¿ªÔÆÌåÓý

Re: Ideas for "bulletproofing" the uBitx


Jack, W8TEE
 

Where do you buy the RD16HHF1's? I've looked on Mouser, Digikey, and Arrow...none. eBay has them, but I wonder about their pedigree.

Jack, W8TEE

On Thursday, December 31, 2020, 11:23:43 AM EST, Jerry Gaffke via groups.io <jgaffke@...> wrote:


If you want to be bulletproof while using RD16HHF1's (Vds-max of 50v)
you definitely want to keep the power supply for the finals down around 12vdc.?
The IRF510's have a Vds-max of 100v, and so can work fine with a 25vdc power supply.
If they have an adequate heat sink.

Looking at the uBitx schematic, I believe we could remove L8 and L9, and instead
supply 12v to a center tap of the T11 primary (many push-pull amps are designed this way).?
The circuit functions the same, but this arrangement is easier to visualize.??

When the bottom FET is turned on, pin 1 of the T11 primary is grounded.
With the center tap at a constant 12vdc, pin 6 of T11 (the other end of the primary)
jumps up to 24vdc.? ?On the next half cycle the top FET turns on, pin 6 of T11?
gets grounded, and pin 1 of T11 shoots up to 24vdc.

So Vds of each FET regularly reaches a max of twice the power supply voltage.
It can be somewhat more because the transformer at T11 is not ideal.
If we push that power supply to 25vdc, then each FET sees 50 volts, the absolute
max for an RD16HHF1.? Not good.

Running without the antenna, the voltages at the primary should remain about the same.
It's like running a 6.3v filament transformer without a load.?
But nothing is ideal, voltage spikes on the FET drains will rise up some.

The old Bitx40 with its single IRF510 in the final is considerably different.
Where the uBitx's T11 works as a "forward transformer", the output transformer T7
of the Bitx40 works as a "flyback transformer".? With no load, the voltages around
a flyback transformer head for infinity.? This is where you really have to worry about
what happens when the antenna is left disconnected, and why you often see a zener
diode placed on the output of an RF amp.? In the case of the Bitx40? operating at 7mhz,
the Miller effect happens to keep those voltage spikes on the drain in check
because they get coupled to the gate as negative feedback to slow things down.
Here's a dimly remembered conversation about this;??/g/BITX20/message/41057

Jerry, KE7ER



On Wed, Dec 30, 2020 at 01:21 PM, Jerry Gaffke wrote:
With all that said, Farhan is right.
If you don't mind spending an extra $20, the RD16HHF1 will help make your rig more bulletproof.

It solves the heat transfer problem because heat from the die is now easily conducted to the tab.?
So the tab (and heatsink) can get much hotter without the die melting.
With the larger delta-T between heatsink and air, there's better heat transfer there too.
Putting in RD16HHF1's may be easier in some cases than dealing with a better heatsink
for the IRF510's.

Also, since the RD16HHF1 tab is connected to the source pin (grounded on the uBitx)
and not the drain pin (connected to the 12v power supply), you don't have to worry about
shorting the tab to ground.

The better heat conduction will help if high SWR is causing high currents through the final.
The RD16HHF1 is quite a bit worse than the IRF510 if the high SWR is causing high
voltages across the FET, but this should not be an issue with a 12v supply.

Jerry, KE7ER

--
Jack, W8TEE

Join [email protected] to automatically receive all group messages.