TH2830 vs TH2832

[KungFuJosh]

Apparently this thing can take screenshots. They should have a dedicated key, oh well. You gotta hold down the period key.

Note: the 100pF cap is silver mica, the 2.2mF and 1.0mF caps are electrolytics.

[KungFuJosh]

Perusing the manual, I confirmed my suspicion that their version number is bullshit. They all say Ver1.0, and the only thing that changes is the date.

[luudee]

That is possible. There is also an internal EEPROM inside the LPC1788 which could have the model ID inside. Only way to find out is to write software into the chip that can dump the internal EEPROM.

The LPC1788 has 512KB of Flash and 4032 bytes of EEPROM.

I'm pretty sure I can dump the Flash (if the security Fuse is not set).

I am not sure about the EEPROM though ....

But I won't be able to do anything until the weekend ...


Cheers,
luudee

[nctnico]

EEPROM access requires to use the internal ROM routines from software in normal situations. I don't know if there is a tool which can be used to read the EEPROM over the JTAG interface.

[luudee]

EEPROM access requires to use the internal ROM routines from software in normal situations. I don't know if there is a tool which can be used to read the EEPROM over the JTAG interface.

Hi Nico,

yes, you are 100% right. This is from the LPC1788 User Manual:



Cheers,
luudee

[luudee]

So, unfortunately, some bad news.

I tried to read the internal Flash of the LPC1788, but could not.

I even created a special adopter, so I can look at the signals with my scope.
I also was able to confirm that the JTAG interface of the LPC1788 was connected
to the 10 pin connector (you can see the traces on the very top layer, going to pins
2 and 4, which are TDO and TDI ...

At this point, I believe that the security fuses have been engaged, and we will
not be able to access the internal FLASH.

If anybody has any ideas/suggestions, I would like to hear them ...

Cheers,
luudee

[KungFuJosh]

More bad news...

The sourcetronic firmwares both have measurement issues. Most measurements are good, but 100KΩ across the board from 100Hz to 200kHz, and 10KΩ at 1kHz are both bad. Off by as much as 50%. I tried ST2832, then ST2830, then TH2830. The only one that accurately measured those was the TH2830 firmware.

@Martin72, can you please test your ST2830 with 100K and 10K resistors at those frequencies and see if it works fine for you? Can you also post your firmware "copyright" date from startup?

I'll post my measurements after I enter them in Excel.

Thanks,
Josh

[KungFuJosh]

Here's the test results on the spreadsheet I setup.

Notes:
The ST42 test frequencies weren't stable, they never settled, so I left the range as the result.
The ST42 signal level is 1Vpp, so it reads 0.3586Vrms.

The failures are pretty consistent in the Sourcetronic firmwares, so I only noted the ST2832 results. The TH2830 results are there for comparison.

I'm not going to bother with L or C measurements for now.

Thanks,
Josh

[Martin72]

Hi Josh,

Could make the measures on monday at work, the last day before my holidays.
Here a snip from the last calibration (May 2023), unfortunately only at 100Hz/1kHz, so I´ll take my new reference board to work and measure with 10khz and 100khz.

[KungFuJosh]

Hi Josh,

Could make the measures on monday at work, the last day before my holidays.
Here a snip from the last calibration (May 2023), unfortunately only at 100Hz/1kHz, so I´ll take my new reference board to work and measure with 10khz and 100khz.

Awesome, thank you. It would also be very helpful if you can get the firmware date when it starts up; whatever the copyright says. I have a hunch they might have supplied outdated firmware files. I don't know if it was on purpose or not. I hope not.

Thanks,
Josh

[Martin72]

I tried ST2832, then ST2830, then TH2830. The only one that accurately measured those was the TH2830 firmware.

To me, stupid as I am, it sounds like the "foreign" firmwares can't find the calibration data, wherever that may be stored.
Or there must be a calibration, which is then stored and then also "found".

[KungFuJosh]

To me, stupid as I am, it sounds like the "foreign" firmwares can't find the calibration data, wherever that may be stored.
Or there must be a calibration, which is then stored and then also "found".

I understand the logic of that thought, but I don't agree for a couple reasons. First, the ST2830 had no issue with recalling data or previous states. Only the patched ST2832 couldn't recall previous states, but it did recall correction data (which is stored separately).

Besides that, I know this issue is related to outdated firmware because:

You may recall we had an issue with a scale on the TH2830 not reading correctly and reported to Tonghui, which they informed us that they had no authorized resellers on AliExpress where we purchased the TH2830. We contacted the reseller and they made contact with Tonghui and we were allowed to receive an updated firmware which corrected the problem. Recall we had a issue with the 1K range and got correct readings using DCR but not the other modes, worked OK below the 1K and above the 1K range, but not at 1K. You could try different ranges to see if this is a similar effect as we had.

