It's really fascinating to get explanations for all those different measurement strategies the meters are using internally when my previous understanding from such topics is that accurate measurements require a lot of knowledge I don't have from multiple different physical phenomena including some quantum stuff.
Thermoelectric effect on resistance measurement and compensation.... Just Wow... You are really helping us to get better in the hardware design understanding... Thank You Sir
I once built a version of this--admittedly very low precision--to find a short circuit in my sister's car. Just an LM317 some resistors, and a switch to select 10 mA, 100 mA or 1A. Coupled with a cheap Radio Shack DVM, I got 1 mohm resolution, and was able to find the short quickly. Fixing the short, on the other hand, took a day upside down on the floor of the car rooting around under the dashboard.
This method is common and can be used even to provide nanoohm resolution, by using higher power current sources (MI have current sources to thousands of amps) and/or better and lower noise detectors. Keithley has specialized setups such as K6221+K2182A to perform this method in so called Delta measurement.
That's why I don't like working on a car, even on my own. Too many covers, latches, screws and whatnot to get to a single wire or contact. Once I finally get it open, it takes me a very long time to put it all back together.
When you perform the zero measurement, be sure to connect your clips with an interleaved connection stack up of HiSense, LoSense, HiDrive, LoDrive so the path of the 1A current doesn't flow through the voltage sensing path. Try a "proper" zero, then try connecting the probes in alternate stack up combinations to see the resultant errors. Also, when you are connecting to 4-terminal resistors, grab the pair of wires at each end with the clips so each of the 4 clip terminals touches only one of the 4 resistor terminals. Always use the AC current mode to eliminate the thermoelectric and offset voltage effects.
Guessing the purpose of the louder buzzer is for go/no go in a loud environment with hearing protection on. Seems to be set up for a factory environment, with the “😀:Over range” indicator in the bottom left.
I've made a portable DIY unit using an 18V Cordless tool battery connected to a DC-DC converter adjusted for CV 3V and CC 10A. With a couple of 2.5mm2 long leads and a DVM mV meter is ideal for chasing the micro ohms in 5kVA 24-48V battery systems or car winch systems etc. I've found 10A is ideal for testing those high-current relays and switches and surprising how relay contacts and high-current switches deteriorate and need replacing or servicing over time. Even testing new high current switches only to find a wide spread in resistance values, and reject some brands due to higher than expected performance.
Hi Shahriar, thank you for your effort! As usual, great content, also glad you discussed the thermoelectric measurement and compensation. It is a simple but elegant solution measuring directly the thermoelectric voltage, regardless of which materials are being used, and of temperature. At the risk of being picky, at 6:09, that 1% 100 milliohms resistor that measured 110.79 milliohms is actually quite off, it deviates 10%. Also, at 6:49, the 1% 10 milliohms is also off. It should deviate a max of 0.1 milliohm, so a max 9.9 to 10.1 milliohms. I also use these cheaper kelvin alligators, and as you mentioned, you can't use them properly with these 4 wire resistors. At these currents, this is probably the main source of error. Especially considering that later, that SMD measuring 303 microohms is within 1%. The SMD shunt benefits more from the kelvin alligator, since their sense terminals are effectively closer to their power terminals compared to the previous ones. But there is the chance that the instrument is out of cal. Or the slimmer chance that both resistors are off.
SourceTronic seem to rebadge Tonghui gear (as do many other companies), so this would be a rebadged TH2516. Based on comments on EEVblog forums the SourceTronic badged items do seem to cost more, but they also provide excellent support.
@@tarheels100 Haha, yeah I did spot that while looking carefully for a "TH2516" label on the PCB silkscreen. I knew it would be Tonghui based on the design and previous experience with their devices. I forgot to mention the PCB logo, but as you say it was kind of obvious.
@@douro20 I don't know if it is still the case, but I recall seeing info on the Tonghui website that indicated they were a subsidiary of BK Precision. Based on how many other companies rebadge the Tonghui models, I wonder if that is no longer the case as it would eat into their own sales. It could also be a result of regional distribution rights which do not mean much for low volume customers.
I was expecting at least some discussion of the accuracy differences between measuring milliohms between the ST2516 and the 4338B. Yes, I did hear the thermal diagnostic advantage at the end, but what about the actual resistance measurements?
@@ickipoo You'll be surprised, it apparently messes up the resonance that allows it to get that loud. Ever tried it when hacking an old smoke detector?
Certainly possible. Trapping the air inside the cavity will quiet it down significantly. But this solution will never fail and I got an awesome huge buzzer out of it! :)
I ran into similar issue a few weeks ago with a very cheap UNI-T multimeter, 96dBA is just too much. I put a tape over that buzzer, which did almost nothing, then I tried to glue piece of wood on it, which reduced that thing to 82dBA, which is some improvement, but it's still too much, I tried almost everything, but replacing that buzzer with the "normal" one was the only solution I found.
