Class E/D of 40 Meters

Welcome and feel free to ask questions about class E, mosfet's, and AM.
You must be registered to post.

Re: Class E/D of 40 Meters

Postby steve_qix » Thu Feb 16, 2012 7:42 pm

Hi Jack,

Ok, I guess I'll jump in here :)

It is most likely working better in class E for a variety of reasons, including the fact that, as the frequency is increased, it is more and more difficult to switch the FETs on and off, as occurs in class D. The timing is also a bit trickier. In class E, the output capacitances, etc. are all part of a resonant circuit so the only thing you have to worry about forcing energy into is the gate.

I use 1:1 (with 1 turn in each winding) transformers because, as you add more turns, the DC component will saturate the cores and cause loss and heating. If you don't run a lot of current relative to the core cross sectional area, you may never get to the saturation point, but as you run more DC current, eventually the cores approach and reach saturation. The cores can also saturate on modulation peaks, because as far as the cores are concerned, modulation is DC. I have built a lot of class E transmitters with more turns and fewer cores, but these were all relatively low power (and I would regard 50 watts per module if you are running 40 or more volts at carrier as somewhat low power).

If you could post your gate waveform with a single IXDD414 driving 2 gates at 7mHz, it would be interesting to see what that looks like.

Regards,

Steve
Visit the class E web site at:

http://www.classeradio.com
User avatar
steve_qix
Class E Guru
 
Posts: 434
Joined: Fri Jan 31, 2003 11:03 am
Location: Townsend MA

Re: Class E/D of 40 Meters

Postby K9ACT » Thu Feb 16, 2012 11:57 pm

steve_qix wrote:Hi Jack,

I use 1:1 (with 1 turn in each winding) transformers because, as you add more turns, the DC component will saturate the cores and cause loss and heating. If you don't run a lot of current relative to the core cross sectional area, you may never get to the saturation point..


This ran very nicely on 80 and 160 at about 200W @ about 95% efficiency. Until or unless I solve the efficiency problem on 40, I am running about 100W using 20V and 7A. There are two stacks of 2 cores so I do not think this is a saturation problem.

One thing I am not clear on is how to run higher voltage without increasing the current. With tubes and a PI net, I would just reduce the loading but this does not seam to work that way. The tuning cap sort of acts like a loading cap and the loading cap doesn't seem to do much of any thing.

>If you could post your gate waveform with a single IXDD414 driving 2 gates at 7mHz, it would be interesting to see what that looks like.

Image

I think the flat top on one is a probe or scope anomaly as they look the same when viewed with the same probe on the same channel.

js
K9ACT
Class E Enthusiast
 
Posts: 146
Joined: Sun Jun 27, 2010 12:03 am

Re: Class E/D of 40 Meters

Postby kb3ouk » Sat Feb 25, 2012 10:45 am

Could you replace the LPF with a tuneable pi-network?. In that case, would the power output of the amp peak as it was tuned through resonance (like a tube amp), or would something like Steve's efficiency meter circuit work here? Because it sounds to me when the filter is tuned to where it needs be, the effiency will be the highest at that point.
Shelby
kb3ouk
 
Posts: 2
Joined: Sat Feb 11, 2012 9:36 pm

Re: Class E/D of 40 Meters

Postby Gilly » Thu Mar 01, 2012 5:18 am

Hi Jack

Thought I might give my twenty cents worth..
Gate waveforms are interesting as they all have a story to tell..
Your one IXDD414 into two 11N90s is giving its best to charge and discharge the gate capacitance but just cannot turn on the FETs long enough to enable high efficiency at 7Mhz...
Will ignore the bottom waveform as probe anomaly as it appears to only just make 10 volts anyway..
If you draw lines down from the 10 Volt mark to ground you can see the duty cycle is well under 40%..
Somewhere in the low 40s gives maximum efficiency..
As this photo of one of my waveforms shows...

Image

What would be interesting to do ... is to remove one FET and see what happens to the waveform..
You would except to see steeper Rise and Fall etc:
Also on 160 / 80M you can get away with some sloppy construction, but on 40M more care is required.
It would be best to have the lead length between the output of the IXDD414 and the gate of the 11N90 as short as possible .. even as close as 1 inch if you can.......
I presume too the waveform is without carrier voltage so under modulation even more demand on the driver is required...

