Some remarks about the measurements with the new CTester running 1.51m, after measuring some components and callibrated profile:
Examples of measuring capacitors on different firmware.
Interesting: you are measuring the same capacitor on each CTester of each photo, right? because in each case the values are different, but on the same range, 'though. However, on the bottom right of the C100 photo it shows a value ten or twenty times higher than the others, like what I'm seeing now with 1.51m or 1.46m vs 1.13k.
So this (re)opens a new/old discussion: in the short term, the best known practice for configuration, to obtain a closer to reality value that gives us the idea on whether to replace or keep the DUT. And in the long term, whether there is something it could be optimized on the firmware
Note: We know that a real LCR Meter and a CTester use different methods for ESR assesment, but the difference is really high in some cases.
For low value capacitors it's better to use an LCR/ESR meter to measure the ESR. The k-firmware does a better job than the m-firmware in that regard, but the measurement error increases as the capacitance value decreases (for both firmwares, Karl-Heinz' documentation tries to explain this). There's another observation. The measurement algorithm delivers ESR results comparable to a 10kHz test signal of an LCR meter. With low capacitance values the ESR results seem to be increasingly converging to the real frequency of the measurement algorithm. For example, the m-firmware measures an ESR of about 5 Ohms for a 100nF film cap. An LCR meter shows 1 Ohm @ 1 kHz and 15 Ohms @ 120Hz.
I see, so in your opinion, the values currently measured by the m-firmware would be closer to the real life usage?
(For my use case, I normally work with switched power supplies).
Got JLC making von Boleslaw J. (boleslaw_43) Adapter_328-644_106_quartzw there still in production, takes like 2-3 weeks before I get them. they or 5 on a part, so 25 total! anyone who want's one, there free for the taking (20) just pay shipping (self return stamped envelope) C1 100nf, R1 680 ohm, C2 C3 22p
I see, so in your opinion, the values currently measured by the m-firmware would be closer to the real life usage?
(For my use case, I normally work with switched power supplies).
For electrolytic capacitors, the difference in ESR readings is minimal. For film ones, the k-firmware readings are closest. Shown in the previous message. For ceramic capacitors, see for yourself...
It Looks Good. Someone will sell this?
I suppose carl1961 is talking about a new version of the following adapter:
Adapter to upgrade the T-Tester with ATMEGA328 DIP28 to ATMEGA644 TQFP.
Idea and implementation by Boleslaw Jamroz from Poland,
I have only developed a circuit board for the cumbersome threading adapter
that simplifies the construction much ;-)
https://www.mikrocontroller.net/topic/transistortester-avr?goto=7398047#7381379
I see, so in your opinion, the values currently measured by the m-firmware would be closer to the real life usage?
The ESR values for low capacitances match an AC test signal of around 600 Hz, while for higher capacitances it's roughly 10 kHz. However, it's about comparing ESR vales at the same test frequency. Let's say you're repairing a buck converter running at 250 kHz, but your LCR meter's highest test frequency is 100 kHz. So you have to stick with the 100 kHz and compare a possibly bad electrolytic to a good one at that frequency.
Sure thing, but for my hobby I don't have a real LCR (high frequency) meter, so I'm using the pulse-train-based budget CTester clones (k-firmware, m-firmware) to get a rule of thumb on whether I shall replace or not a capacitor (most of times electrolitic for through-hole power supplies).
What got my attention is the tenfold difference estimation on each firmware branch, I haven't paid attention to that until know (as the modded m-firmware clones ran quickly from my hands). If you tell me you're trying to assimilate the values to those we would obtain from an LCR, that would be enough reasoning for me. But it you tell me "k-firmware does a better job on that regard" (moreover for lower capacitance, IIUC) I wonder why you chosed to go that path? Of course there should be reasons for that, but I also wonder why we would compromise for less accuracy if we know better, again, if that's the case.
Anyways, what about the Vloss and/or Iloss, if the values are accurate enough, we could judge based on that, or better on the the whole-picture, I suppose.
I suppose carl1961 is talking about a new version of the following adapter:
Adapter to upgrade the T-Tester with ATMEGA328 DIP28 to ATMEGA644 TQFP.
