very accurate resistance & capacitance measurements

[K6JEK]

"All capacitors and resistors in the PS-2 are bridged to within .1%"

.1% yikes.  I thought my cheapie LCR bridge was accurate but it's not close to .1%.  What is?  Is there anything a mere mortal might get his hands on?

[KM1H]

I rebuilt one of those ages ago with a HP LCR meter we had where I worked. Im sure it wasnt even close to .1% but the end result when I plugged it in passed muster.

A couple of the other engineers said to quit wasting time and use a couple of IC's.

Carl
KM1H

[k4kyv]

I have a little ECG capacitance meter, an old Heathkit capacitor checker and a Fluke DVM.  The capacitance meter will measure picofarads down to the single digits and all the way up to hundreds of microfarads, and the accuracy seems very good at all ranges.  It works great for checking out the capacitance of bread slicers, and balance in split stator capacitors.  The only drawback is that great caution is required not to fry it by inadvertently trying to measure the capacitance of a charged capacitor, particularly electrolytics, since they can rebound and build back up a charge even after they have been discharged with a dead short.  I have noticed that oil caps can do the same thing.

I use the Heath checker for leakage tests only.  Once a capacitor checks good on the Heath tester, I switch the Fluke meter into the microamp mode and measure the actual leakage current.  A good capacitor will show less than 0.2 microamps leakage at its rated working voltage.

The Fluke seems to be very accurate on the ohms scale.  I periodically compare its readings to the nominal values of various 1% precision resistors in the junkbox, and it has always been spot on.

Just as the majority of old carbon composition resistors have drifted off value, most old wax paper capacitors have so much leakage that they would be better classified as resistors, than capacitors.

I don't always discard leaky capacitors, though.  For example, a cathode by-pass capacitor that still checks close to its nominal capacitance will work fine even with considerable leakage, since the low resistance of the cathode resistor swamps out any leakiness in the capacitor.

Where you don't want a leaky capacitor is the coupling capacitor between stages, as many 75A-4 owners have found out.  Even a leakage resistance of hundreds of megohms or a leakage current of microamps down into the single digits can cause the grid bias of a tube to shift a surprisingly amount.

[Jim, W5JO]

This piece is about as good as a small pocketbook can afford.  To measure to .1% would require the equipment to be checked against a standard two times more accurate.  Where would one find that.

http://www.aade.com/lcmeter.htm#specs

1% is quite good.  I have one of these, a couple of digital meters by somebody and a Sprague TO 6.  All work fine but reading the Sprague gives more error than I can see.  Are you sure the spec you are looking at is not 1% and the . is just an unintended spot on the paper? 

[WQ9E]

For very accurate resistance measurement, one of the older well built Wheatstone bridges is the ticket.

[Pete, WA2CWA]

"All capacitors and resistors in the PS-2 are bridged to within .1%"

.1% yikes.  I thought my cheapie LCR bridge was accurate but it's not close to .1%.  What is?  Is there anything a mere mortal might get his hands on?

What's a PS-2 

[K6JEK]

Thanks for the tips everyone.  What's a PS-2?  It's the audio phase shifter in a Central Electronics 100V.  It's a completely passive network of resistors and capacitors with unusual values.

[w3jn]

The AADE L/C meter is not a bridge.  Hence, it is relatively useless for finding bad components such as leaking capacitors, capacitors with high ESR, or coils whose Q has been wrecked due to leakage, bad cores, etc.

L/C Meter IIB works by measuring the shift in frequency caused by inserting an unknown into it's oscillator tank circuit. A PIC16C61 micro-controller measures the frequency before and after. It then computes the value of the unknown using a floating point math package and displays the result on a 16 character intelligent LCD display.

While it's certaily useful for sorting components and finding the value of new ones where you're pretty sure the component is good, it lacks the essential ability to display Q or D factor.  Also, note the caveat on testing inductors intended for 60/120Hz - won't do it.  Many bridges will test at various frequencies and this is essential if you're gonna test RF components and stuff like AC xformers, filter caps, etc.

Hamfests sometimes cough up really good digital RLC bridges.  I saw a GenRad DigiBridge go at Manasses hamfest for $10.  I got a HP 4261A for $50.  They're the cat's ass for ease and effectiveness.  The older manuual bridges such as Don's work FB too, but are harder to use, and lack the 4-terminal test lead capability that the more professional instruments use.

[KM1H]

Yep, used in the 100V and 200V. I had both but the 100V Id had since 65 was getting reports of very poor SB suppression by 83 or so. As mentioned the values used were all non standard carbon comp and block micas but of the old style components which allowed room for various combinations of smaller modern parts to fit.

Looked like hell but it worked.

The original 200V PS2 was still doing OK when I sold both rigs in the late 80's.

Carl
KM1H

[Pete, WA2CWA]

"All capacitors and resistors in the PS-2 are bridged to within .1%"

.1% yikes.  I thought my cheapie LCR bridge was accurate but it's not close to .1%.  What is?  Is there anything a mere mortal might get his hands on?

In both my 100V and 200V product brochures, and in my 100V manual (didn't check the 200V manual), it says, "The new PS-2 is a ten element type unit constructed of heat cycled components matched to .1% accuracy." Doesn't necessarily mean any of the components are ".1%". They could have been sampling and drawing components with 1% or even 5% tolerance. Of course, after 50 years, values could be all over the place. I rebuilt my 100V PS2 back in the late 90's with 1%, 5%, resistors, and several low value 5 turn PC type pots (to tweak the values closer to each other).
I had to do something similar to the RF phase shifting network in the Heath SB-10 back in the 80's.

[WD5JKO]

Hi,

  I am a CE 20A user, and I set up the PS-2 on that rig with a scope. The approach is illustrated at this link:

http://www.web-books.com/eLibrary/Engineering/Circuits/AC/02468.png

another image with article:
http://images.google.com/imgres?imgurl=https://pantherfile.uwm.edu/msw/www/AudioOsc/lissajous.jpg&imgrefurl=https://pantherfile.uwm.edu/msw/www/AudioOsc/index.html&usg=__bmX964A0JIrOaTVIS3h2tJ9r9kM=&h=329&w=634&sz=45&hl=en&start=6&um=1&itbs=1&tbnid=rK8kmTvWg0ExoM:&tbnh=71&tbnw=137&prev=/images%3Fq%3Dlissajous%2Bpattern%26um%3D1%26hl%3Den%26sa%3DN%26rlz%3D1C1GGLS_enUS300US303%26tbs%3Disch:1


   To do this you need a good low distortion sine wave source into the Mic input (or line input), a good dual trace scope, and two identical 10:1 properly compensated probes. Go for a perfect circle Lissajous pattern over 300-3000Hz (or higher if possible). Set scope for X:Y mode.

   On my simpler 20a this approach is better than trying to get the exact R or C in there and then praying that it works good enough.

   I also find that the same two probes with a 3 pf series cap is a good way to setup the 9 Mhz phase shifter 1st attempt.

Jim
WD5JKO