apache blew second el34

[NR5P]

I got this apache some time back and have been going through it fixing it up.  RF section is working great now.  I did the w3scc mod to it while I was working on it.  I thought I was all ready to put it on the air while adjusting the gain (with w3scc change I only have the one gain control) there was no registration of my voice on the meter all of the sudden current jumped and shorted a tube.  Got two new el34s and went through and checked the audio section and I figured that it was the ancient tube just was fixin to go.  Well same thing happened again.  grrr. I'm i'm goin outside to do some gardening and do some checkin later. 

[John K5PRO]

Have you solid stated the rectumfinders and raised the HV to > 800 VDC? What screen voltage were you running on the 34's? Were they running cherry?

[AB3FL]

When I had my Apache, I had the same problem.  Even with the original rectumfiers, my HV was over 800V.  The first set of EL34s were toasted.  The only place near me that had EL34s was Guitar Center.  I got a set of Grove Tubes from them.  I am not sure which one they were, but they were made in Slovak Republic so I assume they were actually JJs.  They were larger than the original EL34s and worked great and held off the higher voltage fine.  I think it was these, but not quite sure: http://www.guitarcenter.com/Groove-Tubes-Gold-Series-GT-EL34-M-Matched-Power-Tubes-103145629-i1125784.gc

I do remember them saying Slovakia or Slovak republic on them

[WD5JKO]

Those Groove tube EL-34's might be fine. I also present the Teslovak E34L. It boasts a 30 watt plate dissipation, and the tri-metal plate that resists red spots just like the 6L6GC that takes about 100 watts before it glows red. See the GE Ham news about the 6L6GC overloaded..

Here is the E34L from Penta Labs:

http://www.pentalaboratories.com/pdfs/EL34S.pdf

I would look at the Apache, and make sure the modulator idle current is no more than 25 to 30 ma per tube without audio drive. At 800v, at 30ma, you got 24 watts Pd. The high Gm tubes like these are VERY sensitive to a rise in screen voltage. So if nothing is regulated, then you will see a large swing in idle current when varying the AC line from 115 to 125vac with a variac. If you regulate the G1 bias alone, then things get worse since a variation in line voltage would change the screen voltage and this would normally be in part offset by a change in grid bias. This is why some of the circuits of hi fi audio amps might use a screen regulator, but NOT use a regulated G1 bias supply.



Jim
WD5JKO

[WQ9E]

With the current taking off like that it sounds like you might have a stability problem or RF getting into the audio section.  I would carefully check the modification work to make sure nothing is wrong there before sacrificing any additional tubes. 

Do you have proper grid bias at each EL-34 (no driver transformer or grid resistor issues?).

[W2PFY]

Another thing to check it for a carbon path from the tube socket plate lug to the control  screen grid. I think the screen is right next to the plate. It happen to me so I will assume it could happen to others.

[The Slab Bacon]

Also check the biass rectumfryer. IIRC the Apache (at least the earlier ones) used a selenium stack. If it has gone intermittant, it WILL bite you in the but!! Old stacked selenium rectifiers are notorious for breaking down under a load.

[NR5P]

sorry I haven't had a chance to look at it or get back on here until today.  screen is 385 and plate is 880v this is open though with no tube so I guess the plate should drop under load, a lot of people say there apaches run higher than what is supposed to.  I haven't solid stated the rectifiers.  I'm going to go back over my work again and check the whole schematic over again. 

[NR5P]

oh and i checked for the carbon path.  only dark trace I see is in between the plate and ground which is also next to eachother where it got hot from the short.  But it is the same socket that shorted this time as last time which is interesting

[WBear2GCR]

replace socket - put a yabba dabba doo of silicone rubber around one or both pins... should be good to go forever.

                     _-_-bear

[NR5P]

ya im going to replace both tube sockets while I'm at it.  I think thats probably my problem.  Everything else seems to check out.

[NR5P]

anyone know where I can find a replacement socket that will bolt straight in? 

[WD5JKO]

I cannot speak about where to find a socket, but 880V on Pin3, with 380 on pin4, and pin 2 grounded seems like trouble to me. Yes I know that is unloaded, but the spacing is pretty tight. Also no matter what goes on with the socket, running these tubes beyond 25-30 ma per tube at idle will quickly destroy them. Remember that .03 amperes X 800v = 24W. The max Pd rating of a EL-34 is 25w. The E34L I linked to can do 30w. If you run the idle current too low, the distortion quickly rises.

Look at the data sheets at this web page for various versions of the EL-34:
http://www.drtube.com/tubedata.htm

They show -39v Grid1, for 800V plate, 400v G2, and an idle current of 25ma per tube....

So how did Heath run these tubes?

Jim
Wd5JKO

[W2PFY]

try to find some porcelain sockets. I don't know why they weren't used in there in the first place?

Ebay will have them.

[The Slab Bacon]

I hate those tubes!! Pull them out, along with the sockets. Put in a couple of 5-pin sockets, and change them to 5933s (shorty 807Ws). They will be much happier at those voltages than the EL-34s. You can run them AB1, or if the driver has enough scrote, AB2. The problem will go away forever!! With the plate connection on the top of the bottle, it cant short to anything or jump across the socket pins anymore.

(You have to use the 5933s in an Apache as there is not enough room between the socket holes for full sized shoulder bottled 807s)

The 5933s can also be found cheap in junk boxes at hamfests for next to nothing.
They are basically a 6L6 variant with a plate cap, and have no problems at 800v on the plates.

                                                      The Slab Bacon

[NR5P]

Thank you all for your help.  It was the burned spot on the socket.  I replaced the sockets and works now.  I'm going to order another set of tubes probably mullards from tube depot.  I tested it with two different tubes just to make sure it would work.  Does anyone happen to know how much the pentalabs tubes are?  I emailed them but no response.  thanks.