The AM Forum

http://www.w8ji.com/qst_the_care_and_feeding_of_a_3-500zg_filament.htm

Hmm, I looked at that article and thought:

1.  At last, QST has something about equipment other than some solid state POS in a fish can powered by a 9 v. battery.
2.  The author had a valid point about high service v. ---> possible high fil. voltage and worth checking out.
3.  His method for lowering his worked I guess, but surely there is a better way than 12 feet of wire.
4.  I have a separate filament supply so I could use a variac.
5.  My service v. is 246 v.
etc. thoughts I can't recall now, but at no time did I think,

Oh, how dangerous, they'd better come up with an article on how to work on equipment without removing any covers.

I guess now that they have been spanked by the "authority" we can forget about ever seeing anything with vacuum tubes in QST ever again.  Maybe the same for tuners.   Someone might get an owie if he sticks his hand in one while transmitting.   It will be the author's fault for not telling the reader how to work on it with the cover on. 

I don't recall reading anything about a hand inches away from B+ so I'll have to go back and read it again.

Anything that thins the herd of the appliance operator, and CB type operator is fine with me. Less QRM thanks to Old Sparky ;D

Nice to see Rauch following in Measures' footsteps with rebuttals.

What would they say about Ashtabula Bill's rigs and shack?  Bill's longevity suggests that exposed HV extends one's lifespan.

I did see some bogosity in that article. The 3-500 runs at nominal 5v, not 4.8v, and Tom is correct in that it's not filament burn-out that's the problem, but loss of emission over time due to destruction of the thoriated tungsten filament. I didn't care for  the copper wire resistor either. But I agree with the essence of the article: monitor the filament voltage at the tube pins and try to keep it as close to the nominal voltage as possible.

Maybe QST should stay with technical articles on building a QRP cw rig in a cat-food tin or an LED "On The Air" sign.

I guess the issue has to do with some AL80 layout specifics not made clear in the article (I don't own one) and the v. measuring instructions given by the author.   There were some general warnings about h.v. though.  Well, the Correspondence section will be interesting in about two months.

Not if the readers all get their tickets via Skype.  The content will have to be about who makes the smallest rubber duck antenna, or even non-radio articles.  They already have an "Eclectic Technology" column which seems to be about computers.  I'd rather read YL News and Views, but that seems to have mysteriously disappeared.

I'm sure as heck not the high power tube expert here. My hands-on experience with hollow state devices is pretty much limited to audio gear back in the good ol' days and some time spent on the Gates FM BC rig at college. But, I read voraciously and I'm getting ready to put a tube rig on the air shortly.

Between the QST article and the rebuttal by W8JI, there were a couple things that just didn't sit right with me. First off, I don't have a problem with working on equipment with the covers open and the power on. How else are you supposed to test it?? I guess I may have become somewhat inured to the excitement by working on my planetarium's star projector which has 40 KV to the anode of the CRT. Sure, the hair on the arms stands at attention, but it's way better than the time I got knocked out of my chair by 650 VDC on a Novice rig I bought half-finished back in '66. I've gotten smarter since then. I've got 36.5 KV inside the video projector in the dome too, and that hasn't killed me yet either. If you survive your early days, you learn to work safely around HV.

I've read a bunch of the Eimac material and never saw that 4.8 VDC was Vmax for 3-500ZG filaments. My mind was raising great big question flags, but I hadn't had time to go check the source material when Don posted the link here. As I remembered while reading, filament voltages are generally given as an integer voltage, plus or minus a tolerance. I'll definitely dig the Eimac spec sheets out of the PDF archive and see what the truth is.

The thing that really annoyed me the most was that WA2HMM's solution to the problem, as he perceived it, was a wad of THHN house wire stuffed in a corner of the amp. At least QST had the minimal decency to have an Editor's Note stating that a commercially available power resistor would have done the same job. No kidding, Sherlock? You really think so?

This article is just a continuation of QST's ongoing tendency to publish articles which do not embody what I consider good construction techniques. I've always believed that "if it works good, it should look good". Many of the articles in QST have photos that look like a bowl of multi-colored pasta was dumped in a dairy box or some other field expedient container. If I was unfamiliar with amateur radio, and my first exposure to the hobby was some of the construction projects in QST, it would be my last exposure to amateur radio. I am constitutionally incapable of doing work that looks that bad.

