Applent AT826 LCR

[ot1]

Has anyone purchased the Applent AT826 LCR meter (100kHz) and is happy with it ?   The YouTube tear down video of the cheaper AT825 (10kHz) shows that it has input protection and shielding unlike the GenRad DE-5000 which doesn't appear to incorporate any input protection.

[iDevice]

I'm in the same boat.
I've read some negative returns about completely off readings but it may have been just a bad copy.
The fact is, aside Mike's tear down and review of the 825, there isn't much info about these meters.
It's almost as if noone is buying them...

[ot1]

I talked to GenRad and the DE-5000 (1 yr warranty) is not repairable, also talked to Agilent and the U1733C (3 yr warranty) is not repairable  so after their warranties are up ... You would be buying a new meter if you damaged it with a charged cap.  There are no specs for how much overload the 1733c can handle, nor does tech support know at the moment.  So really all these handheld Chinese meters are throwaways.  I wish fluke had an LCR because I bet it would have a replaceable fuse like their DMMs.  So is the AT826 worth the $360, I haven't found it cheaper yet or by itself without clips.

[mos6502]

The DE 5000 is unbeatable, even if it were double the price. It's also very unlikely to fail since there's only two ICs in it and not much else. Certainly no electrolytics so I can see it lasting a long, long time.  By the way, don't buy the GenRad version, you're just wasting money. The DE 5000's OEM is DER EE, and you can get the same model on eBay for around $120 from Japan.

I would stay away from the Agilent because it's way overpriced for the features it has and the service seems to be subpar:

https://www.eevblog.com/forum/testgear/agilent-u1733c-lcr-meter-flashing-error/

I haven't heard of a defective DE 5000 yet.

An LCR meter is for measuring passive components, so any input protection is completely pointless. All it would do is cost money and distort the readings. Just resist the urge to poke the test leads into a wall outlet and you'll be fine.

OK, so I can already hear people whining about charged caps. Where do you find a charged cap? Do you charge your caps before you put them in your parts bin? If you're repairing stuff, you have to desolder the cap anyway to get a proper reading, so if it was charged, you'd already get zapped while you're desoldering it.

[TurboTom]

This thread seems to be dead for a long time so why not revive it? 

The reason for this is that just a few days ago I picked up an Applent AT826 LCR meter. For some reason I'm a friend of the more complex analog circuitry - maybe because I 'll be able to understand it more easily than the highly integrated more recent designs with the Cyrustek chipsets. Moreover, the additional digit of resolution of the AT826 model over the AT825 for me outweighed the higher price tag.

Now what's funny: Mike assumed in his video teardown of his AT825 that the unpopulated footprints of the PCB might be reserved for the 100kHz AT826 device -- well, they are not. The AT826 appears to be virtually the same hardware as the 825 (and presumabley also the 824), the only difference apparently is the firmware.

The meter works nicely and provides very stable readings, dithering only by two or three LSBs in the lowest capacitance/inductance ranges. The probes look rather "homebrewn" but they do their job well and are rather beefy. What I really didn't like -- at least on my specimen -- is the soldering quality of the PCB. Many of the left-hand terminals of the horizontally mounted resistors and caps (as referenced to the silkscreen) are barely soldered at all. I found two decoupling tantalum electrolytics that had been reflowed only on one side - the other terminal was up in the air! Yet, the meter was working nevertheless. Meanwhile I reworked the questionable parts and I'm confident that reliability will be improved a lot. Applent definitely needs to work on their QC and invest in more modern reflow equipment (or learn to use it properly).

Anyway, I like the meter a lot since it's compact, light-weight and comfortable to use. I don't know if Applent offers firmware updates - so far I didn't find any. But nevertheless there might be a way to "upgrade" the lower frequency versions to 100kHz. Yet, I'm not sure how much calibration will be necessary to get it accurate and if it's worth the hassle, especially since the price difference isn't that dramatic anymore.

Please find attached a photo of the PCB (sorry for the poor quality) and one to show the poor soldering quality (look especially at R621).

If anyone has more detailed questions or needs higher resolution / quality photos, please let me know. I cannot promise to provide immediately since spare time is rather limited but I'll try my best.

Cheers,
Thomas

[tridentsx]

There is a jtag port there right next to the nuvoton arm, why don't you hook up a probe and read the contents of that chip ?

[EU1]

Hi Thomas,

Did you tried to measure small capacitors and coils, like 1 pF or 100 nH?
Also does it measures large resistors (more than 100 MOhm) with acceptable precision?

[TurboTom]

Hi EU1 -

the instrument works pretty well and has a very good resolution of the low end range for capacitance and inductance which was basically the reason for me to get it. At 100kHz test frequency, the resolution is 1nH and 1fF, respectively. The least significant digit "dithers" by approx. two to three counts. Obviously, in this range, the influence of positioning of the probes is considerable. For example, if you do an "open" calibration, you should place the probes approximately at the distance from each other at which you expect to measure the (small capacitance) capacitor. When measuring inductances, a straight piece of wire of 5mm clearly and reproducably changes the reading. I cannot tell for sure how accurate the results are in this very low L / C range but simple tests like doubling the length of a straight wire doubles the inductance reading appears to be quite resonable.

