Final Cage

Final Cage Repair

If you have good final tubes, a good driver tube, good ALC, good HV caps, screen grid voltage is right, and good cathode and screen grid resistors, but there is still low output power,  what could it be?  Well if all those things are good, by the process of elimination, the only thing left is in the final cage itself.  There’s a bunch of high voltage ceramic caps, a few chokes, and a band switch in there, and anything gone bad, will give you problems.  First, let’s look at the most common problem.

Problems with low power, that seem to go across all bands, on the order of maybe only 40–70 watts out, usually have to do with 3 components.  These are the plate coupling cap,  HV choke bypass cap, and a small choke that feeds the cage from the HV supply.  They are C7, C8, and L1 all in the cage and all are attached to the big upright coil choke between the two 6146’s.

  If you have this problem, you should change out C7 and C8 no matter what they look like, as they can be bad and still look fine.  They may even measure OK on a capacitance meter, and still be bad at RF frequencies.  If L1 is bad, it most likely will look burned or charred in color.  All the RF power goes through C7 which puts a big demand on this one little cap.  It should only be replaced with a Class 1 ceramic, not your everyday, .001 3kv cap.  These have gotten very hard to find.  The cap we provide in our final repair kit is a mini doorknob type, designed exactly for this kind of use.  It is a .001 at 10kv and will likely last a long time.

Take a look at the below schematic.  It’s for an 830S but all the hybrids use the same circuitry, even though the part numbers may be a little different.  It is the schematic of all the tuned network components in the cage.  If you’ll compare this schematic to the 830S  service manual you will see that this one has a few changes.  The manual is wrong on several important connections.

Corrected schematic of the final cage. Note the numbers across the top that indicate which part of the coil is shorted out for each band. With the corrections, it’s possible to figure out how the tank circuit works.

  Replacing C7, C8, and L1 will repair 95% or more of units that are experiencing this kind of low power out.  The more stubborn examples may have problems in the tank tuned circuitry and switching.   Usually it will be the ceramic disc capacitors in the cage, many of them very hard to reach.

  Now, if you suspect there are other things wrong in the cage, maybe the band switch or some of the ceramics, here’s how you can check them out without having to take things apart and dig around in there. First, a simple check that gives you an idea of what the band switch is doing. 

BEFORE YOU DO ANYTHING IN THE CAGE, DISCHARGE THE HV CAPS!!  Short the top of the tall choke between the 6146’s to ground.  Do it twice!!

The Load control is a 3 gang variable capacitor.  Only one gang is used on 24 band and above.  Two sections are paralleled on 21 and 18, and all 3 sections on 14 band and below.  Connect one lead of your ohmeter to chassis ground, and put the other lead on the rear most section, where the isolated non moving plates are.   On 24 band, the rear section should show a ground (low resistance).  As you switch from 24 to 21 band, the middle section should ground out on your meter.  The section toward the front of the radio should ground out as you switch down to the 14 band.   If these are right, then you’ve just checked a small portion of the band switch.  Why do the sections get grounded out?  Because of L5, which is a short as far as the ohmeter is concerned.  More below.

The way the final section works is that the final tubes have to output into a resonant circuit.  All these coil taps and ceramic caps are switched in and out of circuit to cause the output section to resonate on the desired band, with some adjustment room left over to tune your antenna as part of the circuit.   If any caps are not right,  resonance will not be achieved, and you will get low or no power out on that band.  It’s rare for these caps to be blown, but it happens, mainly on rigs that have had a QRO mod (higher power) in the past or maybe operating into a high SWR.  All the power has to go somewhere, and if it won’t go into the antenna it ends up in the final section, damaging the tubes, or some of these other components.

Here’s how to check the whole section.  With a few components removed from the circuit, to isolate it, capacitance can be measured for each position of the band switch.  If the capacitance value is wrong, then something is wrong in the circuit, either a bad cap or the  band switch.  Capacitance meters use a signal in the 20-40 khz range and the coil (L3 and L4) will have no effect at these frequencies, effectively putting all the caps in parallel.  Thus all values will add and we can easily check to see if everything is working correctly.

To do this, C7 must be removed, which you should replace anyway.  The other components are L5 and R24, and they must have one leg lifted.  If you remove the fan, and the rear 6146, L5 is right there on the floor of the cage, going from a terminal strip to ground, right in front of the fan.  I find it easiest to cut the ground side leaving enough lead so it can be resoldered later.  R24, a 10K resistor, but marked 10 on the schematic, also must be lifted. Now, the tank circuit is completely isolated.

Now, connect one lead from the capacitance meter to the chassis, and the other to the top of the big coil (junction of L3 and L4).  Turn the Load control so the plates are fully meshed, and the Plate likewise.  As you switch through the bands, you should get capacitance readings similar to the below chart.  If you get these readings, you can be reasonably sure that the band switch and all the ceramic caps are good.  These caps are not passing 100 watts or more, like the coupling cap, C7, and so if they read good, they probably are.

Band       Plate         caps         Load         caps         Total      (measured)
28             220            0              430            100         750         750
24             220            0              430            100         750         750
21             220            0              860            100        1180       1150
18             220            0              860            100        1180       1150
14             220            0              1290          100        1610       1560
10             220            0              1290          100        1610       1560
7               220            0              1290          370        1880       1840
3.5           220          100            1290         1150       2860       2720
1.8           220          512            1290         3090       5112        5110

The measured results are from my own personal 830S.  Notice that there is good agreement on all bands except for 3.5.  It is off by 140pf.  It could be that C9 on switch section S22-2 is bad somehow, either the ceramic cap itself, or the band switch.  The 1.8 band is pretty much spot on, and C9 is part of that total too, so go figure.  I think it’s more likely that Kenwood changed the cap values for the 3.5 and 1.8 band, and it never got on the schematic.

However this particular 830S tunes the 3.5 band just fine and has good power out.  So, I’m not going to mess with it.  But from this example you can see the kind of readings you get.  Keep in mind also, that ceramic caps are typically + or – 10% or worse (in 1980) so that can be a factor.

Now, if you don’t get the right readings on your C meter, you have to go back to the schematic and figure out which caps being good or bad in each band could give you the reading you are getting.  The  band switch  terminals are marked for which band on the schematic, so it can be done.  I did that in my own example above, and identified C9 as the likely culprit for the wrong reading on 3.5.

The coil is also being shorted out one tap at a time by the band switch.  It is possible something could be wrong with that portion of the band switch and would not show up in this test.  The only way I could see to check it would be to unsolder the coil taps and check the leads with an ohmeter as you switch through the bands.

This has been a solution to a dilemma for us in our shop.   What do you do when you don’t know what is happening in there?  One could very easily take it all apart and find nothing wrong.  This won’t be the final solution to this problem, but certainly will help in figuring it out, for much less effort invested.

Incidentally, I did my best to get the schematic right, but there could be additional errors.  If you find something else, please let me know.

 back to top

Comments are closed.