Obviously, W8JI and WA2HMM have differing views on the filament management of a 3-500ZG. I'm going to dig in the archives and come to my own conclusion. I'm motivated since a 3-500ZG may be coming to live at my house soon. One thing is for sure. I will certainly treat QST articles with a bit more skepticism in the future. I'm not sure who's right on this issue yet, but if there's that much area for doubt, I figure it behooves me to question all of QST's articles and do my own research.

Rob (K5UJ) is probably right. This may be the last article on hollow state devices that we ever see in QST. That'll make more room for drivel and swill about the digital modes and QRP rigs in Altoids tins and continuous QuaRMtests. Oh yeah, and interoperability and $65K Ecomm vans.

My cynicism runneth over.

ldb
K5WLF

Edit: I just remembered that Measures is the guy that's got the "main street at high noon" running feud with ARRL over amp design stuff. Why are they permitting him as a referenced expert in an article about amps? I'm puzzled.

When I read that, what jumped out at me was the article's first sentence or two. I lingered on that for the longest time, trying to figure out why that choice of words was allowed past the editors. (And I see I'm not the only one.)

The second thing that jumped off the page was that ridiculous wad of heavy wire. If he felt the need to reduce the filament voltage (which was a bad idea in the first place), a smaller gauge wire would have looked a heck of a lot neater, don't you think?

Since Tom was once almost killed by high voltage, I can understand why he feels the way he does. He's right! The HV on a low-current CRT PS is probably not going to kill you, but contact with a high-current amplifier power supply like that--even though the voltage is much lower--has a high likelihood of killing someone.

Anyway, nobody's perfect. Sounds like the editors of QST had a deadline and the tech editor was out that day. (Can you say, "teachable moment"?) Something tells me that kind of thing won't happen again in QST for awhile.

Just FWIW, I have run the 4X1 in my 4X1 rig for the last 10+ years at between 6.75 and 7V. It is a coffee stained pull from a shaker table that I bought at the Fredneck fester for $25. I notice no noticable difference in performance, plate current, or outpoot if I vary the voltage from around 6.75-7.5V. (or filament brightness).

I figger that loafing it along at 6.75-7V just makes it a little easier on the thorium coated filament. So far it has worked, I have 10 years and bazillions of hours on it with no loss in outpoot. Just to prove to myself that I wasn't crazy, I put my brand new spare in just for testing purposes and found there to be no difference in performance, so I put the old one back in.

It sure looks purty with the plate glowing cherry red! ! !    ;D  ;D

Bill Orr would have had fun responding to that little QST piece :)

Life of filamentary cathode transmitting tubes is increased by managing voltage over the life of the tube but it isn't as simple as the QST author states.  Proper adjustment involves identifying the point where the emission begins to drop below the point necessary to support desired plate current and then voltage is increased to a bit beyond the point that restores required emission.  This voltage will need to be adjusted over the life of the tube-something best done with a Variac instead of cutting pliers for reducing the length of wire :)  

Sustained operation with insufficient voltage will damage the tube.

I have worked on a lot of live equipment and at times that is by far the best way to make the required measurement or adjustment.  Some people should never go near live circuits but those are often the same people who could somehow damage a cast iron anvil with a foam rubber mallet.

I can understand articles that provide a point of view that may be controversial but printing incorrect specs that can be easily verified (4.8 V filament max) is a sign of carelessness.  I also question how many QST readers own a meter that is sufficiently accurate to make the called for measurement.  AC voltage is generally the lowest accuracy of any function of a typical meter and even if it met its specs when new (doubtful for the typical cheaper meter) it can easily drift over time.  Just because a DMM display is easily readable to .001 volt doesn't mean that is an accurate reading.  

I suspect the 4.8V was due to poor proofreading and meant to be a minimum.

