AN8008 US $19, 9999count, 1uV, 0.01uA, 0.01Ohm, 1pF resolution meter

[Fungus]

Do the meters actually work differently or is it just a misprint on the range selector?

[ciccio]

Reading Mark Hennessy post above they work differently, and from  my post, my photo and the photos usually found about everywhere there are at least two different silk screens.
I think that what happened may be:
1) they design the correct multimeter with correct firmware and front panel graphics and they sell the first batch,
2) later, somebody make a mistake in programming a lot of said meters, enabling Hz and Duty Cycle on V position,
2b) they sell them , including one to Mark Hennessy,
3) customers complain and the Boss call Enginereing
4) two decisions are taken: a: program new boards with correct firmware and b: design and print modified panels for already programmed boards in stock 
5) somebody in production mixes everything.

I've seen similar stories many times

[Neytron]

My AN8002 measures frequency in the "VDC" position but only up to 99.99Khz (I do not know why!), to measure further I need to change to the Hz% position.

To add the Hz% function to the "VDC" position, write in the direction 0xAF = 0x12

[Nusa]

My AN8002 measures frequency in the "VDC" position but only up to 99.99Khz (I do not know why!), to measure further I need to change to the Hz% position.

To add the Hz% function to the "VDC" position, write in the direction 0xAF = 0x12

The AN8008 "manual" (the folded bit of paper that comes with it) describes the difference between the two positions as "high voltage low frequency" and "low voltage high frequency". Presumably it's the same if enabled on the AN8002.

[GreyWoolfe]

I received my Aneng An-8008 today.  Since I am only interested in low\$\Omega\$ measurement and I didn't have an FRU handy, I checked with some low resistors I had.  Here is the result:

1R1 measured 1R13
5R1 measured 5R13
9R1 measured 9R18

All the resistors are 5% tolerance and I didn't use the OEM probes.  I replaced them with a set of Frankie's Brymen leads.  For my use, 15R is good, 30R isn't.  The meter is more than capable.  As I previously posted, this meter will never see voltage of any kind.  If the device on site will not power up, because of SLAs, it will be immediately replaced and repaired at home where I have much better meters that aren't afraid of mains voltage.  The other huge plus is the small size.  It is about 1/2 the size of my Velleman meter.  I find the stand is good enough, I can press the buttons without it falling over.  I really like the back light.  It is quite bright.  The leads are more flexible than one would think for the price.  They are almost as flexible as the Brymen leads.  There was a difference in measurement with them, however.  I only checked the 1.1\$\Omega\$ resistor and it measured 2.2\$\Omega\$.  That was out enough for me.  I never had any intention of using the original probes, Dave got me hooked on gold plated and I really like the Brymen leads, especially for the price.

[jackenhack]

Hi!
I was annoyed that the meter took several seconds to stabilize when doing measurements, so I knocked down the noise on the DC rail from 30.8 mV P-P to 3.6 mV P-P. Now the meter stabilizes immediately. Here's the modifications I did. https://www.jackenhack.com/aneng-an8008-modify-for-better-accuracy-faster-readings/. I've seen people mentioning problems with measuring low mA values. I didn't try ampere metering on my meter, only DC before modifying it, but when using an electronic load, I can go step by step in 0.1 mA up from zero without problems. No problems with µA either. How does the problem manifest itself?

[floobydust]

Same cap mods I did back in July https://www.eevblog.com/forum/testgear/an8008-us-$19-10000count-1uv-0-01ua-0-01ohm-resolution-meter/msg1261300/#msg1261300

[GreyWoolfe]

Hi!
I was annoyed that the meter took several seconds to stabilize when doing measurements, so I knocked down the noise on the DC rail from 30.8 mV P-P to 3.6 mV P-P. Now the meter stabilizes immediately. Here's the modifications I did. https://www.jackenhack.com/aneng-an8008-modify-for-better-accuracy-faster-readings/. I've seen people mentioning problems with measuring low mA values. I didn't try ampere metering on my meter, only DC before modifying it, but when using an electronic load, I can go step by step in 0.1 mA up from zero without problems. No problems with µA either. How does the problem manifest itself?

