision resistors makes a big difference in performance and in cost. The fact IET only uses one one-turn trimmer and the CEM use 2 multiturns also is remarkable, multiturns are a cheap and more unstable way to avoid expensive precision components or hand-picking components.

Besides that there are things like routing, connectors, quality of pcb material, proper use of guarding ( very important) and things like Leaking current due to dirty boards or chips ( that is why i never open this sort of instruments unless for repair) and that could be the reason for the potting on the IET chipset  too !
Bob Pease once has written about how you ruin a precision chip by touching it with your fingers. The leakage current can increase with a factor as big as 100. )
In my Fluke 332 are pcbs that have a sticker not to touch them because they are ultrasonic cleaned. My Fluke 720 was out leakage specs at 1000V by a factor of 1000, after two times cleaning using ipa and measuring it hanging on a rope i got it good again.

I designed and build a fF meter ( meaures impedance but ESR is neglectble for the result in those regions) with 1 fF usable resolution. It had 6,5 digit readout using a TRMS meter. The last 2 digits were unstable until I added guarding and cleaned everything with ipa, even cleaning between the front connctors improved performance.

If it was so easy, just using a chipset, every multimeter using a LM399 would be the same too ?

I think cyrustek makes a hell of a lot mor LCR chipsets as for instance AD makes LM399 or LT makes LTZ1000 Vrefs. And both those you can buy in grades so why not the cyrustek chipset ? Also things like burn in time by the manufacturer of the meter make a difference, maybe here too. There also is a crystal. There is a big difference in stability between cheap and good crystals, and a stable frequency is important for accurate measurements. And IET also offers NIST calibration for the DE-5000 as an option ( but that takes 2 weeks and a lot of cash extra)

I also do not understand why the masstek needs 12V and does not work on 9V if it used the same chipset.
An other strange thing, i can not find real datasheets of this chipset. Only the (end-user based) sales brochure they call datasheet, so it could be possible there are different versions. If they statet accuracy < 0.3 % It also can be they can deliver chipsets with accuracy < 0.3 and not all chipsets are better then that. It can be possible they also have 1 % or worse sets. We will never know, and assuming is the mother of all.......

So all these factors make the meter too, but buyers only look, "hey, it is cheap, they use that chipset too, so it must be as good as the IET", even advise to use the IET manual. yes, sure, dream on, like with many multimeters, it uses the same cabinet, or some dare to say, almost the same, so it must be the same meter. The chinese are more clever then some buyers. if I build a car out of wood and cardboard, looking like a RR with a RR engine in it, it is still not a RR ;-)

It even could be possible some chipsets are fake ( I do not say it is, but there are plenty examples of fake Chinese components and instruments so it is possible)

[mzzj]

What do you mean ?
I just grabbed a hand full of MKT, MKP, MMK ect from the caps drawer, measured them all in Cs, D and ESR ( not that that is an interesting number but its the buzz word today ;-) ) on 100, 120, 1K, 10K and 100K.
Results are like I would expect them. A ( very) small decrease in capacitance due to ESL and a rise in D with frequency. ESR high at 100Hz and decreasing with frequency.
I have not compared these to my vna or GR1620 but when I bought the meter I tested that with my test and standard caps and it is very accurate. Better as the specs tell.

Just wanted to see some results out of curiousity ie. how high Q values it will show for a good quality caps.

Peaktech: 2x300nF parallei pulse polypropylenes shows Q overrange  with 100hz and 120Hz frequency as pretty much expected. With 1k measurement frequency results vary somewhere between Q=1800 ... to over range.