Rauch can often be ignored as about the only time he comments is in attack mode when something doesnt agree with HIS views which have often been proven wrong over the decades.
There are many more better qualified engineers out there.

a fight to the death between Measures and W8JI Friday night covered by the farce network

The number one bogus-osity of the whole thing is cheap amplifier design that uses one big transformer for both filament and plate voltages.  That's OK for BC receivers, tube type TVs and smaller transmitters like the DX-60 and Eico 720, but once you graduate to serious transmitting tubes (anything beyond tubes in the class of the 807 and 6146), you need a separate filament transformer.

The best way to maintain proper filament voltage is to use a transformer that actually puts out about 15% above nominal voltage at normal line voltage (and includes primary taps for 105, 115 and 125 volts), and use a rheostat in the primary to reduce the measured voltage at the tube socket to the desired filament voltage. That's how the BC-610 does it.  The advantages of this method over a variac are that the rheostat is smaller, cheaper and lighter in weight than a small variac, but more importantly, the series resistance in the primary circuit serves as a current limiter to absorb some of the surge when the rig is first turned on, and voltage is applied to the cold filament. 

That's the same reason why, unless a series primary resistor is used, one should never use a vastly over-rated filament transformer.  If you use a 10-amp transformer to feed a 10-amp filament, the built-in voltage drop in the transformer will absorb some of the surge.  If you use a 50-amp transformer to excite a 10-amp filament, the tube suffers the full surge every time the filament is turned on, and this just might cause premature failure of the filament. Most of the damage is caused by the magnetic field during the surge; it physically twists,  compresses and contorts the filament structure.

I once noticed a similar phenomenon with the rf  choke I was using in the plate circuit of the transmitter. It was one of those multi-section pi-wound chokes, and the sections had all clustered together to close the gaps between sections. I replaced the choke, and within a few weeks noticed the same thing again. That's when it occurred to me what was going on.  The solution was to use a larger rf choke, and for added protection, I coated the whole thing with a heavy application of coil dope.

Back to the filament voltage problem, a solution I have thought about but never actually constructed, would be to use a variac coupled to a  heavy duty rotary switch as the main power switch in the transmitter. The variac would feed all the filament transformers and power supplies that turn on as soon as the filaments are lit.  It would work like the volume control in old tube type BC receivers, that used the "on/off volume control". You would turn off the transmitter by rotating the variac fully to the left to disengage the switch. This would  gradually reduce the voltage to everything before cutting power.  The rig would be turned back on by rotating the  control clockwise, first with the "click" as the switch engages, then, the variac is rotated to the right until the filament voltmeter reads full filament voltage.

During stand-by periods, the filament voltage could be reduced to about 80% normal, which according to the RCA book, greatly extends tube life. Letting the transmitter sit idle for hours at a time with full filament voltage is a waste of tube life.  The only drawback is making sure to remember to run the filaments back to full voltage before transmitting, since low filament voltage can do more damage to the thoriated-tungsten filament than does over-voltage.

A quick comment. I found out the hard way...

Years ago I used a pair of 6.3VAC filament 6146's as an RF plate modulated final. I didn't have the correct 6.3VAC filament transformer, so used a 12.6V AC xfmr on a Variac. The 6146 filaments were in series.

It worked FB until one day I decided to change these 6146 tubes around with some others in my junkbox.  While the tubes were still hot, I replaced one with a cold 6146. Within seconds the filament in the hot one opened up. I figgered it was a random bad tube. I then replaced the hot, bad one with a new cold one. Two cold tubes. The tubes worked fine.  Then I repeated this and replaced one hot one with a cold one and the hot filament opened up.  I did this a THIRD time with voltage meters reading the fil pins and was shocked to see a third tube filament pop. I was YELLAFIED!

What was happening?  I soon realized that with the tube fils in series, the hot one hogged all the voltage drop until the cold one warmed up. Thus the hot one had 11V++ volts on it and popped during the initial current surge. A cold filament is a lower resistance until it warms up.

This problem does not happen when using fils in parallel.   Some linears run series fils, so be careful when swapping tubes. Best to let them cool down first and/or use some kind of series current limiting system for warmup. Or use a Variac and bring them up slowly every time.

In addition, many times two fils are not matched well and there will usually be unequal fil voltage drop across the tubes. Series fils is not really the best idea man has ever thought of... ;)

T

When I got my box of 4CX3000As I measured the heater resistance at 5 volts operating because I didn't have a spec for the tube. I figured it was at least 5 volts, turned out to be a 9 volt tube.
The new tubes had a lower resistance than the old brown fin tubes.

