The AM Forum
March 19, 2024, 12:59:11 AM *
Welcome, Guest. Please login or register.

Login with username, password and session length
 
   Home   Help Calendar Links Staff List Gallery Login Register  
Pages: [1]   Go Down
  Print  
Author Topic: Class D vs Class E Question(s)  (Read 16364 times)
0 Members and 1 Guest are viewing this topic.
W1KSZ
Member

Offline Offline

Posts: 247


« on: September 08, 2017, 09:44:18 PM »

In my brief research, I haven't seen anything that details the differences between Class D and Class E.

They seem to be similar, except that the few articles I read show Class D as an Untuned Output, while
Class E has a tuned Output. That would seem to me that Class E circuits are less sensitive to mismatch
than Class D. Am I correct in that assumption ?

Also I notice that Class D circuits run off the AC Line (no Xfmr) while Class E circuits seem to have a
rather large power Xfmr in them.

At least that's what it looks like, am I off track on that ? Where can I find the specific differences
between the two classes ?

I would like to build one in the 250 - 300 Watt range, but I need to get more info before I jump in.
 
Thanks for any help on this,

73, Dick, W1KSZ


Logged
kb3ouk
Member

Offline Offline

Posts: 1640

The Voice of Fulton County


« Reply #1 on: September 08, 2017, 09:53:10 PM »

The power supply really has nothing to do with the difference between the two classes, that is just how the circuit designer decided to supply the voltage that the circuit runs at, most class D and E stuff I've seen runs at 40 to 50 volts DC, but there was a class D H bridge schematic on here years ago that was using straight rectified AC to make 160 volts DC. The real different has to do with the waveform, I believe class D is more like a square wave that is rich in harmonics, and the untuned output is actually a low pass filter that is filtering off all those harmonics. Class D transmitters have to run into a matched load. Class E output I believe isn't as full of harmonics as the class D waveform, its slightly more like a sine wave, but it needs the tuned output to match to the load and to keep the efficiency up.
Logged

Clarke's Second Law: The only way of discovering the limits of the possible is by venturing a little past them into the impossible
W1KSZ
Member

Offline Offline

Posts: 247


« Reply #2 on: September 08, 2017, 11:45:51 PM »

The difference in the power supply must be due to different FETs being used.

OK, so, taking into account not many antennas are resonant over a wide range
or usually, not resonant where you want to use them, it looks like Class E would
be a better bet.

I found one site that sells pc boards and parts for Class E, guess I'll look into
them in more detail.

tnx for the reply,

73, Dick, W1KSZ
Logged
VE3ELQ
Contributing
Member
*
Offline Offline

Posts: 258


« Reply #3 on: September 09, 2017, 08:46:56 AM »

Class D and E both operate in switch mode where the switches are FETs, usually in pairs in co-phase. They are either full "ON" or full "OFF" so do not operate in the linear portion of the drain conduction curve as in class "A" where they would become in essence variable resistors making lots of heat resulting in very low efficiency.  But FETs are not quite perfect switches. They take some time to transition from OFF to ON and back again passing through their linear range where they will generate some heat. They also have some Drain to Source resistance (RDSon) of a few millohms which when fully "ON" also makes some heat as current squared over resistance. This can be reduced by paralleling multiple smaller faster FETs. So the faster the FET and its driver, and the lower the RDSon, and the lower the drain current the higher the conversion efficiency will be.  Power switching FETs come in many flavors and current ratings. The older silicon FETs are relatively slow, the newer Silicon Carbide (SIC) are much faster and the latest galium nitride (GAn) are extremely fast.

Class E is switch mode and employs a series tuned output circuit which when tuned slightly off resonance reflects a portion of the waveform back to the FET drain which opposes and thereby cancels the applied B+ just long enough to allow the FET to fully switch ON thus reducing the power loss as it transitions through the linear range. Its a clever trick which works very well with plenty of them on the air. But it requires some variable output components and the series variable cap sees very high voltage so needs to be appropriately rated, usually requiring a vac variable.

