[Amps] LCR meter.

[Jim Thomson]

Date: Tue, 30 Mar 2010 19:21:27 -0700
From: "Jack/W6NF" <vhfplus at gmail.com>
Subject: Re: [Amps] LCR meter

>
> > Here are some real numbers I just measured on a network analyzer
> > calibrated at the plane of the core being measured.
> >
> > Core = T94-2 powdered iron u=10, 22 turns #18 occupying about 80% of core
> >
> > Msmt. freq.     L(measured)
> >
> > 50KHz           3.06uH
> > 100KHz          3.86uH
> > 200KHz          4.53uH
> > 500KHz          4.51uH
> > 1MHz            4.50uH
> > 5MHz            4.50uH
> > 10MHz           4.58uH
> > 20MHz           4.89uH
> > 30MHz           5.52uH
> > 40MHz           6.80uH
> > 68.1MHz         parallel resonant freq.

###  I tried using the inductance measuring capability of the MFJ-259B..  with dismal
results.   A  10 uh commercially made roller inductor,  wound   with 1/4"  tubing... measures
exactly 10 uh  on my B+K 875-B  lcr meter. Same deal on my 50 uh roller.  When the mfj
was used, the same coil measures a LOT less than 10 uh.... as low as 5 uh.....depending on
what freq the mfj is on. Even with the mfj set  to 1.8 mhz.... the  same coil is way  <  10 uh !!!  

##  If I design  a simple PI net  for a tube amp using GM3SEK's  software.. and it  spits out
various values  for  C1... L... C2.... and then build it, using the B+k  875B.... everything  comes out
dead on.    And that's using a resistor  between  anode and chassis, to simulate plate load  Z....
[and mfj on output of C2].    Tune and load tweaked till mfj reads 1:1 SWR. 

##  If instead, the MFJ 259 is used  in it's inductance measuring mode, to tap the coil  for the various
bands... the inductance, comes out on the HIGH side  vs the 875B.   What happens, is the C1  and C2
caps both have to be DECREASED  in value to compensate for the too high uh used on each band. 
This of course, results in decreased  loaded Q.  

## The mfj inductance measurements  are just the opposite of the  results  in the above tabulated 
chart, which used a network analyzer.  The mfj-259  shows lower uh values.. as freq goes UP.  

##  the B+k  875A/B  uses  1 khz  for the measuring freq.   I measured some other commercially made
coils  with it.. and the 875 is dead on with the stamped value of the commercial coil. 

## OK, now maybe the commercially made coil maker used a similar lcr meter  to wind his coils  with.
Now if the coils are actually out to lunch in value... and a 10 uh  coil is actually 6 uh on the  mfj-259...
 OR  is actually way higher than 10 uh  per the  network analyzer above.... then  the  give away/ proof
would be that the tune  + load caps would be WAY off value  from the  spread sheet. 

##  well, the tune /load / coil  measurements  are dead on.. when the 875A/B  is used.. and compared
to the PI  spread sheet.  All other  test gear  throws the pi net way off. 

## I checked a LOT  of vac caps with the 875.. and also  stuff like  HEC  ht-50-58-59  TX  doorknobs,  cardwell
air variable  tune and load caps, broadcast variables,  HV lytics , etc.   Even tried plane jane coax , and measured 
pf  between  center conductor and braid... and it's all dead on.    If I want a screw up.. either the mfj-259  or the network
analyzer used in the above example  will  do the job of messing things  up.  

##  we also used the 875  to  build all the  various tuned inputs over the years, both the  type that have 5-9 x 
individual  tuned inputs.. and also the tapped  coil  +  2 x broadcast variable types.   The  LC networks  were also 
built using the 875 [ used to stack either  2 x identical yagis.... or  3 x identical yagis]   Ie:  50 ohms in... and  either
25 ohms out.. or  16.66 ohms out.  We  used Bing's  Rf software  to design the various LC networks, and then used
the 875 to build em.   They came out dead on the 1st time round. 

##Until  I find something that works  better, which is unlikely, I'm going to stick  with what I know  currently works.
This is  for  160-6m  stuff.......  nothing higher   than 6m. 

later....... Jim  VE7RF