QST Antenna Tuner Measurements
Tom Rauch
The measurements of efficiency in the QST review of antenna tuners are
incorrect. An MFJ-259 analyzer was used to measure VSWR changes when
the load was varied 2 times and 1/2 of the tuned value. Also
correction factors were not applied to the MFJ-259. The 259 is NOT
capable of the resolution required for this type of test, it is NOT a
lab type instrument.
The theory was correct ansd the intentions were good, the instrument
was just used far beyond its design limitations.
The 259's limitations are two fold. The detector is never perfectly
centered at 50 ohms, and the output is not spectrally perfect.
Measurement of high Q circuits requires better than -30dBc generator
noise and harmonic suppression, and perfect bridge compensation. An
instrument like that would cost $10k ++, not $189.95. If I thought my
assortment of 259's would fit that application I'd never have spent
money for an HP Impedance Analyzer, Wiltron RF Analyzer, and Harris
and HP multi-channel vector voltmeters.
The mean value VSWR change the 259 must measure for a 2:1 load
impedance change to indicate correct loss and efficiency follows:
2.0 = 0 dB =100%
1.9=.31 dB =93%
1.8=.669dB=85.7%
1.7=1.09dB=77.8%
1.6=1.60dB=69.2%
1.5=2.22dB=60%
1.4=3.00dB=50.1%
In other words, an SWR reading of 1.8, when SWR is really 2.0, would
indicate a loss of 14.3%, or efficiency of 85.7%, if loss was really
zero. A VSWR reading error only from 1.9 to 1.8 means an 8% error in
loss measurements!
Using a Wiltron RF Analyzer, HP Network Analyzer, and Harris RF
Voltmeter I measured the following AVERAGED efficiency for five
unaltered MFJ-989C tuners (load/ efficiency) on 1.800 MHz (worse
band):
12.5 / 80.9%
25 / 82.9%
50 / 88.3%
100 / 89.3%
200 / 89.9%
400 / 90.1%
800 / 91.2%
ALL 5 tuners were within one percent of each other. I have only one
other tuner type available for testing, and that one measured nearly
seven percent different at 50 ohms. The measurements are probably all
over the place for accuracy.
73, Tom
Roy Lewallen
In article <331ba39d...@netnews.worldnet.att.net>,
W8J...@worldnet.att.net (Tom Rauch) wrote:
>The measurements of efficiency in the QST review of antenna tuners are
>incorrect. An MFJ-259 analyzer was used to measure VSWR changes when
>the load was varied 2 times and 1/2 of the tuned value. Also
>correction factors were not applied to the MFJ-259. The 259 is NOT
>capable of the resolution required for this type of test, it is NOT a
>lab type instrument.
>
>The theory was correct ansd the intentions were good, the instrument
>was just used far beyond its design limitations.
Tom,
I strongly recommend that you write this up and send it to the ARRL. I
suggest sending it to Paul Pagel, with a request that it be published in
"Technical Correspondence".
Roy Lewallen, W7EL
Tom Rauch
On Wed, 05 Mar 97 06:41:48 GMT, w7...@teleport.com (Roy Lewallen)
wrote:
>Tom,
>I strongly recommend that you write this up and send it to the ARRL. I
>suggest sending it to Paul Pagel, with a request that it be published in
>"Technical Correspondence".
>
>Roy Lewallen, W7EL
In 1984 (or so) when QST reviewed the AL-1200, they called me and
asked questions about the power supply. They were getting something
like 2900 volts under load, when the manual specifies 3300 volts.
I asked if anything in the supply was getting hot, or if there was
ripple on the output, and they said no. I indicated they could only
have a power line regulation problem, but they replied they "had a "X"
thousand dollar voltage reglator in the line" and something that
expensive "couldn't be bad".
Despite the disagreement, they published review results with 2900
volts of HV on an amplifier designed to operate at 3300 volts.
They later took the PA to W1AW, and found the HV was 3300-3400 volts.
Mark Wilson called and asked ME to write a technical correction for
QST, and they would publish it in Tech Corresp. I refused, because it
was their mistake....NOT mine. I felt it would look like I was trying
to make excuses for a bad review.
(During that fiasco, I was told they really aren't a "lab". I
certainly agree with that.)
They never corrected the AL-1200 review, despite knowing full well
they made an error. They pubished incorrect data even after being
warned something was drastically wrong with the test setup.
They did something similar with Richard Measures, and his garbage
about parasitics. They made the mistake, published his articles, and
gave him a foothold to spread misinformation.
When they learned they made a mistake, they re-printed comments from a
letter I wrote. That was done, by merging my comments along with those
of a few others, in technical correspondence.
The ARRL is left in the clear while Measures blames me and a few
others for correcting what the ARRL should have never even published.
I've had enough of that type of thing. They need to learn to be big
boys, and pick up after themselves.
73 Tom
Scott Ellington
> ...An MFJ-259 analyzer was used to measure VSWR changes when
>the load was varied 2 times and 1/2 of the tuned value....
>
>The theory was correct and the intentions were good, the instrument
>was just used far beyond its design limitations.
>
Doesn't that assume the input impedance of the tuner is proportional to
the load impedance, and remains resistive? I don't believe that's true
in
general.
--
73,
Scott K9MA
sdel...@facstaff.wisc.edu
Tom Rauch
Hi Scott,
The method is a good method, and it's use to ESTIMATE loss was a great
idea, but it has some small built in errors.
As an additional point, the impedance will never stay resistive,
because the tuner is not a 1/4 wl transmission line. That's ok,
because if the reactance change won't bother anything a large amount.
The bigger problem is very tiny amounts of SWR change indicate a large
change in tuner loss. The smallest amount of harmonic energy (say even
-25 dB), parallax in the tiny MFJ meter, or calibration mistake would
shift efficiency estimates by several percent.
Remember, the change from 1.8592:1 SWR to 1.7293 SWR is a two times
change in loss. Do YOU have an SWR meter that can read that change
without error, on many bands and with many load conditions??
I think that's a job for a more precise instrument or a more direct
method.
73 Tom
Scott Ellington
Tom Rauch wrote:
>
> The bigger problem is very tiny amounts of SWR change indicate a large
> change in tuner loss. The smallest amount of harmonic energy (say even
> -25 dB), parallax in the tiny MFJ meter, or calibration mistake would
> shift efficiency estimates by several percent.
>
> Remember, the change from 1.8592:1 SWR to 1.7293 SWR is a two times
> change in loss. Do YOU have an SWR meter that can read that change
> without error, on many bands and with many load conditions??
>
> I think that's a job for a more precise instrument or a more direct
> method.
