Also of note,
totally independent analysis. Many others have written QST with
revised 10/11/2011 11AM est
Regretfully the October 2011 QST, on page 40, contains
a fatally flawed filament voltage management article.
Without question, the QST article gives
life-threatening advice. The article specifically instructs readers to remove a
protective cover and intentionally defeat a safety interlock
by placing the
reader’s hand only inches from lethal voltages. The article further instructs readers to look away from their hand while high voltage is active.
never a reason to defeat a critical safety protection system and activate lethal
voltages with protective covers removed. Anything determined
with the cover removed can be more safely accomplished with the cover in place,
or with power removed. This is true for
Besides safety, the article’s instructions will almost certainly
lead to improper equipment operation or tube life problems. Nearly
all instructions in the QST article, as well as all technical information in
that article, are
wrong for any brand or type of amplifier. Following the article, readers are more likely to degrade emission purity and tube
life than improve either.
- The article claims a 3-500Z filament is rated at 4.8 volts
maximum. This is, without any question, false.
Actual filament voltage range is specified as 5.0 volts nominal, plus or
minus .25 volts. This means specified normal filament voltage
operating range is from
4.75 volts to 5.25 volts. “4.8 volts maximum” is not true, and easily
verified as false. The correct voltage information was readily available to
the author and QST in tube data sheets.
- The article does not correctly correlate filament
voltage to powerline voltage. Filament voltage or power transformer taps
should never be adjusted without knowing minimum and maximum powerline
voltage range. At the very least, knowledge of maximum
powerline voltage is critical. Ideally, filament voltage should center on
the nominal filament voltage and kept within allowable ranges as powerline
voltage varies. Some manufacturers have already allowed for that in
- The article uses less-than-ideal metering, a 2%
average-response AC meter corrected to indicate RMS, connected at the wrong
place, to verify an adjustment the article claims is critical for tube life.
The correct meter would be an accurate true-RMS meter.
- The article does not allow for transformer and line
voltage drop under maximum power supply load and temperature.
Voltage should always above minimum ratings under the lowest
power line voltage condition at the highest amplifier power demand.
- The article predicts tube replacement reduction of ten
times over the life of the amplifier for an attempted 0.5 volt change. The estimated improvement is
unverifiable by any reliable source, and makes no sense.
Please be aware none of this is
about a specific product or person. It is about the article in general, as the
article may be applied to every system in the world.
Following for the article’s instructions guarantees
filament voltage below published minimum filament voltage (the article measures at
the wrong point) at least part of the time. Your expensive tube’s emission-life could be reduced, and your signal’s bandwidth greatly increased, if you
attempt to follow the QST article. The article and QST also needlessly place you
at risk of injury or death by instructing you to remove a cover and use a rod to
defeat a safety interlock. Never do this in any amplifier, no matter what brand.
Advice to defeat interlocks with your hand just inches from exposed lethal
voltages shows a total disregard for user welfare and safety.
This article comes with no
benefit at all to readers. Following the article can actually harm the
reader’s equipment by decreasing tube life. It also hurts all of us sharing the
bands, because reduced emission increases “splatter”.
|Note: To ensure proper context,
the article’s exact wording is shown in “picture
Safety Concerns First
The article says:
Those are good warnings if someone absolutely
around high voltage, but are also incomplete. It is not necessary, nor wise, to
inches from unprotected 3000 volt circuits, especially alone. The ARRL Handbook has much more complete
should anyone ever have no choice other than working around live high voltage
circuits. Article safety falls apart here:
The ARRL Handbook, unlike the QST article, provides reasonable
advice. The ARRL Handbook tells readers to “always follow the equipment manual”
as well as many other important safety guidelines. Removing a protective cover
with power connected is never safe. There is never a
reason to tell people to defeat safety interlocks,
while placing their hand inches from lethal voltages! Telling readers to remove
a protective cover, connect the amplifier to mains, and defeat a safety
interlock is an inexcusable lapse of common safety sense.
- Never activate lethal voltages with protective covers
- Never defeat safety interlocks in any system, ever.
Safety devices are there for a reason!
All through electronics and electrical testing and service industries, universal advice is
never remove covers and disable interlocks.
In some cases, defeating safety devices, or instructing others to defeat safety
devices, can actually bring criminal charges. I do not believe that is the case
with equipment of this type, but it shows defeating safety devices is serious. Here are
Electronics and Electrical Engineering Laboratory NIST guidelines that, at least in part, the article might
high voltage safe distance off site link
Article Technical Accuracy Errors
Mistakes begin with the first sentence.
