On Fri, 02 Sep 2011 09:38:05 -0700, Jim Lux
wrote:

On 9/1/2011 10:11 AM, Jeff Liebermann wrote:

For a given physical size antenna, high gain antennas imply narrow
bandwidth and critical construction. On the other foot, low gain
antennas, such as the biquad, is fairly broadband, and therefore not
particularly critical to construct. What's fun is to attach the
antenna to a reflection coefficient bridge or directional coupler,
http://pe2er.nl/wifiswr/
http://www.qsl.net/n9zia/rlb/texscan.png
http://www.qsl.net/n9zia/rlb/
sweep generator, and oscilloscope to look at the VSWR curve. Then try
moving things around. On my crowded workbench, location of the
antenna relative to the highly reflective test equipment make a huge
difference. The changes do not really have a big effect on antenna
operation, but they certainly present a different picture as compared
to the nice clean curves on the data sheets.

You're building a CW radar, basically.
That's how near field ranges work, too.

Yep. Also known as a proximity detector and possibly a really bad
interferometer. I can see cars driving by, people moving around in
the house, trees swaying in the wind, and the opening and closing of
doors and windows. These are all easy to identify on the sweep
because they all move around. Only the major dip in the VSWR curve,
near resonance, remains fairly stable.

Back to MFJ bashing... On my desk is yet another MFJ-259B for
repair, probably with the two blown shottky diodes that I previously
mentioned. One of the local hams heard me talking about the problem
over the local repeater, and decided that maybe I could fix it for
him. Usually, they wait until the week before Field Day for such
repairs but he's installing a tower next week and wants it fixed
yesterday. Sigh.



--
Jeff Liebermann
150 Felker St #D http://www.LearnByDestroying.com
Santa Cruz CA 95060 http://802.11junk.com
Skype: JeffLiebermann AE6KS 831-336-2558

Jeff Liebermann wrote:


Back to MFJ bashing... On my desk is yet another MFJ-259B for
repair, probably with the two blown shottky diodes that I previously
mentioned. One of the local hams heard me talking about the problem
over the local repeater, and decided that maybe I could fix it for
him. Usually, they wait until the week before Field Day for such
repairs but he's installing a tower next week and wants it fixed
yesterday. Sigh.


Have you discussed this with Martin Jue? I think it's rude to bad-mouth
a fellow ham behind his back. I don't see how the diodes are a problem,
if you follow directions. If you live somewhere dusty or snowy and dry
enough to make static, use a gamma match or an UnUn or some other means
to keep your antenna at DC ground. I have been using germanium diodes
for 50 years and can't remember frying one in a small signal RF
application. What does Martin say about bleeder resistors?

I have a 4" pigtail around the ground lug and a male banana plug on the
end that lives in the middle of the SO-239, unless I am making
measurements. I use the banana plug as a shorting stick to neutralize
any residual capacitive charge in the device/coaxial cable under test.
Center conductor to cable ground. If I am especially concerned I'll
groung the 259B to my extensive safety ground system. When dealing with
little diodes you always want to make sure none of the sparks go through
them. Ground, ground and ground.

On 03 Sep 2011 13:56:30 GMT, dave wrote:

Jeff Liebermann wrote:


Back to MFJ bashing... On my desk is yet another MFJ-259B for
repair, probably with the two blown shottky diodes that I previously
mentioned. One of the local hams heard me talking about the problem
over the local repeater, and decided that maybe I could fix it for
him. Usually, they wait until the week before Field Day for such
repairs but he's installing a tower next week and wants it fixed
yesterday. Sigh.

Have you discussed this with Martin Jue?

No. I don't own an MFJ-259B. This is my 3rd(?) repair for what
appears to be exactly the same problem. I don't see how calling MFJ
will prove anything as they are apparently aware of the ESD problem.
See quotes from manual below.

I also reverse engineered the MFJ-1800 antenna, and also decided that
it has a problem. I have not called MFJ on these issues. Too busy
and too lazy.

I think it's rude to bad-mouth
a fellow ham behind his back.

Would you prefer I keep it secret and not tell fellow hams how to fix
it and why I think they blow up? I believe that I clearly labelled my
guesswork as conjecture and not fact. If hams were only allowed to
discuss things that are absolutely certain, the airwaves would be
silent.

I don't see how the diodes are a problem,
if you follow directions.

The units that are failing are not mine. I have no control over how
they are used.

As I vaguely recall, one failed while connected to some HF wire
antenna, the 2nd failed while plugging in a variety of calibrated
loads on the bench, and the most recent failed while attached to a
mobile HF antenna. It was difficult to determine the exact cause of
each failure because the unit did not just die, but instead started
producing insane readings. In all cases, the user thought something
was wrong with the antenna or loads, not the MJF-259B.