>
> 73 Tom
I agree that the SWR meter error is a major problem, but the
nonlinear behavior of SWR vs load resistance also introduces
a significant error in loss measurement, for the same reason. For
example, consider a simple L-network tuned to match 250 to 50 Ohms. If
the load resistance is changed to 500 Ohms, the SWR is close to
2:1, but if the load resistance is 125 Ohms the SWR is 3:1. So it
would seem that even with a perfect SWR meter, determining the
tuner loss requires some fairly complicated calculations.
Dan K6MHE
[SNIP]
Tom,
Was there much variation between the units tested? You give the
averaged values but, I was wondering if they were all "pretty close"
to one another or if the swing was great between them.
73,
Danny, K6MHE
William E. Sabin
Tom Rauch wrote:
>
> On Wed, 05 Mar 1997 14:14:17 -0600, Scott Ellington
> <sdel...@facstaff.wisc.edu> wrote:
>
>
> >Doesn't that assume the input impedance of the tuner is proportional to
> >the load impedance, and remains resistive? I don't believe that's true
> >in general.
>
>
> because the tuner is not a 1/4 wl transmission line. That's ok,
> because if the reactance change won't bother anything a large amount.
>
I did a simulation yesterday of Witt's approach, using various lengths of
*lossy* xmission line, using Mathcad (see QEX August 1996). When I change
the load from Z0 to 2Z0 and to Z0/2 the input impedance of the line does
indeed become *substantially* reactive (just as it should). Nevertheless,
the Equation 5 that Witt presents does predict the dB line loss fairly
accurately (5%) in several cases that I simulated. My software gives the
exact value of loss to compare with Witt's results.
The use of xmission line for this is OK. A Transmatch (I do like that
word) at *one frequency* is mathematically equivalent to a certain length
of lossy xmission line that has a certain value of Z0.
Witt's Equation 5 involves the inverse product of the *magnitudes* of two
reflection coefficients (expressed in terms of SWR). The factor of 9 in
the denominator so far eludes me. It may be an empirical "fudge factor"
but I'm not sure about that.
The accurate reading of SWR is important and if the generator has much
harmonic content that could cause errors. The percentage of harmonic
content would be something to look at. These harmonics would produce
little error in a resistor measurement, but in an LCR circuit they could
be not so good.
Bill W0IYH
at...@imap1.asu.edu
William E. Sabin (sab...@crpl.cedar-rapids.lib.ia.us) wrote:
snip
: I did a simulation yesterday of Witt's approach, using various lengths of
: *lossy* xmission line, using Mathcad (see QEX August 1996). When I change
: the load from Z0 to 2Z0 and to Z0/2 the input impedance of the line does
: indeed become *substantially* reactive (just as it should). Nevertheless,
: the Equation 5 that Witt presents does predict the dB line loss fairly
: accurately (5%) in several cases that I simulated. My software gives the
: exact value of loss to compare with Witt's results.
: The use of xmission line for this is OK. A Transmatch (I do like that
: word) at *one frequency* is mathematically equivalent to a certain length
: of lossy xmission line that has a certain value of Z0.
: Witt's Equation 5 involves the inverse product of the *magnitudes* of two
: reflection coefficients (expressed in terms of SWR). The factor of 9 in
: the denominator so far eludes me. It may be an empirical "fudge factor"
: but I'm not sure about that.
snip
Hi Bill,
The factor of 9 comes because he changes the reflection coefficient
at the output to +1/3 and -1/3. Here is a "derivation" of Witt's
equation using an S matrix to describe the tuner.
First look at the case where the tuner is matching 50 ohms to 50 ohms
so we are putting a unit signal into port 1 and taking it out of port
2. Since we adjust the tuner for a perfect match, no signal comes out
of port 1 when we put power into port 1, the S matrix must look like:
out = S * in
[0 ] = [0 s12] [1]
[s12] [s12 s22] [0]
where |s12| will be close to 1 and s22 will be small if the loss is small.
The fractional power output will be |s12|^2. So the tuner loss in dB is
L = -10*log(|s12|^2)
If we change the reflection coefficient, rout, on the
output, then the signal into port 2 will be rout*out, and we will measure
some reflection coefficient, rin, on the input port 1,
[rin] = [0 s12] [1]
[out] [s12 s22] [out*rout]
Multiplying the matrix gives two equations:
rin = s12*out*rout
out = s12+s22*out*rout
and solving the 2nd equation for the signal out of port 2 gives:
out = s12/(1-s22*rout)
and plugging into the first equation gives the reflection coefficient at
input port 1:
rin = s12^2*rout/(1-s22*rout)
Now we want |s12|^2 to calculate the loss. Calculating the magnitude of the
input reflection coefficient gives
|rin| = |s12|^2 |rout|/|1-s22*rout|
Witt then makes rout = 1/3 and -1/3, and takes the geometric
mean of the magnitude of the reflection coefficient rin. This gives
the result
|rin|_av = |s12|^2/(3*sqrt(|1-s22^2/9|))
Plugging into Witt's formula, we get
L = -10*log(3|rin|_av) = -10*log(|s12|^2) + 5*log(|1-s22^2/9|)
The -10*log(|s12|^2) is the correct answer, and Witt neglects the
|1-s22^2/9| term which should be small since s22 is already small,
and it is now squared and divided by 9.
Other resistive output impedances go through exactly the same way with
the output port normalized to that impedance.
73 Kevin w9cf
dick sander
W8J...@worldnet.att.net (Tom Rauch) wrote:
>Hi Scott,
>
>The method is a good method, and it's use to ESTIMATE loss was a great
>idea, but it has some small built in errors.
>
>As an additional point, the impedance will never stay resistive,
>because the tuner is not a 1/4 wl transmission line. That's ok,
>because if the reactance change won't bother anything a large amount.
>
>The bigger problem is very tiny amounts of SWR change indicate a large
>change in tuner loss. The smallest amount of harmonic energy (say even
>-25 dB), parallax in the tiny MFJ meter, or calibration mistake would
>shift efficiency estimates by several percent.
Tom, I'm with you so far.
>Remember, the change from 1.8592:1 SWR to 1.7293 SWR is a two times
>change in loss. Do YOU have an SWR meter that can read that change
>without error, on many bands and with many load conditions??
Tom would you explain how/why there is a two times change in loss from such a
small change in SWR?
>I think that's a job for a more precise instrument or a more direct
>method.
>
>73 Tom
Thanks, Dick Sander -K5QY
William E. Sabin
Scott Ellington wrote:
>
> For
> example, consider a simple L-network tuned to match 250 to 50 Ohms. If
> the load resistance is changed to 500 Ohms, the SWR is close to
> 2:1, but if the load resistance is 125 Ohms the SWR is 3:1.
Not correct. The SWR is 2:1 in both cases. Verified with Mathcad
calculations. An interesting point also, the reflection coefficients are
"negatives" of each other. for example .2-j.267 and -.2+j.267. Their
magnitudes, therefore the SWRs, are equal.
