Large discrepancy in measured and specified capacitances when soldered onto PCB?

Thread Starter

SiCEngineer

Joined May 22, 2019
366
Hi,

Not sure if best to post in general chat or power electronics, but:

I am having some significant issues when measuring my power converter capacitances with a hand-held LCR meter than I expect from the specification. I have a series-parallel combination of capacitors that should give an effective capacitance of 30nF. I have attached a schematic.

However, when the components are placed on the board, they are each reading a capacitance of approximately 17.5nF. I was confused because this significantly higher value isn't just one component, but all of them. I also have a 330uF electrolytic capacitor at my power converter input alongside some 0.1uF and 1uF decoupling capacitors, but the total input capacitance when measured with the LCR meter is about 20uF. It is a combination of electrolytic and smaller MLCC capacitors, but no more than 20% tolerance from the specification.

I also have a 120R resistor, which is 1% tolerance, showing up as 98R. Some more 1nF capacitors, 5 in parallel, which is reading a capacitance of >8nF. I have tested a different set of capacitors, also C0G from KEMET with an impedance tester, which confirms a capacitance of 10nF. I have not tested the ones soldered on my board yet, but plan to do so next week when I am back at work.

Can anyone suggest any reasons why a hand-held LCR meter could give such wildly different readings when components are soldered to the PCB? Some kind of layout issue, perhaps? Any suggestions are welcomed.

Note: I tested the 10nF C0G capacitor with the hand-held LCR, it is showing 10nF as expected, and agrees with the impedance tester, when not soldered to the PCB. So this leads me to believe that there is something else going on "inside the bonnet" when they are soldered which starts to effect their effective capacitance?

Thanks,
SiC
 

ronsimpson

Joined Oct 7, 2019
1,751
There is a thought that soldering the parts damaged them. (I think not) Please remove some of the parts and measure them off the board.

I think the meter is not happy with something on the board. (meter problem)
 

Thread Starter

SiCEngineer

Joined May 22, 2019
366
There is a thought that soldering the parts damaged them. (I think not) Please remove some of the parts and measure them off the board.

I think the meter is not happy with something on the board. (meter problem)
I also believe meter issue is more likely but it does seem to measure *some* components correctly. It is much more consistent when off the board. I did think it could be damage but I don’t think every single capacitor could be damaged, maybe just one or two, so I was surprised to see that every capacitor showed the same value of capacitance despite it being >70% above stated tolerance for a C0G part. I also didn’t solder the parts myself but got a technician at my university to do it. So I would hope they would be well soldered. Thanks for the reply!
 

KeithWalker

Joined Jul 10, 2017
2,095
When a component is measured in-circuit, all the other components on the board that are connected directly and indirectly through series, parallel and combined circuits to it will effect the value measured. If you need to know the accuracy of the components you are using, measure them before you connect them in circuit. When assembled, just test the functionality of the circuit to make sure it is doing what it should.
 

Thread Starter

SiCEngineer

Joined May 22, 2019
366
When a component is measured in-circuit, all the other components on the board that are connected directly and indirectly through series, parallel and combined circuits to it will effect the value measured. If you need to know the accuracy of the components you are using, measure them before you connect them in circuit. When assembled, just test the functionality of the circuit to make sure it is doing what it should.
So it can effect the measurement but not the component itself? I had this suspicion but I was only measuring across one of the capacitors to ground, so thought it would still only measure the capacitance between those terminals. Actually the circuit works OK as far as I can tell - but it’s almost as if the power delivered is less than it should be because of the lower resonant frequency caused by the higher “measured” capacitance. I’ll measure them off board next week and get back to this thread! Thanks!
 
Top