If you live somewhere dusty or snowy and dry
enough to make static, use a gamma match or an UnUn or some other means
to keep your antenna at DC ground.

Attach a high impedance voltmeter to a wire antenna blowing in the
wind and note the DC voltage produced. In my area, the humidity
rarely goes below about 30%, so static build up should not be a
problem when attached to an antenna. Currently, the humidity is
80-90% (morning fog), but when we get the hot dry winds from the
desert, the humidity will drop sufficiently low to cause problems for
a few daze.

What I believe is killing the diodes is not RF. It's the user
building up a static charge on plastic seat covers, synthetic clothes,
plastic carpet, etc, and discharging it into the antenna connector
when plugging in the antenna connector. Incidentally, one of my
customers with chronic equipment failures was traced to a negative ion
generator, which produced impressive high voltages on nearby object.

I have been using germanium diodes
for 50 years and can't remember frying one in a small signal RF
application.

How many of these germanium diodes were directly connected to the
antenna connector?

Apparently you missed my previous rant on the topic. See the
schematic extract of the RF section at:
http://802.11junk.com/jeffl/crud/MFJ-259B-RF-section.jpg
Notice the directly connected diodes. The diodes in question are
Avago HSMS-2820 zero bias shottky diodes.
http://www.avagotech.com/docs/AV02-1320EN
15V Max PIV is rather low. It won't take much voltage at the antenna
go exceed 15V. The two 47K resitors going to 0.01uf bypass caps make
an effective ground to any fast risetime voltage spike at the antenna.
An important clue is that BOTH D3 and D4 appear to be blown each time,
which implies an external failure, not a component failure.

What does Martin say about bleeder resistors?

I don't know. I haven't discussed this or any of my allegations with
him or MFJ support.

I have a 4" pigtail around the ground lug and a male banana plug on the
end that lives in the middle of the SO-239, unless I am making
measurements. I use the banana plug as a shorting stick to neutralize
any residual capacitive charge in the device/coaxial cable under test.
Center conductor to cable ground. If I am especially concerned I'll
groung the 259B to my extensive safety ground system. When dealing with
little diodes you always want to make sure none of the sparks go through
them. Ground, ground and ground.

Yep. That's a good way to provide some protection. However, there's
no protection while you're juggling connectors when you run the risk
of a static discharge to the center of the coax connector.

I don't recall reading such a procedure in the user manual. However,
there are plenty of warning:
http://www.mfjenterprises.com/pdffiles/MFJ-259B.pdf
In section 4.1:
WARNING: NEVER APPLY EXTERNAL VOLTAGES OR RF SIGNALS TO THE
ANTENNA CONNECTOR.
and in 5.1:
WARNING: NEVER APPLY RF OR ANY OTHER EXTERNAL VOLTAGES TO THE
ANTENNA PORT OF THIS UNIT. THIS UNIT USES ZERO BIAS DETECTOR
DIODES THAT ARE EASILY DAMAGED BY EXTERNAL VOLTAGES OVER A
FEW VOLTS.
and in 5.2:
WARNING: NEVER APPLY EXTERNAL VOLTAGES OR RF SIGNALS TO THE
ANTENNA CONNECTOR. PROTECT THIS PORT FROM ESD.

Clear enough. It would appear that MFJ is fully away of the fragile
nature of the input circuitry.




--
Jeff Liebermann
150 Felker St #D http://www.LearnByDestroying.com
Santa Cruz CA 95060 http://802.11junk.com
Skype: JeffLiebermann AE6KS 831-336-2558

On Sat, 03 Sep 2011 09:25:31 -0700, Jeff Liebermann
wrote:

I have a 4" pigtail around the ground lug and a male banana plug on the
end that lives in the middle of the SO-239, unless I am making
measurements. I use the banana plug as a shorting stick to neutralize
any residual capacitive charge in the device/coaxial cable under test.
Center conductor to cable ground. If I am especially concerned I'll
groung the 259B to my extensive safety ground system. When dealing with
little diodes you always want to make sure none of the sparks go through
them. Ground, ground and ground.

Hmmm.... The problem might be the SO-239 connector. When you plug
something into that connector, it connects the center pin first, and
then the ground. If it were replaced by an N connector, the ground
would make contact before the center pink, thus offering some added
protection. I'll see if the connector can be replaced.

http://www.w8ji.com/mfj-259b_calibration.htm
Most Likely Failures

Other than manufacturing errors, the detector diodes clearly
stand out as the most common problem. They are the most easily
damaged devices in the analyzer. If you have a sudden problem,
it is most likely a defective detector diode. Diode damage almost
always comes from accidentally applying voltage on the antenna
port.

