The title is pretty self-explanatory I guess. I'm trying to measure resistance and capacitance without desoldering.

 

No.

When you try to measure a component in a circuit, you are not measuring just the single component. You are measuring the effect of the entire circuit connected to the single component.

You only need to desolder and lift off one leg of a 2-legged component such as a resistor, capacitor, or diode.

 

No.

When you try to measure a component in a circuit, you are not measuring just the single component. You are measuring the effect of the entire circuit connected to the single component.

You only need to desolder and lift off one leg of a 2-legged component such as a resistor, capacitor, or diode.

Thank you for the desoldering tip. I've read with an esr meter it's possible to measure the equivalent series resistance of a capacitor while connected. Why aren't there instruments for capacitance, or even measuring resistors?

 

Some meters say they measure in circuit. That depends on many factors. Say you are measuring a 0.1uF cap on the 5V supply. There are 30 of these caps across the 5V supply and some 10uF caps and maybe more. The meter will measure all of them as one. In that case the answer is no.
When a meter measures it must put out some voltage. If a meter puts out 5V to measure it will turn on some of the transistors and ICs. Some meters measure with a voltage that is below 0.5V so most silicon will not respond and mess up the measurement.
Measuring in circuit is very problematic but might work.

 

Thank you for the desoldering tip. I've read with an esr meter it's possible to measure the equivalent series resistance of a capacitor while connected. Why aren't there instruments for capacitance, or even measuring resistors?

Think about it. Say you have a circuit that has a 235 Ω resistor that is made up of two 470 Ω resistors in parallel.

Now you use your Acme In-circuit Resistance Measurement to measure the resistance of one of the resistors.

Next, you want to us it to measure the effective resistance of both resistors in parallel.

What would you do different in order to get a reading of 470 Ω the first time and 235 Ω the second time?

 

In a circuit is a circuit test. Component tests are component.

 

Here are many inexpensive LCR tweezers on sale these days. Their test voltage is 0.1-1V, and the measurement frequency is 100Hz-100kHz. You can buy one and try it out.
The ZOYI ZT-MD2 supports 0.1V and 100kHz.

 

If doing component tests in a circuit, most times the test results will be skewed, because of the circuit stuff around it, more so when there's passive components around the dut. Always best to have a schematic on-hand to validate dut test results.

 

To minimize or eliminate the influence of other components, LCR meters maintain a low measurement voltage of 0.1-0.3 V. Schottky diodes with a low voltage drop can be particularly problematic. To facilitate measurement conditions, it's best to switch to 0.1 V.

 

The title is pretty self-explanatory I guess. I'm trying to measure resistance and capacitance without desoldering.

Hi,

I'll try to make this as clear as possible...

There are two cases:
1. You can measure a single component value without desoldering anything.
2. You cannot measure a single component value without desoldering anything.

So sometimes you can and sometimes you cannot measure without desoldering something.

To determine if you have #1 or #2 you have to know the circuit in detail.

For example, with two resistors in parallel WHEN YOU DO NOT KNOW there are two in parallel, when you measure one resistor you are actually measuring the effects causes by both resistors, and thus you will ALMOST never get the right reading. This is case #2.

On the other hand, if you have a resistor in series with a capacitor, once the capacitor charges up you can get a decent reading on the resistor as long as the cap leakage is not high relative to the value of the resistor. This is case #1.

Now with the first example you almost never get the right reading. You may get CLOSE if the two resistors vary widely in resistance. If one is 10 Ohms and the other is 10k Ohms, you will measure mostly the 10 Ohm resistor. What you really measure though is both resistors (the parallel resistance).

The bottom line is you have to know the circuit or else it is hard to be sure.

A 3rd case might be where you can read the value of the component (like 1k) and you actually measure that value (1k) with a meter. In most cases you can assume that you are measuring just that one resistor, but it's still a chance you are taking.

When you have a schematic you can study that and see what components you can measure in-circuit and which ones you need to desolder. Yeah, desoldering is a pain in the neck and you risk lifting the pads and traces and losing contact between top and bottom traces or even more with multilevel circuit boards. If the traces lose contact with any vias, it's hard to figure out sometimes and may be hard to fix because you can't solder vias that are buried deep into the PC board material.

 

My answer is "yes BUT"!!! And only some of the times.
The "BUT" is that the measurement is neither accurate nor simple. BUT it works much better if you can see the whole circuit. Then it is sometimes simple to understand the effect of the rest of the circuit.
I have successfully located open and shorted diodes in an un-energised power supply, and located open resistors as well.