The author complains about an open filament, but Eimac’s
filament voltage management program is not about open filaments or reducing open-filament failures. The
Eimac program is about
gradual emission loss over extended operating hours in AM/FM broadcast service. You can read the
actual Eimac filament voltage reduction program by
but you won’t find anything in the program about the author’s complaint, an
Open filaments are caused by mechanical shock, thermal shock,
poor initial materials, bad solder connections in tube base pins (common with
import 3-500Z tubes), or bad welds in the filament.
Open filaments are primarily a function of off-on thermal cycles,
mechanical shock, or poor tube assembly or filament processing.
The most common cause of open 3-500Z filaments are
cold solder joints in tube base pins. The article neglects all common 3-500Z failures. The article focuses on one of the least likely
3-500Z failures, an internal open
filament. It tells readers if a non-fix is applied for a non-problem the
amplifier will use ten times fewer tubes over the amplifier’s life!
The article then says:
The author wrote “According to Eimac”, but
came from AG6K, not Eimac. AG6K has no connection to Eimac or any
tube manufacturer, and thus does not determine tube ratings. (ARRL members
see this link)
Below is what Measures actually claims at
We can see QST and the author extracted the above words from “bullet 3” below:
|Circuit Improvements for the TL-922by
Richard L. Measures, AG6K
In its factory-stock form, the TL-922 has the following problems:
1. The filament inrush-current is 48 peak amperes. This exceeds Eimac®’s
maximum allowable rating for the 3-500Z.
2. When operated from a stiff 240vac source, the inrush-current through
the on/off switch is so great that it will eventually cause switch
3. The filament-voltage is typically 5.3V measured between pins 1 and 5
on the 3-500Z sockets when the amplifier is operated from 120V/240V,
60Hz. 5.3V exceeds Eimac®’s maximum filament-voltage rating. According
to Eimac®, each 3% increase in filament-voltage, above what is needed to
achieve full power output, will reduce the useful emission life of a
directly heated amplifier-tube by one-half. The bottom-line is that at
5.3V the useful emission life of the 3-500Zs will be reduced to about
one-sixth of what could be realized if the filaments were operated at
Note that Measures does not link to an Eimac specification, he
just tells us “According to Eimac”.
Note: The light-red text boxes below this point are
actually what Eimac does say, and all of this information was readily available
for many years.
Also of note, the article’s actual reference (AG6K) claims
1/6th life for a change from 4.8 to 5.3 volts, but the article claims 1/10th.
All of this disagrees with Eimac.
Even worse, the article modifies Measures’ self-created 4.8 volts
“full power output voltage” to a
maximum filament voltage rating”. Neither the author nor Measures knows how much voltage is “needed” to produce
full output, because neither measured it! Minimum safe
operating voltage can only
be determined through intermodulation measurements with SSB amplifiers, and
minimum voltage varies from tube-to-tube, band-to-band, and with tube age. For
the benefit of readers, I actually measured minimum filament voltage two
Minimum Filament for IMD
These measurements were made with new tubes. The target IMD was -35dB PEP 3rd order,
with 3000 volts anode voltage. The target power was 850 W PEP. This power level gives
100 watts headroom for acceptable performance as the tube ages. Filament voltage
is rounded to nearest
|Min filament V RMS
|PEP @ -35 dB IM3
With a variety of fresh new tubes in the same test system, we can see minimum
acceptable filament voltage varies as much as 0.75 volts! It is impossible to
set a minimum filament voltage without actually testing tubes. We will see, in
the following text, Eimac agrees with that!
What Eimac Actually Says (not what people say
This may seem obvious to some, but it needs mentioning. The
first requirement, before adjusting filament voltage, is
knowing filament voltage! Eimac’s Filament Management paper tells
potential filament management program
users the following:
Are all tube elements metered in the transmitter?
Elements should be metered for both voltage and current, and meters
should be red-lined to define operation within safe limits. Modern
transmitters may incorporate a microprocessor controlled circuit to
monitor all pertinent parameters. In addition, the following controls
are necessary if effective filament voltage management is to be
Power output metering for an FM transmitter or a distortion level
meter for AM equipment
Accurate filament voltage metering; an iron-vane instrument
is preferred over the more common average responding RMS calibrated type
but true-RMS digital meters are acceptable.
The filament voltage measurement must be made at the tube socket
terminals and filament voltage control should be capable of being
adjusted in 0.1 V increments.
The author used the least preferred type of meter. The
author’s meter is a common 2% tolerance (on AC) average responding meter corrected to RMS. It is
not the “acceptable true RMS meter”!
Further, in the photo above the author measures at the wrong point. Voltage at
measurement point is ~2% higher than voltage between tube pins. This is because
of resistances in the intentionally light gauge wire, and the long socket clips.