[PA4TIM]

I measured a few caps.
Philips HQ 344 MKT 2.2 uF 10% 250V, I could not find a datasheet but I type this now I'm not rolling on the floor anymore from laughing, found a post on some forum of some idiot who was playin-in these caps in a CD player and amplifier.    So I'm sorry, i just hooked them up cold, should you think I leave then on my bridge for a couple a weeks for final measurement 

4 of those soldered together. That adds some copper resistance

A GR 1424A 0.25 % 1-10 uF polyprop standard as reference 1.001,906,233 uF at 1kHz, D = 0.000204 Xc = 158.852133912 ESR = 0.0324 Ohm measured on a 0.01% GR-1620 at 1kHz
The percentage below is the deviation to the 1424A as measured on the GR1620 (and also checked on sone other precision bridges and meters so it will be pretty close)

GR1608A, 0.1 % bridge at 1005Hz
GR1424A at 1kHz 1.0015 uF and D = 0.00025  Xc = 158,91656 ESR = 0.039 Ohm Q = 4000 (-0.04% well within specs)
MKT at 1 KHz 2.239 uF D= 0.0035 Xc = 70.7294 ESR = 0.247 Ohm  Q=285.71
4MKT at 1KHz 8.837 uF D= 0.0044 Xc = 18.01006 ESR = 0.079 Ohm Q = 227.3

DE-5000
GR1424A
at 100 Hz 1.0028 uF  D = 0.000 Q = OL ESR = 0.03 Ohm -90 degrees
At 1 KHz  1.0029 uF D = 0.000 Q = OL ESR = 0.03 Ohm -90 ( + 0.099% well within specs)
1MKT
100 Hz 2.249 uF D= 0.001 Q =  661.3 ESR 1.07 Ohm
1 kHz 2.238 uF D = 0.005 Q = 185.6 ESR = 0.371 Ohm
ESR at 100 kHz = 0.078 Ohm

4 MKT parallel
100 Hz 8.865 uF D = 0.001 Q = 671.1 ESR = 0.30 Ohm
1 kHz 8.839 D= 0.005 Q = 185.6 ESR = 0.097 Ohm
ESR at 100 kHz = 0.020 Ohm
.

[SeanB]

I have an old board with a few film caps on them, the largest is an ICW  PW10uF250V unit, which came up as a pretty close to perfect capacitor when I checked it on my new MS5308. It was run in for a few years with a 5A or so 15kHz drive as it was in a tag detector as part of the output LC filtering network.

[Wytnucls]

Yes, the LCR meter does measure ESR in Rs mode, but with the required testing frequency of 120Hz for electrolytic caps, one only gets one extra digit of resolution.
Testing with these caps at a higher frequency, would give one more digit of resolution still, but a false ESR, which in turn would give a false D to enter 120Hz datasheet tables.
Here is a test, run on an new Daewoo RSS electrolytic cap of 100uF 100V:
Measured at:
120Hz     C:94.81uF     D:0.042    ESR:0.6 Ohm     Rs:0.62 Ohm       Theta:-87.3
1kHz       C:90.71uF                    ESR:0.21 Ohm     Rs:0.214 Ohm      Theta:-81.2
10kHz      C:83.2uF                     ESR:0.15 Ohm     Rs:0.155 Ohm       Theta:-38.3
100kHz    C:18.67uF                   ESR:0.14 Ohm     Rs:0.145 Ohm       Theta:34.6

Calculated ESR (120Hz) = D /2 Pi f C = 0.588 Ohm

Datasheet: Max D: 0.080

[PA4TIM]

Not all datasheets state D at 120Hz, 1KHz is also often stated. For many decades 1KHz was the standard, every single frequency bridge was 1 KHz, multi frequency once had 10 KHz extra, the adjustable ones started somewhere between 20 and 200Hz when capacitance increased ( in the 60's a 1000 uF cap was a big one, now we use 1000 uF / 1A as minimum) 120 and 100 Hz became more important.  It is not false ESR, it it the right ESR but like I told often, you still know nothing. And you wrote a while back what I think, D tells you more.

You get higher resolution using the Rs function as primair display. And why you need the resolution ? If it is giving you zero for ESR, D, and -90 for Theta you know it is a good one, who cares if it is 0.005 Ohm. If you want this sort of figures you need a lab grade C meter for component research, an impedance meter, VNA or lab grade bridge. Not a handheld LCR, alltough the DE-5000 does a hell of a job to prove some handhelds can almost compete with these lab grade instruments. But that is not strange if you think IET/GR has decades experience in building worlds best LCR meters/bridges. Their GR1620 is still a standard. My GR1608 is 0.1% but the specs on later model were much better. As far as I know not because of improvements but because it later turned out those specs were to much on the save side.