I have bookmarked your page.  I have one and am planning on getting another.  I have no smd caps here so it's time for a Mouser order as I need some other stuff also.  It's a shame that the local surplus store only sells SMD passives in full reels.  I don't want that many if they even have the proper values.

[stj]

you could use a broken harddrive.
they usually have 10 or 22uf ceramics on them.

[JohnPen]

In a fit of enthusiasm I decided to add extra decoupling to my AN8008 3v rail.  Thanks to Jackenhack I decided to progressively add a 1000pfd, a 0.01ufd, a 0.1ufd and finally a 4.7ufd. With each addition I observed the impact on the 3v rail with a scope.  The first 2 capacitors had very little effect but adding the 0.1ufd definitely improved things. Finally adding the 4.7ufd really improved the stability of readings and startup on the volts range was an almost immediate to zero (This was with no probe leads attached as pickup in the leads has an impact on the readings).   I left all the additional capacitors in place as well as the untouched original 100ufd  and the meter is now much more stable than before.  A very  worthwhile modifcation.

[jackenhack]

In a fit of enthusiasm I decided to add extra decoupling to my AN8008 3v rail.  Thanks to Jackenhack I decided to progressively add a 1000pfd, a 0.01ufd, a 0.1ufd and finally a 4.7ufd. With each addition I observed the impact on the 3v rail with a scope.  The first 2 capacitors had very little effect but adding the 0.1ufd definitely improved things. Finally adding the 4.7ufd really improved the stability of readings and startup on the volts range was an almost immediate to zero (This was with no probe leads attached as pickup in the leads has an impact on the readings).   I left all the additional capacitors in place as well as the untouched original 100ufd  and the meter is now much more stable than before.  A very  worthwhile modifcation.

Great to hear. All props goes to floobydust who put me on the right track.
The 0.1 µF capacitor definitely improved the high frequency noise, but to knock down the 1.1 KHz spike, I had to replace the 100 µF cap with a 1000 µF one. I guess the voltage reference have more problems with the higher frequency noise than the big ones at 1.1 KHz. Anyway, the voltage rail looks like when a CPU is doing micro-sleeps to save power, something is definitely injecting noise to the 3 V rail.

[GreyWoolfe]

you could use a broken harddrive.
they usually have 10 or 22uf ceramics on them.

The couple of hard drives I have floating around here are still good and are company property.  The only bad ones I had were also company property and I had to ship them to our main office for destruction.  The caps are cheap enough and I have other stuff i need for another couple of projects, so I will have an order that's worth it for the shipping.

[stj]

a quick cross-link to a discovery of interest in the 8001/2 thread.
https://www.eevblog.com/forum/testgear/aneng-an8001-6000-count-true-rms-multimeter/msg1322321/#msg1322321

[exe]

progressively add a 1000pfd, a 0.01ufd, a 0.1ufd and finally a 4.7ufd.

I think it's better not to mix more than two types of caps on the rail as they resonate. Detailed explanation: .

I ended up installing one 1uF polyester cap in addition what was already installed in my DMM. HF-noise mostly gone, still, there are dips 6.3mV p-p. I don't have a good electrolytic at hand .

One thing I don't understand is how this charge pump works. I'd expect to see two big caps, not one. So, it could worth upgrading another cap as well.

[kde]

One thing I don't understand is how this charge pump works. I'd expect to see two big caps, not one. So, it could worth upgrading another cap as well.

Why two big caps? You need just one big cap to maintain the output voltage (as I understand it is connected to pin 51 and ground) and one small cap to "pump" the big one (it is connected between pins 52 and 53). Since this pump capacitor is quite small, you can charge and discharge it very quickly with relatively low current. On the other hand, if you change the pump capacitor to a bigger one, the charge current increases too and it can generate more interfernce for the ADC. So I don't think it's a good idea to increase its value. Although it could be quite harmless to increase the other (big) one.
Regarding the principle of operation itself, I think it first charges the pump capacitor to the battery voltage, then IC re-connects it so that it's voltage is added to the battery and charges the result capacitor with this doubled voltage. Then the process repeats.