Since you purchased from an authorized reseller have them contact Tonghui for the updated firmware. Hopefully this may correct your issue.

I'm looking forward to seeing what version your Sourcetronic firmware is. I'm confident that it's newer than the versions I have.

Thanks,
Josh

[Martin72]

Hmm...you might be right.
Lately a thread about a handheld lcr tester from east tester, it has problems with measuring resistance and this is a bug that needs to be solved.
OK, I take my reference to work and measure the resistance under various frequencies and record the version number.
Interesting thing !

[luudee]

So I have been following Josh's resistor measurements, and did some myself,
on my TH2832:

On My SDM3065, I did a 4 wire measurement of a random resistor. It reported it
to be 100.36 KOhm.

But my TH2832, reports:

200Khz:  93.36 KOhm
100Khz:  98.70 KOhm
 10Khz: 100.18 KOhm

But, then, I looked at the calibration report for resistors (see attached).
It looks like resistor measurements are only specified up to 10 KHz (and below).

So, may be what you are seeing, Josh, is normal operational behavior ?!

PLEASE SEE POSTS BELOW FOR CORRECTION !!!

Cheers,
luudee

[nctnico]

Did you do a open/short calibration? What are the accuracy specs for that resistance value at those test frequencies?

[luudee]

Did you do a open/short calibration? What are the accuracy specs for that resistance value at those test frequencies?

Hi Nico !

Indeed, I haven't calibrated it since I took it apart, thank you for keeping me honest!  ;-)

So, after calibration, things look much better (same resistor):

200 KHz: 100.35 KOhm
100 KHz: 100.33 KOhm
 10 KHz: 100.36 KOhm

Please see my previous post for accuracy specs in the Calibration certificate.

Cheers,
luudee

[nctnico]

 

One thing to keep in mind when using an LCR meter is that the reading depends on measuring amplitude and phase. At some point these measurements can no longer be made with a reasonable accuracy. Most LCR meters have a graph or formula in the manual that tells you what kind of accuracy to expect for a given component value. Looking at the calibration results, you can see they don't bother measuring 100k Ohm at 100kHz and the 30k Ohm allows for a larger error margin compared to the measurement at 10kHz.

[2N3055]

One thing to keep in mind when using an LCR meter is that the reading depends on measuring amplitude and phase. At some point these measurements can no longer be made with a reasonable accuracy. Most LCR meters have a graph or formula in the manual that tells you what kind of accuracy to expect for a given component value. Looking at the calibration results, you can see they don't bother measuring 100k Ohm at 100kHz and the 30k Ohm allows for a larger error margin compared to the measurement at 10kHz.

This.

For instance, capacity of electrolytic capacitors is routinely measured at 100/120 Hz. You NEED to measure them at that frequency to compare capacity to nominal.. At 100kHz 1000uF represent miliohms and is hard to measure.

Example diagram from old Digibridge:

[mawyatt]

Another thing to keep in mind here regarding 100KHz measurement of a 100K Resistor, is at +-0.1% that's the equivalent of ~+-0.016pF (16fF) capacitance change.

Remember this is in the entire setup including instrument, DUT, any cables, fixtures and whatnot, and quality OEM Kelvin cables help but don't eliminate this potential error source. Cable bending, placement and the position of the clip tips have effects, also Kelvin techniques do not circumvent DUT shunt effects such as distributed capacitance between Kelvin clip tips which are not shielded wrt a ground shield. This is why we prefer for LCR measurements, especially high impedance and upper frequency ranges, using a test fixture specific for the type DUT, rather than Kelvin leads, even the fixture for leaded DUT like the TH26048A has proven more reliable (repeatable) than Kelvin leads in our experience.

So extreme care is necessary for precision repeatable measurements of high impedance devices at 100KHz and above for these lab grade LCR meters. On our Hioki IM3536 for example, there's an entire 13 page Chapter (10.6) dedicated to just accuracy. It's very complex with 5 variables in addition to the DUT measurement parameter and range selected. These variables coefficients are for Level, Measurement Speed, Test Cable Length, DC Bias, and Temperature, and each of these variables coefficients is specified for 7 different frequency ranges from DC to 8MHz.

Best, 

[KungFuJosh]

That's all (at least basically 😉) understood, which is why I used SMD resistors on an SMD test fixture. I let the meter run for over an hour before applying correction (manual says at least 30 minutes).