Wiring in the AL80 series is intentionally minimally sized to
limit filament inrush current. This protects the tube, and should not be
The author uses a 2% tolerance meter, measures at a point 2% higher than the
filament-pin voltage, does not specify if he measured at minimum or maximum line voltage,
and the meter is not even true RMS. In combination, all of these combine to
produce a totally useless measurement technique.
The author “carefully” adjusts voltage to a useless arbitrarily-determined
4.8-volt meter value with someplace around 4% worse-case accuracy, measured with no
load on the transformer. All of that is
wrong. The tube filament could easily wind up more than 10% low in
necessary operating voltage, which can be more
harmful to life than being too high in filament voltage! It also creates
unnecessary splatter or distortion.
This is a good article on how NOT to
adjust voltage, because the article completely disregards everything important or critical!
Rather than saying what someone says, let’s look at Eimac’s words. Unedited, here is what Eimac actually publishes regarding voltage adjustment:
|When a noticeable change occurs in the output
power or if the distortion level changes, the derating procedure must
stop. Obviously, operation at and beyond this point is unwise since
there is no margin allowed for a drop in line voltage. The voltage
should be raised 0.2V above the critical voltage at which changes are
observed to occur. Finally, recheck power output or distortion to see if
they are acceptable at the chosen filament voltage level. Recheck again
after 24 hours to determine if emission is stable and that the desired
performance is maintained. If performance is not repeatable, the
derating procedure must be repeated. Continuing the Program The filament
voltage should be held at the properly de-rated level as long as minimum
power or maximum distortion requirements are met. Filament voltage can
be raised to reestablish minimum requirements as necessary. This
procedure will yield results similar to those shown in the illustration
(Figure 8), to achieve as much as 10% to 15% additional life extension.
When it becomes necessary to start increasing the filament voltage in
order to maintain the same power output, it is time to order a new tube.
Filament voltage can be increased as long as the increase results in
maintaining minimum level requirements. However, when a voltage increase
fails to result in meeting output level requirements, filament emission
must be considered inadequate and the tube should be replaced. Dont
discard it or sell it for scrap! Put it on the shelf and save it. It
will serve as a good emergency spare and may come in handy some day.
Also, in AM transmitters, a low-emission RF amplifier tube can be
shifted to modulator use where the peak filament emission requirement is
not as severe.
The above is directly from the Eimac Filament
Management program! Adjusting for minimum voltage requires measurement of
distortion or some other parameter to ensure enough reserve emission and
linearity. Eimac also does not claim a large life extension, you can read
the text for yourself.
Actual Eimac emission data, useful if an amplifier is a
successful filament voltage management candidate, appears here:
In the small text above, Eimac says “filament voltage management allows
extended tube life when accompanied by a continuing housekeeping program“.
How many Hams are going to measure distortion weekly, and readjust voltage, or
regulate their power lines?
There are many more conditions that
exclude amateur use of filament management!
The above Eimac graph tells us before entering a filament
management program, we should first run the tube 200 hours. For a typical Ham, that’s
about 6 months of operation to stabilize emission.
The Eimac graph does indicate filament emission life
is dependent on voltage, but also shows too little voltage (stars) is
much worse than too much voltage (dashed line). It also shows end-of-life voltage
increased to obtain full benefits.
actually tells us just less than 5% extra voltage over 100% (center of NOMINAL
filament voltage range)
results in about 40% reduction in life. It actually shows an ~8% increase from
absolute minimum results in about 3 fold increase in emission life. This
is for AM/FM broadcast use under special conditions, not amateur use! It clearly does
not agree with the article.
Most important, we don’t have AM/FM transmitters with regulated filaments
operating 24 hours a day. It is unlikely to nearly impossible that we will ever
see and life change, because amateur tubes very rarely fail from too many
Here is what Eimac says about filament systems similar to
those used in amateur amplifiers:
NOTE: If the filament voltage cannot be
regulated to within ± 3%, the filament should always be operated at the
rated nominal voltage specified on the data sheet.
It should be noted that there is a
danger to operating the emitter too much on the cold temperature side.
It may become poisoned. A cold filament acts as a getter; that is, it
attracts contaminants. When a contaminant becomes attached to the
surface of the emitter, the affected area of the emitter is rendered
inactive, causing loss of emission.
Eimac warns us, if we cannot keep voltage within 3%, we should
use the nominal voltage on data sheets. They further warn us
filaments operated too cold (besides causing splatter or low output),
will cause premature emission failure.
Both Measures, and the article by repeating Measures, ignored Eimac’s
publications related to filament voltage.
The author’s Fluke 77 digital meter is rated at 2% accuracy on
AC, and is not a true-RMS reading meter.
