What causes tubes to go gassy? 4-400's in broadcast TX

[VE7 Kilohertz]

I am currently working on getting my CCA1000D broadcast TX on 160M and am having a dickens of a time getting it actually working on it's original frequency.

I have the complete history of the TX's maintenance right from when it was new in 1972, (which is just awesome) and I notice around 1990 they were having a terrible problem with tubes going gassy after only 3 months (4-400's). This went on for about 2 years and they didn't make any reference to finding the problem, just kept changing the tubes and swapping in pulls and getting warranty replacements. Seems that the TX went into overload and reset itself, they would attend the site and find purple haze in the tubes, swap, reset the bias and it would be fine again for a while.

So, what I am wondering is....could whatever caused the tubes to go gassy still be a problem and perhaps be causing me this grief in getting any RF out of the beast. What causes tubes to go gassy after only 3 months?

Here is what I am seeing on the meters for various tube elements.  This is all in relation to the RF section, readings are for the pair, we'll do the audio later. There is normal grid drive, 20mA, the screen current is quite high, around 80mA but should be around 40-50 I think, cathode Ik seems low at around 50mA each tube, almost like they are cut off, and there is absolutely no plate current or RF out. Plate supply is set to low power at 1800V and there is plate voltage on the PA tubes (confirmed by oops with my fingers).  :shock: The tube line up is 12BY7A OSC, followed by 6146 IPA then straight into the grids of the 4-400's. I pulled the xstal and all the meters drop like they should, so there is RF there. However, I found it odd that I couldn't even hear the OSC in my shack RX like I can with the Bauer or any other RF source that I have going like generators etc. It seems like the OSC is working but at the same time, there is no RF coming out of it???  BTW, yes, the 4-400's are just fine, tested them in the Bauer.  

I have done all the normal "look for broken connections, burnt parts etc." with nothing found.

Help! How can the tubes have all this activity and yet not a peep on the output?

I am going back to study the schematic and finish the log book.

This is just like a good mystery novel....

Cheers all,

Paul
VE7KHz  hopefully getting on 160M tonight with some help from my friends.

[W2VW]

Are you sure that there is plate Voltage on the plates? Almost sounds like the screens are acting like plates with the extra current. Can you temp a 200K or so power wirewound across the plate to ground and see if it does indeed heat up with the plate supply on?

[VE7 Kilohertz]

Hi Dave,

Okay. Good idea. I got zapped last night on the PA plate caps after I shut her down, the bleeder still hadn't dropped it below 300VDC. Now, it's still possible there is voltage there but if it's a flakey connection with not enough Ior V to run the tubes...

Thanks

Paul

[Bacon, WA3WDR]

A class B or class C tetrode or pentode will draw extra screen grid current when driven, if the plate is not loaded.  The reason is that without a plate load, plate current will pull the instantaneous plate voltage way way down, so it won't be very attractive to electrons.  With the plate not attracting them much, the screen will attract many more electrons than usual, because its voltage is still high.  As a result, extra screen current will flow.

The screen can get really hot because of this.  If it gets hot enough, it can start emitting electrons too... and that can increase the screen voltage and make the situation worse.  The tube can go into thermal runaway.  And even though the plate is unloaded, it can dissipate a lot of power, because whatever power it produces is not going to a load.  That means more heat.

Metal has trapped gas in it, and if the metal gets hot enough, that gas will leak out.  The metal itself can evaporate, and any anti-emissive coating can evaporate too, making the situation worse yet.  The heat can cause gas to leak out of the glass as well, and the glass-metal seals around the tube pins can crack and start to leak.  The result is a gassy, ruined tube.

Maybe there is a VHF parasitic that produces no output at lower frequencies passed by the output tank.  The result can be pretty much the same.

I think the problem is that the transmitter has some tricky intermittent in its output coupling circuitry.  That intermittent must have kicked up once in a great while and overheated the tubes, and cooked out some gas inside them.  Station engineers never found the cause, so it could be a very tricky problem that you won't find the first time, so probably you should make some sort of protection system that shuts the TX off if the screen current goes high, or if there are other trouble indications.

[W2VW]

Hi Dave,

Okay. Good idea. I got zapped last night on the PA plate caps after I shut her down, the bleeder still hadn't dropped it below 300VDC. Now, it's still possible there is voltage there but if it's a flakey connection with not enough Ior V to run the tubes...

Thanks

Paul

Build a Jesus stick and install it before doing anything else.
What Bacon sez could be right on the money. It may help to build a resistive load which equals your plate load impedance and see what happens if you pump some R.F. backwards into the output network.

[VE7 Kilohertz]

Doh!!! :roll:

Colour me stupid. Well, it took you guys to put me on the right path. Turns out the high voltage I was seeing on the plate was just space charge off the grid??  I ahhh......well..........mmmmmm.........kinda missed one high voltage wire going to the PA that snuck down in behind the inductor while I was making all the connections the night I put in the xfmrs.

Powered it up and it is working great on 1240KHz and I even have the osc. and IPA working now on 1810KHz and am just starting to work the PA over to get it tuned up. Once I get her going on 1810 will continue up the band to 1945 and should be able to get on tonight.

Look for you later and chat on 3870 as well.

Cheers

Paul