You definately said a mouthful there! ! !  In all of the years that I have been fixing radios, I have found/had only a very few filament failures in transformer powered parallel filament radios. The series string radios (like All American 5s) and older series string TVs, this was a very common occurence!

Yeah, in working on AA5s, I read somewhere on the web to never put a cold tube in a hot radio or a hot tube in a cold radio because of this.  If you have a metered variac and use it to turn on an AA5 rx you can see the inrush on the current meter.  It is surprisingly large and the meter isn't fast enough to really get it. 

Tom - very interesting!  Never make that connection (pun).  That make a good case for me to go in and rewire all my fils in the Bendix TA12 for either 12 volt operation and use 1625s in place of the 807s or replace the 12SK7 tubes with the 6 volt equivalent 6SK7s and wire the fils in parallel.  Right now the fils take a 28 volt source to power up. Some creative combination of series connections are made in the filamenet chain.

Would the other alternative be to just make a practice of going off to a coffee break before powering up a xmtr after replacing one in a series wired filament chain? Just thinking out loud.

Al

If I ever get to meet W8JI, I'll have to ask him who pissed in his Wheaties!

Here's all the reader needs to come away from the article and this thread with:

1. Don't exceed or otherwise materially stray from manufacturer's component ratings if you want acceptable trade-off between component life and performance.

2. The QST article author could/should have been more careful with his interpretation and presentation of technical data.

3. The article author should have relied solely on, and clearly referenced, manufacturer published data only.  You don't cite what Eimac supposedly says, then reference an article by a third party...particularly one as controversial as AG6K.

4. The author's implementation of his solution to filament over-voltage is an example of a "ham-hack."  Other posts in this thread have covered this issue quite well.

Jim
W2BVM
"Anything worth doing is worth doing to excess."

Jim, where'd you get that tie?  I like it.

Jim, where'd you get that tie?  I like it.
[/quote]

From my son, the one who drew my portrait some years ago.  It doesn't bear any resemblance to me, but I just love the picture!

Yes, filaments are managed in broadcast transmitters. But just because broadcasters do this, doesn't mean that hams operating their tubes intermittently would also benefit by doing the same thing.

No, I didn't read the QST piece. I don't get QST.

With a new or rebuilt tube, Eimac has always recommended running it at 100% rated voltage for a week or two to 'season' the jug. Then, while carefully watching peak modulation, running at full power and modulation, you can reduce filament voltage just to the point where you see positive mod peaks being impaired. Then turn it back up a hair. But you have to check this from time to time, it's not static. Also, in some cases, power line regulation becomes an issue.

But doing this is bloody silly in ham service, which is *not* 24x7. No ham runs their jugs at 100% power for 8,700 hours a year in transmit mode.

A broadcaster might get 10%-20% longer usable life out of a transmitter tube with proper filament management, but I firmly don't believe this to be the case for tubes run in typical ICAS.

The tube manufacturers all specify normal voltage limits for filament operation. That's where they should be run and not be screwed around with...Unless one enjoys a worthless activity.

Bill

Well stated Homie!!!
Which is essentually what Tom says at the conclusion.


I'm no real fan of Tom's, but the jest of his response is mostly right on. His delivery and people skills could use some work!

The Kenwood hybrids all use series connected 6146's and I havent heard of the hot-cold problem but most usually swap in pairs I guess. Ive rewired the TS-830 used to test amps for the 12V industrial versions since they are far cheaper.

The KW TL-922 amp and a few other brands run their big jugs in series also. I never did understand the reasoning since trying to get an equal didtribution at high current is damn near impossible.

Many amps are running filaments at overvoltage and even 5% high which is in "spec" WILL shorten emission life. Ive been careful to run all mine at around 3% low and longevity has been the result.

Transmitter in a cat food can - BRILLIANT!

Despite what Tom says, I don't see safety as an issue in the QST article. The writer covers that pretty thoroughly:

...let me caution you.  High voltage is dangerous, and may be lethal. Meaning it probably will kill you - switch to safety!  If you are not familiar with working around high voltage, get help from someone who is.

