¿ªÔÆÌåÓý

  1. Qex
  2. Topics

Class-E Amplifier Woes

jim conrad
 

My Class-E amplifier (Sokal, Jan/Feb 2001;?; etc) investigation has struggled to find suitable devices for the final and its driver. ?The challenges include:

  • Common BJT PAs cannot switch fast enough to minimize the time spent in the linear region (so much charge, so little time)
  • MOSFET devices switch fast *if* the driver can supply the peak gate current (this can be measured in amps with some MOSFET finals!)
  • No discrete driver of my design (including those with Schottky diodes and other accelerators) has rise/fall times for operation in the upper-HF region
  • The only setup I've gotten to work is paralleled 74ACT00s driving the BS170 (widely used in many QRP projects)
  • But the BS170 Vds(Max) is only 60 V, nor does Pd(Max) provide much headroom for operation at 13.8V
  • Class-E Vds can easily exceed 60 V
A similar QRPLabs thread suggested using paralleled 74act244 buffers, good for 600 mA dynamic current. ?Also suggested for the final was the FDT86256 MOSFET which has a 150 Vds(Max) rating.

Does anyone here have a better suggestion for the driver or finals? ?Do you have Class-E design experiences to share?

--
JimC (KQ7B)
 

¿ªÔÆÌåÓý

Jim, I built a push-pull class-E amp some years ago (for driving an 813-based home brew AM rig).

In case my own experiences might be useful to you, my notes can be found via the link below.


Best regards,

- Jeff, k6jca


On Jun 18, 2023, at 12:34 PM, jim conrad <conr2286@...> wrote:

?My Class-E amplifier (Sokal, Jan/Feb 2001;?; etc) investigation has struggled to find suitable devices for the final and its driver. ?The challenges include:

  • Common BJT PAs cannot switch fast enough to minimize the time spent in the linear region (so much charge, so little time)
  • MOSFET devices switch fast *if* the driver can supply the peak gate current (this can be measured in amps with some MOSFET finals!)
  • No discrete driver of my design (including those with Schottky diodes and other accelerators) has rise/fall times for operation in the upper-HF region
  • The only setup I've gotten to work is paralleled 74ACT00s driving the BS170 (widely used in many QRP projects)
  • But the BS170 Vds(Max) is only 60 V, nor does Pd(Max) provide much headroom for operation at 13.8V
  • Class-E Vds can easily exceed 60 V
A similar QRPLabs thread suggested using paralleled 74act244 buffers, good for 600 mA dynamic current. ?Also suggested for the final was the FDT86256 MOSFET which has a 150 Vds(Max) rating.

Does anyone here have a better suggestion for the driver or finals? ?Do you have Class-E design experiences to share?

--
JimC (KQ7B)

jim conrad
 

On Sun, Jun 18, 2023 at 03:25 PM, Jeff Anderson wrote:
Building an 80-Meter Class E/F RF Amplifier...
Jeff,

Your 80m amp inspires several questions...

  • Is the purpose of R8/R9 to protect U8/U9 from excessive peak current, or to mitigate ringing, or...?
  • Did you find the symmetry adjustment (R3/R6) necessary or was it more just-in-case?
  • Did you ramp up/down the PA ON signal to minimize key-clicks? ?I've been thinking of doing something similar to minimize clicks when the PA turns on/off.?
--
JimC (KQ7B)

jim conrad
 

WRT my third question, I just figured out that the RC time constant of Q1's gate drive provides a soft ramp for the PA.
--
JimC (KQ7B)

 

Hi Jim,

R8/R9 seemed to reduce the ringing, but I never proved this definitively.

The symmetry adjustments were included as a "just in case" feature, but I never did any further work to determine if it was actually needed, or not.

The amp's purpose really is just to provide RF for an AM signal.? So, apart from from turning on and off at the start and end of a transmission, the amp isn't being keyed as one might do for CW operation.? Never the less, there is ramping provided via C1, R1, and R2.

- Jeff, k6jca


On Sunday, June 18, 2023 at 08:42:06 PM PDT, jim conrad <conr2286@...> wrote:


On Sun, Jun 18, 2023 at 03:25 PM, Jeff Anderson wrote:
Building an 80-Meter Class E/F RF Amplifier...
Jeff,

Your 80m amp inspires several questions...

  • Is the purpose of R8/R9 to protect U8/U9 from excessive peak current, or to mitigate ringing, or...?
  • Did you find the symmetry adjustment (R3/R6) necessary or was it more just-in-case?
  • Did you ramp up/down the PA ON signal to minimize key-clicks? ?I've been thinking of doing something similar to minimize clicks when the PA turns on/off.?
--
JimC (KQ7B)

jim conrad
 

After obtaining a non-smoking dummy load for my little transmitter, here's where the investigation wound up:

  • Breadboard implementation (no ground plane)
  • Class D operation into a resistive load (no Class E tank circuit)
  • Using an SI5351A->MCP14A0601->RD15HVF1 power chain
  • Using a 1 ohm resistor with an RF bead for the gate resistor (for dampening ringing on the breadboard)?
  • Initial testing performed at 1 mHz with Vds configured for 9V, threshold current limited at 0.5A reduced actual supplied Vds to 8.3V
  • Rise/fall times are attractive (measurement limited to <10 nS)
  • BUT... MCP14A0601 dissipation is excessive (it heats up rather quickly)
Discussion: ?The MCP14A0601 data sheet () equations 4-1, 4-2 and 4-3 offer insight into the overheating. ?This chip has quick rise/fall times and is likely fine for an efficient inverter operating at 10kHz but does not appear up to the demand of an HF power chain (assuming I've correctly applied equations 4-1, 4-2 and 4-3). ?At this point in this investigation, the most attractive power chain I've constructed for the 5..10W range with potential to operate in the upper-HF region is Hans' Class D design for the QDX ().

Humor: ?Everytime I think I have a better idea, something smokes :(

73... KQ7B




--
JimC (KQ7B)

 

Just a moment for a few memories of my old friend Nathan..

I .Often Visited Nat at Design Automation in 1980s

He was a brilliant engineer, and teacher, with a wide range of interests, especially power electronics topology.

Nat collaborated with other engineers in East Europe and Asia.

He created a CAD program for class E design and wrote the book on class E.

Nat recieved an IEEE award for his class E invention.

As ¨¤ consultant, Nathan Sokal was my mentor.

RIP Nathan, old friend...

from Jon in Paris

1 - 7 of 7