Re: The new uBITX boards are here

Ashhar Farhan
 

Jerry, 

This is science at work. We often get unexpected results. Those are the ones that pump up our adrenalin, the game is afoot, Watson! Something unexpected has shown up. Often times, in our bid to functionally decompose, we ten to ignore that the results from string up all the blocks together is also a result of how well they keep working to the spec under changed conditions. 
There is some interactivity between the driver output behaviour and the FET finals. This needs investigations on a rainy day.
- f

On Mon, May 28, 2018 at 9:35 AM, Jerry Gaffke via Groups.Io <jgaffke@...> wrote:
As I say, it's curious.
And I still haven't figured any of this out.
My simulations often gave results that I found counterintuitive.
Never did graduate to messing with actual parts.

In the fourth paragraph of post 22597, Allison warns against too high of a drive impedance.
Though that likely pertains mostly with DC supplies well over 12v.
 
In post 41227 I report simulating the uBitx final in LT spice.  When driving the FET's with 
a 1:1 transformer I saw the input impedance drop from around 50 ohms at 7mhz
down to around 12 ohms at 30 mhz.  So 220 ohms seemed out of the ballpark. 

The Vishay datasheet says the IRF510 has an input capacitance of 180pf typical
when Vds is held constant at 25v.  I assume it's much more in this amp due to the Miller effect.

But I can't argue with what works. 

Jerry


On Sun, May 27, 2018 at 07:54 pm, Ashhar Farhan wrote:
The load on the Mosfet gates is of 130pf in parallel with the gate resistor : 220 ohms in this case. The reactance of the gate capacitance would be about - j50 towards, the higher end of the HF spectrum. Paralleling a resistor with the the gate produces a load impedance for the push-pull driver. As Allison noted, our driver chain does not have enough juice at the higher end of the spectrum, the higher impedance decreases the drive requirements. I came upon the 220 ohms experimentally. I tried values from 10K downwards. More than 1K rested in in-stability. The output starts tampering off below 1K and rapidly so below 470 ohms. I tabulated the results inthe logbook. I don't have access to it now.
- f


Join BITX20@groups.io to automatically receive all group messages.