...I unplugged the power cord and removed the cover... with the use of my meter's clip leads connected to the two black leads  from the  filament choke.

Keep one  hand in your pocket.

...plug the amplifier back in, and place the POWER switch to ON.

With an insulated rod, depress the INTERLOCK switch and measure the filament voltage...

Release the INTERLOCK switch, place the POWER switch to OFF, and unplug the amplifier.

Wait at least ½ hour before touching anything in the amplifier... Using an insulated tool, carefully short the plate cap to the chassis...



That looks to me like he covers safety precautions pretty well.  I would say his recommendations are a little over cautious as far as the half-hour wait.  If the rig is unplugged, the HV is shorted out with the jesus stick, and then re-shorted out before touching anything associated with the HV, plus checking the voltmeter reading, I'd say it is safe to proceed.  If there is an open bleeder, chances are the capacitors won't discharge enough in a half hour to make them any safer to touch.

That doesn't mean that I don't have other issues with the article.

Anybody see this? http://www.eham.net/ehamforum/smf/index.php/topic,77559.0.html
Interesting, the discussion that Tom had with the ARRL about that article. Page down to Reply #10. Wow.

It would have been far simpler to have added an expanded scale filament voltage meter to the front panel (or added a function for it to an existing meter) and just used a variac on the amplifier, for cryin out loud. Oh I'm sure that would have deprived someone of a few watts. I read the article but was shocked at the wad of wire. Its not a professional way to do that. It is something I might do in the field, but I can't imagine why.

I think I figured this whole thing out. It is the October issue and therefore a Halloween Horror Story to frighten engineers and technicians!

That thread on Eham is amazing! JI holding court with his usual collection of groupies with a few others thrown in. As JI pretty much gets laughed off forums he keeps moving down the ladder.


Pure BS. He has his head wrapped around a Chinese 3-500 and totally ignores the many decades of US built tubes that seem to run forever when treated properly. Even an Eimac 3-500Z needs to be regettered regularly by getting the anode up to a bright red or dark orange color. With the hundreds of SB-220 and other 3-500 amps Ive repaired or converted to 6M since the 70's Ive seen far more simply weak tubes than outright failures.
The Cetron 572B, 813, 4-250/400, 4-1000 and others are more examples.


I read the article the day QST arrived and did a double take and just considered the source and the publication. As far as the interlock goes I just use a clip lead to bypass; you cant do much of anything with the cover off such as looking for a gassy tube. If Ameritron had used the type of interlock that the plunger could be pulled up to lock and activate it would probably be safer than jury rigging.
Anybody that has worked on amps for a long time has developed their own ways to stay alive. Ive managed since 1956.

.

Same BS JI is famous for



Not just RCA but the groupies have limited reading experience.



Enough already, we already know your experience is limited to a small low end company that has a long history of design issues and failures.


JI then tries to marginalize the above. Anyone with a basic reading ability can see that Eimac has set the guidelines for long life yet our favorite bozos jump on the term theoretical.

That thread has played its course, no real engineer would bother joining it. And JI has shown his true side..its all about the money

Carl

Same here. He seems to have some technical expertise, but I find him very acerbic in some of his responses, especiallly if your response is not how he would have expressed it.

I know of one occasion when a person on QRZ asked about the choice of impedance matching networks for a particular application and a number of repondents including myself answered.

Since my response was not how Tom "JI" would have expressed it, it was criticized and we had a short exchange. I had given in my references many Broadcast texts about Impedance Matching Networks but he never repsonded to, nor apparently looked at those texts, but always pointed to his website.

  
Phil - AC0OB

 "My political ideal is democracy. Everyone should be respected as a person and no one should be deified." A. Einstein.

Just wondering if anyone has ever determined if input drive power contributes to filament heat in directly heated cathode (filament) tubes such as the 3-500Z? With the filament power at a nominal 70W or so, and drive power in GG service at 80W, where does the 80W of power go? I am sure that much gets dissipated in the grid, but all of it? 