Idea and implementation by Boleslaw Jamroz from Poland,
I have only developed a circuit board for the cumbersome threading adapter
that simplifies the construction much ;-)
https://www.mikrocontroller.net/topic/transistortester-avr?goto=7398047#7381379
Yes, he posted the grbl files for v106 the v108 is SMT crystal page 32 along with the files for AY-AT ComponentTester-1.51m_mod_files_for_AY-AT.zip Page 31 for mega644_T7_AY_AT-Mod_v113k.zip Lot's of good reading from the designers! a "must to read" to know better about these testers
FWIW, I ran the ESR test through a series of film Capacitors. The trend seems logical to me, what I was discussing was the differences (including slightly different slope).
FWIW, I ran the ESR test through a series of film Capacitors. The trend seems logical to me, what I was discussing was the differences
At the beginning of the work on upgrading the k-firmware, I compiled similar graphs. I hoped to find the dependence of ESR on capacitance, but when analyzing different types and models of capacitors, I realized that the idea was a dead end. A larger capacitance may have a higher ESR value. Here is an example of measurements for capacitors of 1 µF and 0.47 µF.
It's always a good idea to have the micro socketed.
And a few spares around too.
Just got one of these:
https://www.ebay.com/itm/142849416622?var=441847192858
Seems to be a new variant of the AY-AT board (8MHz) with Atmel chip. Once calibrated it is very accurate... surprisingly accurate if you ask!
That was the description of the one I got from Amazon, but mine has the ATP32F172K8T6 Chinese chip. Same clear DIY case.
At the beginning of the work on upgrading the k-firmware, I compiled similar graphs. I hoped to find the dependence of ESR on capacitance
I read some post you made about this topic on mid-2022, and I tried to retake that line to prove a)there's a relationship, and more important to me, b)the difference between k-firmware and m-firmware approach.
For the first test, I selected some Capacitors from the same vendor, same batch, same voltage range (63V-100V), and the results were very nice IMO. For the second test, I took a blind approach (like the practical scenario for the repair shop): different vendors, different ages, different voltages, so I expected high discrepancies, which in fact ocurred but not enough to discard the idea IMHO. I suppose if somebody repeat the exercise with same brand and voltage Capacitors, a more evident trend can be obtained. Lastly, I notice the measurements differences between k-firmware and m-firmware, the values are more similar for this escenario (being the m the more consistent), but the slope of m-firmware is higher, confirming the observations of the first test.
Just got one of these:
https://www.ebay.com/itm/142849416622?var=441847192858
Seems to be a new variant of the AY-AT board (8MHz) with Atmel chip. Once calibrated it is very accurate... surprisingly accurate if you ask!
That was the description of the one I got from Amazon, but mine has the ATP32F172K8T6 Chinese chip. Same clear DIY case.
The kit includes 2% 1/8W resistors. I have checked them all and had to substitute some of them because they were beyond the 2% limit. I bet they're factory rejects to reduce costs...
The chip seems to be a legit ATMEL but nowadays who knows...
As I said in my previous port, once calibrated it is quite precise. I compared results with my DER DEE-5000 for resistors, capacitors and inductors and with my Peak Atlas DCA55 for semiconductors.
I fried my TC1 by charged capacitor. I have replaced the damaged components. The DIP MCU is also damaged. Is any firmware available for TC1 with atmega 328p? Attached picture is similar to my unit. I am not an expert, it would be great if there is hex file available....
Some types of film capacitors have specific frequency characteristics. For comparison, I showed these characteristics in comparison with the readings of the k-firmware. The points on the graphs correspond to frequencies of 1,5,10,20,40,60,80,100 kHz
Absolutely, that's why is important -if we want to find trends- select same type of Capacitor, same Voltage range, and if possible, same vendor and batch (which I did for my first graph and not for the second).
For low value capacitors it's better to use an LCR/ESR meter to measure the ESR.
I fully support that opinion. The use of a transistor tester to evaluate the quality of ceramic, film and other similar types is purely indicative and can very often misjudge such capacitors. There's no point in citing specs and graphs. It's easy enough to compare a few of the different types I listed above with a professional LCR meter to realize this. The difference between real ESR and tester readings can be 10-1000 times! Capacitance and ESR measurements a transistor tester are purely bonus in nature and are only good for monitoring ELECTROLYTIC capacitors.
I personally find the tan(delta) or D (dissipation factor) a more convenient measure, as it is independent on the capacitance. Also, given it is specified in the datasheets, it becomes a more certain method to evaluate the state of a capacitor.