You mean: you measure the resistor using a (large) current, that heats up the resistor. Is that being compensated? No it is not. But my guess is that this effect will be tiny as the power dissipation inside the resistor will be small. For example, the 10 mOhm resistor was measured using 1 A, that means P=I^2 * R = 1^2 * 0.01 = 10 milli Watt, which is not a lot. The resistor will warm up but not by a lot (how much? See datasheet for thermal resistance). Also, the resistors Shariar is using here will have a very small temperature coefficient meaning their value will change very little, how much can be found in their datasheet.
Usually autozero in DMM means that internally meter will alternate between measuring your input signal and internal short to remove drift offset errors. Turning on/off current and measure residual TEMF is a different feature, usually called offset compensation or OCOMP in most bench top DMMs. Few meter go a step further and reverse current instead of turning off which can be important if you trying to measure resistances very accurately and maintain same power applied to DUT instead of pulsing current on and off.
I assume I am missing something, BUT, why the hell is it pushing high current all the time, and not pulsed (something like 20uS at 1% duty cycle), ALL THE TIME! Surely for the INSANE $ thing thing cost they could have precision op-amps stable in that time!
This is a DC measurement on purpose. Even the OVC runs as a very slow rate and can be turned off. As soon as you switch the current you will make measuring the resistance of large inductors impossible as the reactance begins to come into play.
Wow, that’s four pieces of content in two weeks! Outstanding. Brother, thank you for all the effort you put into this. I really appreciate it.
That transformer is a beauty for sure.
I bet Pooch is glad you took out that giant buzzer. 😺
Oh yes!
It's really fascinating to get explanations for all those different measurement strategies the meters are using internally when my previous understanding from such topics is that accurate measurements require a lot of knowledge I don't have from multiple different physical phenomena including some quantum stuff.
Thermoelectric effect on resistance measurement and compensation.... Just Wow... You are really helping us to get better in the hardware design understanding... Thank You Sir
I once built a version of this--admittedly very low precision--to find a short circuit in my sister's car. Just an LM317 some resistors, and a switch to select 10 mA, 100 mA or 1A. Coupled with a cheap Radio Shack DVM, I got 1 mohm resolution, and was able to find the short quickly. Fixing the short, on the other hand, took a day upside down on the floor of the car rooting around under the dashboard.
This method is common and can be used even to provide nanoohm resolution, by using higher power current sources (MI have current sources to thousands of amps) and/or better and lower noise detectors. Keithley has specialized setups such as K6221+K2182A to perform this method in so called Delta measurement.
That's why I don't like working on a car, even on my own. Too many covers, latches, screws and whatnot to get to a single wire or contact. Once I finally get it open, it takes me a very long time to put it all back together.
I literally shed a tear seeing so many uploads. I’m grateful for subscribing to this channel
When you perform the zero measurement, be sure to connect your clips with an interleaved connection stack up of HiSense, LoSense, HiDrive, LoDrive so the path of the 1A current doesn't flow through the voltage sensing path. Try a "proper" zero, then try connecting the probes in alternate stack up combinations to see the resultant errors. Also, when you are connecting to 4-terminal resistors, grab the pair of wires at each end with the clips so each of the 4 clip terminals touches only one of the 4 resistor terminals. Always use the AC current mode to eliminate the thermoelectric and offset voltage effects.
And now I know how OVC works - thank you!
Wonderful. I love experiments after the repair/teardown.
Gorgeous looking transformer !...cheers.
Guessing the purpose of the louder buzzer is for go/no go in a loud environment with hearing protection on. Seems to be set up for a factory environment, with the “😀:Over range” indicator in the bottom left.
I mean, they might also like Muons, so good that they clarified that.
I've made a portable DIY unit using an 18V Cordless tool battery connected to a DC-DC converter adjusted for CV 3V and CC 10A. With a couple of 2.5mm2 long leads and a DVM mV meter is ideal for chasing the micro ohms in 5kVA 24-48V battery systems or car winch systems etc. I've found 10A is ideal for testing those high-current relays and switches and surprising how relay contacts and high-current switches deteriorate and need replacing or servicing over time. Even testing new high current switches only to find a wide spread in resistance values, and reject some brands due to higher than expected performance.
Why would you reject brands that gave higher performance?
Hi Shahriar, thank you for your effort! As usual, great content, also glad you discussed the thermoelectric measurement and compensation. It is a simple but elegant solution measuring directly the thermoelectric voltage, regardless of which materials are being used, and of temperature.
At the risk of being picky, at 6:09, that 1% 100 milliohms resistor that measured 110.79 milliohms is actually quite off, it deviates 10%.
Also, at 6:49, the 1% 10 milliohms is also off. It should deviate a max of 0.1 milliohm, so a max 9.9 to 10.1 milliohms.
I also use these cheaper kelvin alligators, and as you mentioned, you can't use them properly with these 4 wire resistors. At these currents, this is probably the main source of error. Especially considering that later, that SMD measuring 303 microohms is within 1%. The SMD shunt benefits more from the kelvin alligator, since their sense terminals are effectively closer to their power terminals compared to the previous ones.
But there is the chance that the instrument is out of cal. Or the slimmer chance that both resistors are off.
