KF1Z 's new 75 meter class-E

Post your photos here and write-ups you may have

Postby frank carcia » Fri Jun 15, 2007 12:47 pm

Your antenna tuner is reflecting the harmonic energy so I would not worry about it unless the reflected power is bothering the rig and you would see it in the waveform. Think about it an AB1 linear has a tank Q of 10 to 15 and a class e tank is 4 to 8 so you will be letting some harmonics out.
I see a little reflected power from my 160 meter tuner but well under 1.5 to 1.
Maybe you need a faster scope to see this fuzz since it is less than 2% of your output power.
frank carcia
Senior Contributer
 
Posts: 765
Joined: Fri Jan 31, 2003 10:48 am
Location: Enfield Ct. / Niantic Ct.

Postby frank carcia » Fri Jun 15, 2007 12:50 pm

The new transformer may be causing more reflected power because you are driving the network at a higher Z which is reducing the Q. This is the price of the extra efficiency I bet.
frank carcia
Senior Contributer
 
Posts: 765
Joined: Fri Jan 31, 2003 10:48 am
Location: Enfield Ct. / Niantic Ct.

Postby kf1z » Fri Jun 15, 2007 1:16 pm

Well, ok...

It IS bothering the rig though.....

That's what's causing the imbalance between modules.... I'm assuming...

Since all is ok fine into the dummy load...

No problem until I go to the Ant.......



I wonder if the "extra" power I realized by using the 1:2 transformer is actually this harmonic stuff ??

I guess I'll have to go back to the 1:1, and look for this stuff, and find out...
I'd rather have lower efficiency, than crap....


Maybe I am seeing something on the scope....
(It's a 100 mhz scope by the way..... my 200mhz scope bit the dust last spring....)



Here's my drain waveform running at 40vdc in and 8 amps....

Didn't bother to get a pic of both modules, but they are just about identical.....

Image
User avatar
kf1z
Senior Contributer
 
Posts: 558
Joined: Sat Mar 18, 2006 7:07 am
Location: Ely, NV.

Postby frank carcia » Fri Jun 15, 2007 1:31 pm

Just add a bit more L and raise the Q. You need to do that anyway to hit the low end of the band. Try putting some beads on your coax near the rig to isolate the reflected power that could be back on the shield.
The module that is screwing up look at the secondary of the transformer. Is it the side terminated to ground or the side going to the tank? Just interested.
Make sure you have a good ground on the tuner.
I use a scope that can see a 1 gHz sine wave easily. Your waveform looks very good though
frank carcia
Senior Contributer
 
Posts: 765
Joined: Fri Jan 31, 2003 10:48 am
Location: Enfield Ct. / Niantic Ct.

Postby kf1z » Fri Jun 15, 2007 1:49 pm

Right, ok I'll wind a new coil... as you say, I was going to do that anyway...

As for the beads.... thanks for mentioning that, I was going ask if that would help.. use the same type 43 beads as for the transformers ok ?



As for the module that's screwing up, it's hard to say....

What happens is one module raises in current slightly, and the other drops..

But, I'd say since the module that looses current, makes a more radical change (looses anywhere from 1/2 amp to 1 amp.... that's up to 25% of it's total)
Then that one is probably the one that's at "fault"...
And that is the one that is terminated to ground.....
User avatar
kf1z
Senior Contributer
 
Posts: 558
Joined: Sat Mar 18, 2006 7:07 am
Location: Ely, NV.

Postby frank carcia » Fri Jun 15, 2007 2:05 pm

I figured it was the one connected to ground. I read somewhere this is a common problem with BB amplifiers and to use a balun after the transformer but never tried it on a class E rig because that would add leakage inductance.
Yes use the same cores.
I would tune up the rig into the dummy load then connect to the tuner and set the tuner up knowing the rig was set up into 50 ohms. Your tank settings should be close to each other on the dummy load and antenna tuner or the tuner isn't matching to 50 ohms. A RF ground loop due to reflected power could also be the problem.
Remember the tuner is adding a lot of Q to the final so I'm not surprised it is reflecting a bit of power.
fc
frank carcia
Senior Contributer
 
Posts: 765
Joined: Fri Jan 31, 2003 10:48 am
Location: Enfield Ct. / Niantic Ct.

Postby kf1z » Fri Jun 15, 2007 8:54 pm

Ok, so,

I tried increasing the inductance of the series L....

No, no good... couldn't get rid of the ringing on the drain (wasn't TERRIBLE) but not as good as before....

Only increased it to 6.3uH


So, for a test (of course, what else!) I went back to the old 1:1 transformer....

Back to a shade over 280 watts out ... about 88% efficiency....


BUT, so far no problem with imbalance, and no great vswr problem...
Reads basically the same as the rice box....

Just a little tiny bit more reflected power ( hard to tell...crossneedle meter...)
And one module might be reading very slightly more current than the other on the antenna vs the dummy load....


So right at the moment, it appears that the 1:2 transformer at least accentuated the harmonic output, though probably wasn't the cause..

Make sense?



So, it was a good experiment.....

Now, I wonder what happens if I play with the inductance of the transformer?

If I load on another pair of ferrites to each transformer? for a total of 8 on each,.....

I'll wait until later to do anything else.... want to test some more, and make sure of where I stand, before I make anymore changes!
User avatar
kf1z
Senior Contributer
 
Posts: 558
Joined: Sat Mar 18, 2006 7:07 am
Location: Ely, NV.

Postby KD3CN » Sat Jun 16, 2007 5:18 am

Hi Bruce and Frank,

Just wanted to complement you both on this excellant thread. Great info for new builders (and others).
Thanks!