Why are the diodes so sensitive? In order for the detectors to be
accurate within a fraction of a percent (one bit), detector diodes
must have very low capacitance and a very low threshold voltage.
This means the diodes, through necessity, must be low-power
zero-bias Schottky microwave detector diodes. The same
characteristics that make them accurate and linear cause the diodes
to be especially sensitive to damage from small voltage spikes.
ALWAYS discharge large antennas before connecting them to the
analyzer! Never apply external voltages greater than 3 volts to the
antenna port!

and

Because the detector is broadband and because it is dc coupled
to the antenna, any external voltage across the input port
causes measurement errors. It is the accumulated voltage of
multiple sources that is most important, not the strength of
any individual signal. Because of that, large antennas should
be tested at times when propagated signals in the range of the
antenna's response are at minimum strength.

W8JI designed the MFJ-259b.



--
Jeff Liebermann
150 Felker St #D http://www.LearnByDestroying.com
Santa Cruz CA 95060 http://802.11junk.com
Skype: JeffLiebermann AE6KS 831-336-2558

On Sat, 03 Sep 2011 09:51:00 -0700, Jeff Liebermann
wrote:

(...)

K0TO is collecting voltage measurement of the MFJ-259b for the purpose
of identifying blown diodes:
http://www.k0to.us/HAM/MFJ%20Diode%20measurements/Gather_MFJ_Data.htm
Estimates of the correct values:
http://www.k0to.us/HAM/MFJ%20Diode%20measurements/MFJ-259B%20Test%20Point%20Voltages.htm

Incidentally, his schematic at:
http://www.k0to.us/HAM/MFJ%20Diode%20measurements/11-17-sch_mfj259b-BW.pdf
looks better than most of the scans I've seen.

--
Jeff Liebermann
150 Felker St #D http://www.LearnByDestroying.com
Santa Cruz CA 95060 http://802.11junk.com
Skype: JeffLiebermann AE6KS 831-336-2558

On 9/3/2011 1:32 PM, Jeff Liebermann wrote:
On Sat, 03 Sep 2011 09:51:00 -0700, Jeff
wrote:

(...)

K0TO is collecting voltage measurement of the MFJ-259b for the purpose
of identifying blown diodes:
http://www.k0to.us/HAM/MFJ%20Diode%20measurements/Gather_MFJ_Data.htm
Estimates of the correct values:
http://www.k0to.us/HAM/MFJ%20Diode%20measurements/MFJ-259B%20Test%20Point%20Voltages.htm

Incidentally, his schematic at:
http://www.k0to.us/HAM/MFJ%20Diode%20measurements/11-17-sch_mfj259b-BW.pdf
looks better than most of the scans I've seen.


Good resources. Thanks, Jeff.

John - KD5YI

On Sat, 03 Sep 2011 09:25:31 -0700, Jeff Liebermann
wrote:

The two 47K resitors going to 0.01uf bypass caps make
an effective ground to any fast risetime voltage spike at the antenna.

Hi Jeff,

I'm sure you are perfectly aware of the single point of failure in
that generality.

Few Caps exhibit 0.01uF (when so marked) to transients (where it is
presumed they will exhibit 1/2*pi*f*c reactance to the risetime).

When we (silverbacks) got into this game, (the preferable) mica caps
were available, snipped out of the nearest sacrificial TV or radio.
Trying to read those several styles of color coding was the biggest
hurdle, but I had plenty in my junk-box.

Ceramic is ubiquitous, now, and far from choice in these matters,
unless you do deep research (maybe).

You got any favorites that respond to this?

73's
Richard Clark, KB7QHC

On Sat, 03 Sep 2011 11:31:51 -0700, Richard Clark
wrote:

On Sat, 03 Sep 2011 09:25:31 -0700, Jeff Liebermann
wrote:

The two 47K resitors going to 0.01uf bypass caps make
an effective ground to any fast risetime voltage spike at the antenna.

I'm sure you are perfectly aware of the single point of failure in
that generality.

I am? Ummm... well, I guess so.

Few Caps exhibit 0.01uF (when so marked) to transients (where it is
presumed they will exhibit 1/2*pi*f*c reactance to the risetime).

True. They all have some internal resistance to overcome. However,
that's negligible resistance when compared to that of a static blast.
Static electricity has lots of potential (volts), but is only able to
deliver small amounts of current. That's why we don't get
electrocutes by the potential (voltage) difference between our head
and our feet. Dividing the large voltage, by the tiny current,
results in a fairly substantial source resistance. I'm too lazy to
look it some real numbers, but I'm sure it's in mega ohms. The 47K
resistance, and whatever ESR the 0.01uF contributes, has little effect
on the energy delivered to the shottky diode.