Bill W0IYH
William E. Sabin
at...@imap1.asu.edu wrote:
> Sabin wrote:
> : Witt's Equation 5 involves the inverse product of the *magnitudes* of two
> : reflection coefficients (expressed in terms of SWR). The factor of 9 in
> : the denominator so far eludes me. It may be an empirical "fudge factor"
> : but I'm not sure about that.
atkes wrote:
>
> The factor of 9 comes because he changes the reflection coefficient
> at the output to +1/3 and -1/3.
The stuff below, from another post, anticipates the .333 reflection coeff
for the lossless network. Mag(0.2 +/- 0.267) = 1/3.
Sabin wrote:
>> An interesting point also, the reflection coefficients are
>> "negatives" of each other. for example .2-j.267 and -.2+j.267. Their
>> magnitudes, therefore the SWRs, are equal.
So the loss is approximately the ratio of the refl. coeff. of a lossless
network to the refl. coeff. of a lossy network. Witt takes the geometric
mean of two such measurements.
Bill W0IYH
Tom Rauch
There was no detectable difference from unit to unit of tuners,
including one that was outside all winter in a dog-house.
The tuners are closer to each other than the repeatability of my
measurements are.
To be clear, it is the overall policy set by management I object to,
and not the people involved. People just aren't found who are nicer
and more concerned with accuracy than Frank Witt or Rick Lindquist.
The only thing that upsets me is the unwillingness of an editor to
take responsiblity and properly correct and accept responsibility for
errors.
In other words, I was miffed at the AL-1200 review deal, that feeling
was renewed by the poor handling of the Measures mistake that caused
nothing but work and grief for me for years.
I really didn't want to post anything here, but too many people were
accepting the tuner measurements as correct when they are not.
Once something incorrect gets in print, it takes on a life of its own.
better to stop it early than wait seven years for a response.
73 Tom
Tom Rauch
On 6 Mar 1997 22:17:04 GMT, san...@aud.alcatel.com (dick sander)
wrote:
>>Remember, the change from 1.8592:1 SWR to 1.7293 SWR is a two times
>>change in loss. Do YOU have an SWR meter that can read that change
>>without error, on many bands and with many load conditions??
>
>Tom would you explain how/why there is a two times change in loss from such a
>small change in SWR?
Hi Dick,
Because what we are talking about is the SWR improvement due to loss
or attenuation between the load and the measurement point.
It takes a lot of loss to make a small improvement in SWR.
For example, to reduce a 2:1 SWR to 1.5 :1, the return loss must
change from 9.542 dB to 13.979 dB. This is a return loss (two way
loss) change of 4.437 dB.
The one way loss (or loss by attenuation) between the load and source
must increase by 1/2 of that difference, or by 2.2185 dB, to make a
2:1 SWR look like 1.5:1 at the source.
Adding 2.2185 dB of loss makes a 2:1 load SWR read 1.5:1 at the
source.
2.2185 dB is 40% power loss. A tuner input SWR reading of 1.5, when
the load is changed from perfect to 2.0, is about 40% loss.
That's a neat way to measure loss, IF everything is near perfect. Of
course the tuner adds other complexities, because its loss does not
remain precisely the same as the load impedance changes, but this is a
fair approximation of loss. The main rub I see are many little subtle
errors creep in, plus it demands a clean and accurate measurement
instrument.
I like the MFJ-259, but not that much.
73 Tom
David L. Thompson
W8J...@worldnet.att.net (Tom Rauch) wrote:
>The measurements of efficiency in the QST review of antenna tuners are
>incorrect. An MFJ-259 analyzer was used to measure VSWR changes when
>the load was varied 2 times and 1/2 of the tuned value. Also
>correction factors were not applied to the MFJ-259. The 259 is NOT
>capable of the resolution required for this type of test, it is NOT a
>lab type instrument.
I have the MFJ 259 and it sure replaces most other antenna measurement
devices (GDO, Noise Bridge), but agree its not a lab instrument.
I think the comments about which run hot and handle power is the most
important in a high power environment. I have seen an MFJ 986 arc
over from the capacitor to a mounting screw nearly 1" away trying to
load into a 400 ft loop on 160. The loop had been tuned to 1:1 easily
with 50W. This with a Clipperton L in the low power position at 350W.
Generally you want a wide ranging system on 160/80 (maybe 40) and
something that will tune out (for the transmitter!!!) 3 to 5:1 SWR on
20 to 10. The two best that I have found are the Ten Tec 229 (not
sure of the new number but aside from the new front panel its the
same.) and the Xmatch. The Xmatch is a delight on 160 or 80..gosh I
even loaded my 20 meter beam on 160. It never gets hot with my 8877
(single hole), but is somewhat touchy on 15 (my KT34XA has a bad cap
and 3:1 SWR on 21.025). The Ten Tec main weakness is the antenna
switch (replaced it entirely)...wish I could put my Hofi 5 position
switch in rather than the wafers. They also use 3KV 470pf caps that
can be switched in. The 3KV rating is impressive until you realize
its in Dc volts not RF AC volts! I had up the 400 foot loop several
years ago and ran in the CQ WW 160 contest. The caps gradually
changed value during the contest and then two blew! I threw in two
big transmitting micas (good ole Sangamo ) as a temp replacement.
After the contest K4DLI and I figured that based on the antenna feed
point match and the ladder line we had about 1 amp of RF at the tuner.
Bet those tuners don't dare give you a RF AMP capability rating.
For the casual operator any one of the tuners will do. But beware of
published ratings and false specs. Sounds almost like the wild
antenna specs that finally caused QST to stop publishing gain/FB
figures.
Now I wonder if the counter in my 259 is also a lab instrument?
Dave K4JRB
at...@imap1.asu.edu
: Bill W0IYH
Yes, I agree. You can even get a firm bound on the error.
Since the output power must be less than the input power, the
value of |s22|^2 less than 1 - |s12|^2. If you plug that into the
error term, you get
Error(in dB) < |5*log(8/9+|s12|^2/9)|
and if we write |s12|^2 = 10^(-L/10) where L is the tuner loss in dB,
Error < 5*log((8+10^(-L/10))/9) for small values of L, this goes to
L/18, annd the fractional error decreases as the loss increases. The
percent error in the loss is bounded by 6 percent, almost exactly what
you indicated in your previous post. Taking L to infinity, shows the
error is bounded by 0.26 dB. For some other values of loss, the Error
bound is
Loss(dB) Maximum Error(dB)
L<1.0 .06 L
1.0 .05
2.0 .09
3.0 .12
6.0 .19
10.0 .23
infinity .26
73 Kevin w9...@ptolemy.la.asu.edu
William E. Sabin
at...@imap1.asu.edu wrote:
>
> You can even get a firm bound on the error.
