Consequences of a high filament voltage??

[KX5JT]

Hey there AMers.  You all may recall my recent acquisition of a Henry 2K4 floor console amplifier.  It's the one that had a hum which I resolved with a proper filament choke.  Everything was smooth sailing until last week.  

I had been in a pretty long roundtable so the amp was heating up pretty good.  She is rated for continuous duty so I wasn't not too worried about that, the blower works FB.  During a somewhat buzzardly transmission, maybe 3 minutes into it, a raspy eerie squeal emanated from the tube area, my watt/swr meter when crazy.

Here's what I think may have happened.  The amp is a 1971 model and everything in the manual states 220 vac for the mains.  My line voltage is at 242 vac, the "3000 v B+" is indicating "3400 volts" on the plate meter.  No biggie, some amps run a pair of 3-500Z's at 3800 volts.  BUT the filament voltage at the socket is measuring 6 volts ac, not 5 volts ac.

Could this contributed to a filament getting too hot and shorting intermittently?  I have read it does contribute to shortened tube life.  If I'm going to spend the cabbage on new HV rectifiers and tubes (which I will), I want to not have it occur again in a few months!

That's right, the overload relay and then the 15 amp built in circuit breaker started tripping and I have found one of the two HV rectifier blocks to be bad.  One side of one checks shorted both direction.  Each block represents two diodes in series and two blocks form the full wave bridge rectifier.  

Those rectifiers and good 3-500Z's are going to cost me a few months of saving up to replace.  How can I get my filament voltage lowered short of buying an expensive variac to run the amp on?  I can think of resistors but heck they would have to be able to handle a lot current.  Bucking transformer?  I'm not sure how that works but I heard of it.

John KX5JT

[WQ9E]

John,

I missed your earlier posts but when you replaced the filament choke your replacement probably has lower resistance than the original.  It doesn't take much resistance to make a big difference in voltage drop given the current flow through the choke.  The amp is designed with the voltage drop across the choke in mind when the filament transformer is chosen.  Your higher line voltage might lead to a slight increase but not a 20% increase. 

Next problem, the AC range is the least accurate on most volt meters.  Make sure that you are measuring the filament voltage at the socket with the tubes plugged in with an ACCURATE volt meter.  Adjusting the voltage where it is too low will also damage the cathode assembly of your tubes.

To use a transformer to buck the voltage, the primary is connected across the line (the primary must be rated the same as the voltage is is connected across).  Then the secondary of the transformer is put in series with the primary of the transformer you wish to "buck".  Test it with the tubes out as you have a 50/50 chance of connecting it in boost instead of buck mode.  If the voltage went up reverse the wiring of the bucking transformer secondary.  Typically a 12 or 24 volt filament transformer is used depending upon how much you want to reduce the voltage-the secondary should have a current capacity sufficient to carry the primary current of the transformer you are controlling.

Alternatively, you could just run additional wire between the transformer and choke to provide the additional drop.  It must be sized large enough not to heat up excessively while dropping the voltage.  Refer to the resistance per foot tables in a handbook to calculate how much wire is needed.  I used this method to fine tune the voltage to my homebrew Russian tube amp.

[WA1GFZ]

If the tubes still light up and are not shorted to the grids they may be OK. But yes the life will be limited at 6 volts. I don't know the rig but if the heater transformer is not part of the high voltage you could put a resistor in the primary. Or if it is an open frame transformer add a few primay turns with teflon insulated wire. You need a space between the winding and outer core to sneak extra turns in.

[KX5JT]

Hmm... the resistance of the filament choke! Maybe so.

Here are some pics of the amp.  The suspected toob had been showing plate color more one ONE side of the plate!  Upon inspection, there is a darken area on the inside of the glass that corresponds to the side that showed more color.  Also, when last handling the tube I noticed a "springy" loose sound on that toob. 

See the darkened side...
http://www.flickr.com/photos/79319481@N00/4515561628/

John KX5JT

[KX5JT]

I'm actually considering rewinding the toroid in a bifilar fashion.  I kept the toroid of course.  TFE wire?

[WQ9E]

John,

I see that Henry uses the same transformer for both plate and filament so using a Variac or bucking transformer is not that practical.  Henry also put the tube filaments in series which I really don't like with these power tubes.

If the choke you have in is working OK, I would suggest using an additional length of wire between the transformer and choke to reduce the voltage. 

Your bad tube probably has a damaged grid or misalignment of the grid structure.  Although the grid structure is fairly rugged you do want to make sure the loading is sufficient to prevent excessive grid current. 