IET will not overspec instruments. Their specs have beem always more a worst case number. While for most Chinese the specs are like for instance minimal 1 % ( so if you are lucky it is 1% but it can be 5% too, or the state typicl, avarage ect, most times not even tracable. IET can provide A2LA accredited calibration with uncertainties on the DE-5000.  Not sure you are going to get that from an
handheld on ebay.  As well if you have questions who do you call if you are buying a Uni-T? I had a stupid question about my DE5000 ( stupid becuse it was in the manual, but I mailed them and got the answer within a few hours ( and not a RTFM style answer)

Yes IET is more expensive.  There will always be people that purchase
just based upon price.  But I think service and support as well count
for something.  For example you can find the LCR meter guide on their website
http://www.ietlabs.com/notes/lcrprimer/?___store=default as well as other
material at http://www.ietlabs.com

But it is easy to transfrorm D. If you meaure D at a 1KHz bridge, the D reading is for 1KHz only. If you feed that bridge 100Hz the "D reading /10" is the D at 100Hz.
But you know how to calculate and interpret D so I do probbly not understand your question.

[Wytnucls]

That would explain why the LCR meter defaults to 1kHz on start up, instead of 120Hz.
I am surprised that IET equipped their handheld with a Cyrustek chipset, instead of developing their own, considering their vast experience with LCR meters.
I'll agree that you get better and faster support from companies like IET, but after sale service and customer relations are getting better in China, Hong Kong and Taiwan. The main obstacle is still the language barrier, especially on the technical side.
I am in Changhai at the moment and will visit UNI-T's offices in the next few days for a look at their latest meters. I also hope to have some time with the UT612 for a comparison with my own CEM LCR meter.

[robrenz]

Yes IET is more expensive.  There will always be people that purchase
just based upon price. 

[Fanboy mode]

At $335.00 the DE-5000 is roughly $100.00 cheaper than the Agilent U1733C.  The Agilent has an extra digit of resolution and a much more sophisticated user interface. But I think the extra digit came from marketing not engineering since it has no 4 wire measurement and the additional digits error goes thru the roof on the lowest scales   Even though the banner spec of the 1733 is 0.2% versus the 0.3% of the 5000, the 5000 is equal or more accurate on many frequency/range combinations.

An example of how the extra digit is from marketing:
The IET DE-5000  will give you +/- 1% +3 counts on the 20 Ohm range at 1kHz  Calculated % of value error would  be:
10 Ohm  = 1.%
1.0 Ohm  = 1.3%
0.1 Ohm   = 4.0%
0.01 Ohm  = 31.0%
0.001 Ohm = 300%

The Agilent U1733C will give you +/- 0.7% +50 counts on the 2 Ohm range at 1kHz  Calculated % of value error would  be:
10 Ohm  = 0.75%
1.0 Ohm  = 1.2%
0.1 Ohm   = 5.7%
0.01 Ohm  = 50.7%
0.001 Ohm = 500%
0.0001 Ohm = 5000%

[/Fanboy mode] 

[kripton2035]

@PA4TIM,
How much did you pay for your DE5000 ? price + shipping + vat ?
did you buy it by IET directly ?
thanks.

[robrenz]

Here is PA4TIM's data from post #38 reorganized a little.

[staxquad]

It is realy measuring ESR in Rs mode, in all frequencies. Rs = ESR, besides that, I checked readings agains the D of my GR1620 ( and tested it for other ranges against my standards) and it is realy ESR, not Z like on most ESR meters.

Problem is the test done by Robrenz does not prove it measures ESR. ( but it does !) But it proves the accuracy, and I do not think our meters are the only two, a compare would be nice but difficult, i o not think someone will buy two of them and I'm not Dave, the manufactures do not send me standard caps and coils or instruments as a test ;-)

A Z meauring ESR meter will also show an acurate increase because they operate at 100kHz and Xc is neglectble there for larger capacitors, ( datasheets give Z at 100 kHz, the times I saw ESR it was at 1 kHz together with D, what is stated always at 1 kHz or 120Hz) that is why many peope have so much faith in their ESR meters. But that test is not the way. Measuring DF or phase angle and transformimg that to ESR is the right way. The IET DOES measure real ESR in all C ranges and in Rs.

