Throttling down 4X1 tube filament power in standby?

[K1JJ]

To save power and generate less heat, I'm thinking of a scheme to put the 4-1000A's filaments to sleep during standby.   These are directly heated cathode, so NOT prone to getting contaminated at lower voltages like indirectly heated filaments, yes?  I understand these tubes are "instant on."

By using a 120V to 240 AC isolation transformer with taps, I could switch from 110 or 240 (or 120 / 60V)  using a relay. I was thinking of switching the 4X1 filament transformer outputs from the normal 7.5V -  OR use 3.75V AC to go to sleep.  1/2 voltage  equates to 1/4 power.  They would stay thermally warm at 1/4 power.

Here's how it wud work:  When the PTT is keyed, the 7.5 fil voltage comes up immediately on the first sequencer step.  This would give one second before use.   Using a time delay, the 7.5V full fil voltage would stay on for 5 minutes, regardless of use. (or as long as the PTT is keyed)  When 5 minutes are up, the fil voltage would drop back to 3.75V (1/4 power) and stay there until the next key up.   This would work FB in both old buzz round tables and fast break-in. In OB mode the tubes would rest while receiving -  while in break-in they wud be ready for action as long as a key up occurred within 5 mins.

There will be a lot more thermal cycling, but doesn't his happen anyway with plate power?

Any ideas?

T

[Opcom]

might be better to ramp it up and down even though it is an instant-on.
7.5V/21A is 157W. Is that causing issues like HVAC loading etc?

A nice analog soft start/stop almost forgotten is to insert a bridge rectifier's AC terminals in series with the AC coming in, and put a big TO-3 transistor across the + and - of the bridge. The base current will control the transistor smoothly. An LDR and a small pilot lamp running under-voltage for long life could make a good isolated control. It could be a way to regulate the voltage if that's important, or, since there is always a voltage drop across a bridge and a junction, a relay could be engaged to short across the bridge and recover the last 1.5-2VAC when the transistor's as On as it can get. Maybe it does not matter but the old analog way sounds smoother than snapping the voltage up and down.

[KK4YY]

Tom,

7.5V x 21A = 157.5 x 4 tubes = 630W
Transmitter power on for 4 hours a day.
Cost of electricity 20 cents / kwh.

630W x 4 hours a day x 365 days = 919.8 kwh x 20 cents = $183.96 annual cost @ full voltage.
$183.96 / 4 = $45.99 annual cost @ half voltage (quarter power).
183.96 - 45.99 = $137.97 annual savings without ever transmitting.
With a 50% duty cycle in standby annual savings = $69.98.
YMMV.


Then there's this:

EXTENDING
TRANSMITTER TUBE LIFE
CPI Microwave Power Products Eimac Operation APPLICATION

"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."

Full document found here.
http://www.w8ji.com/Vacuum%20Tubes/CPI%20AB-18%20New%20v2.pdf

I think this applies to directly heated cathodes, as I read it. Maybe you better check.

-Don

[K1JJ]

Tnx for the idea, Op. That seems like a softer start.

Don, I appreciate the time you took to look up that data. I'm still not sure if they are including a directly heated cathode.  Someone who knows chemistry could probably decipher what they mean there in the CPI text you posted. 

I still like the general idea, but am now leaning toward dropping it if there is any chance of shortening the life of the tubes.

T

[KK4YY]

Yes, I thought the CPI app note was ambiguous, at least to my addled mind. A shame, really, as it's hard to find a higher source than the manufacturers own recommendations.

My favorite idea for a filament soft start is a variac. It's as soft as you turn the knob and also allows control of the cathode emission to maximize filament life. A digital voltmeter may be needed to get enough resolution to make this practical. The "Oops, I forgot to turn the variac down before turning the rig back on" syndrome can be avoided with a microswitch, tripped from the variac shaft, wired in series with (and latching) the main power contactor.

I suspect most used 4-1000As are broadcast pulls and may be on the low emission side by the time they fall into amateur hands. This is a way to squeak out the last drop of emission before mounting the tube on a piece of polished oak and displaying it on the mantlepiece (or listing it on eBay as "untested").