[KX5JT]

Yes, that is the one thing so far I really don't like about the amp, the filaments are in series.  The filament source voltage is taken off another secondary on the plate transformer.  I wonder if I should really try to just wind the original toroid but in the proper bifilar fashion.  The problem there would be not knowing how much wire/turns to do.  Anyway the measurement I took at the sockets were with no tubes installed so I could be spinning my wheels and maybe the tube was just ready to short anyway.  The pair of tubes were Eimacs and they were not purchased together.  The previous owner told me he bought the amp with one Eimac and one Chinese, and the Chinese tube was soft, so he purchased a supposedly NOS Eimac to replace it. eBay.  How can any Eimac 3-500Z really be trusted to be a NOS!?  I'm considering a pair of Amperex but man they are quite expensive.

It's going to be a while before I can afford the tubes and rectifiers so I will likely try to pull the large components out of the bottom and clean up all that dust in there while saving up for the new components.  Also, I'd like to check the choke filter and it's oil filled cap in parallel.  My meter will check the capacitance, but I'll will be left to a resistance check on the choke.  I suppose I should also make sure it isn't shorting to ground too.

John KX5JT

P.S. I sure have learned a LOT in the last 18 months since becoming an AM operator!  

[WA1GFZ]

The tube running hot may actually be doing more work and the other one may be soft. Don't replace them until you are sure they are bad. You could always wind or buy a heater transformer?? A 200VA variac core would work and just use the winding as the 120VAC primary. Then just wind enough secondary turns with teflon insulated wire to make 10 volts. #12 or #10 should be fine for 15 Amps if you keep the tubes in series. Heck there are plenty of 10 VAC transformers kicking around. Then you can keep the high voltage up. A second transformer will provide a better stable voltage on the heaters.

[KM1H]

You can test one tube at a time by disconnecting the plate lead to the dummy tube and lifting the RF drive wire to that filament. Use 50W drive in key down CW, not AM.

The other option is jury rigging a 5V @ 15A transformer.

Carl
KM1H

[wa6mtz]

I wonder if one could use a stud mount diode in series with one lead of the filament voltage. That will reduce the voltage by about 0.6 volt. It would be pulsating DC but wonder if that would make any difference?

[WA1GFZ]

You could put a bridge in series but diode noise may create crud. Maybe a cap as long as it doesn't peak charge too high. DC heaters not a bad idea.

[WQ9E]

With light bulbs DC operation does result in shorter life due to uneven evaporation of the tungsten filament.  I wouldn't want to chance that reduction in life for high value power tubes.

DC makes sense for low level circuits requiring absolute minimum hum (vintage scope preamps for example) but I wouldn't run a power tube cathode/heater assembly on DC unless it was vintage mobile.

[WD8BIL]

I see that Henry uses the same transformer for both plate and filament so using a Variac or bucking transformer is not that practical.

Sure it is!
His line voltage, as stated, is 242VAC. The amp is set to run on 220VAC. A variac on the line set to 220VAC will bring ALL the voltages to within design levels.
I'd just get a 20 Amp variac and be done with it.

[WQ9E]

Looking at the Henry 2K4, specified input voltage is 115 or 230 so the 20% increase in filament voltage is not from a 5% increase in line voltage.  I believe that it is either due to the new filament choke, an AC meter that is reading high, or a combination of both.

In my neck of the woods high current 240 volt variacs are about as expensive as modulation reactor chokes.

[KX5JT]

Well all of this may actually be the result of an intermittent grid to plate short inside one of the 3-500Z's.  After discussing how all of this came about with several others, I now have a clearer picture.  One side of the plate on one tube had been showing more color.  Understand that this amplifier and it's tubes traveled from Lampasas, TX to Maurice, LA on Valentine's Day this year.  The tubes did travel in their boxes separately.

Well, the amplifier went to SQUEALING after a long keyup.  After unkeying and reducing drive it was okay.  I started an investigation the next night.  Someone suggested to me that maybe it was a bandswitch arcing.  That tube showing color on one side of the plate SHOULD have clued me, but I'm really a novice you know... learning EVERYDAY.  Well I pulled the rf deck and looked under the sockets, at the bandswitch and everywhere.  WHAT I missed was that TUBE.  Everything looked okay.  I put it all back together and that's when she never powered up without tripping the relay overload and after a couple tries, the 15 amp breaker in the amp. 

The next day after taking it all apart again, I noticed a "springy" sound in that one tube when I picked it up.... and a darkened color on the glass on the inside, corresponding to the side where it was showing more color....

Of course I still measured 6 volts unloaded on the sockets.  I am curious to see what the voltage is with the tubes in place now.  But now I am looking at tubes and rectifiers.  Quite a bit of money and I'm not willing to just put brand new ones in without knowing for sure whether I actually solved it all.  A friend is going to send me some old soft pulls and a FW Bridge board for testing purposes.  If the amp seems okay with those temporary components, then I'll consider springing for the new replacements. 

AM operations force us to learn, sometimes the hard way!