The four-wire probes, though they appear to be of simple, even primitive design, serve their purpose very well and despite their appearance, I cannot tell anything bad about them. I hardly used the SMD tweezers so far so no comment on them. What I like a lot is the possiblity to measure very small resistances (preferably in the low frequency range of course) accurately. That's something a classic multimeter completely fails at. High resistances are not that much a domain of this pure AC meter. Capacitive effects appear to affect the readings at high resistances (>10MR). Anyway, the device won't measure above 100MR. For these high values, I anyway use a high voltage insulation tester (megger).

I don't know if Applent will ever offer a firmware update, but one thing I would like to see addressed is to eliminate the "garbage readings" when there's no DUT connected, like negative resistances, capacitances or inductances. Also, quickly jumping readings that may indicate open probes and the influence of some minor static electricity could be easily ignored. This doesn't affect measuring accuracy but makes the product appear a little "unfinished".

Otherwise, I didn't regret the purchase...so far

Cheers,
Thomas

[Carrington]

Hi Thomas,

For example, if you do an "open" calibration, you should place the probes approximately at the distance from each other at which you expect to measure the (small capacitance) capacitor.

It does not come with a compensation block?
http://cp.literature.agilent.com/litweb/pdf/16000-95103.pdf

I hardly used the SMD tweezers so far so no comment on them.

These SMD tweezers are a agilent 16334A clone, and they are not cheap.



Cheers.

[EU1]

Hi EU1 -

the instrument works pretty well and has a very good resolution of the low end range for capacitance and inductance which was basically the reason for me to get it. At 100kHz test frequency, the resolution is 1nH and 1fF, respectively. The least significant digit "dithers" by approx. two to three counts. Obviously, in this range, the influence of positioning of the probes is considerable. For example, if you do an "open" calibration, you should place the probes approximately at the distance from each other at which you expect to measure the (small capacitance) capacitor. When measuring inductances, a straight piece of wire of 5mm clearly and reproducably changes the reading. I cannot tell for sure how accurate the results are in this very low L / C range but simple tests like doubling the length of a straight wire doubles the inductance reading appears to be quite resonable.

The four-wire probes, though they appear to be of simple, even primitive design, serve their purpose very well and despite their appearance, I cannot tell anything bad about them. I hardly used the SMD tweezers so far so no comment on them. What I like a lot is the possiblity to measure very small resistances (preferably in the low frequency range of course) accurately. That's something a classic multimeter completely fails at. High resistances are not that much a domain of this pure AC meter. Capacitive effects appear to affect the readings at high resistances (>10MR). Anyway, the device won't measure above 100MR. For these high values, I anyway use a high voltage insulation tester (megger).

Thank you for the information, it was really helpful for me to make a decision. Just ordered one

[Hydrawerk]

What the hell? There are about 35 ICs in the AT826. If the soldering is crappy, how reliable can this be?

I prefer my DT-9935. (Yes, it has no external power supply and no 4-wire test leads.) It has acceptable pcb quality for the price. https://picasaweb.google.com/106264218831814439783/CEMDT9935LCRMeter#

[EU1]

I don't know if Applent offers firmware updates - so far I didn't find any. But nevertheless there might be a way to "upgrade" the lower frequency versions to 100kHz. Yet, I'm not sure how much calibration will be necessary to get it accurate and if it's worth the hassle, especially since the price difference isn't that dramatic anymore.

I also noticed that there are different opamps on AT825 and AT826 photos...

[EU1]

Has anybody measured its accuracy at 120Hz?
At 100/1K/10K/100K my unit seems to be pretty precise. It looks like the actual Z and Thd errors are less than a half of the specified (I haven't checked all the ranges though).

But at 120 Hz Thd accuracy is barely within the specs or rather even slightly out of the specs at the range 1. Test report includes only 100/1K/10K/100K frequencies, so I'm not sure is it calibrated at 120Hz at all?

Update:
Thd means Theta, which accuracy is specified in the section 9.2.3 of the user manual. I can't insert the letter itself because it seems that the forum doesn't support Unicode.

[TurboTom]

I can confirm this behavior on mine as well. The "metric" ranges are spot-on but 120Hz appears to be not as accurate. Maybe Applent uses the same set of calibration constants for 100Hz and 120Hz and prefers to have the 100Hz range accurate?

Anyway, since this instrument is pure AC, measurements sometimes need some interpretation since all losses (in magnetic cores, dielectrics and so on) manifest themselves as ohmic resistances in the result. This may lead to funny (and sometimes not quite easily understandable) readings when changing measurement frequencies. Moreover, stray capacitance and resonances sometimes also lead to unexpected results.