Incidentally, if the source resistance of the static blast was much
less, then the diode would not simply be fried. It would probably
explode.

When we (silverbacks) got into this game, (the preferable) mica caps
were available, snipped out of the nearest sacrificial TV or radio.
Trying to read those several styles of color coding was the biggest
hurdle, but I had plenty in my junk-box.

Dumpster diving in Henry Radio's trash can in West Smog Angeles was
one of my favorite after skool exercises. Salvaging old TV chassis
and dead tubes were the grand prizes. Silver mica caps came a close
second.

Ceramic is ubiquitous, now, and far from choice in these matters,
unless you do deep research (maybe).

Ceramic is cheap. I was a big fan of porcelain caps from AVX in big
power amps. If you wanna handle current, there's nothing better.
Silver mica would get hot, ceramic would explode, and everything else
was either too big or too expensive.

Incidentally, I don't think they make 0.01uF silver mica caps. The
biggest I played with were in antenna tuners at 4700pF (or should I
say uuF for nostalgia purposes).

You got any favorites that respond to this?

I don't understand the question. Favorite what?

--
Jeff Liebermann
150 Felker St #D http://www.LearnByDestroying.com
Santa Cruz CA 95060 http://802.11junk.com
Skype: JeffLiebermann AE6KS 831-336-2558

On Sat, 03 Sep 2011 15:51:10 -0700, Jeff Liebermann
wrote:

Few Caps exhibit 0.01uF (when so marked) to transients (where it is
presumed they will exhibit 1/2*pi*f*c reactance to the risetime).

True. They all have some internal resistance to overcome.

It goes beyond that. Extrapolating from power applications hides the
defects of ceramic.

At HF/VHF and above, successful applications comes from throwing uF
solutions at pF problems. Ceramic's performance reveals inductive
reactance above 1-10 MHz. ESR also exhibits the same turn-around in
the same frequency range. Ceramic temperature coefficient is (Y5V)
goes into the toilet in weather that most of the south and eastern
seaboard has seen this summer. XR7 voltage coefficient causes
capacity to plummet at the voltages you offer for static. Over time,
ceramics lose capacity for simply having been in service for a while.

Aside from that, they work fine.

I was a big fan of porcelain caps from AVX in big
power amps. If you wanna handle current, there's nothing better.

However, those ceramics are 1,000 times (min.) larger than what you
have recommended. They serve an entirely different agenda.

AVX discusses these issues in much the same terms (for those larger
caps too) at:
http://www.avx.com/docs/techinfo/mlc-tant.pdf

Incidentally, I don't think they make 0.01uF silver mica caps. The
biggest I played with were in antenna tuners at 4700pF

Where there is every chance that one silver mica head-to-head with the
ceramic actually exhibit better performance (protecting the diodes).
Perhaps with the scarcity of silver mica, however, 10uF ceramics would
make do (it is not like any precision is demanded to force a selection
of 0.01uF which is boilerplate recommendation from the 50s).

73's
Richard Clark, KB7QHC

On Sat, 03 Sep 2011 22:29:49 -0700, Richard Clark
wrote:

At HF/VHF and above, successful applications comes from throwing uF
solutions at pF problems.

That's not a problem. In order to get obtain decent bypassing across
5 octaves of bandwidth (2-30MHz), one needs to have multiple capacitor
values and types in parallel. The self-resonant characteristics of
the capacitors is the limiting factor. At some frequency, every
capacitor, and its associated lead inductance, will exhibit an
impedance dip commonly known as series resonance. Below this
frequency, the capacitor will look ummm... like a capacitor. Above
this frequency, it will be more like an inductor.
http://www.ecircuitcenter.com/Circuits/cmodel1/cmodel1.htm

None this has anything useful to do with the 0.01uf caps in the
instrument. The diodes are in series with 47K resistors, which are
much larger than any inductive reactance that the 0.1uf bypass
capacitor might present. Since the MFJ-259b only works well up to
maybe a 10:1 VSWR or 5Kohms, the 47K is sufficiently larger than
whatever reactance is presented by the 0.01uf to make the capacitor
characteristics to not be an issue.

While component selection and circuit design are interesting topics,
the current problem is MFJ design quality, MFJ-259b, ESD protection,
and chronic detector diode failures.





--
Jeff Liebermann
150 Felker St #D http://www.LearnByDestroying.com
Santa Cruz CA 95060 http://802.11junk.com
Skype: JeffLiebermann AE6KS 831-336-2558