> Since the output power must be less than the input power, the
> value of |s22|^2 less than 1 - |s12|^2. If you plug that into the
> error term, you get
>
> Error(in dB) < |5*log(8/9+|s12|^2/9)|
>
Another source of error is in the accuracy of the SWR meter. I analyzed a
situation with Mathcad where a 1 dB attenuator (typical transmatch loss)
precedes the load.
The power into the attenuator is 1500 W.
The correct value of SWR for RL=25 or RL=100 is 1.720
The 1 dB loss is predicted very accurately by Witt's formula. It does not
appear to be an approximation, at least to three decimal places, using
very accurate values (3 decimal places) for the 1 dB pi attenuator.
The actual power loss in the 1 dB attenuator is 308 W.
Now suppose the SWR meter is off by +/- 5%.
For 5% low reading the "indicated" loss is 416 W.
For 5% high reading the "indicated" loss is 207 W.
The error of approx. +/- 100 W out of 1500 says that the SWR meter should
be within perhaps 2% or so for reliable results.
Also, the value 1/3 can be improved from the formula
G = (RL - Z0) / (RL + Z0)
where G is refl. coeff. of the lossless network, RL and Z0 are measured
values of load resistance and system Z0.
Bill W0IYH
Scott Ellington
William E. Sabin wrote:
>
> Not correct. The SWR is 2:1 in both cases.
>
Bill is correct. My error.
William E. Sabin
R. L. Measures
In article <331f7e8d...@netnews.worldnet.att.net>,
W8J...@worldnet.att.net wrote:
> On Thu, 06 Mar 1997 15:29:09 GMT, l...@mcn.org (Dan K6MHE) wrote:
>
> >[SNIP]
> >
> >Tom,
> >
> >Was there much variation between the units tested? You give the
> >averaged values but, I was wondering if they were all "pretty close"
> >to one another or if the swing was great between them.
> >
> >73,
> >Danny, K6MHE
>
> There was no detectable difference from unit to unit of tuners,
> including one that was outside all winter in a dog-house.
...snip...
Mr. Rauch:
> The only thing that upsets me is the unwillingness of an editor to
> take responsiblity and properly correct and accept responsibility for
> errors.
>
> In other words, I was miffed at the AL-1200 review deal, that feeling
> was renewed by the poor handling of the Measures mistake that caused
> nothing but work and grief for me for years.
>
...snip...
The ARRL technician who tested the AL-1200 in the ARRL Lab told me that
during testing he observed internittent tank arcing, that appeared to be
VHF parasitic related. However, this information was not mentioned in the
QST review. It seems to me that you got somewhat more than a fair deal,
Mr. Rauch. OTOH, the 3CX1200A7 has about 8 times as much feedback-C as an
8877, so it is not easy to make VHF- stable.
RE: the 'Measures mistake' that caused you nothing but years of work and
grief: You and Measures (me) finally got the public debate you pined for,
Mr. Rauch. During this debate, on November 28, 1996 you posted a reply
wherein you finally got around to doing some calculations for a suppressor
that was made with resistance-wire versus one that was made of copper-wire.
You used an ESR of 5-ohms. When you realized that the Rp
(parallel-equivalent R) of a copper-wire suppressor was much higher than
the Rp of a similar resistance-wire suppressor, you cancelled your post and
refused to talk about why. The bottom-line is that the VHF voltage gain of
a HF amplifier is basically Mu*Rp. More VHF Rp means more VHF voltage
gain.
When Wes, N7WS, measured copper-wire suppressors versus resistance-wire
suppressors on a HP impedance analyzer in December, he found that the
copper-wire (W8JI) suppressor had more than 1.6 times as much VHF Rp as a
resistance-wire (AG6K) suppressor.
€ Common sense ought to tell virtually anybody that such a debate was
unnecessary. OK, so you want to build a lossy L/R circuit to suppress
amplifier oscillations. Are you going use very unlossy copper or even
unlossier silver--- or are you going to use lossy resistance-wire?
€€ IMO, any and all of Mr. Rauch's comments deserve careful consideration.
On page 71 in the September, 1994 issue of *QST* magazine, Tom Rauch, W8JI
(a.k.a 'Charles Rauch' and Charles Thomas Rauch, Jr.) said (regarding
intermittent VHF parasitic-oscillations in HF amplifiers):
"...there is no basis in amplifier tube theory or actual experience to
support such conclusions. They are not supported by design theory or the
experience of recognized experts in the RF amplifier community that include
Eimac, Siemens, ETO, Henry, and Ameritron. ..."
If you guessed that Ameritron's recognized 'expert' is none other than
Charles Thomas Rauch, Jr., W8JI, congratulations.
Those who are reluctant to accept the word of self-recognized 'experts' on
this subject might try reading section 5.8, Parasitic Suppression
Techniques, by H. F. Dittrich in *Tubes for R. F. Heating*. N.V. Phillips
Gloeilampenfabrieken, Eindhoven, The Netherlands, copyright Oct. 1977.
Also, see page 72 in the 1926 Edition of the *Radio Amateur's Handbook*.
The 'such conclusions' to which Mr. Rauch refers can be found in the
September and October, 1990 *QST* magazine article "Parasitics Revisited"
by this author.
--
--Rich-- ag6k, 805-386-3734
Tom Rauch
>I think the comments about which run hot and handle power is the most
>important in a high power environment. I have seen an MFJ 986 arc
>over from the capacitor to a mounting screw nearly 1" away trying to
>load into a 400 ft loop on 160. The loop had been tuned to 1:1 easily
>with 50W. This with a Clipperton L in the low power position at 350W.
The biggest problem with power handling not using the maximum
capaciatnce and minimum inductance required to match the load.
>Generally you want a wide ranging system on 160/80 (maybe 40) and
>something that will tune out (for the transmitter!!!) 3 to 5:1 SWR on
>20 to 10. The two best that I have found are the Ten Tec 229 (not
>sure of the new number but aside from the new front panel its the
>same.) and the Xmatch. The Xmatch is a delight on 160 or 80..gosh I
>even loaded my 20 meter beam on 160. It never gets hot with my 8877
>(single hole), but is somewhat touchy on 15 (my KT34XA has a bad cap
>and 3:1 SWR on 21.025).
The X match cures this problem by FORCING the customer to set one
capacitor to maximum capacitance, the user has no choice.
It is a simple T network like any other T network tuner, and actually
has a slightly lower Q inductor than the 989, but the amount of
capacitance is much higher. That lowers the voltage in the tuner.
The fixed capacitor on one side of the T forces you to use maximum C
to match the load.