[exe]

Why two big caps?

Ah, last time I designed a charge pump the frequency was 50Hz (from mains)... I see, in this case this is not necessary.

[hughtmccullough]

Thanks to floobydust, jackenhack and JohnPen for their information on decoupling capacitors.

I notice you did your tests on the voltage measurement.  Did you see any improvements on the resistance ranges?  For instance, when you short the probes it normally takes several seconds to settle to the nominal value for zero resistance.  Does this improve on your meters now that you have upgraded the decoupling? 

[jackenhack]

Thanks to floobydust, jackenhack and JohnPen for their information on decoupling capacitors.

I notice you did your tests on the voltage measurement.  Did you see any improvements on the resistance ranges?  For instance, when you short the probes it normally takes several seconds to settle to the nominal value for zero resistance.  Does this improve on your meters now that you have upgraded the decoupling?

I'm almost afraid of saying it, because it could be wishful thinking, but I think the continuity tester is quicker. Could be that I also cleaned the test probes with isopropanol.

Haven't tried the resistance.

Edit
Tested shorting the lead and it takes around 10 seconds to stabilize at 70 milliOhm.

[JohnPen]

Unfortunately I didn't check the before settling time for the resistance range.  However now a short of the probe leads, the one's supplied with 8008, resulted in a reading of 60 milliohms after 5-6 seconds.  It was rather critical on how well one shorted the leads together as the probe plating tends to cause some contact variability affecting the final reading.  With regard to the continuity, again not checked before, but it does seem virtually instantaneous now.  Sadly my continuity check sound level is one of the quieter ones but it is usable so I will leave that alone.

[kde]

Does this improve on your meters now that you have upgraded the decoupling?

I'd done the same mod and I say no, it does not improve resistance at all for me. The measurement on shorted probes takes 10 - 15 seconds to settle down. Although it may depend on the batteries condition - I put a couple of new duracells in the meter, so they could be good enough with stock decoupling capacitors.

[hughtmccullough]

So, those results show that improving the decoupling doesn't speed up the response when measuring resistance.

I tried measuring a 390 ohm resistor using the AN8008 and at the same time measured the voltage across the resistor using another meter.  As soon as I connected the AN8008, the voltage across the resistor settled in less than one second to 32.60mV while the resistance value slowly converged towards 389.9 ohms.  So, the reference current through the resistor clearly settles quickly and has a value in this case of about 0.085mA.  That means the delay must be in the ADC measurement process.  Now, if another meter can read the voltage quickly and it the AN8008 can measure other voltages quickly, why would there be such a lag in the resistance measurement?

[kde]

That means the delay must be in the ADC measurement process.  Now, if another meter can read the voltage quickly and it the AN8008 can measure other voltages quickly, why would there be such a lag in the resistance measurement?

What if they average several successive ADC readings to get the final value? Some kind of software LPF.

[floobydust]

The DMM firmware has several digital filters with algorithm for the LCD displayed value.

If the input signal is moving around, the "fast" average is used.
Once the signal settles down, the "slow" average is used, for high accuracy. Takes a few seconds.
Both quantities are always being acquired, and the "slow" average" is primed (i.e. filter does not start from 0) by the "fast" value to speed response.
The "volatility" of the measured value determines which value to display. There is also hysteresis to avoid the LSD or rollover bouncing around.

It's one of the main reasons this meter has 9,999 count resolution - the firmware filters out a lot of A/D noise.

[001]

NEW MODEL 8009 RELEASED JUST NOW

New tread here https://www.eevblog.com/forum/testgear/latest-aneng-9999-dmm-invented!-an8009-now!/

[burkm]

The multi-million $ question is, is the new AN8009 just another model in the Aneng lineup or the successor to the AN8008, replacing it ?