PW amplifier power supply

[k7mdo]

Well, I sorted through what I have in the "heavy metal" warehouse.  I have all of the parts shown in the attached schematic.

I think it will be enough for the single 4D32 amplifier discussed in the "PW" thread.

I drew up the schematic and am looking for critics of my handiwork? 

I have a good steel chassis that will hold the transformer... think I will make the power supply completely independent of the amplifier so that if things don't go as planned I can just reuse the supply for some other project.

I also have two choices it seems as I have several 4D22's as well as a couple of 4D32's and filament transformers for either.... not yet sure which one to choose but that will come after the supply is complete.

73, Tom

[K1JJ]

Hi Tom,

The center tapped 800 V transformer as configured will put out only 400 VAC, or maybe 600VDC under load. The tube rectifiers will drop the voltage even more. The choke will also drop some voltage.

I would use a full wave bridge and get over 1000VDC. The 4D32s need that much to give the peak overhead you desire with that big audio capability the PW rig has.   Otherwise you will be dialing the carrier way back to compensate.

Use solid state diodes for less heat and much better rectifier efficiency. The heat waste comes from somewhere and it's from the voltage drop across the rectifier tubes.  

Also, I would add some more capacitance. 4 uF is barely enough to give you the regulation you desire for a clean class B linear amplifer, especially for a cap input circuit. Good idea using the choke.  Be sure the transformer can handle a fullwave bridge config. Some are not designed to carry the extra voltage when the C.T. is not grounded.

Might as well put the primary 115VAC windings in parallel, assuming they are identical, for less IR drop - can't hurt.

T

[k7mdo]

Thanks Tom for the thoughts... I only settled on that circuit because I had all the parts in the stockpile. 

I will re-think the circuit in light of your concerns... I had thought the final voltage would be closer to 800 vdc but in reality I would not mind getting rid of the rectifier tubes....

I think I can avoid the high in rush to the transformer easily enough... etc.

Tom

[K1JJ]

Assuming you go with a full wave bridge:

You could add a Variac to have full adjustment of the HV. This is always good for testing and experimenting with parameters. ALL of my power supplies from big to small use Variacs.

Another method would be to use a choke input config and short out the choke with a relay that can handle the voltage across it - when desired to up the HV, making it a cap input filter.  Put as much filter C on the PS output as you can.

You could also adjust the HV by putting the two 115V primary windings in series or parallel for a X 0.5 or  X 1 AC xfmr output. (Assuming it's fed with 115VAC source)

A Variac would do it all, but a combination of the latter two would work and give some more flexibility.

The choke input config would give better regulation than C input - which is what is needed with that class B amplifier current requirements due to peak power X5 or X6 AM audio cyclic drawdown.

T

[KA2DZT]

I don't thing that circuit will work with the CT grounded.  You need two filament xfmrs and single plate rectifier tubes.  Won't work with 5R4s.

Get rid of the tube rectifiers.  Use SS diodes in a FWB circuit,  you need at least 8 diodes even 12 three in each of the four legs.  Don't worry about the inrush current.  The choke helps to limit that.  Measure the DC resistance of the xfmr secondary.  You could add a small resistor in series after the FWB and before the first capacitor.  You need more capacitance.  Inrush current is related to the amount of input capacitance.  4ufd is not much, put both 4uf in parallel on the input and use as much capacitance as you can find on the output.

Like Tom said, with a FWB the CT has half the DC voltage which you can load or just leave unloaded.

Also, like Tom said,  you MUST put both primaries in parallel or you'll exceed the VA rating of the one primary alone.

Fred

Correction to the above.

Jim has pointed out (in the next post) that the xfmr is 1600vct.  If it is 1600vct then your circuit will work as drawn with the CT grounded.

[WD5JKO]

Guys,

   Maybe I just don't understand....

Looking at the posted schematic, isn't that transformer 800v each side of CT? If so, that is 1600vct. With a FWB rectifier, capacitor input filter, the unloaded DC voltage could be around 2.25 KV

Just asking....is that what you guys are suggesting?

Jim
Wd5JKO

[KA2DZT]

Guys,

   Maybe I just don't understand....

Looking at the posted schematic, isn't that transformer 800v each side of CT? If so, that is 1600vct. With a FWB rectifier, capacitor input filter, the unloaded DC voltage could be around 2.25 KV

Just asking....is that what you guys are suggesting?

Jim
Wd5JKO

Jim,

Good point,  not sure,  it may be 1600vct.  If it is 1600vct then his circuit will work as drawn.

Fred

[K1JJ]

Guys,

   Maybe I just don't understand....

Looking at the posted schematic, isn't that transformer 800v each side of CT? If so, that is 1600vct. With a FWB rectifier, capacitor input filter, the unloaded DC voltage could be around 2.25 KV

Just asking....is that what you guys are suggesting?

Jim
Wd5JKO

Good point, Jim.

However, from a previous thread about the PW amplifier he mentioned:  "The other transformer has plenty of current capability but the plate rating is closer to 800 VAC with a CT."

So I figured the transformer was really 800VAC CT and he simply mislabeled the schematic drawn above.

So let's see... Tom/ MDO, is that transformer 800V or 1600 V CT?  (Measure the winding from end to end, no CT.)

If it's really 1600 VAC CT, then a full wave rectifier using the CT would be best.

T

[k7mdo]

1600 with CT is correct.  Sorry about that earlier post that led folks down the wrong path.  When looking through the transformer stash this one "jumped out" at me. So I grabbed it and started working up the diagram.

I have a great filament transformer as well but going solid state really appeals from construction simplicity.

I will start laying out my HV caps, etc to see what I might need to put this together.  I just felt the 5r4 would use up some of those parts out of the bins. 

Thanks for the thoughts, 73, Tom

[Opcom]

As drawn the only caveat that I saw is that the filament ought be on and warm for 10 seconds before the possibility of drawing >100mA from the tube (each). That should not be an issue in a transmitter. The plates wiring is good because it will keep peak current lower in the cathode than wiring each tube as a single plate rectifier. The voltage drop should be 40V or so but since there are two tubes it may be more like 20. This is from the RCA JEDEC registration data.