A lot more instruments should have little smileys, that inform you of going over range. A lot more friendly then just a red led
SourceTronic seem to rebadge Tonghui gear (as do many other companies), so this would be a rebadged TH2516.
Based on comments on EEVblog forums the SourceTronic badged items do seem to cost more, but they also provide excellent support.
Nice, good to know.
What tipped you off, the large "Tonghui" printed on the PCB? 😁
There are some B&K Precision products which are made by them, like their very popular portable ESR meter.
@@tarheels100 Haha, yeah I did spot that while looking carefully for a "TH2516" label on the PCB silkscreen. I knew it would be Tonghui based on the design and previous experience with their devices. I forgot to mention the PCB logo, but as you say it was kind of obvious.
@@douro20 I don't know if it is still the case, but I recall seeing info on the Tonghui website that indicated they were a subsidiary of BK Precision. Based on how many other companies rebadge the Tonghui models, I wonder if that is no longer the case as it would eat into their own sales. It could also be a result of regional distribution rights which do not mean much for low volume customers.
I can only dream of owning such a device in my lab... maybe one day, maybe one day. Peace.
Very interesting
wow you have amazing skill Sir
ohm my gosh
If still too loud, stick a self adhesive label over the buzzer's hole.
nice, four-wire measurement.
I was expecting at least some discussion of the accuracy differences between measuring milliohms between the ST2516 and the 4338B. Yes, I did hear the thermal diagnostic advantage at the end, but what about the actual resistance measurements?
According to the PCB,it's from an instrument company of China named Tonghui(同惠),well-know for its LCR meter. Maybe this is an OEM/ODM product.
But what is its output voltage during measurement? It may happen that it is too high for some components.
It’s designed to remain very low. Below 40mV.
The voltage is technically "almost zero" because you are measuring low ohms.
It's from Tonghui.
是的
HI
DO HAVE ANY MULTIFUNCTION CALIBRATOR
It seems that this is a ratiometric measurement that will make possible to cancel the emf by ovc.
What is the value of the resistance when the probes are shorted together? شكرا على الفديو
The value is "null". Because it compensates for the resistance of the cables and connections, and is nulled (zero'd) before each measurement.
It would be nice to measure that 4-wire resistor with the kelvin clips open, moving the measurement plane inside the resistor.
WTF!! That's not a buzzer. That's an alarm siren!
dry circuit testing?
I would have tried taping over the buzzer to make it quieter before resorting to modifying the board.
110dB is rather more than a bit of tape can handle. Think jackhammer, rock concert or putting your head against a car horn...
@@ickipoo You'll be surprised, it apparently messes up the resonance that allows it to get that loud. Ever tried it when hacking an old smoke detector?
Certainly possible. Trapping the air inside the cavity will quiet it down significantly. But this solution will never fail and I got an awesome huge buzzer out of it! :)
I ran into similar issue a few weeks ago with a very cheap UNI-T multimeter, 96dBA is just too much. I put a tape over that buzzer, which did almost nothing, then I tried to glue piece of wood on it, which reduced that thing to 82dBA, which is some improvement, but it's still too much, I tried almost everything, but replacing that buzzer with the "normal" one was the only solution I found.
What about the temperature that will raise making the resistance shift value?
You mean: you measure the resistor using a (large) current, that heats up the resistor. Is that being compensated? No it is not. But my guess is that this effect will be tiny as the power dissipation inside the resistor will be small. For example, the 10 mOhm resistor was measured using 1 A, that means P=I^2 * R = 1^2 * 0.01 = 10 milli Watt, which is not a lot. The resistor will warm up but not by a lot (how much? See datasheet for thermal resistance). Also, the resistors Shariar is using here will have a very small temperature coefficient meaning their value will change very little, how much can be found in their datasheet.
hahaha that buzzer is HUGE
Believe it or not but I deal with 50µΩ +-2% shunts at work. So there totally is a need for something with this resolution!
I thought dmms with auto zero do the same thing to zero out TE voltages or am I mistaken ?
Usually autozero in DMM means that internally meter will alternate between measuring your input signal and internal short to remove drift offset errors. Turning on/off current and measure residual TEMF is a different feature, usually called offset compensation or OCOMP in most bench top DMMs. Few meter go a step further and reverse current instead of turning off which can be important if you trying to measure resistances very accurately and maintain same power applied to DUT instead of pulsing current on and off.
@@xDevscom_EE thanks for the detailed explanation!
Id just love to have some somewhat good meters that arent from 1969 or 1971 and half defective.
I assume I am missing something, BUT, why the hell is it pushing high current all the time, and not pulsed (something like 20uS at 1% duty cycle), ALL THE TIME! Surely for the INSANE $ thing thing cost they could have precision op-amps stable in that time!
This is a DC measurement on purpose. Even the OVC runs as a very slow rate and can be turned off. As soon as you switch the current you will make measuring the resistance of large inductors impossible as the reactance begins to come into play.
@@Thesignalpath Yep, I knew I was missing something, DAAAAH!!!! :)