Parts Express component measurements

[arc2v]

Just wanted to share something with the group regarding crossover components.

I had my test rig up in the theater room last night and after all the acoustic testing of the drivers in different baffles (I'm working on an OB speaker), I figured I'd test my passive crossover components once and for all.

I use Speaker Workshop and a modified Wallin / Claudio jig specifically designed around my M-Audio mobilePre and ECM 8000.

After calibrating, I got around to testing around 30 resistors, 20 capacitors, and 15 inductors. The resistors ranged from 1 ohm to 20 Ohms (within the range of my test jig). Capacitors were 1 uF to 50 uF, with most being between 1 and 20. And the inductors were from 0.025 mH to 2.2 mH, with almost all of them below 1 mH.

Resistors and caps were Parts Express Dayton brand. The caps were the cheap metalized poly. The inductors were the Jantzen 18 and 20 gauge varieties.

What I found astounded me. Almost every component measured spot on its nominal value. Occasionally the DCR of inductors were a little off, but everything was correct to the 3rd significant digit (usually the hundreths decimal place).

So despite the 5% tolerance listed on a lot of these, they were much tighten. Something to keep in mind before springing for the 1% ones.


The only disappointments were the electrolytics. The two I tested were 1 uF, 10% tolerance and both measured 1.2, or 20% out of spec.

So just FYI, but something to think about for your next passive crossover.

AC

[lochness]

test them again when they are really hot (been running for a few hours) or really cold (as bout as cold as your house would get with no heat during winter...assuming you live in a place that gets real winters). Also test them after they have time to burn in for a few days. I can guarantee that they will vary from their rated specs. Even the + 1% components vary over time, use, and conditions.
Honestly unless your going for exotic components that you can't order from places like Digikey you're better off paying the extra few cents to extra few $ for the + 1% components.

[Jay_WJ]

Originally posted by arc2v
What I found astounded me. Almost every component measured spot on its nominal value. Occasionally the DCR of inductors were a little off, but everything was correct to the 3rd significant digit (usually the hundreths decimal place).

So despite the 5% tolerance listed on a lot of these, they were much tighten. Something to keep in mind before springing for the 1% ones.

I had the same experience. Dayton components are very consistent and accurate. As for the cap, you don't need to pay more for the precision 1% tolerance caps. The standard one is sufficiently accurate.

I found that Solen caps are much less consistent.

Jantzen inductors? Amazing! Inductance is very, very accurate.

Madisound standard air cores are also accurate.

[zigzagflux]

Originally posted by lochness
Even the + 1% components vary over time, use, and conditions.
Honestly unless your going for exotic components that you can't order from places like Digikey you're better off paying the extra few cents to extra few $ for the + 1% components.

I don't follow your reasoning. If both 1% and 5% components vary over time, why buy 1% ? He has already stated that the 5% components match much better then 1% as new. As would the 1%. They start from the same baseline.

Typically, the only thing that separates a 5% from 1% component is measurement by the manufacturer. At least with WW resistors, capacitors, and inductors. In fact, you will find that most 1% resistors do NOT fall within 0.1%. They will often fall within the outside regions, with exact resistors being rare.

Any of my designs that require tight matching typically require better than 1%, so I find myself buying many and hand matching after a burn-in process. 5% or 1% as purchased makes no difference.

[owdi]

I have found measuring the resistance of inductors is very tricky, because it's usually low enough that any flaws in your jig make a big difference. You said the test range of your jig is 1 - 20 ohms, I bet most of your inductors were well under 1 ohm.

About a month ago I was measuring the impedance of a 16 awg erse steel core inductor. It was supposed to be 2.7mH, but when I measured it, it was 2.4mH. Shaking my head I proceeded to unwind it to my desired 2.2mH, and found the impedance started to rise! Eventually it measured at 2.6mH, before it started to fall again. My jig measures everything else close to spec, so I don't know how to explain this.

Dan

[augerpro]

The 5% Dayton caps are incredibly tight tolerance. I've also found their 10watt resistors to be spot on also, better than the more expensive Mills resistors.

[lochness]

I don't follow your reasoning. If both 1% and 5% components vary over time, why buy 1% ? He has already stated that the 5% components match much better then 1% as new. As would the 1%. They start from the same baseline.

Because the 1% won't vary as much over time and use. Also not everybody is patient enough/needs to have 100% exact values that they will buy a bunch burn them in and go from there. If you're going that rout then going with the 5% and greater is worthwhile.

[c2cthomas]

Hi Guys,

My experience working as a calibration tech in a Precision Measurement Equipment Lab (PMEL) taught me a couple of things about component measurements. To accurately measure precision components, you need precision equipment. Generally, the equipment required needs to be an order of 10 times more precise than whatever you are testing. 1% resistors would need to be measured with a piece of equipment that had .10% accuracy - and is calibrated to national standards. When you get into measuring very low values (ohms, volts, capacitance, inductance, etc.) things like lead resistance and capacitance get involved and the test scheme and the test procedure starts to get really exacting. I won't get into those details but there are books and published papers available around the net if that is if interest to you.

As far as audio goes not a big problem - use a decent meter / scope and you should be close enough. Heck - most of the stuff needs to be "tweaked" anyhow! That's half the fun of things!

[zigzagflux]

In general, your reasoning is correct, but it should be noted that often times it is not the absolute value that is critical, but the relative value. For instance, I don't care if the plate resistors in my LTP are exactly 47.00K. I am only concerned that the two are similar in measurement; so 45.91K and 45.92K is excellent matching.

As well, a particular meter might not have 0.1% accuracy, but when using a 1% accurate meter, and measuring two resistors 10 seconds apart, 0.1% repeatability is possible. So I could use a 1% meter and obtain better than 1% matching. I won't, of course, obtain 0.1% accuracy - only matching. Then again, a 1% meter doesn't have more than 3 digits, so I can't rely on 0.1% matching, but the general principle still applies - matching better than specified accuracy.

[c2cthomas]

Hi Zig,

I agree in general with you in regards to matching component values in this type of application. The process you are using is time proven to be an effective method. I especially agree with "burning in" the components to be matched and have found that 72 hours is a good method for this (that's from my old QC days).