In GG service, much of the driving power is passed on through to the output, to combine with the tube's output.  I recall the FCC defined input power to a GG amplifier, as the DC input to the final plus the DC input to the driver stage.  The load through the tube acts like a swamping resistor that helps maintain a constant load on the driver stage, reducing distortion, so that the class-B GG circuit results in a cleaner signal than does the conventional grid-driven circuit.

I believe the FCC went to the additive feedthru power awhile after they went to 1500W out as a few hams were really pushing it with some of the old bottles that needed 300-500W of drive. There were some other tricks they used to interpert the regs.

The earlier amps were always advertised as 1000W CW or 2000W PEP INPUT. Flip the CW/SSB switch for an extra 3dB on CW ::)

Ive had forum battles with JI going back to the mid 90's as did many others. He was always right even when the CE of several other commercial and ham amp companies with much better reputations, Eimac engineers, and EE Professors said otherwise. His web site is regularly changed without attribution when he is forced to recant. It still has plenty of BS on it and not just on amps.

With the output power definition, transmit power is the total rf power input to the antenna.  It doesn't matter whether some of that power comes from the driver stage or all from the final.  That was one of the arguments the feecee used when they proposed the change.

But they shot themselves in the foot with output power by adopting the bogus one-size-fits-all p.e.p. definition (feecee magic), which allows some modes to run more actual power (up to 6 dB) than others, plus probably the majority of the "question pool" community has little clue what p.e.p actually is other than "that what the meter says".

If I ever get to meet W8JI, I'll have to ask him who pissed in his Wheaties!

More likely in his scotch

Thanks Carl and Don for the technical comments; they have been very educational for me.  For one thing, I got in a dispute with Bob K4TAX over something a few years ago (long since forgotten by me) but I remember his giving his power with the additive feedthrough method (g.g. amp) and I thought he was nuts.  Well it turns out I was the one who was an idiot.  I owe him an apology.

The feedthrough power of a grounded grid amp is probably the closest thang we get in this whirl to a "free lunch."

Besides the addition of drive power to the output, that power also acts as regenerative (negative) feedback to make the signal cleaner. Just like an unbypassed cathode resistor used in an audio amp.  Very stable config, used extensively at VHF for voltage amps too.

Of course, the downside is that the GG config is harder to drive than a grid driven stage, but WTF. The grid driven stage would also be harder to drive if some RF negative feedback was introduced, so there... ;D

I love GG for linear service, as do the Eimacs of the whirl as shown by their tube designs.

T

Wonder if anyone has ever tried push-pull grounded grid in a class B modulator.  It would take more driving power (at least 50 watts of audio to drive a modulator for a KW), but maybe you could use a large class AB1 power amplifier, like a quad of 6L6s or larger in push-pull/parallel, without worrying so much about the internal (plate) resistance, and the coupling between windings of the driver transformer wouldn't be such an issue.

It seems like such a simple solution to a long-perplexing problem, but I never have heard of anyone doing that, so is there some drawback I am overlooking?

I did. In 1995 and 6 I had a 4-1000 modulated by a pair in cathode driven audio service. Used 1:1 mod iron @ 6 KV. It was a kluge and I didn't have the test gear to prove it's worth.

you would be feeding the drive audio into the filament.  running ac on filaments would probably introduce hum, have to have dc filaments.

2 seperate fil transformers. Feed center tap of each with audio. The audio transformer "secondary" center tap is the cathode return to b-.

Mine was not that simple because there were not too many off the shelf several hundred watt 4 ohm to 60 ohm center tapped transformers available to me at the time. I landed up using more transformers than a KA2DZT transmitter!

Today, I'd just grab a core from Antec and mod it. 

oops, class B needs two tubes, i was thinking of it being single ended for some reason, which makes me wonder, could you have a grounded grid single ended audio amp, and in that case would it need to have dc fed filaments?

Single ended class B works for RF leen-yar service because the fly-wheel effect of the tank circuit fills in the missing half of the rf sine wave.  For audio, it has to be push-pull in order to accommodate both halves of the audio waveform, above and below base-line.

so it couldn't be run as another class? like a regular class a audio amp can be single ended.

Some very early receiving tube manuals listed single ended Class B specs. Dont remember distortion details.

Dave,
I just bought a 1KVA Antek power transformer for my HPSDR final. They make very nice stuff.