While it's good information, it's not relevant here. The devices are clearly capable of reading 100KΩ at 100kHz and 200kHz:

200 KHz: 100.35 KOhm
100 KHz: 100.33 KOhm
 10 KHz: 100.36 KOhm

The buggy ST2832 firmware failed at a lot more than just high freq.
ST2832 100KΩ:
100Hz = 49.5052
1kHz = 67.6866
10kHz = 78.2236
100kHz = 98.1972
200kHz = 91.5881
ST2832 100K DCR = 100.057
ST2832 10kΩ @1kHz fail = 6.78268

Same hardware, 100KΩ, switched back to TH2830 firmware:
100Hz = 100.013
1kHz = 100.013
10kHz = 100.005
100kHz = 99.8795

[Martin72]

You NEED to measure them at that frequency to compare capacity to nominal.. At 100kHz 1000uF represent miliohms and is hard to measure.

And still there a people in the forum who don´t want to understand this or can´t.
It makes no sense measuring µF at 100khz, but they do and ranting about the "crappy meter"...

@nctnico:

This were my thoughts after reading the cal-protocol and wonder why they did´t test anything above 1kHz/10kHz, although the meter goes up to 100kHz.

[KungFuJosh]

One thing to keep in mind when using an LCR meter is that the reading depends on measuring amplitude and phase. At some point these measurements can no longer be made with a reasonable accuracy. Most LCR meters have a graph or formula in the manual that tells you what kind of accuracy to expect for a given component value. Looking at the calibration results, you can see they don't bother measuring 100k Ohm at 100kHz and the 30k Ohm allows for a larger error margin compared to the measurement at 10kHz.

@nctnico:

This were my thoughts after reading the cal-protocol and wonder why they did´t test anything above 1kHz/10kHz, although the meter goes up to 100kHz.

I think it's got more to do with either laziness or reducing the labor cost of calibration, and/or possibly whatever is actually required for specific cal standards.

The Tonghui cal didn't measure less than 100pF at 1kHz, they didn't measure capacitance below 100Hz at all, and they didn't measure beyond 1nF, 10nF, 100nF at any range besides 1kHz. That's clearly not about machine capability.

[nctnico]

One thing to keep in mind when using an LCR meter is that the reading depends on measuring amplitude and phase. At some point these measurements can no longer be made with a reasonable accuracy. Most LCR meters have a graph or formula in the manual that tells you what kind of accuracy to expect for a given component value. Looking at the calibration results, you can see they don't bother measuring 100k Ohm at 100kHz and the 30k Ohm allows for a larger error margin compared to the measurement at 10kHz.

@nctnico:

This were my thoughts after reading the cal-protocol and wonder why they did´t test anything above 1kHz/10kHz, although the meter goes up to 100kHz.

I think it's got more to do with either laziness or reducing the labor cost of calibration, and/or possibly whatever is actually required for specific cal standards.

The Tonghui cal didn't measure less than 100pF at 1kHz, they didn't measure capacitance below 100Hz at all, and they didn't measure beyond 1nF, 10nF, 100nF at any range besides 1kHz. That's clearly not about machine capability.

It is. See the graph 2N3055 posted. The manual of your LCR meter has similar graphs or at least a formula that can be used to derive such a graph.

In the end an LCR meter is designed to measure components accurately but can only do so if there is a measurable voltage across the DUT.

If you want to measure components over a wide frequency range, then a network analyser is a more suitable piece of equipment although the accuracy will be lower. For example: when doing Power Distribution Network measurements, you are measuring capacitors ranging from hundreds of pf to tens of uf and trace inductances (nano Henries) over a wide frequency range (up to several hundred MHz) in order to check the impedance of the power distribution network on a circuit board.

[Martin72]

The manual of your LCR meter has similar graphs or at least a formula that can be used to derive such a graph.

And it´s nearly always the same, I think because of the same measure principle.
Here the diagram of the ET4410, where I must confess that I have problems to read it right - too many crosslines...

[2N3055]

The manual of your LCR meter has similar graphs or at least a formula that can be used to derive such a graph.

And it´s nearly always the same, I think because of the same measure principle.
Here the diagram of the ET4410, where I must confess that I have problems to read it right - too many crosslines...

Yes, nomograms like this really are better when colors are used....

But they exist as a handy reminder what exactly the numbers are on top of user intuition where limits are approximately.

Basically, check what are limits of your LCR meter when measuring R. Those will be limits of measurement circuit. Then you know if test frequency is high, your L must not be higher impedance than what R you can measure. Or C cannot be smaller impedance then what R you can measure.
On the same token, if frequency is too low for small L it will be represented by very small impedance, or small C will be very high impedance...

Low frequency: good for large C and large L
High frequency : good for small C and small L


This nomogram is just nicely plotted that...