Class D is also switch mode and employs a low pass output filter matched to a specific load resistance, usually 50 ohms. There is no reflected pulse so it requires fast FETs and drivers to achieve high efficiency. The latest SIC and GAn FETs work very well on 40M and below in class D.

The power supply voltage has nothing to do with the mode D or E. But an RF deck designed for high voltage low current will have improved efficiency due to the lower drain current squared times RDSon losses.

My latest AM transmitter runs class D with a 3 pole low pass filter, using 6 fast SIC FETS with very fast drivers and operates at 160V (under carrier load)  into the modulator, also class D, with 75V at carrier. It makes 400W 1700W PEP and runs cold.  The B+ is rectified 240AC split phase with no power transformer.  It requires an L tuner to match the antenna to 50 ohms.

Hope that helps.

73s  Nigel  VE3ELQ
Logged

A "Tube Free" zone.
W1KSZ
Member

Offline Offline

Posts: 247


« Reply #4 on: September 09, 2017, 11:00:20 AM »

Your second and third paragraphs are what I was looking for, now I understand.

Thanks for the reply,

73, Dick, W1KSZ
Logged
K6IC
Contributing
Member
*
Offline Offline

Posts: 745


« Reply #5 on: September 09, 2017, 04:48:24 PM »

Hi Dick,

You have probably seen WA1QIX Steve's Class E Forum:
http://classe.monkeypuppet.com/index.php?sid=1f90e0dd4dde8aa722282b140a615cfa

AND,   Steve has a site on building,   testing and using Class E transmitters:
http://www.classeradio.com/

Lots of great info on these two sites,   all about Class E.

Good Luck,   Vic
Logged
W1KSZ
Member

Offline Offline

Posts: 247


« Reply #6 on: September 09, 2017, 05:30:10 PM »

Saw that site, lots of stuff.

Tnx, Dick, W1KSZ
Logged
steve_qix
Contributing
Member
*
Offline Offline

Posts: 2599


Bap!


WWW
« Reply #7 on: September 12, 2017, 11:31:33 PM »

Good comments here.

The choice of what to use somewhat is determined by how you like to do things.

First, if you do not have an antenna tuner, use class E - you can tune into a fairly wide range of antenna loads.

If you want something as small as possible, class D would be the way to go.  Class E transmitters are bigger, all other things being equal, because of the output tuning components.  Class D tends to be less expensive to build, again, because you typically don't have any variable capacitors.

In my experience, class E is "easier" to get going because the output network is very forgiving with respect to exact component values in the tuned circuit.

On the power supplies - personally, I like to see a power transformer because it just makes everything so much easier.  The entire circuit is isolated from AC line meaning easy testing and troubleshooting (if you have trouble).  The modulators are as simple as it gets.

Ultimately, it is a personal choice.

Regards,  Steve
Logged

High Power, Broadcast Audio and Low Cost?  Check out the class E web site at: http://www.classeradio.org
wa1knx
Contributing
Member
*
Offline Offline

Posts: 451



« Reply #8 on: September 13, 2017, 04:08:52 AM »

      Way back in the 80/90's with the first fet rig I got going,  I went with class D as I was using a T match.  Among problems
it over came was that it was indestructible.   By keeping the Totem pole fets in some known state, ie either switching
carefully with timing, or on or off and adjusting for delays esp Toff delay, it kept changes to the load from destroying parts.
I could short or open the output.  worst case is the devices got warm. No hv peaks.  I remember taking the output
to a shorting wire sparking it up and down a file, no failures.  Huge gate capacities, slow Ton,Toff,Tdly they were
lousy parts but I got into the mid 90's efficiency.  I got 100-150w out of that beast!
Logged

am forever!
W1KSZ
Member

Offline Offline

Posts: 247


« Reply #9 on: September 13, 2017, 10:57:38 AM »

It's becoming clear to me that if I have a resonant antenna, then Class D is OK, but if I have a dipole,
only resonant at one freq, then Class E is a better bet.