>The Ten Tec main weakness is the antenna
>switch (replaced it entirely)...wish I could put my Hofi 5 position
>switch in rather than the wafers. They also use 3KV 470pf caps that
>can be switched in. The 3KV rating is impressive until you realize
>its in Dc volts not RF AC volts! I had up the 400 foot loop several
>years ago and ran in the CQ WW 160 contest. The caps gradually
>changed value during the contest and then two blew! I threw in two
>big transmitting micas (good ole Sangamo ) as a temp replacement.
>After the contest K4DLI and I figured that based on the antenna feed
>point match and the ladder line we had about 1 amp of RF at the tuner.
>Bet those tuners don't dare give you a RF AMP capability rating.
It isn't the amperes delivered that are the problem, it is the VAR
power in the tuner.
The current in the tuner can be several amperes when deliveing only a
few hundred watts to the load, and the voltage can be several kV, if
you use less C than required to match the load.
A tuner that will handle kilowatts of power can be made to arc at a
less than 100 watts IF the capacitors are set too far open.
73 Tom
Dan K6MHE
Tom,
While out mowing the back yard I was thinking about the results you
obtained. It would seem to me that if the highest loss occurs on their
lowest frequency range at the lower impedances then would it not be
unreasonable to assume that at the other extreme would be on the
highest band (10 meters in this case) at the higher impedances.
If that is so, have you measured the tuners at the high frequency end
of their ranges ?
73
Danny, K6MHE
Tom Rauch
Hi Danny,
No I haven't. It took days to measure only one band accurately on a
sample group of tuners. I ran out of time to do more.
I did receive an E-mail from Frank Witt, the person who measured the
tuners for the ARRL. Let me summarize his NEW measurements of a 989
production tuner's efficiency on 1.8 MHz. These are tenative of
course:
Load Mine His Published
12.5 80.9% 74% 62%
25 82.9 86 71
50 88.3 91 80
100 89.3 91 82
200 89.9 92 80
400 90.1 92 78
800 91.2 91 78
73, Tom
K1BQT
Hi Tom--
Boing, boing, boing....
Rick K1BQT
W6KKT
On Sat, 08 Mar 1997 19:27:38 GMT, W8J...@worldnet.att.net (Tom Rauch)
wrote:
>I did receive an E-mail from Frank Witt, the person who measured the
>tuners for the ARRL. Let me summarize his NEW measurements of a 989
>production tuner's efficiency on 1.8 MHz. These are tenative of
>course:
>
>Load Mine His Published
>12.5 80.9% 74% 62%
>25 82.9 86 71
>50 88.3 91 80
>100 89.3 91 82
>200 89.9 92 80
>400 90.1 92 78
>800 91.2 91 78
>
>73, Tom
Hi Tom, have I missed a thread? Is he using a new test method? Why
such discrepancy between the new test and published figures?
73, Jesse, W6KKT
Tom Rauch
On Sun, 09 Mar 1997 00:24:32 GMT, w6...@frazmtn.com (W6KKT ) wrote:
>Why such discrepancy between the new test and published figures?
I'm not totally sure Jessie. It was the same method.
By making only one pass without crosschecks, its impossible to find a
measurement error. That's why I like to repeat the same measurements
at another time, and try to measure the same things two or three
different ways.
I used the SWR method, measured coil Q, voltage across the load (on a
very good expanded scale meter), and direct transmission loss. The
only method that wandered all over the place was the SWR method. It
was consistantly 3 or 4% off the other methods, no matter how careful
I was. Now I'm not saying it looks bad on paper, or is unusable, I'm
just saying when I use that method I have a time getting results that
agree with three other methods that do closely agree.
I'd sure like to know what's going on. If we don't know it's broke,
we'll certainly never fix it!
73 Tom
Reynolds
Tom Rauch <W8J...@worldnet.att.net> wrote in article
<3321b72d...@netnews.worldnet.att.net>...
> On Sat, 08 Mar 1997 14:03:17 GMT, l...@mcn.org (Dan K6MHE) wrote:
>
> >Tom,
> >
> >While out mowing the back yard I was thinking about the results you
> >obtained. It would seem to me that if the highest loss occurs on their
> >lowest frequency range at the lower impedances then would it not be
> >unreasonable to assume that at the other extreme would be on the
> >highest band (10 meters in this case) at the higher impedances.
> >
> >If that is so, have you measured the tuners at the high frequency end
> >of their ranges ?
> >
> >73
>
> Hi Danny,
>
> No I haven't. It took days to measure only one band accurately on a
> sample group of tuners. I ran out of time to do more.
>
> I did receive an E-mail from Frank Witt, the person who measured the
> tuners for the ARRL. Let me summarize his NEW measurements of a 989
> production tuner's efficiency on 1.8 MHz. These are tenative of
> course:
>
> Load Mine His Published
> 12.5 80.9% 74% 62%
> 25 82.9 86 71
> 50 88.3 91 80
> 100 89.3 91 82
> 200 89.9 92 80
> 400 90.1 92 78
> 800 91.2 91 78
>
> 73, Tom
Tom,
Does this mean that the published figures for other bands are also not
accurate?
Did Frank Witt mention whether he re-tested the 989 on other bands?
I am not asking because I own one. It is that I am not yet a believer of
the
results as published.
73, Paul, W2CW
R. L. Measures
In article <19970308212...@ladder01.news.aol.com>, k1...@aol.com
(K1BQT) wrote:
The amazing thing is that even through Tom Rauch knows his theories didn't
wash in the VHF parasitics debate, and knows his VHF suppressor design
produced the highest/worst VHF-Rp in Wes Stewart's [N7WS] measurements with
a Hewlett-Packard Model 4191A RF Impedance Analyzer, Tom still has to
pretend in public that his self-recognized "expert" pronouncements were
'absolutely, positively, 100%' correct---and Tom's lap-dogs still have to
yap their approval. -chortle-
RE: "EXPERTS": At OJ's civil trial, there was a defence witness who
claimed to be a "photography expert" who testified under oath that a
certain photo of OJ was fake---and we know what happened to his
credibility.
What kind of an amplifier expert rejects standard AC Circuit-Analysis and
scientific measurements whenever they indicate that his dictums didn't
wash? Apparently, the kind you apparove of, Rick.
Boing...........
--
--Rich-- ag6k, 805-386-3734
Tom Rauch
On 9 Mar 1997 11:36:28 GMT, "Reynolds" <reyn...@capital.net> wrote:
>Does this mean that the published figures for other bands are also not
>accurate?
I don't know for sure.
If they are off for one band, my guess is they are likely off for
other bands. Theory indicates the efficiency should be higher as
frequency is increased, up to a ceratint limit, and then it should
decrease. The optimum band should NOT be at one end, and is certainly
not on 160.
>Did Frank Witt mention whether he re-tested the 989 on other bands?
So far just at 1.8 MHz. I hate to open the can of worms of the other
bands. The Tucker T-3000 hasn't been re-tested, and the one I measured
here seems to test WORSE than the published results, but I need to
spend more time with it to be certain. I only made one measurement
pass on it.