Maybe others might want to comment as well...


Wayne
Gilly
Class E Enthusiast
 
Posts: 167
Joined: Tue May 15, 2007 6:59 pm
Location: Mt.Eliza Victoria Australia

Re: Class E/D of 40 Meters

Postby steve_qix » Thu Mar 01, 2012 8:42 am

Thank you Wayne! Very well described :D
Visit the class E web site at:

http://www.classeradio.com
User avatar
steve_qix
Class E Guru
 
Posts: 434
Joined: Fri Jan 31, 2003 11:03 am
Location: Townsend MA

Re: Class E/D of 40 Meters

Postby Diego-CX4DI » Thu Mar 08, 2012 6:37 am

Some Guys require a source for ferrite cores material 61, please check this:

http://www.cwsbytemark.com/index.php?ma ... th=206_227

I was purchase this items several times with very good results.

73's
Diego
CX4DI
cx4di@adinet.com.uy
Class E with PWM modulation
Diego-CX4DI
Member
 
Posts: 43
Joined: Fri Sep 30, 2005 10:22 am
Location: Montevideo-Uruguay

Re: Class E/D of 40 Meters

Postby K9ACT » Thu Mar 08, 2012 9:13 am

This is simply a data point from my seat of the pants engineering but I see no difference in results between using 61 vs 43 material in the output transformer on 40 meters.

I made the switch in both my configurations and there was no measurable difference in the characteristics of either.

One has single drivers driving single fets and the other has single drivers driving 2 fets.

As it stands now, I have a 40m Class E that works well at 100W which seems about all that is needed on 40.

The other one is driving my 8000's on 40 and loafs along with only 12V DC. The same supply does all. This is of course, RF only ... no modulation.

I guess the case is closed for me unless someone has some concrete ideas for 40 meters.

js
K9ACT
Class E Enthusiast
 
Posts: 146
Joined: Sun Jun 27, 2010 12:03 am

Re: Class E/D of 40 Meters

Postby steve_qix » Fri Mar 09, 2012 11:51 pm

Class E definitely works very well on 40 meters - I have a 16 MOSFET transmitter up there operating at around 90% efficiency. Use the same modulator as the 80/160 meter RF amplifier.

Getting the drive right is very important. Once you've got that nailed down, the class E output network is pretty straight forward. I use type 43 material on 40, but that's about as high as I would recommend going with 43.
Visit the class E web site at:

http://www.classeradio.com
User avatar
steve_qix
Class E Guru
 
Posts: 434
Joined: Fri Jan 31, 2003 11:03 am
Location: Townsend MA

Re: Class E/D of 40 Meters

Postby K9ACT » Sat Mar 10, 2012 10:17 am

Let's talk about the output network a bit.

As has been stated, my RF decks are W1VD Class D designs in which I have replaced the LPF with a tuned tank circuit ala QIX schematic.

His called for 5uH for 80 and I started with 3uH for 40 and it worked out of the box after a fashion.

It does not exactly work like a Pi net on a tube rig as the tune and load functions seem swap places depending on how they are set but it gets the job done.

What I find curious is that it seems to make very little difference what the inductance is beyond a certain minimum.

Below the minimum, the signal has distortion but from around 3uH to 15uF, the only difference I can see is where the tuning caps end up. The output level, efficiency and distortion are the same as at 3uH.

Any thoughts?

js
K9ACT
Class E Enthusiast
 
Posts: 146
Joined: Sun Jun 27, 2010 12:03 am

Re: Class E/D of 40 Meters

Postby steve_qix » Tue Mar 13, 2012 6:27 am

The output network values will have an effect on the efficiency and also the voltage across the series capacitor.

The larger the inductor, the more voltage you will get across the series capacitor. Also, the overall efficiency will be better with a smaller inductor, to a point. I have conducted a lot of experiments with this, and it is true in every case I've tested.

Ideally, the output will "peak" when adjusting the loading capacitor. This is generally the optimal point of operation - right at the peak or slightly on the higher capacitance side of the peak.

