Author Topic: $20 LCR ESR Transistor checker project (Read 3475463 times)

MosherIV · « **Reply #2925 on:** November 19, 2016, 10:28:05 am »

Quote

have one of those, and when I power it from a adjsutable power supply, it says "Bat. 10V OK". But i'm supllying it 9V, at least that's what's my multimeter (uni-t ut61E) says...

Is it something that could be a problem or it doesn't matter?

No problem, it is ok and bear in mind that the values are for indication inly (ie could be way off as you have found)

They have used a cheap zener diode instead of a voltage reference, which leads to the inaccuracy.
Another way to word around it is to run the calibration routine.
(If you replace the zener diode with a voltage reference, you must run the calibration routine)

madires · « **Reply #2926 on:** November 19, 2016, 01:12:21 pm »

If they've used a cheap zener as external reference all meassurements would be affected. Another cause could be the voltage divider for the battery voltage and some parts in the power section, i.e. a reverse voltage protection diode and a PNP. You can adjust the resistor values of the divider and the voltage offset in the source code.

stj · « **Reply #2927 on:** November 19, 2016, 02:43:34 pm »

that unit uses a TL431

madires · « **Reply #2928 on:** November 19, 2016, 04:08:02 pm »

I'd suggest to use a LM4040AIM3-2.5 (SOT-23) or LM4040AIZ-2.5 (TO-92). But that depends on the 5V LDO. The external voltage reference makes only sense if its tolerance is about 10 times better than the LDO's. A cheap TL431 got 1 or 2%.

stj · « **Reply #2929 on:** November 19, 2016, 06:46:18 pm »

the regulator on that unit is a Holtek HT7550,
it's got good specs.

hapless · « **Reply #2930 on:** November 19, 2016, 08:00:49 pm »

Quote from: siri_ on November 19, 2016, 01:03:50 am

Hi allI

have one of those, and when I power it from a adjsutable power supply, it says "Bat. 10V OK". But i'm supllying it 9V, at least that's what's my multimeter (uni-t ut61E) says...

Is it something that could be a problem or it doesn't matter?

Thanks!

Sounds very much like a low accuracy resistor in the voltage divider. Does this also happen when you connect a battery?

Quote from: MosherIV on November 19, 2016, 10:28:05 am

No problem, it is ok and bear in mind that the values are for indication inly (ie could be way off as you have found)
[...]

I wouldn't say that. If it's way off, the automatic shutoff will not work right, and when running on rechargeable battery power it could mean discharging the battery beyond its limit or using the unit with voltages below the expected 5V on V_CC.

upsss · « **Reply #2931 on:** November 19, 2016, 09:00:49 pm »

Hi, where can I find a compiled firmware for this tester with a 16MHz crystal?

Thanks

pepe10000 · « **Reply #2932 on:** November 20, 2016, 09:00:20 am »

Quote from: upsss on November 19, 2016, 09:00:49 pm

Hi, where can I find a compiled firmware for this tester with a 16MHz crystal?

Thanks

It's not complicated, you can compile it yourself.

If you need it urgently I enclose one that I have compiled myself version k (698) for crystal of 16MHZ, English language, without Pull-UP resistors and with measurement of activated quartz crystals.

Unzip with WinRar.

A greeting

upsss · « **Reply #2933 on:** November 21, 2016, 04:04:23 am »

Quote from: stj on October 09, 2016, 05:17:07 pm

16MHz does more accurate tests and can also test crystals and resonators upto about 10MHz

I am in the process of upgrading my crystal to 16MHZ and I am still confused. In the Karl-Heinz ver. 1.12k manual in the OP_MHZ configuration it says that "The 8 MHz operation is recommended for better resolution of capacity and inductance measurement", so what is it?