Not because of their operating characteristics, but just an SWR issue. Using an Antenna Tuner just
further complicates the issue.

Time to start haunting Flea Markets for big caps !!

Thanks for all the replies,

73, Dick, W1KSZ
Logged
KD6VXI
Contributing
Member
*
Offline Offline

Posts: 2644


Making AM GREAT Again!


« Reply #10 on: September 13, 2017, 11:10:48 AM »


Time to start haunting Flea Markets for big caps !!

Thanks for all the replies,

73, Dick, W1KSZ

Or eBay for some commie vac variables!

--Shane
KD6VXI
Logged
steve_qix
Contributing
Member
*
Offline Offline

Posts: 2599


Bap!


WWW
« Reply #11 on: September 13, 2017, 11:41:49 AM »

It's becoming clear to me that if I have a resonant antenna, then Class D is OK, but if I have a dipole,
only resonant at one freq, then Class E is a better bet.

Not because of their operating characteristics, but just an SWR issue. Using an Antenna Tuner just
further complicates the issue.

Time to start haunting Flea Markets for big caps !!

Thanks for all the replies,

73, Dick, W1KSZ

The caps don't have to be THAT big  Smiley  How much power are you going to start out with?

A very proven design that has been implemented many, many times is the 8 MOSFET class E RF amplifier.  This will run 45V @ 10 A quite nicely.  Of course, you can run less power - just not more.

For 80 meters, the tuning capacitor requirements are reasonably modest.  You need a variable cap that's good for around 4500V at 500pF.  These are fairly common.

The loading cap can be an old broadcast variable or similar.  You should have 1500-2000pF available.  I often use those 4 gang variable caps that come out of old HP signal generators.


If you build the above RF amplifier, the modulator for this is reasonably straight-forward.  You can use toroids for the two PWM filter inductors.  The power supply requirements are not bad with PWM.  Modest current - 4.5A at somewhere between 120 and 130V will be sufficient.  Simple capacitor filter.

The design is reasonably well documented, and the is a good pool of folks to answer questions if something comes up.
Logged

High Power, Broadcast Audio and Low Cost?  Check out the class E web site at: http://www.classeradio.org
YB1AHY
Member

Offline Offline

Posts: 34


me


WWW
« Reply #12 on: September 14, 2017, 05:59:42 AM »

      Way back in the 80/90's with the first fet rig I got going,  I went with class D as I was using a T match.  Among problems
it over came was that it was indestructible.   By keeping the Totem pole fets in some known state, ie either switching
carefully with timing, or on or off and adjusting for delays esp Toff delay, it kept changes to the load from destroying parts.
I could short or open the output.  worst case is the devices got warm. No hv peaks.  I remember taking the output
to a shorting wire sparking it up and down a file, no failures.  Huge gate capacities, slow Ton,Toff,Tdly they were
lousy parts but I got into the mid 90's efficiency.  I got 100-150w out of that beast!

I also have class D (Dual H Bridge) using PWM modulator supplied direct from AC. I standardize input output filter to 50 ohm using T topology and using analog drive (transformer). Everything was running good without filter and connected to using dummy. What happen when i inserted to filter, the RF Transformer became extremely hot. What i suspect is the large power of harmonic reflected back to transformer.

Until now, i cannot manage that problem, so never use it since then, i still keep it and waiting for solution.

The good thing from class D: No matching network, just brute force filter, and also we can limit the peak voltage without destroying FETs (without considering the same time 'ON' on high side and low side).

Any idea on my problem? Thank you.

73s Agus (YB0DJH) now YB1AHY
   
Logged
KD6VXI
Contributing
Member
*
Offline Offline

Posts: 2644


Making AM GREAT Again!