It sounds as if your output network is functioning properly. Are you getting good efficiency? If the gate drive is good - the proper voltage (around +12V) and the waveform is correct (not too pointy), *no* overlap in the "on" times (if you have a push pull arrangement of any kind - single ended or using the same output transformer), and the on time is less than the off time - usually 40% to 60%, everything should work quite well !
Visit the class E web site at:

http://www.classeradio.com
User avatar
steve_qix
Class E Guru
 
Posts: 434
Joined: Fri Jan 31, 2003 11:03 am
Location: Townsend MA

Re: Class E/D of 40 Meters

Postby K9ACT » Tue Mar 13, 2012 11:03 am

steve_qix wrote:Ideally, the output will "peak" when adjusting the loading capacitor. This is generally the optimal point of operation - right at the peak or slightly on the higher capacitance side of the peak.


That is most interesting because that is exactly what I stumbled into. I adjust the tuning for a low level of current and then peak the loading for max RF and then go back and adjust the tune to the current I want.

>It sounds as if your output network is functioning properly. Are you getting good efficiency? If the gate drive is good - the proper voltage (around +12V) and the waveform is correct (not too pointy),

I think the "too pointy" is the problem because of the single drivers on two fets but I think I can improve it some more now that I understand the tuning. The efficiency seems to increase with voltage but I have been reluctant to go much above 20V because I was not able to control the current. It was about 80% last night at 20V and 6A.

I will tinker some more with it. I am going to run it on the net today.

For anyone in range, the Noon Time 'Forum (1700 GMT) is on or near 7.295 now. Timtron checked in last week.

Thanks,

js
K9ACT
Class E Enthusiast
 
Posts: 146
Joined: Sun Jun 27, 2010 12:03 am

Re: Class E/D of 40 Meters

Postby steve_qix » Tue Mar 13, 2012 6:33 pm

K9ACT wrote:
steve_qix wrote:For anyone in range, the Noon Time 'Forum (1700 GMT) is on or near 7.295 now. Timtron checked in last week.


I'll listen around for you tomorrow. If I hear anything, I'll check in with my 40 meter class E rig.

Regards,

Steve
Visit the class E web site at:

http://www.classeradio.com
User avatar
steve_qix
Class E Guru
 
Posts: 434
Joined: Fri Jan 31, 2003 11:03 am
Location: Townsend MA

Re: Class E/D of 40 Meters

Postby K9ACT » Tue Mar 13, 2012 7:59 pm

The sky fell down today after about 30 mins. Condx were normal when we started but suddenly everyone but one station vanished.

Hopefully, things will be better tomorrow and we will listen for you.

Several stations reported distortion and FMing on my sig but switching to my DDS VFO made it all better.

Not sure why this xmtr is getting into my homebrew VFO but the 8000's don't seem to bother it.

Don't recall if this was yours or Jay's VFO design but I added a buffer stage to it and never had a problem before. Always something to putz with.

js
K9ACT
Class E Enthusiast
 
Posts: 146
Joined: Sun Jun 27, 2010 12:03 am

Re: Class E/D of 40 Meters

Postby Gilly » Mon Mar 19, 2012 5:10 am

Hi Jack

The photo is of a IXDD414 directly into the gate of a single 11N90 with no Vds......
The Crystal freq was 7.505Mhz....
Length about 1 inch...
The bottom waveform is the input to the IXDD414 from a 7404...
Maybe should have used a 7414 but couldn't find any lying around...

Image

Its fairly close....
Just to give you an example etc:


Wayne
Gilly
Class E Enthusiast
 
Posts: 167
Joined: Tue May 15, 2007 6:59 pm
Location: Mt.Eliza Victoria Australia

Re: Class E/D of 40 Meters

Postby K9ACT » Mon Apr 16, 2012 7:38 pm

Most interesting QSO the other night. I scrambled to get on the air when I heard W1VD calling CQ on 40 meters.

I happened to be on my VD Class D rig which I had converted to a sort of Class E and needless to say, this was the topic of conversation. I suspect he got a kick out of hearing one of his creations on the air.

Small talk aside, what came out of the conversation was the fact that he also has been unable to make the design work on 40. The problem seems to be in the output transformer. He has solved the problem using what he calls a "transmission line transformer" but has not had the time to document this.

So, switching to the Class E type output network seems to have been a good idea as it works well enough, even with single drivers for FET pairs. I can live with 100W on 40.

js
K9ACT
Class E Enthusiast
 
Posts: 146
Joined: Sun Jun 27, 2010 12:03 am

PreviousNext

Return to Class E Discussion

Who is online

Users browsing this forum: No registered users and 1 guest

cron