Gyro · « **Reply #2934 on:** November 21, 2016, 10:11:46 am »

At the time that Karl-Heinz wrote that, the choice was 1MHz or 8MHz.

upsss · « **Reply #2935 on:** November 21, 2016, 04:40:11 pm »

The 1.12k manual mentions all 3 frequencies, 1MHz, 8MHz and 16MHz with the quote I mentioned above!

madires · « **Reply #2936 on:** November 21, 2016, 06:46:42 pm »

The measurements of inductance and low value caps is based on measuring time. As higher the MCU clock as higher the resolution of the time measurement. And it also impacts the range limits. With an 1 MHz MCU clock the inductance measurement's lower limit might be around 100µH. With 8 MHz it's about 10µH. These values are for the standard measurement method.

upsss · « **Reply #2937 on:** November 21, 2016, 07:22:48 pm »

So what is the resolution and accuracy of the measurements with 16MHz versus 8MHz clock?

madires · « **Reply #2938 on:** November 21, 2016, 08:49:17 pm »

16 MHz provides twice the resolution of 8 MHz. Take the equations for the measurements and calculate the impact of the higher time resolution. The accuracy with 16 MHz might decrease a little bit in case the measured time is at the mid of a clock cycle of the 8 MHz clock (jitter).

PointyOintment · « **Reply #2939 on:** November 21, 2016, 10:11:15 pm »

Anybody know how these devices measure the reverse leakage current of diodes? I checked a diode with one last week and it said "Ir = 2 nA"! How??

madires · « **Reply #2940 on:** November 22, 2016, 11:56:00 am »

Very simple

A 470k ohms resistor as current shunt and an ADC resolution of about 1mV when using the 1.1V internal bandgap reference: 1mV / 470kOhms = 2nA

ardhuru · « **Reply #2941 on:** November 24, 2016, 06:20:25 pm »

Hi guys,

I'm new to AVRs, so please excuse me if I ask some stupid questions.

I wanted to make an absolutely minimal tester (USB powered, no LCD, UART output thru a USB-to-serial adapter, no crystal and Atmega8 based.

I chose ver 1.06k and compiled it for 8MHz, internal, uploaded it with an USBasp and all seemed well.

BUT, I find that the unit auto-repeats for a few cycles (during which it reads components very reliably), and then simply hangs.

I tried it on another Atmega8, very similar results; works for a few cycles, and then hangs. At this point, the only option is to interrupt the power, and then it works for a few cycles of testing again.

Could anybody point me in the right direction?

Many thanks in advance.

madires · « **Reply #2942 on:** November 24, 2016, 07:05:15 pm »

Please use the trunk version (1.12k) and read Karl-Heinz' documentation (see POWER_OFF option).

ardhuru · « **Reply #2943 on:** November 25, 2016, 07:13:33 am »

A thousand thanks! Brilliant!

Instead of using an actual pullup, I just recompiled by enabling the internal pull-ups and it worked flawlessly.

Which brings up the question, why is it optional/configurable? Are there any advantages to disabling the internal pullups?

I havent tried the version 1.12k per your suggestion yet, as I need to fit the code into an Atmega8, with the UART enabled, (and disabling the LCD related code) and I already use up 99.9% flash space with ver 1.06k, so I suspect the 1.12k would not fit.

Thanks again.

ardhuru · « **Reply #2944 on:** November 25, 2016, 07:37:25 am »

Okay, got the answer to that question in Chapter 5 of the manual. I havent seen much degradation in measurements, so I guess I'll stick with the internal pullups enabled.

lpc32 · « **Reply #2945 on:** November 27, 2016, 11:05:53 am »

Quote from: madires on October 17, 2016, 02:08:47 pm

A few posts ago flywheelz reported some funny ESR business with low value film caps and different MCU clock rates. I'm able to reproduce that and even have some extra fun.

So what's the recommended version for decent ESR measurement?

madires · « **Reply #2946 on:** November 27, 2016, 01:48:43 pm »

The latest k and m firmware versions

The issue is only related to film caps about 180 - 220nF, IIRC.

lpc32 · « **Reply #2947 on:** November 27, 2016, 02:50:22 pm »

You mean newer versions fix the clock speed dependency?

madires · « **Reply #2948 on:** November 27, 2016, 03:46:49 pm »

There's no fix for that yet (m-firmware). Unless you need to measure the ESR of a 180 - 220nF cap, you don't have to worry about this.

lpc32 · « **Reply #2949 on:** November 29, 2016, 12:13:51 am »

Thanks.

By the way, what were the changes in "trendy" vs "classic"? Is it now continuing on a single line based on just "trendy"?
And why, actually, does Karl-Heinz work on a separate branch instead of merging back to your mainline?


EEVblog Main Site	EEVblog on Youtube	EEVblog on Twitter	EEVblog on Facebook	EEVblog on Odysee

Author Topic: $20 LCR ESR Transistor checker project (Read 3475463 times)

Share me