All this offers no power consumption savings, which is the point of your post, but it should be kind to your tubes and save on replacement cost. I know it would bring a tear to my eye to buy four of them!

-Don

[W3RSW]

I read the plate amps/ fil. Wattage chart to be for thoriated filament tubes simply because I ran the calc. For 8874's, 8877's etc and the ratios found for those indirectly heated tubes were off the chart.  Further the literature says its for thoriated tubes only in bold type.

E.g. Say an 8874 running .5 amp peak and say .25 amps ave yields 0.25/(6.3*3) = .013
The chart maxes out at .009

A 4x1 running 0.5 amp ave. Ip falls right smack in the middle of the "Good" range.

The PDF is very recent as tubes go, year 2010. Interesting and verifies a lot of older recommendations.

[K1JJ]

Hi Rick,

So in summary, are you saying the data says it's not a good idea to run a 4X1 at reduced filament power EVEN when it is sleeping without drawing plate, grid or screen current?

T

[W3RSW]

Yeah Tom, sometimes we like to play with things to death.  
I think cycling the filament adds variables beyond design intent, but, hey, I know you like to experiment, so if you have bottles to spare then you may be on to something. Further Eimac is saying that filament voltages too low poison the emission.

Also if like many hams you build more than you run, then you may never get enough real world cycles to secure enough data to be sure your more complex circuitry is worth it.
It may detract life rather than add it.  Additionally if your additional circuitry fails, will it fail high or low?  One way your ok, the other your tube is rapidly toast or both?

From what I've read (realizing that Eimac is writing for a 24/7 world) they recommend setting filament voltage as constant as possible at a point under bogey where emission plus headroom is sufficient to meet peak power output AND have given an empirical means to judge if the tube is even" one of the chosen."

I think in our ham applications, a simple way to vary and monitor the filament is good, e.g. A Variac and meter of voltages at the socket.  Then run the 95%- headroom trials for each tube at installation.  Additionally I'd install a step start in ham service since we turn our stuff off and on so much.  I'd put the resistor or bucking transformer in the primary of the fil
xformer  and relay switch these.  

Another idea for turning on the fil. Is a two way switch where a resistor is bypassed on "full up" position of the switch. This relies on manual timing but the worst that could happen would be to apply full voltage immediately....

I have in my 813s xmitter  a sliding variable resistor always in line to fil. Xformer that acts as a combo. Voltage reducer and surge suppressor. It was set once and more or less forgotten. I think a set of taps from the sockets  to a jack out the rear would be enough to check on it once in a while. Can't remember if I did that.

Added later: pix of underside; large slider to left is the fil. primary R. one side directly to screw terminals of transformer.  Guess I'll have to add the sense lines to fil. socket terminals.  Hey, check out that neat plexiglass grid circuit mounting. I'd forgotten about that.

[W3RSW]

Thinking more about cycling filaments, think what 100TH's , 250Th's, 807's, etc. went through in WWII, "the big one."

I can't imagine a more severe enviroment and test bed. I'll bet All Eimac's derivative tube types, especially after WWII carry the best of Eimac's war time experience in their design and ruggedness right down to the materials and hanging structure of their filaments.

I'll bet the tetrodes, etc. for the T-3 got a good workout too.

I Wonder (outside of advertisements) if there's some literature on the carry-through of Eimac's war experience in the design and reliability of their more modern tubes. 

[K1JJ]

OK on all Rick -

Yes, I'll bet WWII did result in better tubes.

Your homebrew rig pictured....  please describe it.


What cornfuses me about doing the fil thang is I was told by someone who has engr experience with power tubes that the indirectly heated cathode got contaminated with fil voltages that were run lower than spec. After all, they depend on the fil for gettering too.  (8877, 4CX-350, etc)

But he said the directly heated cathodes did not get contaminated by lower fil. (4-400, 813, 4-1000A, etc)   And, they get their gettering from running the plate red, as we all know. He thought the slight risk may be thermal cycling, maybe causing seal failure or whatever. But, when a ham makes lots of short transmissions, doesn't the plate dissipation cause lots of cycling anyway?

This is just a bell and whistle feature I cud do without, but it would be cool to have if it was OK wid the tubes...