I wish I had more time. I figure I would need 20-40 hours per tuner to
do a good job using multiple methods.
73 Tom
aa...@gate.kc5aug.ampr.org
I have seen lots of measurements and opinions about the article
written in QST and would like to put my two cents worth in.
Nowhere in any of the discussion here, and in the article is there an
indication of how the tuner was tuned. "I tuned it for lowest swr at
the frequency and proceeded to measure it."
There is no indication of the Q, bandwidth, inductance or capacitance
settings, so how can one measurement from source#1 compare to
source#2, or the QST measurement?
On a differential tuner such as the MFJ-986, there is only one place
where the inductance and capacitance will cancel, giving a proper
match. If everyone was using this type of setting, then everyone
would have valid measurements to compare. Everyone will have the
same Q, bw, etc.
On standard T-tuners, you can have a large variety of settings which
will bring the system into resonance with a 1:1 swr. My general
tuning proceedure is to set both input and output capacitances to
maximum, and turn the roller inductor until the swr starts coming
down.
On higher bands, it is possible to have both input and output
capacitors set to maximum, giving maximum power transfer and lowest
loss.
On the lower bands, once I dip the swr with the inductor, I then
change the output capacitor, leaving the input(transmitter) capacitor
at maximum. This setting generally gives the lowest losses in the
tuner, and maximum power transfer to the antenna (system).
Unfortunately, it's possible to tune the system using maximum
inductance and a much lower amount of capacitance to give a 1:1 swr
for a given frequency and antenna system. So, any measurements taken
such as power lost in the tuner will be VERY different from the
maximum-Q setting.
So, what am I saying?
You might ALL be right! Don't discredit QST just because you came up
with a different figure, until you know all the facts.
Tom Stone
m...@mfjenterprises.com
William E. Sabin
aa...@gate.kc5aug.ampr.ORG wrote:
>
> Unfortunately, it's possible to tune the system using maximum
> inductance and a much lower amount of capacitance to give a 1:1 swr
> for a given frequency and antenna system. So, any measurements taken
> such as power lost in the tuner will be VERY different from the
> maximum-Q setting.
>
Did you mean "minimum-Q setting" ?
Bill W0IYH
Wes Stewart
In article <measures-ya0240800...@news.vcnet.com> meas...@mail.vcnet.com (R. L. Measures ) writes:
>From: meas...@mail.vcnet.com (R. L. Measures )
>Subject: Re: QST Antenna Tuner Measurements
>Date: Fri, 07 Mar 1997 17:57:36 -0800
Good Grief here we go again...
I thought a stake had been driven through the heart of this topic.
For those who don't read rec.radio.amateur.homebrew, some background.
A debate was ongoing for some time when I volunteered (to my everlasting
regret) to perform some measurements on parasitic suppressors using
laboratory equipment. I posted the results of these measurements on my home
page (since removed). Along with the data, I clearly stated that the material
was copyrighted by me but could be used or quoted for NON-COMMERCIAL purposes.
Mr. Measures ignored this claim and proceeded to post excerpts of the
material, which he erroneously believes support his position, on his commercial
home page. I wrote Mr. Measures and politely asked him to remove the material
and any references to me from his commercial, as I do not support his position
and resent the inferences that I do. In contravention to my request, Mr.
Measures continues to post the information and make claims, such as what
follows, that can lead the unknowing to believe that I and the data favor his
claims. Simply put, this is untrue.
> When Wes, N7WS, measured copper-wire suppressors versus
resistance-wire>suppressors on a HP impedance analyzer in December, he found
that the>copper-wire (W8JI) suppressor had more than 1.6 times as much VHF Rp
as a>resistance-wire (AG6K) suppressor. >€ Common sense ought to tell
virtually anybody that such a debate was>unnecessary. OK, so you want to
build a lossy L/R circuit to suppress>amplifier oscillations. Are you going
use very unlossy copper or even>unlossier silver--- or are you going to use
lossy resistance-wire?
In the foregoing, Mr. Measures carefully omits the fact that the copper or
silver-plated wire inductors are shunted with LOSSY PARALLEL RESISTORS. The
data that he so fondly refers to also show that the HF Q of a high Q inductor
shunted by a resistor is higher while at VHF has essentially the same Q as the
nichrome device. In other words, unless you are selling nichrome wire, there
isn't a dime's worth the difference. If you ARE selling nichrome wire, then
there is twenty bucks or so worth the difference.
[snip]
>--
>--Rich-- ag6k, 805-386-3734
N7WS
Tom Rauch
On 9 Mar 97 17:50:58 GMT, aa...@gate.kc5aug.ampr.ORG wrote:
>Nowhere in any of the discussion here, and in the article is there an
>indication of how the tuner was tuned. "I tuned it for lowest swr at
>the frequency and proceeded to measure it."
Mr. Witt certainly knows how to adjust the tuner, and he did so
correctly. We exchanged setting information, and our knob setting
agreed.
And Tom, I think you mean minimum Q, not maximum. Minimum operating Q
guarantees maximum efficiency, maximum operating Q produces the
highest losses.
73 Tom
Lou or Annette Hinshaw
W8J...@worldnet.att.net (Tom Rauch) wrote:
>The ARRL is left in the clear while Measures blames me and a few
>others for correcting what the ARRL should have never even published.
>I've had enough of that type of thing. They need to learn to be big
>boys, and pick up after themselves.
>73 Tom
The arrl is a religion and religions, by definition, NEVER make
misteaks.
73 WB0NRU
Prolific inventor,
Lousy Salesman,
Incompetent businessman.
(Typical)
R. L. Measures
> W8J...@worldnet.att.net (Tom Rauch) wrote:
>
> >The ARRL is left in the clear while Measures blames me and a few
> >others for correcting what the ARRL should have never even published.
€ Yeah, Rich Measures, AG6K, was so clever that he was able to trick QST
into publishing the Oct. 1988; March, 1989; and the Sept./October 1990
articles on intermittent VHF parasitic oscillation, which Mr. Rauch says
are technically incorrect and should never have been published. .........
Look ma, everybody's out of step except Johnny!
> >I've had enough of that type of thing. They need to learn to be big
> >boys, and pick up after themselves.
€ I did not bring it up, Tom. It was our. self-recognized 'amplifier
expert', who is appently still posessed by delusions of grandeur.
> >73 Tom
>
> The arrl is a religion and religions, by definition, NEVER make
> misteaks.
>
> 73 WB0NRU
€ Amen, brother Lou, Amen. -chortle-
--
--Rich-- ag6k, 805-386-3734
R. L. Measures
> W8J...@worldnet.att.net (Tom Rauch) wrote:
>
> >The ARRL is left in the clear while Measures blames me and a few
> >others for correcting what the ARRL should have never even published.
