Print

[N3QE]

If there is someone who has knowledge of the test criteria and the actual conditions under which plate dissipation is determined by manufacturers, that information would be appreciated.

CCS and ICAS ratings are largely determined by economics rather than the laws of physics.

For sweep tubes like the 6MJ6's in your AL84, the ratings in the receiving tube handbooks are generally realistic for several-year-life use in TV sets. They are rated at 30 watts each.

[KM1H]

The insulators for those caps are readily available from Oren Elliott Products.
http://www.orenelliottproducts.com/

Lou probably has dozens of them that fell under the bench

[N3QE]

I don't think there's any problem gluing the insulator back together, or fabricating a new one from ceramic or high-temperature plastic/fiberglass. That particular part of the amp tank does not get crazy hot and polycarbonate might do just fine - it's incredibly easy to cut and form polycarbonate. It will not have the dimensional stability of ceramic for sure but that's rarely an issue for the variable capacitor in an amplifier tank.

I have many old variable capacitors where the insulator has a crack in the neck like yours, and I just be careful with supporting it. Didn't even bother glueing it (if it's only cracked on one side, or gravity keeps everything in place.) Been using them like that in high-voltage sections of homebrew tuners for many years.

Don't get me wrong, you do have to be careful to not let anything arc over and I've had a lot of accidental arcovers especially at legal limit in my homebrew tuners. e.g. placed too close to cinderblock wall and drew a giant arc. But never a problem across plastic of that size.

[W1QJ]

The insulators for those caps are readily available from Oren Elliott Products.
http://www.orenelliottproducts.com/

Lou probably has dozens of them that fell under the bench

I did at one time.  but as Inspector Gluesu' once said "Not anymore"

[HAMHOCK75]

The epoxy was sufficiently cured to do an alignment of the the capacitor plates. Alignment is not perfect but I decided to reinstall the plate capacitor to do a low power test using an HP synthesizer as the source just to be sure there were no other problems. 

The AL-84 tuned up on all bands 160-15M. This one does not have the 10M option installed. Power gain was about 10 dB from 160 to 20M but dropped to 5 dB on 15m. The load capacitor was at maximum except on 15 and 20M. That was expected at low drive levels since the plate resistance is higher at low drive. The Yaesu FT101 finals show the same behavior.

[HAMHOCK75]

Regarding buying capacitor parts from Oren Elliot Products. Here is their answer to my query,

Dear sir --

We don't sell parts to our capacitors, only complete capacitors.

[KM1H]

I guess things changed when Oren died and his brats took over. Oren sold me whatever I wanted, including the bearing grease, as I was repairing amps and rebuilt the variables on a regular basis.

How many insulators do you need?

[HAMHOCK75]

None now. The structural epoxy seems to work fine. The ceramic has no stress on it from what I can tell. The threaded rods slipped right in. I just tightened the nuts.

[HAMHOCK75]

Start testing the AL-84 with a Yaesu FT100D transceiver. In CW mode, about 50 watts drive on 20, 40M gave just under 300 watts output with plate voltage dropping from 1KV to 880 volts. Plate current was 0.6A so 528 watts input power gave about 300 watts output or about 56% efficiency. Load capacitor was still nearly entirely meshed. I had expected a decrease in load C with higher output power.

[HAMHOCK75]

Here is my analysis of what maximum output power should be in CW mode.

where

Pdiss = the tubes plate dissipation spec.
Ppl = plate power during transmit
Popr = plate dissipation in operate mode but with no RF
Pin = DC power in
Po = RF power out

assuming a 50% duty cycle

Popr/2 + Ppl/2 = Pdiss  or

Ppl = ( Pdiss-Popr/2 ) * 2

for example, if Popr = 0 and Pdiss = 30 watts then Ppl = ( 30 - 0 ) * 2 = 60 watts. In other words, the plate is half the time dissipating 0 watts and half the time it is 60 watts for an average of 30 watts.

The AL-84 has a Popr of about 12 watts/per tube because of the bias current of around 12 mA. In this case assuming 6LQ6's with Pdiss = 30 watts

Ppl = ( 30 - 12/6 ) * 2 = 48 watts or 192 watts for four tubes. So assuming an efficiency of around 60%.

Pin - Po = Pdiss but since Po = 0.6Pin we can write

Pdiss = Po/0.6 - Po = Po ( 1.67-1 ) = Po * 0.67 or

Po = Pdiss/0.67

if Pdiss = 192 watts then Po = 192/0.67 = 286 watts
Pin = Po/0.6 = 476 watts

The 6MJ6 has a Pdiss = 35 watts so Ppl = ( 35 -12/2 ) * 2 =  ( 35 -6 ) = 58 watts. For four tubes Pdiss = 232 watts

Po = 232/0.67 = 346 watts
Pin = Po/0.6 = 577 watts

Increasing bias current for better linearity reduces Po so as not to exceed Pdiss

[HAMHOCK75]

Discovered something not on the schematic. The AL84 has a series resonant trap for 11M at the input. It is attached to the input side of R1 ( 22 ohm, 50 watt  resistor ) and ground as shown in the photo below circled in green.

[HAMHOCK75]

Also discovered that W8JI, who designed this amplifier wrote on another web site on February 24, 2004,

The AL84 grounds the grids, and has a negative feedback resistor to improve IM performance. So they are not similar.

300 watts is OK for four tubes on CW. You could get more, but tube life would be shorter.

[W8JX]

If you limit a sweep tube amp to a max of 500 watts INPUT power (not output power) and you should get good tube life. Where they get a bad reputation for eating tubes is pushing them well beyond that.

[HAMHOCK75]

I decided to examine the input impedance of the AL84 to see if it could cause a solid state rig driving it to fold back. The answer was a bit different than expected. The input impedance at the cathode was calculated based upon the voltage drop across R1, the 22 ohm 50 watt input resistor. Based on this the input impedance at the cathode was calculated as about 45 ohms which is in series with R1 bringing the input impedance at the rear panel input connector to about 65 ohms. This is an SWR of about 1.3, but the story gets more complicated. As drive power is increased the cathode impedance stays about the same, but harmonics are starting to appear flowing back towards the driving source. Those harmonics are seen by the SWR bridge of the driving rig as reflected power so it will start to read SWR as higher than 1.3 when in fact the SWR did not change. The image below shows drive at about 6 watts. Test frequency was just above 14 MHz.

The yellow trace is the input to R1, the blue trace is the output side of R1 which goes to the cathodes.



At 28 watts drive, there is a significant increase in harmonic content. R1 does a reasonable job of “swamping” the harmonics from backing out into the driving source.



This is the driver output into a 50 ohm load showing little distortion.



Also, discovered why there is a series resonant trap at the input. It's an FCC requirement from 47 CFR 97.317(a).

(3) Exhibit no amplification (0 dB gain) between 26 MHz and 28 MHz.

That might be drawback for amateur use. Normally there is a capacitor to ground at the input to R1 which should also help short harmonics. A series resonant circuit will pass the harmonics unless the harmonics fall in the CB band.

[VK6HP]

Very nice investigation which underlines some of the complexities of real systems.  Your results are similar to those pictured in my 19th edition (1972) of Orr's "Radio Handbook" (Sect. 7.7).  The case for a tuned input remains as strong as ever

I guess you don't have to leave the 11m trap in place if you're a licensed amateur - at least that'd be the situation here in VK.

73, Peter.