EMRFD Message Archive 10105

Message Date From Subject
10105 2014-06-27 11:45:35 m5evt Class E amplifier design

I am currently working on a Class E amp design, but have some queries if anyone else has any experience.


I currently have the N7VE presentation on these designs, the NA5N pdfs and the QEX article. I have a spreadsheet of the calculations in the QEX article and I am working from those values. I am trying voltages lower than 8 V, so have gone back to re calculate values rather than using "off the shelf" values from other 12 V designs.


A couple of questions:


-My understanding of the VCC-Vo is that Vo is the Vgs value? So in the case of a BS170, it's Vcc-2.1 V.


-I have done some digging around for how to wind a high Q coil, bits of info come up. Anyone have anything to share for some rules of thumb? For 10 turns spread out on a FT37-43, what sort of Q should I typically expect?


I'm using the same coil with a tap in as Dan N7VE used for impedance matching to amp to 50 ohms. Does the value of L2 (this is the coil on the output of the amp that most texts refer to as L2...) need to be adjusted for this series L on the impedance matching?


My initial attempts are around 50 % efficient and not kicking out as much power as I hoped.


Best regards,


Matt M5EVT.

10106 2014-06-27 13:53:31 Graham Haddock Re: Class E amplifier design
Matt:

I don't have the articles in front of me, but several comments:

1.) VCC-Vo is more likely Rdson x Id.  In other words, the difference between
the Voltage supplied and the Voltage output is a loss caused by the drain to source resistance
times the current you are pushing through the drain.  This can easily be 1.5 to 2.5 Volts
in a power FET, so your efficiency will drop rapidly at lower Voltages, like Vcc = 8 Volts.

2.) Type 43 material is a ferrite, and is generally not suitable for tuned circuits. It is
great for broadband transformers and chokes, but the inductance walks around significantly with
frequency, applied current through the coil, and temperature.  Since class E amplifiers
are rather precision tuned devices, I don't think you can depend on an inductor wound on 43
material to stay where you want it to be, from the standpoint of inductance. I would go get
a high mu powdered iron core, instead. Something like type 2 material.

--- Graham / KE9H

==


10107 2014-06-27 15:21:40 Chris Trask Re: Class E amplifier design
>
>2.) Type 43 material is a ferrite, and is generally not suitable for tuned
>circuits. It is great for broadband transformers and chokes, but the inductance
>walks around significantly with frequency, applied current through the coil,
>and temperature. Since class E amplifiers are rather precision tuned devices,
>I don't think you can depend on an inductor wound on 43 material to stay where
>you want it to be, from the standpoint ofinductance. I would go get a high mu
>powdered iron core, instead. Something like type 2 material.
>

Type 43 material is also horribly nonlinear. Not that it matters much for switching amplifiers, but it does matter a great deal for linear amplifiers, even small signal. You're much better off using type 61 or a suitable form of powdered iron.


Chris Trask
N7ZWY / WDX3HLB
Senior Member IEEE
http://www.home.earthlink.net/~christrask/
10108 2014-06-28 02:40:28 jorschei Re: Class E amplifier design

Matt,


Look at the site of Class E design program from Tonne Software

There is a  E class calculator for this type of amplifiers.

I am also interested in anQRP- E class PA from low voltage (one Li-ION cel accu) bud I think dat an DC-DC up coverter is nessesary to have some output.


Best regards  Joris  PE1KTH



 

10109 2014-06-28 05:24:45 dixonglennb Re: Class E amplifier design
Hi Matt,

Sounds like fun, hope you are successful.  Some things I have found: 

My last batch of BS170s from Fairchild had a threshold voltage of 1.6v. You can measure this by putting a high resistor (1M) in the source to gound, tying the gate to the drain, applying voltage and measuring the source to drain voltage.  This will be the threshold voltage.  Your square wave drive needs to go well below and above this value.  I switch the gate with a 74AC family device DC coupled and running at 6VDC, the max Vcc allowed for those parts.  About 5.5V is the minimum gate voltage you need to get the BS170 all the way on.  A single 74AC output can drive two or even three BS170 gates up to 30 MHz.

 Paul, N5AN's document has nice graphics and though his class E seems to differ a little from others, following his recipe should get you pretty good efficiency.