I was told that a filament step start was not all that critical for indirectly heated cathodes, but was important for directly heated.  So the 4X1 will need a fil step start, no doubt.

T

[W3GMS]

Tom,
I did some internet searching and looked at W8JI's website on this subject.  For those that have interest his site can be found at:

http://www.w8ji.com/filament_voltage_life.htm


Here is the part that talks about filament voltage from his site:

"Filament Voltage

Most tube data sheets specify an allowable range of filament voltage. This voltage range, without question, satisfies warranty requirements and assures proper performance as related to the filament. We can move voltage outside that range only if the manufacturer approves, or if we properly verify performance. Most important and often overlooked, there are minimum and maximum voltages! While generally not mentioned on amateur radio forums and web pages, excessively low filament or heater voltages can actually be more deleterious than needlessly high filament or heater voltages.

Minimum voltages are especially critical in oxide-cathode tubes. While operation above allowable range deteriorates the very long term emission life to perhaps 60% or so for every 5% above nominal voltage, operation below safe minimum voltages will usually destroy the tube in surprisingly short order. This includes tubes like the 8877/3CX1500A7, 3CX5000A7, and 3CX800A7. I've occasionally seen amplifiers that have been modified to reduce tube filament voltage with repeated short tube life. One amateur amplifier's 8877 was set at 4.2 filament pin volts, and went through a new tube in less than one month of casual amateur operation. The brand new tube, because of low voltage poisoning, had no warranty.

Directly heated tubes (those without heater warm-up time) are more tolerant of low voltage, but not immune. In some cases, a filament management program can extend directly-heated power grid tube, but not always. One private website, and the October 2011 QST Magazine (parroting that private website), make grossly exaggerated claims of increased tube life through reduction of filament voltage to an arbitrarily created value. Unfortunately, not only are life increase promises greatly exaggerated; the target voltage is selected incorrectly, and the article's measurement methods are wrong. The result is SSB bandwidth and tube life can be compromised from excessively low filament voltage.

To comply with tube filament voltage management, the filament must be regulated within +- 3% and a host of other criteria must be met. Here is what Eimac says, in Eimac's own unedited words:

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.

If we cannot meet all criteria, we should use nominal rated voltage. Nominal rated voltage is given in every tube manufacturer's data sheet, and cannot be arbitrarily rewritten to a new value at whim."


Joe, GMS

[w1vtp]

<SNIP>

This is just a bell and whistle feature I cud do without, but it would be cool to have if it was OK wid the tubes...


I was told that a filament step start was not all that critical for indirectly heated cathodes, but was important for directly heated.  So the 4X1 will need a fil step start, no doubt.

T

I think most commercial amp makers deliberately use fil transformers that barely provide the required fil power.  So, when AC is first applied, the fil transformer saturates and results in a limit to the startup inrush current.  That might be enough for you Tom

Al

[W3RSW]

Good reference Joe.  Agree.    And right Al, just comfortably make the voltage.

And Tom,

Your homebrew rig pictured....  please describe it.

The whole thing is almost a copy of your 813x813 stuff and was described here and pictured back in '08 I think.  I'll have to dig up the link. Remember when I corrected the meter polarities on your schematic and you said, "oh, yeah, all currents run one way in the tubes,"  etc. or something like that?

If you mean only the filament circuit, it's simply a Raytheon 10V @ 20 amp. fil. xfor. /230V primary, way too big for a pair of 813's, hence the 50 ohm, 50 watt var. resistor on the input side.  So I just didn't bother to step start something which is always in line anyway, hence the dual purpose of adding real resistance to low resistance windings on startup as well as reducing overvoltage at the sockets.

So yes, tube voltages are pretty critical for a lot of reasons.  Tubes don't last forever, are consumable, and Eimac and other good designers knew that.  Didn't meat to sound trite, but "thas a fac."  Most complete tube spec sheets add twice as much design caveats and warnings as the actual tube characteristics.  

In addition to Joe's, There's a lot of stuff on SSB linear sites (gasp) about the newer indirect heated power tubes, their shelf life, how to gently bring them up to date after sitting a long time, glass vs. ceramic, whether it's even worth it, NOS vs. brand new, etc.  Some of it sounds pretty accurate to me.  I have to take someone's word for stuff, I suppose, since I don't design or build the darn things.