But nevertheless, the longer I own this instrument, the more I like it since it permits "views" on components that are just not possible with a plain multimeter. ...and the display simply is terrrific! 

Cheers,
Thomas

[EU1]

Thank you Thomas.
Their test report includes only 100 Hz, 1 kHz, 10 kHz, and 100 kHz, and these ranges are calibrated very well.
Does your meter shows higher capacitance and much higher D at 120 Hz?

I used unknown Chinese 220 nF polypropylene capacitors. I don't know the manufacturer, but they are really good despite the fact that they are cheap. Below are some measurements.

4 caps in parallel:
100 Hz: 869.9nF, D=0.0003 (Thd=-89.95 degrees)
120 Hz: 870.4nF, D=0.0005 (Thd=-89.92)
1 kHz: 869.7..869.8nF, D=0.0000 (Thd=-90.00)

2 caps in series:
100 Hz: 108.16nF, D=0.0000..0.0001 (Thd=-90.00..-89.99)
120 Hz: 108.25nF, D=0.0013 (Thd=-89.92-89.93)
1 kHz: 108.10nF, D=0.0002 (Thd=-89.97)

4 caps in series:
100 Hz: 54.34nF, D=0.0003 (Thd=-89.98)
120 Hz: 54.39nF, D=0.0020 (Thd=-89.89) - too bad. By specs Thd accuracy is +/-0.1 degrees for this impedance range.
1 kHz: 54.31nF, D=0.0002 (Thd=-89.97)
10 kHz: 54.31nF, D=0.0003 (Thd=-89.98)
100 kHz: 54.31nF, D=0.0024 (Thd=-89.86)


I'm impressed by 1k/10k/100k readings (and I'm impressed even more by these cheap SX PP caps).
120 Hz test frequency looks useless, especially when compared with other test frequencies. Anyway, it's not a great loss.

[EU1]

I prefer my DT-9935. (Yes, it has no external power supply and no 4-wire test leads.)

What about real phase angle/D accuracy? Per the manual, DF accuracy when measuring a capacitance is +/-0.0500 + few percents. There is another parameter - "D value Accuracy", which is +/-0.0090...0.0100. I don't know which one is correct, or what is the difference between these parameters?

[TurboTom]

Okay, I tested a few caps from my "bits and pieces bin" as well, not as systematic as you "EU1", but anyway...

1. High precision Siemens Styroflex 22nF 630V 2,5%
 100Hz: 22.007nF Thd -90.00°
 120Hz: 22.032nF Thd -89.86°
 1kHz: 22.001nF Thd -89.99°
10kHz: 22.002nF Thd -90.00°
100kHz: 22.017nF Thd -89.96°

2. Power MKP Siemens 0.68µF 630V 10%
 100Hz: 710.7nF Thd -89.98°
 120Hz: 711.1nF Thd -89.92°
 1kHz: 710.6nF Thd -90.00°
10kHz: 710.4nF Thd -89.98°
100kHz: 708.0nF Thd -89.77°

3. Precision Styroflex (brand/voltage unknown) 10nF 2,5%
 100Hz: 10.125nF Thd -90.00°
 120Hz: 10.136nF Thd -89.96°
 1kHz: 10.122nF Thd -89.99°
10kHz: 10.122nF Thd -89.99°
100kHz: 10.121nF Thd -89.99°

4. Something really ugly - Siemens MKL 0,33µF 63V 20%
 100Hz: 340.68nF Thd -89.50°
 120Hz: 340.63nF Thd -89.49°
 1kHz: 335.11nF Thd -89.18°
10kHz: 326.75nF Thd -88.80°
100kHz: 316.08nF Thd -88.03°

5. And finally - the "mother" of all capacitors (at least in my basement) --
AVX Power MKP 3200µF 1000V 10% (always stored with a shorting wire around its terminals for good reason)
http://download.siliconexpert.com/pdfs/2011/2/1/5/9/0/864/avx_/auto/fflc.pdf
 100Hz: 3155µF Thd -89.95°
 120Hz: 3155µF Thd -89.93°
 1kHz: 3220µF Thd -89.4°
10kHz: -------
100kHz: -------

The 0.1µF digit was dithering by about 10 counts so I didn't report it here. I guess round about that order of capacitance the useful range of the instrument is reached.

So the 120Hz range is also not as accurate as the "decimal" ranges though the effect appears to be more prominent in inductance mode. Due to the lack of precision inductors in my basement, and the higher dependance on frequency of their values, giving accurate figures here is more difficult.

The dissipation factor is simply calculated from the phase angle (D = 1/tan(Thd)) so the accuracies follow the same relation. Btw, "Thd" and "Thr" are the same, only one time specified in degrees and the other time in radian.

Cheers,
Thomas

[EU1]

Okay, I tested a few caps from my "bits and pieces bin" as well, not as systematic as you "EU1", but anyway...

I don't have small polystyrene capacitors, but I've tries with other caps, and my results were almost exactly the same. Almost the same capacitance and angle error at 120Hz for each capacitance value