New Coil in Linear...

[W7SOE]

After a few failed attempts at winding my own air core coils (Fimo failure) I finally produced one usable version.  I installed it in the amp in place of the under-rated roller inductor.  That one shown signs of previous flashover and really zorched when I added a second 813 to the amp.

I still need to add a bandswitch but I have the coil temporarily tapped for 80m now.  The attached photo (cell phone    )shows it tapped at the end as I started finding the right tap location.

No more zorches!

I fired it up into a dummy load.  I can get lots of power out according to my Swan meter but I think the power readings are suspect (high).  I know my plate meter is off but I have my Fluke in series to get a good current reading.  The meter is in series with B- , not the filament transformer center tap.

For example at 800W on the Swan the plate current is 306mA.  With 3200V on the plates that is 979 watts in and a plate dissipation of 179 watts.  But I see a dull red glow on the 813 with a steel plate.  This would indicate to me that I am actually exceeding the plate dissipation max for the two of 250 watts.

I see a fairly large difference in power readings on the meter when a switch between the 1000W and 2000W scales, not a good sign.

Rich

[N2DTS]

The coil looks nice!
I have made them out of tubing, and some #6.

I would verify all the meters read correctly, I keep having the dropping resistors in my hv meters change value, and was running the 813 rig at 3000 volts instead of 2000, at .4 amps, and wondered why the meter was showing 800 watts carrier and things had a hot metal smell!
Forced air cooling likely help save the tubes, the graphite plate 813's are TOUGH!

I dont think you would get much over 70% efficient, if things show more than that, you have a problem, are not counting screen and grid currents, or have a magic rig....

Brett
N2DTS


 

[K1JJ]

Looks like you're making good progress, Rich!

On the coil -  if you're tuned on 75M (looking at the picture) you might need to tap the coil for less turns and add more C1 and C2. (plate tuning and loading) This will increase the Q. It might produce a better match to the tube plates too, so worth a try to experiment and watch the relative power output efficiency amongst different taps..

I realize space is tight in that amp, but the next one you build, try to position the coil so that if has at least one diameter's free space around it. Otherwise anything metal will couple into the field and reduce the coil's Q.

Also, keep an eye on that plate coupling cap - it might not make it for long.  I had a 20kv plate coupling cap PARTIALLY crap out on me the other day. Was a bitch to find. Intermitantly, I would key up and the HV supply would blow all the fuses. I thought the tube was arcing internally. I put another cap in and it happened again. I finally substituted in a new coupling cap (last resort) and it worked. The bad cap passed the hipot test for DC, but with RF, it shorts.  So, remember these symptoms if this happens to you. 

** BTW, the path for the DC short was thru the coupling cap, into my SAFETY RF choke to ground.  Without that safety choke, I would have hot DC sitting on the antenna lead with nothing to do but wait...     There's an example where that choke finally did its job, several times, in fact.


Other than that, as long as you get good power out with no arcing, it's time to paint if blue and ship it!  




T

[W7SOE]

Thanks guys,
   Yes it is tight in there.  That was one of the reasons I bought it, it is small. 

Tom,
   It is interesting that you mentioned the blocking cap, it is rated for 2500VDC when I have 3200 on the plates!  I see an rfparts.com order in my future.

I have not narrowed down the perfect tap for 75m yet, do you have a preferred method other seeing how much power I can crank out on each tap?

Rich

[K1JJ]

Thanks guys,
   Yes it is tight in there.  That was one of the reasons I bought it, it is small. 

Tom,
   It is interesting that you mentioned the blocking cap, it is rated for 2500VDC when I have 3200 on the plates!  I see an rfparts.com order in my future.

I have not narrowed down the perfect tap for 75m yet, do you have a preferred method other seeing how much power I can crank out on each tap?
Rich

Yep, time is ticking on that 2500V cap.... :-)


As for the coil taps.  Go back to a previous post of mine about using the pi network calculator to arrive at a plate inductance.  Use Q=12  and about 4K for the 813 pair plate impedance in linear.  Output = 50 ohms

Once the coil inductance is found (in microhenries - uh)  then use the coil calculator based on your physical coil size to arrive at the approx numbers of turns needed for that uh number.    This shud be pretty close, but try a tap or two in and out looking for best power in vs: power out. Sometimes tapping won't make much difference - it's just changes the Q up or down and power remains close as long as the coil is efficient. Your coil uses wire that is on the borderline to being thin, so you might see a difference using a lower Q = more turns.  You want at least Q=12 to store as much energy as possible (RF cycles) which will help linearity and reduce your harmonics too.

T