Here's a pix looking down the top to refresh your mem. about the Raytheon. It was so cool I just had to use it, even though I bought a new 10V@10A from Hammond.  No almost adequate filament voltage here.  

(oh that was an old pic, js clips to tank coil replaced with proper permanent connections, etc.  Coil is #10 wire, 160 through 40 only.)

[K1JJ]

Very nice layout in the 813 rig, Rick.  Nice work.

OK on using the primary dropping resistor as both a voltage adjustment and as an anti- surge device.  My fils are about 7.8V, so I will add a small resistor in the primary to get them down to 7.5V.  Not much, but something.   I didn't want to take up the space of two Variacs this time, so the resistors will do.


Joe, tnx for the additional info.  They suggest +- 3% fil regulation for the directly heated cathodes, so guess I will not bother with the sleep technique.

T

[w4bfs]

RCA released a ham note or something like that concerning standby .... it went something like a 5% voltage reduction for up to 15 to 30 min, nominal voltage for less time and turn off fil for more time in standby

W5CBS knows some about fil voltg on directly heated BIG tubes ... the tubes he worked on had access portholes for fil replacement and a vacuum pump to keep the vac good .... he said they operated at just enuff fil voltg to make full mod pwr and just a bit more .... this technique maximized fil life

what makes this tricky for directly heated thoriated tungsten carbidized fils is the relatively small operating delta t for the filament between efficient emission and carbide evaporation (around 2400 degrees)

I had an application at reduced fil power (5.6v) with a Svetlana 811-10 at 50W static dissipation for 100's of hours with no notable change in characteristics .... it sounded pretty good too

[K1JJ]

RCA released a ham note or something like that concerning standby .... it went something like a 5% voltage reduction for up to 15 to 30 min, nominal voltage for less time and turn off fil for more time in standby

John,

Based on RCA's notes, 5% is not all that much power reduction.... about 31 watts.   Also, 15 to 30 minutes of standby time to use the technique would not work for break-in very well.

So based on this, I will keep the fils on at 7.5 volts all the time.

It was worth considering.

T

[W3GMS]

Very nice layout in the 813 rig, Rick.  Nice work.

OK on using the primary dropping resistor as both a voltage adjustment and as an anti- surge device.  My fils are about 7.8V, so I will add a small resistor in the primary to get them down to 7.5V.  Not much, but something.   I didn't want to take up the space of two Variacs this time, so the resistors will do.


Joe, tnx for the additional info.  They suggest +- 3% fil regulation for the directly heated cathodes, so guess I will not bother with the sleep technique.

T

Tom,

Maybe put a selector switch label "old buzzard transmission--OBT" and when flipped to the other position it could be labeled "Break In ---BI".  So when your in a round table that's not break in style, just kill the filament completely.  I have a very gentle inrush circuit on my filament supply and I often just shut it off while listening to others in the round table.  Then when its your turn, you sequencer would handle the rest!  I have been doing that for over 12 years with the same pair of 3-500Z without any exhibited harm to the tubes.  They still have good emission and put out full strap.   

Joe, GMS       

[K1JJ]

Tom,

Maybe put a selector switch label "old buzzard transmission--OBT" and when flipped to the other position it could be labeled "Break In ---BI".  So when your in a round table that's not break in style, just kill the filament completely.  I have a very gentle inrush circuit on my filament supply and I often just shut it off while listening to others in the round table.  Then when its your turn, you sequencer would handle the rest!  I have been doing that for over 12 years with the same pair of 3-500Z without any exhibited harm to the tubes.  They still have good emission and put out full strap.  

Joe, GMS      

That's a pretty good idea, Joe.

Yes, since the tubes are "instant" turn on, cycling them completely on and off once in a while won't hurt. Especially if a very soft inrush circuit is used as you suggested. The thermal cycling is the only issue, but my gut feeling is the tubes will outlast me.  I will look into that once I get these arcing problems fixed.

Tnx.

T

[W2VW]

If you use the final in wintertime and enjoy the glow it is 100% efficient and cannot be improved.

20 dollars and a handful of thoriated tungsten.