John, KX5JT

[WQ9E]

John,

The only valid filamentary cathode voltage measurement is with both tubes in place. All you need is .06 ohm of resistance within the secondary circuit (secondary winding, choke, wiring) to drop .9 volt with 15 amps of current draw with the two tube cathodes in series.  

I imagine you will find your voltage is close to correct with the proper load on the transformer.

If you have a decent scope, you can use it to double check your meter accuracy on AC.  The scope calibration should remain constant whether it is reading AC or DC and you can use a standard carbon zinc C or D cell as a fairly accurate DC source for checking the scope calibration.  A new carbon zinc should be about 1.57-1.58; a zinc chloride will be 1.60-1.63.  Put 3 or 4 new cells in series to provide a reasonable working value and adjust the variable input attenuator of the scope (scope coupling set for DC) to establish a convenient calibration point.  Switch the scope coupling to AC and display the input from a 6.3 volt (nominal) filament transformer.  Calculate the peak voltage displayed from the graticule lines on your scope.  Multiply this peak value by .7 and that is what your AC meter should display when connected to this same source.

[WA1GFZ]

pick up some 4-400 pulls and drop them in

[ke7trp]

Parasitic oscilation was the squeeling...  It took a 500 out.   Thats my guess.  The amp should really not produce more then about 300 watts on AM.

I have about ten 3-500s here. Untested. Some came as spares with amps I have sold over the years. If you need one, Let me know.

there is no way that 242 volts on the mains is causing the amp that much trouble.  If you are worried about tube life, turn the amp off when you are not going to use it. Its instant on. You can flip the power switch and when you get HV, Key down full power. No warm up needed.

C

[W2VW]

Looking at the Henry 2K4, specified input voltage is 115 or 230 so the 20% increase in filament voltage is not from a 5% increase in line voltage.  I believe that it is either due to the new filament choke, an AC meter that is reading high, or a combination of both.

In my neck of the woods high current 240 volt variacs are about as expensive as modulation reactor chokes.

A 125 volt variac will do the job.

[KX5JT]

Parasitic oscilation was the squeeling...  It took a 500 out.   Thats my guess.  The amp should really not produce more then about 300 watts on AM.

I have about ten 3-500s here. Untested. Some came as spares with amps I have sold over the years. If you need one, Let me know.

there is no way that 242 volts on the mains is causing the amp that much trouble.  If you are worried about tube life, turn the amp off when you are not going to use it. Its instant on. You can flip the power switch and when you get HV, Key down full power. No warm up needed.

C

The manual states she can be loaded up to 350 watts carrier AM.  I have actually been running it between 200 and 250 depending how good I can get the AM envelop to look on the scope.  The DX-60 will not key up 25 watts of carrier to give back 350.  I was setting at between 12 and 15 watts to give back the 200  to 250 and still have what looked like nice near 100% modulation.

I'm still thinking that the plate on that one tube that was showing a lot of color on one side and now has blackened the glass is a big clue.   This I noticed but everything seemed okay otherwise.  That is until the squeal. 

A friend, Ken, KA5RHK offered to send me a couple of 4-400 pulls that are soft but will provide my with a way to put something back in to troubleshoot.   Another friend, Sammy, W5DPP offered to send me a HV rectifier.  He said he has owned a LOT of Henry amps in the past and has blown the rectifiers before, usually something else doing it (like a tube) but occasionally they just go out.

So once again the AM community is helping me out!  I love this hobby!

John KX5JT

[KB2WIG]

Some var I recall seing a few ckts utilizing a 120V variac on a 240V line..... Maybee somthing to gogle.


klc

[W2VW]

Parasitic oscilation was the squeeling...  It took a 500 out.   Thats my guess.  The amp should really not produce more then about 300 watts on AM.

I have about ten 3-500s here. Untested. Some came as spares with amps I have sold over the years. If you need one, Let me know.

there is no way that 242 volts on the mains is causing the amp that much trouble.  If you are worried about tube life, turn the amp off when you are not going to use it. Its instant on. You can flip the power switch and when you get HV, Key down full power. No warm up needed.

C

I'm still thinking that the plate on that one tube that was showing a lot of color on one side and now has blackened the glass is a big clue.   This I noticed but everything seemed okay otherwise.  That is until the squeal. 

It is. The filament probably has a dead segment. Ken KC2UDZ just went through this with a 4-1000.

A further clue of this malady would be a change in drive impedance. In other words the SWR between the exciter and amplifier would change.

Last time I heard high pitched whining from an R.F. amp was an exchange of circulating current between a plate choke and DC blocking capacitor because the choke was near resonance. Take a good look at these parts to make sure they have not overheated.

[DMOD]

Have you checked the insulated post that feeds HV to the plates via the parasitic plate chokes?

The screw that clamps the plate connections with the top end of the plate choke (at the stand-off insulator) is known to come loose and cause unequal current sharing and buzzing. Also, this srew and connections should be Deoxed and retightened.

Phil - AC0OB