« Reply #13 on: September 14, 2017, 10:21:31 AM »

Your solution would be to use a duplexes at the output.

From the output you go into a low and high pass filter.

High pass filter output dumps to a 50 ohm dummy load.

Low pass filter output to antenna tuner or directly to coaxial cable.

Harmonic energy is dissipated (and measurable!) into the dummy load.  RF at the wanted frequency is passed straight through the LP filter.  No harmonic energy is reflected back to the output xformer or amplifier.

This has the added benefit of dropping imd by an appreciable amount usually as well.

--Shane
KD6VXI
Logged
wa1knx
Contributing
Member
*
Offline Offline

Posts: 451



« Reply #14 on: September 15, 2017, 04:01:26 AM »

      Shane has probably hit the nail of the head. in theory perfect switches and transformers wouldn't dissipate real power.
In the circuit I used, the fets were truely totem pole. ie directly connected source to drain and a floating supply. no ferrites
were used, air wound coil T match.  since I had no idea how this would work, everything had plenty of heat sink.  They were
digitally driven hard.   when shorted or left open any real power really dissipated in the fets RDS and linear switching time R.
this solved the high voltage problems I had with a single ended final, where I kept blowing fets.  the T match capacitor really developed high voltage even at 100w or so!   That transmitter was bullet proof, however today I'd go with what is tried and proven in here. E
Logged

am forever!
VE3LYX
Member

Offline Offline

Posts: 769


Crystals are from the stone age


WWW
« Reply #15 on: September 18, 2017, 10:01:17 AM »

Nigel,
I I want to build a low pass filter like in you last rig. What parts should I order? I want to start with that because that is the part that bugs me. Once I get the parts I will stop by for a lesson.
(Radio is doing very well . Running all over the house. She is watching me write you patting my head.)
Don
Logged

Don VE3LYX<br />Eng, DE & petite Francais
VE3ELQ
Contributing
Member
*
Offline Offline

Posts: 258


« Reply #16 on: September 18, 2017, 11:09:59 AM »

Don,  The filter values are quite easy to determine using one of the many filter design applications. My favorite is "Elsie" which calculates the values and shows a response graph and allows fiddling with values to achieve the desired result. The inductors are hand wound (there is an app for that too), measured and trimmed as required and the cap(s) are RF grade 1KV silver mica.

The most important and challenging part is determining what filter you need to do the job.  How much harmonic rejection, how many poles, input and output impedance (they don't have to match), voltage and power rating and other stuff. Best is to visit and bring what you built and we can go over it all, make some measurements and come up with a design. Sounds like you want class D.
Glad to hear the AM kitty "Radio" is doing well.

73s  Nigel
Logged

A "Tube Free" zone.
VE3LYX
Member

Offline Offline

Posts: 769


Crystals are from the stone age


WWW
« Reply #17 on: September 18, 2017, 12:08:44 PM »

I like the filter type output. I built a 2 meter amp in the ladt century like that and was more then pleased.
Don
Logged

Don VE3LYX<br />Eng, DE & petite Francais
VE3LYX
Member

Offline Offline

Posts: 769


Crystals are from the stone age


WWW
« Reply #18 on: September 18, 2017, 09:44:02 PM »

So I thought this thru. I thought why not build your circuit and learn from that since I know it works well. I can go rouge later. First I need to kmow how to do it right. I thought back on my days building race engines and helping others learn. Many were always trying to improve or cheap out. Those who listened and psid attention did well. I decided to draw from that and learn from you.
Don
Logged

Don VE3LYX<br />Eng, DE & petite Francais
Pages: [1]   Go Up
  Print  
 
Jump to:  

AMfone - Dedicated to Amplitude Modulation on the Amateur Radio Bands
 AMfone © 2001-2015
Powered by SMF 1.1.21 | SMF © 2015, Simple Machines
Page created in 0.047 seconds with 19 queries.