A very old test for LCR meters is to add 1 K in series with the capacitor, the meter should still give the same capacity. Many meters fail this test and give the capacitance based on the |Z| meaured.

If accuracy is based on the right R ratio calibration and Open/short calibration. I think there could be the huge difference between IET and other chipset users. The R-ratio calibration is done by the manufacturer and I think the resistors under the analog front end take care of that. And using cheap non precision chinese resistors or high quality precision resistors makes a big difference in performance and in cost. The fact IET only uses one one-turn trimmer and the CEM use 2 multiturns also is remarkable, multiturns are a cheap and more unstable way to avoid expensive precision components or hand-picking components.

Besides that there are things like routing, connectors, quality of pcb material, proper use of guarding ( very important) and things like Leaking current due to dirty boards or chips ( that is why i never open this sort of instruments unless for repair) and that could be the reason for the potting on the IET chipset  too !
Bob Pease once has written about how you ruin a precision chip by touching it with your fingers. The leakage current can increase with a factor as big as 100. )
In my Fluke 332 are pcbs that have a sticker not to touch them because they are ultrasonic cleaned. My Fluke 720 was out leakage specs at 1000V by a factor of 1000, after two times cleaning using ipa and measuring it hanging on a rope i got it good again.

I designed and build a fF meter ( meaures impedance but ESR is neglectble for the result in those regions) with 1 fF usable resolution. It had 6,5 digit readout using a TRMS meter. The last 2 digits were unstable until I added guarding and cleaned everything with ipa, even cleaning between the front connctors improved performance.

If it was so easy, just using a chipset, every multimeter using a LM399 would be the same too ?

I think cyrustek makes a hell of a lot mor LCR chipsets as for instance AD makes LM399 or LT makes LTZ1000 Vrefs. And both those you can buy in grades so why not the cyrustek chipset ? Also things like burn in time by the manufacturer of the meter make a difference, maybe here too. There also is a crystal. There is a big difference in stability between cheap and good crystals, and a stable frequency is important for accurate measurements. And IET also offers NIST calibration for the DE-5000 as an option ( but that takes 2 weeks and a lot of cash extra)

I also do not understand why the masstek needs 12V and does not work on 9V if it used the same chipset.
An other strange thing, i can not find real datasheets of this chipset. Only the (end-user based) sales brochure they call datasheet, so it could be possible there are different versions. If they statet accuracy < 0.3 % It also can be they can deliver chipsets with accuracy < 0.3 and not all chipsets are better then that. It can be possible they also have 1 % or worse sets. We will never know, and assuming is the mother of all.......

So all these factors make the meter too, but buyers only look, "hey, it is cheap, they use that chipset too, so it must be as good as the IET", even advise to use the IET manual. yes, sure, dream on, like with many multimeters, it uses the same cabinet, or some dare to say, almost the same, so it must be the same meter. The chinese are more clever then some buyers. if I build a car out of wood and cardboard, looking like a RR with a RR engine in it, it is still not a RR ;-)

It even could be possible some chipsets are fake ( I do not say it is, but there are plenty examples of fake Chinese components and instruments so it is possible)

I run my MS5308 at 8.55V.

If you don't physically compare the units, the IET against other Cyrustek chipset units, it's just conjecture and opinion to justify paying ~3 times more for about the same result.

The MS5308 is pretty tight with the GR-1608A, actually differs by only .0001uF, (so either your IET Cyrustek chipset is inferior to my MS5308 Cyrustek chipset or your GR-1608 is only a .1% version, or a combination of both - right back at you)

[PA4TIM]

The 12V is from the topic about the ms5308 and a video. It ran on 12V batteries and the adapter that came with it alters measurement results and is  known to kill meters.
But if yours does run on less as 9V, production is maybe not very consistent.

My GR1608 is the 0.1% version but it agrees with standard caps and my GR1620.
I'm told the older 1608's are much better as the specs were, so on later models the specs are changed.

