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

[floobydust]

Lifted the 24C02 EEPROM's SDA, SCL pins and attached wires. Used an Arduino and 3V level translator.
Bytes match a Russian AN8008 dump, aside from the CAL factors.

The DMM IC is full featured but the multimeter's hardware (rotary switch, die-bonding PCB) prevents their use. LCD has all annunciators though. I don't have the exact datasheet for the DMM IC, just its predecessors.

RS-232 output - I can't see the IC TXD pin bonded.
hFE test - needs transistor socket and 3 resistors but could be done from the front banana jacks.
Mid-range amps - not possible as rotary switch does not bypass 99R (uA) shunt. I still hear Dave yelling about this gyp. Could add a switch, lol.
Temperature - I enabled in EEPROM on the mV function, works °C/°F and factory CJC CAL is already set at 25°C fixed.
NCV - could be used with antenna 10MEG resistor added.
Clamp-on CT - could add if you enter CAL factors.

Presently only two buttons: SEL/HOLD and RANGE/BL.
To add MAX/MIN & RELATIVE buttons- I haven't checked if that MCU pin available; they are muxed.

So you can enable the temp function (although shown here it's advertised in Russia; my meter did not have the function). 
Also you can change timer settings for BL and power off etc. or change calibration.
Factory alarms set to 1,000VDC and 750VAC. OL at 1,100VDC and 800VAC.

It was a fun adventure.

[Gandalf_Sr]

Flooby, seems like you're much deeper into the guts of the AN8008 than most of us - well done!  I'd like to make a couple of mods when mine gets here so I'm interested in your thoughts:

1. Change Amps to mA - I know you've stated that it can be done by changing shunt resistors but can the EEPROM also be hacked to give the correct display?  I don't want to measure Amps, just mA and uA.
2. Take out the elCheapo electrolytic Cap on the 3 volt rail and put in 1206 caps as suggested by you a couple of pages back (search he thread for "1206")
Any other low-hanging fruit?

[plazma]

RS-232 output - I can't see the IC TXD pin bonded.

How did you check this?
For one AN8002 unit I removed die potting epoxy to reveal pin 20 pcb pad. I added a pull-up resistor and measured the signal. Nothing was visible. I sacrificed this unit and removed the potting epoxy to reveal the bonding wires and the die. The pin was unbonded.

[bitseeker]

1. Change Amps to mA - I know you've stated that it can be done by changing shunt resistors but can the EEPROM also be hacked to give the correct display?  I don't want to measure Amps, just mA and uA.

+1 Seems the best mod for electronics use.

Any other low-hanging fruit?

Temp? I wasn't sure if that really worked. Some posts said yes, I think one said it didn't.

[ocw]

1. Change Amps to mA - I know you've stated that it can be done by changing shunt resistors but can the EEPROM also be hacked to give the correct display?  I don't want to measure Amps, just mA and uA.

I changed the 0.01 ohm shunt resistor for the AmA jack to 0.095 ohms and now with the leads in the COM and AmA jacks, but the AN8008 rotary switch in the uA position, I measure mA current up to 99.99 mA with 0.01 mA resolution and 0.1 mA resolution up to 999.9 mA with the decimal in the correct location.  999.9 mA is the maximum current which can be read now.  The mA readings are now typically 0.35% high.  I haven't tried to change the meter's software yet.  Others might need a slightly different value shunt resistor.

[Gandalf_Sr]

1. Change Amps to mA - I know you've stated that it can be done by changing shunt resistors but can the EEPROM also be hacked to give the correct display?  I don't want to measure Amps, just mA and uA.

I changed the 0.01 ohm shunt resistor for the AmA jack to 0.095 ohms and now with the leads in the COM and AmA jacks, but the AN8008 rotary switch in the uA position, I measure mA current up to 99.99 mA with 0.01 mA resolution and 0.1 mA resolution up to 999.9 mA with the decimal in the correct location.  999.9 mA is the maximum current which can be read now.  The mA readings are now typically 0.35% high.  I haven't tried to change the meter's software yet.  Others might need a slightly different value shunt resistor.

Thanks, what size is that 0.095 Ohm resistor? (my AN8008 hasn't arrived yet).

[ocw]

Thanks, what size is that 0.095 Ohm resistor?

While a resistive wire resistor like a 0.1 ohm IRC OAR1R100FLF would be best due to its 20 PPM/C temperature coefficient, they are not available in odd values.  Because of that I went with two 0.0475 ohm Vishay RCWH080547L5FNEA 200 PPM/C 1/3W resistors in series.  With the one amp current maximum with that modification they would each dissipate 0.0475 watts and would have a collective burden voltage of 0.095 volts.  You need remove some solder mask to conveniently install the new resistors.

You would probably end up with a lower temperature coefficient using the mentioned 0.1 ohm resistor with a 2 ohm resistor in parallel with it.  This might make it easier to fine tune the resistance to "calibrate" it.  A 10% value change in the 2 ohm resistor value will only change the parallel resistance by 0.43%.  I might try that.

[floobydust]

Missing are the 9.999mA and 99.99mA ranges.