€ Yeah, Rich Measures, AG6K, was so clever that he was able to trick QST
into publishing the Oct. 1988; March, 1989; and the Sept./October 1990
articles on intermittent VHF parasitic oscillation, which Mr. Rauch says
are technically incorrect and should never have been published. .........
Look ma, everybody's out of step except Johnny!
> >I've had enough of that type of thing. They need to learn to be big
> >boys, and pick up after themselves.
€ I did not bring it up, Tom. It was our. self-recognized 'amplifier
expert', who is apparently still posessed by delusions of grandeur.
K1BQT
The title of this thread is "QST Antenna Tuner Measurements". As I
recall, the thread started because QST published some disputed data about
several commercial tuners--data that has, at least in part, been ammended
by the person who generated it. Now I'm reading about Rich Measures'
personal problems, getting a thinly-disguised public opinion pole on
whether or not Tom Rauch is a nice fellow, and god knows what else. Look
out Sally Jesse, you've got competition!
I don't want to be rude. But, wouldn't it make sense to stick to the
topic and save the group therapy for a different thread??
Rick K1BQT
R. L. Measures
In article <n7ws.270...@azstarnet.com>, n7...@azstarnet.com (Wes
Stewart) wrote:
> In article <measures-ya0240800...@news.vcnet.com>
meas...@mail.vcnet.com (R. L. Measures ) writes:
> >From: meas...@mail.vcnet.com (R. L. Measures )
> >Subject: Re: QST Antenna Tuner Measurements
> >Date: Fri, 07 Mar 1997 17:57:36 -0800
>
> Good Grief here we go again...
>
> I thought a stake had been driven through the heart of this topic.
You must have gone on vacation when you suggested that my 22 year old
female helper was licking more than envelopes, Wes.
> For those who don't read rec.radio.amateur.homebrew, some background.
>
> A debate was ongoing for some time when I volunteered (to my everlasting
> regret) to perform some measurements on parasitic suppressors using
> laboratory equipment. I posted the results of these measurements on my home
> page (since removed).
€ TranslationŠ When I realized that my measurements torpedoed my pal, I
pulled 'em, pronto.
>Along with the data, I clearly stated that the material
> was copyrighted by me but could be used or quoted for NON-COMMERCIAL purposes.
>
> Mr. Measures ignored this claim and proceeded to post excerpts of the
> material, which he erroneously believes support his position, on his
commercial
> home page. I wrote Mr. Measures and politely
Tersely is more like it. Like Reid Brandon, W6MTF, Wes presumed that my
mail-order business was a money-making enterprise. From its beginnings in
December, 1988, the suppressor retrofit-kit business has been operated as a
not-for-profit enterprise. I make no profit by paying my helper a bit more
than she would make working at MacDonald's ...........Hmmmm. I wonder if
Mr. Rauch receives monies from MFJ?
>asked him to remove the material
> and any references to me from his commercial,
The only reference to Wes is the requisite Copyright credit. The url
is:http://www.vcnet.com/measures
> as I do not support his position
> and resent the inferences that I do.
There are no inferences other than those imagined.
> In contravention to my request, Mr.
> Measures continues to post the information and make claims, such as what
> follows, that can lead the unknowing to believe that I and the data favor his
> claims. Simply put, this is untrue.
"the unknowing"---chortle. The scientific measurements speak for
themselves. Congrats on the measurements, Wes. Congrats to Hewlett
Packard. The whining is, however, somewhat disappointing.
> > When Wes, N7WS, measured copper-wire suppressors versus
> resistance-wire>suppressors on a HP impedance analyzer in December, he found
> that the>copper-wire (W8JI) suppressor had more than 1.6 times as much VHF Rp
> as a>resistance-wire (AG6K) suppressor. >€ Common sense ought to tell
> virtually anybody that such a debate was>unnecessary. OK, so you want to
> build a lossy L/R circuit to suppress>amplifier oscillations. Are you going
> use very unlossy copper or even>unlossier silver--- or are you going to use
> lossy resistance-wire?
>
> In the foregoing, Mr. Measures carefully omits the fact that the copper or
> silver-plated wire inductors are shunted with LOSSY PARALLEL RESISTORS. The
> data that he so fondly refers to also show that the HF Q of a high Q inductor
> shunted by a resistor is higher while at VHF has essentially the same Q
as the
> nichrome device.
This was the one part of the debate that you never seemed to grasp, Wes..
which is probably why you did not realize the danger your measurements
posed to one of the combatants. However, there's bound to be someone else
who has access to a HP impedance analyzer who would have eventually made
the Rp measurements and resolved the issue. Despair not. In the end, you
are more likely to be remembered for your contribution to radio-science
than for your contribution to slander.
Sure, it's easy to lower the VHF-Q of a copper-wire or silver-strap VHF
suppressor simply by increasing the inductance of the suppressor inductor,
Ls. In fact, by increasing Ls, anyone can make the VHF-Q of copper-wire
or silver-strap suppressor substantively lower than the VHF-Q of a
resistance-wire suppressor. Tom Rauch has said so on many occasions, and
he is absolutely, positively, 100% (thanks, OJ) right. ....... However,
arbitrarily increasing Ls causes the suppressor resistor, Rs, to
incinerate during operation at 28MHz---and with Rs in Kaputville, ---next
stop: Regenerationville. // This matter was discussed in the March, 1989
issue of QST magazine. I'm guessing that if you had read this article, you
probably would have realized the flaw in Mr. Rauch's argument, Wes.
Shortly after posting an article on 28-November, and, presumably doing a
few calculations, Tom Rauch realized---after six years in
Slumberland---that adding even a few ohms of VHF-ESR to Ls by making it out
of resistance-wire had a subtantive effect on VHF Rp. ..... OOPS. ....
Since VHF voltage-amplification is basically equal to VHF-Rp*Mu, it was now
pretty obvious that Mr. F. E. Handy was right in 1926 and Mr. C. T. Rauch,
Jr. was not right in 1996. Mr. Rauch cancelled the 28-Nov. article---but
the Rp cat was out of the proverbial bag.
Mr. Rauch seems to be caught between a large rock and a boulder. He can't
admit that F. E. Handy (and yours truly) was right because he declared
himself to be recognized amplifier expert in the 9/94 QST. Alas, there's
no way out short of an apparently much too painful fessin' up. Six years
wasted, and he is still trying to ride a quite-dead horse. Everybody makes
mistakes, Mr. Rauch. A pernicious problem lies in trying to hide them.
Wes: YOU--of all people--should have known better. Whenever anyone
rejects standard AC circuit-analysis, it's a sure bet that his 'science' is
not going to wash.
BTW, Wes, on the same day you published your VHF suppressor measurements,
Tom Rauch reportedly launched into a 4 hour diatribe on 1850KHz LSB that
was said to be a sure-cure for insomnia. A number of people bailed out and
sought refuge up the band.