About our measurement set up, this tells nothing because the whole setup is bad. That HP decade is 0.25 % +/-3 pF. I have one too.
But that specs are only valid for 3 terminal measurement.

You use a two terminal connection ( the 1608 can make 3 terminal measurements so why this bad setup with the very long cables ? ) Those cables add a lot of capacitance. So the value on the meter or the value on the HP is wrong because you can not measure 40 pF capacitance exact like you do.

But this is not important and and the discussion, there is allready a topic about all the great " features" of the MS o defend it there ;-)

If you buy a IET you know it will match the specs or better. You now the warranty and service is good.
And yes, that is worth the extra ~100 dollar ( it s not 3X the price)  over the ms5308 with a powersupply that can kill the instrument, a cal button you can not reach and have you seen the insde from the MS and the IET in the teardowns and videos ? Maybe you are just lucky you found a good MS but with a brand like that it stays a gamble. I do not like to gamble. And I'm boss over my own wallet, so you do not have to decide for me.

But, hey, i do not care, buy the ms if you want, like I do not want that.

[Wytnucls]

Since the subject is on accuracy, please note that Cyrustek has some other proviso on published figures:
1. Accuracy is only garanteed if the DUT is less than 10cm away from the front end chip (ES51920) pins.
2. If D (Tan d) is higher than 0.1, the published accuracy figures have to be increased by a factor of SQRT (1+D^2)

[staxquad]

The 12V is from the topic about the ms5308 and a video. It ran on 12V batteries and the adapter that came with it alters measurement results and is  known to kill meters.
But if yours does run on less as 9V, production is maybe not very consistent.

My GR1608 is the 0.1% version but it agrees with standard caps and my GR1620.
I'm told the older 1608's are much better as the specs were, so on later models the specs are changed.

About our measurement set up, this tells nothing because the whole setup is bad. That HP decade is 0.25 % +/-3 pF. I have one too.
But that specs are only valid for 3 terminal measurement.

You use a two terminal connection ( the 1608 can make 3 terminal measurements so why this bad setup with the very long cables ? ) Those cables add a lot of capacitance. So the value on the meter or the value on the HP is wrong because you can not measure 40 pF capacitance exact like you do.

But this is not important and and the discussion, there is allready a topic about all the great " features" of the MS o defend it there ;-)

If you buy a IET you know it will match the specs or better. You now the warranty and service is good.
And yes, that is worth the extra ~100 dollar ( it s not 3X the price)  over the ms5308 with a powersupply that can kill the instrument, a cal button you can not reach and have you seen the insde from the MS and the IET in the teardowns and videos ? Maybe you are just lucky you found a good MS but with a brand like that it stays a gamble. I do not like to gamble. And I'm boss over my own wallet, so you do not have to decide for me.

But, hey, i do not care, buy the ms if you want, like I do not want that.

I knew about the power supply having noise before I purchased MS5308.  I blame the noisy power supply, you blame the bystander, MS5308.  I tested the unit for the operating voltage range and operate it at 8.55V from a constant voltage power supply or from NiMH cells x 7 ranging from 10V to 8.55V in it's capacity discharge.

I treat the HP decade value I dialed up as an unknown.  Don't really care about it's specifications.   Two meters are measuring the value and I'm comparing.  Why change the parameters for the GR if it's not the same as the MS5308?   

MS5308 did measure 40.0pF.  I didn't measure with the GR1608A because I didn't find the cabling.   Now that I've found the cabling, I tested that 40pF capacitance and got 40pF with 0 D, just like MS5308.   

With a shorter cable, a few less pF would not change the uF results, but dissipation would decrease.  I re-tested to that effect, D did improve to ~.0005.

I paid $170 from Franky delivered to my door, no extra expenses.     jpb, the originator of this thread paid in all $570 which is 3.35 times more.

And just like you, I claim that my equipment performs much better than it's specifications.

Sure great companies like Fluke, HP, GR, Tek produce more dependable and accurate equipment, but it seems that the Cyrustek chipset has a lot to do with the accuracy and performance of Mastech's MS5308 and IET's DE-5000.