Right now, it has 10mR and 100R shunts only. The hardware (range switch) needs to switch in R24 1R shunt (J4/R14 in the generic schematic) to the input jack. If you look at the PCB you will see nothing connected to R23/R24 junction.

Then you need the MCU firmware to enable this function, and it has separate CAL factors for the ranges. The MCU firmware probably supports this- the predecessor 6,000 count DMM IC has those ranges under function code 0EH, 0DH which is not used in my AN8008 EEPROM dump. The predecessor IC has no 60.00uA (99.99uA) range but the new IC does.
I haven't had time to try it.

Notice the meter has Hz/%duty cycle twice on the rotary switch. I think they screwed up on the design a bit.

[floobydust]

RS-232 output - I can't see the IC TXD pin bonded.

How did you check this?
For one AN8002 unit I removed die potting epoxy to reveal pin 20 pcb pad. I added a pull-up resistor and measured the signal. Nothing was visible. I sacrificed this unit and removed the potting epoxy to reveal the bonding wires and the die. The pin was unbonded.

For lowest cost, I assumed unused pins would not be bonded out. TXD would be very useful for testing.
Here is the region on the PCB I guessed TXD would be brought out. I do not see any traces there, however I do not have the exact die pinout. I'm not using the QFP pinout.
What are you using for a solvent on the encapsulant? TXD may very well be accessible.

[plazma]

RS-232 output - I can't see the IC TXD pin bonded.

How did you check this?
For one AN8002 unit I removed die potting epoxy to reveal pin 20 pcb pad. I added a pull-up resistor and measured the signal. Nothing was visible. I sacrificed this unit and removed the potting epoxy to reveal the bonding wires and the die. The pin was unbonded.

For lowest cost, I assumed unused pins would not be bonded out. TXD would be very useful for testing.
Here is the region on the PCB I guessed TXD would be brought out. I do not see any traces there, however I do not have the exact die pinout. I'm not using the QFP pinout.
What are you using for a solvent on the encapsulant? TXD may very well be accessible.

There is a IC (QFP) footprint under the encapsulant. The die is bonded to the inner end of the pads. So you can carefully remove the encapsulant to reveal the outer end of the pads without harming the bonding wires. I just used hot air to soften the encapsulant and removed it carefully with a surgical knife. The TX pad is the fourth from the top on the left side in your image. Check the DTM0660 datasheet pinout.

[ocw]

Missing are the 9.999mA and 99.99mA ranges.

Changing to the 0.01 ohm resistor to 0.095 ohms adds the 9.99 mA (yes, 9.999 would be better) and 99.99 mA to the 999.9 mA range while removing the 9.999A range with the decimal point in the correct position when the meter switch is left in the uA position but the AmA jack is used instead.

Right now, it has 10mR and 100R shunts only.

My AN8008 does not have a 100 ohm resistor.  It has a 99 ohm resistor in series with a 1 ohm resistor.  The 1 ohm resistor is there for use (with some switching).

[floobydust]

There is a IC (QFP) footprint under the encapsulant. The die is bonded to the inner end of the pads. So you can carefully remove the encapsulant to reveal the outer end of the pads without harming the bonding wires. I just used hot air to soften the encapsulant and removed it carefully with a surgical knife. The TX pad is the fourth from the top on the left side in your image. Check the DTM0660 datasheet pinout.

Ok sounds good, maybe post a pic I have no idea what it looks like under the blob.
Part two to getting serial data output is hitting the RS-232 button. Oh wait we don't have one 
The button matrix is 2x3 mux and "REL/RS-232" button should be PT1.0 pin 42 to EEPROM SDA, both traces are available.
The AN8008 has oddball button assignments in firmware.

[plazma]

There is a IC (QFP) footprint under the encapsulant. The die is bonded to the inner end of the pads. So you can carefully remove the encapsulant to reveal the outer end of the pads without harming the bonding wires. I just used hot air to soften the encapsulant and removed it carefully with a surgical knife. The TX pad is the fourth from the top on the left side in your image. Check the DTM0660 datasheet pinout.

Ok sounds good, maybe post a pic I have no idea what it looks like under the blob.
Part two to getting serial data output is hitting the RS-232 button. Oh wait we don't have one 
The button matrix is 2x3 mux and "REL/RS-232" button should be PT1.0 pin 42 to EEPROM SDA, both traces are available.
The AN8008 has oddball button assignments in firmware.

Here is a picture of the AN8002 IC.

[floobydust]

Thanks for the pics of what is under there     For some reason the LQFP-64 and die versions have different pinouts 

I notice the max. counts is programmable with some people changing upper limit 4000, 6000, 9999 counts etc. in the EEPROM for the 8002, 210, 890's etc.
That means you can crank up the A/D counts, although the Vref on 8002 is not super good.

Some guy's  Arduino sketch to change EEPROM. I did my own sketch, not pretty enough to post.

[tronde]

Why do you think the pinout is different? The Victor VC921 is based on DTM0660 and it is die based. It also has visible pads (they used less epoxy than on the AN8002). The pin numbers matches the data sheet.