For me it is easiest to build the circuit with just the class E resonant circuit (without the following low pass filter) and adjust the resonance with little or no load.  Note that the resonant circuit capacitor and the following low pass filter input capacitor are often combined, and you need to separate them to do this.  A tunable frequency source can help you find the resonant frequency where the current draw is lowest and efficiency is highest.  Once resonance is adjusted to the right frequency (for true class E circuits this is higher than the operating frequency) you can add the low pass filter and the load.   I sweep the low pass filter separately to make sure it is resistive at the operating frequency and will not disturb the resonant effect when it is installed. 

I have found that sometimes a series tuned low pass works better than a parallel tuned one (series starts with an inductor, parallel with a capacitor).  This has to do with the impedances the filter presents to the harmonics of the resonant circuit.

Move to iron powder on your inductor as others have suggested.  I would also use a larger toroid (t50-6 or t50-2) and perhaps a mica or polystyrene capacitor for low loss, though this was not necessary
10110 2014-06-28 06:38:19 jorschei Re: Class E amplifier design

Hi Glenn,


Could you tel us how many BS170 you are using parallel on 6 volt power and the RF output on 6 volt ?


Thanks

Best regards,

Joris PE1KTH

10111 2014-06-29 07:33:51 dixonglennb Re: Class E amplifier design
Hi Joris,

The 6VDC supply is for the driver IC (the 74ACXX connected to the gate(s)).  The supply for the drain can be anything from 0 to 15VDC, including 6V if you wish.   The output power is proportional to the drain voltage you use, and in the circuit described below 6VDC on the drain would give about 3.5 to 4 watts from four BS170s.

I posted a file years ago: 20M XMT.pdf which I see is still in the files section here.  The only changes are L3 (see below) and the supply for the 74AC74 (and the preceding gate as well) is 6VDC (5VDC is a little too low for good efficiency...at the time the supply was actually 5.6VDC even though it says 5V).  I got ten watts output from four BS170's at 96% drain efficiency from this circuit, which is not technically class E, but more a current mode class D amplifier.  I have been unable to do quite as well from subsequent circuits (even duplicating the schematic...maybe I had magic BS170s or a bad calculator:->) but I can get 90% drain efficiency consistently.  You have to supply twice the operating frequency because the 74AC74s divide the F by two to assure 50% duty cycle on the drivers.  The transmitter covers the entire 20M band without having to adjust anything but the input frequency, but (like most high efficiency amplifiers) needs a fairly good impedance match (maybe 2:1 SWR at full 13.8 VDC...lower voltage on the drain allows higher SWR).

You can use this as a CW transmitter (it meets FCC harmonic specs) but you would need to add some pulse shaping on the drain voltage keying so there are no key clicks.  I can give a circuit if anyone is interested.

One of the coils in the schematic calls for #18 wire, which is tough to wind.  #22 works.  The polystyrene capacitors can be replaced by ceramic with a very slight (0.3dB or so) reduction in efficiency.  Also L3 should be more like eight turns instead of sixteen (it is a choke and the value is not critical, but fewer turns reduces VHF ringing, I think).

It works well built in ugly-style construction on solid copper-clad.  I've been hoping someone would someday build this circuit and confirm my results, so if any of you have any questions, please ask!  If any of you want to try the circuit, let me know and I'll post a picture of a similar circuit I just did.

73,
Glenn AC7ZN
10114 2014-06-29 12:16:13 m5evt Re: Class E amplifier design
Thanks to all who have replied.

I've managed to get 0.8 W at 5 V, with 3 BS170s and 76 % efficiency. However, there is still much to experiment with. I'll report back when I've exhausted all avenues of experimentation!

10116 2014-06-29 13:33:52 jorschei Re: Class E amplifier design

Hi Glenn and Matt,

Thank s for the info I found the schematic in Glenn file  and shall study it .It looks as a balance amp, similar  I  found on the site from Matti OH7SV  the JUMA-TRX1 hi  is  using the IRF110 Fets and line transformer in the balance schematic.  See,   http://www.nikkemedia.fi/juma-trx1/

My plan is to make an SMD pocket QRP  CW  transceiver working on one 3.6 volt cell 3000mA  Li-ION accu thad can be charged from an 5.5 solar cell or wall charger cq  USB.

1.5 or 2 watt is ok for me.  Have the BS170 fets in stock and order also the Fairchild  FDT86256 and the ST.com PD85004 fets and want to try them on 3.6 volt  or  (DC-DC step up converter  for the TX amp if the 3.6 volt is not possible)  to see what the output is.  The E amplifier has a good efficiency for this small accu set.

Interested in your experimental  experience.

73'Joris PE1KTH