--
--Rich-- ag6k, 805-386-3734
Joe Subich
In message <3322299b...@netnews.worldnet.att.net> -
W8J...@worldnet.att.net (Tom Rauch)Sun, 09 Mar 1997 03:46:48 GMT
writes:
:>
:>I used the SWR method, measured coil Q, voltage across the load (on a
:>very good expanded scale meter), and direct transmission loss. The
:>only method that wandered all over the place was the SWR method. It
:>was consistantly 3 or 4% off the other methods, no matter how careful
:>I was. Now I'm not saying it looks bad on paper, or is unusable, I'm
:>just saying when I use that method I have a time getting results that
:>agree with three other methods that do closely agree.
Tom,
Since the loads were all pure resistances, wouldn't it be simple and
accurate to simply measure the RF voltage across the load resistor
with a scope (or spectrum analyzer)?
Test RF could be supplied with a low power RF generator (1 Watt should
be enough). The "matched condition" of the tuner could be observed
with a scope connected to a 10 dB or more 50 Ohm directional coupler
at the input of the DUT. One could even confirm input power level
with a good RF milliammeter. Based on other published equipment test
results, the ARRL lab should have all of the equipment necessary to
make those tests.
Any competent technician should be able to read voltage from a
spectrum analyzer to within .2 dB (assuming a linear 2 dB per
division scale like the Tek 7L14 with its minor divisions at .4
dB steps). If I remember correctly, the uncertainty in the results
would be less than +/- 2.5% ... probably better than the other
methods.
73,
... Joe
========================================================================
Joe Subich, W8IK ex-AD8I
<W8...@IBM.NET>
<W8...@AMSAT.ORG>
<sub...@ibm.net>
<sub...@compuserve.com>
========================================================================
R. L. Measures
In article <19970310225...@ladder01.news.aol.com>, k1...@aol.com
(K1BQT) wrote:
Did you read Mr. Rauch's post of 6 March on this thread, Rick?
--
--Rich-- ag6k, 805-386-3734
Tom Rauch
On 11 Mar 1997 04:13:07 GMT, sub...@ibm.net (Joe Subich) wrote:
>In message <3322299b...@netnews.worldnet.att.net> -
>Tom,
>
>Since the loads were all pure resistances, wouldn't it be simple and
>accurate to simply measure the RF voltage across the load resistor
>with a scope (or spectrum analyzer)?
I used that method Joe, with a Harris RF Voltmeter specified at 1%
accuracy. I checked the resistors and tuning of the tuner on a
HP-4191A, and then verified the results on a Wiltron Transmission loss
RF analyzer.
When I was done I used a simple Bird wattmeter with a four inch
mirrored scale to measure loss directly at 50 ohms.
>Test RF could be supplied with a low power RF generator (1 Watt should
>be enough). The "matched condition" of the tuner could be observed
>with a scope connected to a 10 dB or more 50 Ohm directional coupler
>at the input of the DUT. One could even confirm input power level
>with a good RF milliammeter. Based on other published equipment test
>results, the ARRL lab should have all of the equipment necessary to
>make those tests.
The tests were done outside. In a second set of measurements, we now
virtually agree.
I really didn't want this to get into a long discussion, I just
noticed quite a few people took the published measurements as factual
data. That data would influence HOW they decide to construct an
antenna system, perhaps steering people away from the simple efficient
method of using ladder line and a matching network.
Rather than see another rumor get accepted as fact, I thought it was
better to caution everyone. I'm sure in a few months, buried way back
in a Stray or in Technical Correspondence, they might correct all
this.
73 Tom
William E. Sabin
Tom Rauch wrote:
>
>
> I really didn't want this to get into a long discussion, I just
> noticed quite a few people took the published measurements as factual
> data.
I, and I'm sure all of us, think you have made a valuable contribution.
Congratulations Tom, and thanks.
Bill W0IYH
R. L. Measures
In article <5g2m4j$chi$1...@news-s01.ny.us.ibm.net>, sub...@ibm.net (Joe
Subich) wrote:
> In message <3322299b...@netnews.worldnet.att.net> -
> W8J...@worldnet.att.net (Tom Rauch)Sun, 09 Mar 1997 03:46:48 GMT
> writes:
> :>
> :>I used the SWR method, measured coil Q, voltage across the load (on a
> :>very good expanded scale meter), and direct transmission loss.
...snip...
> Since the loads were all pure resistances, wouldn't it be simple and
> accurate to simply measure the RF voltage across the load resistor
> with a scope (or spectrum analyzer)?
>
> Test RF could be supplied with a low power RF generator (1 Watt should
> be enough).
...snip...
The potential trouble with testing a tuner with 1 watt, Mr. Subich, is
that tuners typically handle more than 1 watt. Remember the commercial
"3kW" tuner whose roller-inductor caught on fire at under 1kW in the ARRL
lab test? What kind of an electronics engineer rates his tuner design at
"3kW"---although he rather obviously never tested it at anything close to
the advertised power rating in QST ads?
However, it gets even better. The REPLACEMENT roller-L ALSO caught on
fire in the ARRL lab test. Now any reasonable person would have probably
thought that this electronics engineer would have taken the first fire as a
clue that something wasn't quite right, but no such luck. It's like the
lights are on in the engineering building ...... but no one answers the
doorbell.
The reason that the roller-inductors caught fire is that both were made
with a Delrin® coil-form.
I got booted out of an engineering college after my Junior year, and even
I know that Delrin® has a D-[dissipation]-factor that is unsuitable for use
around high RF potentials.......like it can melt and catch fire, as was
obviously the case.
IMO, the answer to the above question......"what kind of an electronics
engineer.....?" Is: one that has a much higher opinion of his
capabilities than his education would warrant. In one word the answer is:
HUBRISTIC.
Before I was allowed to advertise my VHF parasitic-oscillation suppressor
retrofit-kits in QST, I was required to send a retrofit-kit to the ARRL for
installation in a commercially made amplifier. I complied. A few weeks
later, I got the green light to advertise. /// At one time, the
Trio/Kenwood-donated TL-922 amplifier at W1AW used one of our suppressor
retrofit-kits.
Were MFJ's .QST ads for their "3kW" tuner accepted without foundational
proof that MFJ's product actually did what MFJ said? Does QST 'kow-tow'
to rich advertisers? Has the tail begun to wag the dog?
In a recent technical discussion with an MFJ engineer on another USENET
Newsgroup, it became somewhat apparent that this person was unfamiliair
with algebraic expression, scientific notation and AC circuit analysis.
(see pages 4-5 in the Fall 1996 issue of *Communications Quarterly*).
In German, 'rauch' means 'smoke'.
--
--Rich-- ag6k, 805-386-3734