The DTM0660 chip is sold as 4000/6000 counts with 9999 for capacitance and frequency. You can increase the counts, but it seems like the AC/DC converter fails above some 7500 counts because of the crest factor. 9999 seems to be OK for DC-only.

I guess the main difference between DTM0660 and the new chip in AN8008 is how they implemented the AC/DC converter. It is believed to be a digital signal processor.

The picture of the VC921 is from kazus.ru

[floobydust]

The DTM0660 datasheet gives two schematics for a 3V multimeter; 13.1 QFP and 13.2 with die pinout. The pin numbers are much different, some pins are joined 2-1, at the corners or power pins.

HY12P65 has simple true-RMS DSP built-in as in the block diagram; the MCU at 4MHz cannot do much math on its own, and only 6KB code space.

AN8008 true RMS is rumored to not work with a DC bias on a sine-wave. I have not checked this.

[tronde]

The DTM0660 datasheet gives two schematics for a 3V multimeter; 13.1 QFP and 13.2 with die pinout. The pin numbers are much different, some pins are joined 2-1, at the corners or power pins.

They don't differ that much. The numbers are different, yes, but they follow the same sequence on both. The die has three more pads, but three pins on the 64-pin is connected to two pads each on the die. Else they look similar.

[evava]

Now, I have got the meter.

Continuity slower than I expected, uV resolution has unneeded gap around zero (shows only 0 or -5uV or +5uV  - nothing between), toy like.
No competitor to UT61e IMHO.

Price corresponds fully to usefulness.

Not very happy, wasted money for me.

Thank you all who have been trying in direction data logging, I read your contributions like detective story 

[ocw]

I changed my replacement for the AN8008 0.01 ohm current shunt resistor to a better option.  I put an IRC OAR1R100FLF 0.1 ohm resistor in its place and added a KOA SR733ATTE2R10F 2.1 ohm resistor it parallel with it.  The metric 6332 size of the resistor fits nicely with the existing trace after a bit of trace mask is removed.  Using the meter's AmA jack with the meter's switch in the uA position MY meter now reads XXX.X and XX.XX mA readings better than 0.1% accuracy.  Slightly less accuracy is available due to the resolution limit for X.XX mA readings.  Using the meter that way has the decimal point is in the correct position with a maximum current reading of 999.9 mA.

Due to the 1% resistors involved, other meters might require that a 2.2 or 2.0 ohm resistor be used in place of the mentioned 2.1 ohm resistor to achieve similar accuracy on your meter.  The approximately 5% change in resistor value will only change the parallel resistance by about 0.23%.  So, an even greater change in the value of the 2.1 ohm resistor might be needed.

[Fungus]

MY meter now reads XXX.X and XX.XX mA readings with less than 0.1% accuracy.

Less than 1% accuracy? You mean more than 1% error?

Due to the 1% resistors involved, other meters might require that a 2.2 or 2.0 ohm resistor be used in place of the mentioned 2.1 ohm resistor to achieve similar accuracy on your meter.  The approximately 5% change in resistor value will only change the parallel resistance by about 0.23%.  So, an even greater change in the value of the 2.1 ohm resistor might be needed.

How about usiing a 2 Ohm, 25 turn trimmer? That should do the trick. 

[floobydust]

The 0.01R shunt takeoffs include the (high-current) copper trace on the PCB 
So I would expect drift at high currents are the trace heats up, as it is part of the shunt resistance.
Next time they could try for pseudo 4-wire on that...

[ocw]

Less than 1% accuracy? You mean more than 1% error?

Sorry, my mind was locked into the per cent figure being lower which is obviously greater accuracy.  I have corrected that to:
"better than 0.1% accuracy."
And, yes I mean better than 0.1% accuracy using a hand selected 1% resistor to provide the best accuracy.  I don't think that this model meter warrants any effort for better accuracy than that.

[Fungus]

And, yes I mean better than 0.1% accuracy using a hand selected 1% resistor to provide the best accuracy.  I don't think that this model meter warrants any effort for better accuracy than that.

My comment was a bit of a joke, those 2 Ohm trimmers cost as much as the meter.

[evava]

I wonder if that 8uV gap around zero (meter doesnt show values +-1uV +-2uV +-3uV +-4uV, instead of it only shows 0 or +-5uV or more) could be eliminated or at least reduced?
Is that set in EEPROM, or is it invariably written in firmware?
Is it possible to find out?

Otherwise it seems to be stable enough, I see no need to hide these small values (+-1uV +-2uV +-3uV +-4uV).

[Kalvin]

uV resolution has unneeded gap around zero (shows only 0 or -5uV or +5uV  - nothing between), toy like.

I have two of these meters and the other one shows some intermittent fluctuating 5 uV - 10 uV offset voltage in DC mV range when the probes are short-circuited at room temperature. Now that I test this, the offset voltage displays 0uV, 5uV, 6uV, 7uV and 8uV, but no values between 1uV - 4uV. I guess there is some numerical hack so that the display is rounded down to 0 if the measurement value is below 5uV in order to hide this tiny offset voltage and avoid customer complaints. I can live with this, but it is good thing to keep in mind when measuring these tiny voltages.