The author also did not measure at the correct points. He
measured at the choke, but there is about 0.1 volt drop between the choke and
the tube pins. Filament voltage decreases more when the amplifier transformer and
filament choke warms up, and when the transformer and power line is under
operating load. Here is how voltage actually
Using AL80B instruction manual
125-volt maximum line voltage settings
Line voltage at time of measurement
Article test point
Actual filament pin voltage (cold) RMS meter
Actual filament pin voltage (hot) RMS meter
If the author really set his measurement point at 4.8 volts
using his meter, the
tube filament could be as low as 4.5 volts or less even without normal power line variations. The
author almost certainly set voltage far below allowable minimum voltage.
Eimac’s application note says nothing of the sort; the author is wrong
about maximum voltage. As we see below,
voltage range is 4.75 to 5.25 volts.
5.25 volts is the maximum recommended Eimac voltage. Apparently no one looked at a
readily available 3-500Z data sheet. We also know voltage was
measured at the wrong point with the wrong type of meter.
Eimac’s widely published 3-500Z voltages show a maximum voltage of 5.25
volts, not the claimed 4.8 volts maximum.
Voltage rating is, and always has been, 5.0 + 0.25 and – 0.25 volts.
Further, the article says:
This is a problem because the article has no references to any application
note. I can’t find
anything Eimac publishes that agrees with that, so I have no idea where that
from. That claims disagrees with the only article reference (Measures), and the
sole article reference (Measures) disagrees with himself significantly at various
places in his writings. Even if it were correct (which it isn’t) the claim also
assumes all tube failures are filament emission failures from long term filament
operation, amplifier life will be many dozens of years, and no other failures
Eimac does show, if we look at the graph for AM/FM broadcast above, emission hours until wear-out might
increase about 3 times for that change. A major problem occurs because we cannot apply that data to
different use of amateur service with today’s lower-quality tubes, especially without structured management. It appears someone just
made up some numbers, and assumed every single failure would be because of
emission depletion over many thousands of hours of steady use.
The AL80B transformer produces the correct high voltage when it also
produces the correct filament voltage. If the filament is too high, so are anode and 12 volt buss
voltages. This is a case where the author never
checked with the factory, and apparently never read the manual.
If the author set voltage at his measurement point to 4.8 volts
with a perfect meter, voltage at
would measure about 4.6 volts with normal amplifier output power. There is 0.1 volts
drop in conductors between the pins and his measurement point, and power line
and transformer loading (with a good power line) causes an additional 0.1 volts
drop. This is below tube ratings, and is unsafe.
It also gets worse. The author used a 2% meter
that is not a true RMS meter, so the tube pin RMS voltage might be 4.5 volts or less
even if line voltage never varies.
If the author adjusted his amplifier voltage during a maximum line voltage
period and his power line voltage decreased 3% from increased power grid load, the tube
could be operating at less than 4.35
Using the author’s meter on a perfect power line, we can’t be fully sure what the filament RMS voltage
is, or what it will be later when we transmit or when power line voltage changes.
USA power line voltage can vary as much as 10% and still be within
specification, although in most cases I have measured voltage stays within a 5%
This article is unsafe, uses wrong voltage values as a target, promises more
tube life change than is possible, probably actually decreases tube life,
ignores Eimac, almost certainly increases splatter, and never even considers
power line variations. It is an example of how not to optimize emission life.
The proper way to set filament voltage in the AL80B is to
monitor your power line voltage for a period of time with a reasonably good
meter, and determine the highest line voltage. You then follow the manual,
setting transformer taps for the next voltage wiring above the maximum line voltage
measured. The amplifier’s high voltage meter, with the amplifier running but on
standby, is useful as an alert of
abnormal power line excursions. If high voltage changes outside of normal ranges
for your line voltage and tap settings, verify the mains
voltage has not moved significantly.
In any amplifier, you must know power line voltage extremes,
and filament regulation extremes, and measure at the right point, to properly
set filament voltage. It is much more reliably set if you never go over the
maximum voltage! Do not set for minimum voltage, unless you know what that
minimum voltage really is and you monitor distortion.
In reality, despite the article’s seemingly dramatic claims, amateurs will
find virtually no tube life change from doing a great deal of work and having a
great deal of anxiety, and doing nothing at all. This is because the vastly
predominant tube failures in amateur radio use are tube shorts, with other
failures related to repeated hot and cold cycling, and tube manufacturing
defects. Emission failures
from filament-hour depletion, while a worry in broadcast service, are virtually
unheard of in amateur radio. However, operating a tube with improperly
implemented reduced filament voltage, as advocated in the article, will make it
more likely the emissions from the filament will decrease over time, due to
poisoning of the emissive material.