813 HB Rig Efficiency

[W7SOE]

Threw a HV probe on the plates of the 813 rig (811A's mod'n 2 813s) this morning.  Placed the HV probe upstream of the RF blocking inductor.
I measured 1350V at 300mA plate current.  My watt meter was measuring ~265W into a dummy load.

It it reasonable to conclude that this indicates a 65% efficiency?  Is this a "typical" efficiency for this type of transmitter?

Thanks

Rich

[K1JJ]

Hi Rich,

Maybe you're already aware... but does the 300 ma plate current include screen and grid current too?  Where is the plate current meter inserted? Sometimes, you need to subtract these from the total plate current to get the plate current alone - as with  plate current metering  in the filament CT lead.  This will improve efficiency calculations.   

Reason: A meter in the HV lead gives plate current only, but the meter must be floated and insulated from human contact and ground.  A floating negative PS lead with a 200 ohm resistor to ground is another technique often used to sample plate current at low potential.

If not, we'll talk about it further.

T

[W7SOE]

The plate current meter is in series with the cathodes, from the CT of the filament transformer to ground.  I have attached a rough schematic of the RF deck.  There are separate meters for screen an d grid currents. The grid current was 25mA and I think the screen current was around 60mA.

Thanks

Rich

[K1JJ]

Hi Rich -

As I thought....   

You are reading combined plate, screen and grid in the fil CT since the all go thru that point. 

So, let's recalculate:    300ma - 60 ma screen  - 25 ma grid =  215 ma plate current.

1350V X .215 = 336 W input.

265w output / 336w input =  79% plate efficiency.    Congrats.

Maybe by playing with the L/C tank ratio you can improve it some more and find the sweet spot. Use 10ma more of grid current than recommended to get the highest peaks on the 813s.


T

[W7SOE]

Very good observation!  Very good for two reasons:

Now I understand this
Now I can crank the "plate" current up further!

If I read the 813 spec correctly I can run two 813's in class AB1 up two 300mA, am I understanding this correctly?

Thanks for the help Tom.

Rich

[K1JJ]

Rich,

Wish all problems were this EZ... :-)

Class C / plate modulated is a little different. To set up the 813 plate power out in class C, plate modulated service...

It's really a matter of juggling plate tuning, plate loading, plate L/C ratio,  grid drive, screen current and audio drive to arrive at the highest power out and highest audio peaks at the same time.

Lets say you load the rig up to maximum carrier and then put thru audio and find the audio peaks are flat topping early. (<100% positive)This means you need more headroom and need to try different parameters. 

Much depends on your plate voltage and L/C ratio to give the finals the proper load. Also, you might find that a certain final load is best for your modulation transformer ratio to insure the biggest and cleanest audio peaks for a given drive level.

Since there are so many variables, just keep trying different settings of everything mentioned above until you find a good carrier power level that also lets you put a tone through to give at least 130% positive modulation. You need this headroom to accomodate your asymetrical male voice or flat-topping will occur.   The sweet spot of the rig takes time, but when you hit it, you will know and the rig will really perform.

Generally, you will not go over the tube's dissipation if you are able to get 130% positive and also see no color on the tube plates, if carbon. If metal plates, then a  little color is OK. 

I ran my 813 pair at 3KV and had no problems loading it up to max carrier and audio with good tube cooling. So, at 2500V and lower you will be OK simply looking for the max sweet spots, using recommended screen current and using about 5ma per tube ABOVE the recommended grid current to get max audio peaks.

T

[W7SOE]

Tom,
   This is all valuable advice, thanks.  I will spend some time twiddling the plate, grid and audio drive while watching the power out and the scope.  I'll see  if I can find a sweet spot as you describe.

Rich