 

Here is another "yes BUT" situation.

I once had to diagnose a fault in a computer where there was a short between the supply rails, i.e. between Vcc and GND.
So I measured low resistance, less than 1Ω between Vcc and GND at every IC that was connected to Vcc and GND. Now there were literally hundreds of ICs and capacitors connected between Vcc and GND. The low resistance measurement was indicating that there was a fault condition. I just needed to find the one component that was the culprit.

To make a long story short, there was a decoupling capacitor that was shorting across Vcc and GND.

 

My answer is "yes BUT"!!! And only some of the times.
The "BUT" is that the measurement is neither accurate nor simple. BUT it works much better if you can see the whole circuit. Then it is sometimes simple to understand the effect of the rest of the circuit.
I have successfully located open and shorted diodes in an un-energised power supply, and located open resistors as well.

Hi,

Yeah it depends on a number of things like the failure mode. If a resistor burns up, that's easy to see. If a cap leaks, that's easy to see. If a resistor is marked 10 Ohms and it reads 1000 Ohms across in circuit, you know it's bad, but if it reads 2 Ohms you still can't be sure because of the possibility of other components in parallel with it.

I found a shorted diode in a TV set one time because it read close to zero Ohms, in circuit, and it was too low to be any other component in parallel.

Since I also worked in R&D for a while, I found some unusual problems because of design flaws. The most unusual and kind of hard to figure out was a synchronous binary counter vs an asynchronous binary counter. Synchronous counters need a clock, and if that clock depends on noise and there is no noise for a time, during that time the circuit won't work in that it will not do anything without that clock.
Amazingly, this actually came up TWO times over some 10 or 20 years. One app was an older design of a precision electronic weigh scale, and the other was for a solar panel max power tracker circuit.

 

Hi,

Yeah it depends on a number of things like the failure mode. If a resistor burns up, that's easy to see. If a cap leaks, that's easy to see. If a resistor is marked 10 Ohms and it reads 1000 Ohms across in circuit, you know it's bad, but if it reads 2 Ohms you still can't be sure because of the possibility of other components in parallel with it.

I found a shorted diode in a TV set one time because it read close to zero Ohms, in circuit, and it was too low to be any other component in parallel.

Since I also worked in R&D for a while, I found some unusual problems because of design flaws. The most unusual and kind of hard to figure out was a synchronous binary counter vs an asynchronous binary counter. Synchronous counters need a clock, and if that clock depends on noise and there is no noise for a time, during that time the circuit won't work in that it will not do anything without that clock.
Amazingly, this actually came up TWO times over some 10 or 20 years. One app was an older design of a precision electronic weigh scale, and the other was for a solar panel max power tracker circuit.

Amazingly enough I did run into a problem in such a counter a whole lot of years ago. Counters are mostly either ripple-through for the count, synchronous counters set-up the conditions and all of the stages advance when the next count pulse (clock) arrives. As the frequency rises, conditions develop so that the count may not completely ripple through before the count period is over. At counts of even hundreds of Hz, it would occasionally lose 100 counts. Fortunately a more experienced engineer was able to rapidly recognize the problem. .

 

Amazingly enough I did run into a problem in such a counter a whole lot of years ago. Counters are mostly either ripple-through for the count, synchronous counters set-up the conditions and all of the stages advance when the next count pulse (clock) arrives. As the frequency rises, conditions develop so that the count may not completely ripple through before the count period is over. At counts of even hundreds of Hz, it would occasionally lose 100 counts. Fortunately a more experienced engineer was able to rapidly recognize the problem. .

Hi,

That's interesting because I did a frequency counter years and years ago using ripple carry counters. It was 8 digits! It was 8 digits because I wanted to use it for other types of measurements too. 8 digits means 32 flip flops (ha ha) that have to carry over to the next, so you could imagine the delay that imposed. As the frequency went up, the count would be off by 10 counts at times. That's not real bad, but if I had used synchronous counters it would not be bad at all. Unfortunately, the packages available for the sync counters do not have the same pinout so it's not possible to just swap out the chips for new ones. I did use chip sockets, but the whole thing was hand wired so I'd have to do some rewiring.
Because of the sockets I was able to upgrade to the LS version TTL once they became widely available, and that saved a lot of power so the heat sink ran much cooler.