73, Karl

p.s. Bruce, received the cap yesterday FB, thanks for the fine packing job!
KD3CN
Member
 
Posts: 41
Joined: Sat Apr 08, 2006 5:26 pm
Location: Strasburg, PA

Postby kf1z » Sat Jun 16, 2007 6:50 am

Ok Karl...
I just realized we're up to 4 pages...! :shock:

Well, you know, I'm trying to learn "what happens if..."
Otherwise I would have just let the thing ride at lower efficiency and let it be "good enough".... What's an extra 20 watts anyway..... certainly no one would ever know the difference on a receiver!

I suppose though, it's 20 watts not being dissipated on the heatsink.....?


Maybe later when the thunderstorms move in our area I'll have a chance to get on the air......
Wait, no.... I may get a chance to try adding some ferrite to the output transformer... thats what I meant.... :roll:


bruce

btw: Karl, sorry that cap couldn't find it's way to the PO earlier... :oops:
User avatar
kf1z
Senior Contributer
 
Posts: 558
Joined: Sat Mar 18, 2006 7:07 am
Location: Ely, NV.

Postby frank carcia » Sat Jun 16, 2007 6:45 pm

hi Bruce,
I spent a good part of the day building up my HPSDR breadboard. All set for the the big test when the usb2 board arrives in a couple days.

Everyone is running 6 cores with good luck so I suspect 6 is enough. I use 8 of a slightly lower permability material and it barely got luke warm at 5 KW peak out.
I would improve the bypassing at the DC input of each transformer. I used 20 .005 uf in with a couple orange drops in parallel with them parallel in my rig. Also look at the outputs of the buck regulators they could be getting effected by RF floating around.
I would also monitor drain and gate waveforms to see if there is a change. You might also check the RF on each DC lead at the transformer.
Do you have bypassing at each meter?
As I said before you will see some reflected power because the tuner is rejecting harmonic energy. The only way to avoid that would be to keep the harmonic energy from coming out of the rig with additional filtering.
A class e final likes a low Q tank which will let some harmonic energy out.
I have not played with trying to get high efficiency and low VSWR between the rig and tuner. Guys running long runs of coax to a dipole may never even notice this reflected power.
I figured raising the L would hurt your efficiency so pick your poison bigger C or lower efficiency. Stay with 1:1 and you will need additional C and lower L.
The bottom line with this stuff is circuit layout. The layout needs to be perfect for it to work perfectly. Low z is not easy.
Yes this is an interesting thread......
frank carcia
Senior Contributer
 
Posts: 765
Joined: Fri Jan 31, 2003 10:48 am
Location: Enfield Ct. / Niantic Ct.

Postby kf1z » Sat Jun 16, 2007 10:57 pm

Yep, just turned the scope and rig off acouple minutes ago from some more testing....

8 cores per transformer (this type 43) is too much...
Tried lower L inductor, tried higher....no dice....
Efficiency way down....

So, I went the other way and went to 4 cores pre transformer....

Much better than 8 per....

There is a VERY touchy spot... takes a lot of fiddling with the Series C and loading where the efficiency seems pretty good...

But, the cores got pretty warm....

So, yes, I'll put the 1:1 wiith 6 cores back in.....

Maybe tomorrow I'll put in some more C, and see what happens with even lower L.....



What is the advantage in using so many bypass caps for the DC input?
Don't imaging it's heat..... maybe though..... my big orange drops get just barely warm to the touch.....
In the picture are the poly film that Bill ke1gf, and I guess Brent w1ia are using....
I took them out 'cause they were getting "burn you finger hot".... though that was before I quenched a parasitic problem in the vfo......



There is a little something on the output of those switching regulators... but doesn't seem like much....


I wonder how much problem I may have created by having my tank inductor so close to the rest of the drivers etc..... ??
I was trying to keep the transformer leads as short as possible....but maybe too extreme?


No, I have NO bypassing at the meters (they are in the modulator, btw)
I'm using shunt resistors, and running twisted pair to the meters.....
When I get to putting the modulator in a new box, I'll be running the DC right through the meters....
Always wondered about meter bypassing.......


When I went back to my origional 1:1 transformer, I didn't see much, if any reflected power ( probably a little, 1/10th of what I was seeing with the 1:2 transformer,,)


We'll see what tomorrow brings!
User avatar
kf1z
Senior Contributer
 
Posts: 558
Joined: Sat Mar 18, 2006 7:07 am
Location: Ely, NV.

Postby frank carcia » Sun Jun 17, 2007 8:55 am

Yes 4 cores is not enough and that is why they are getting hot. You could saturate them on modulation peaks.
1:1 into the tank is a lot higher Q so that is why your reflected power is low...and your efficiency is lower too.

Bypass caps. My output Z is about 2 ohms. Take a tube final running at 3000 ohms output Z. Put a .001 at the bottom of the RF choke with an Xc on 75 of about 50 ohms. Do the math.
orange drop caps get pretty ineffective above about 500 KHz due to ESL and that is why they are warming up. I've designed a fair number of filters and find orange drop perform very well for low frequency stuff but you need a better cap in parallel with it to deal with high frequency crud.
The only poly film cap I have ever seen that works good at RF was a custom and I can't buy them. Orange drops not even close.
frank carcia
Senior Contributer
 
Posts: 765
Joined: Fri Jan 31, 2003 10:48 am
Location: Enfield Ct. / Niantic Ct.

Postby kf1z » Sun Jun 17, 2007 11:18 am

frank carcia wrote:orange drop caps get pretty ineffective above about 500 KHz due to ESL and that is why they are warming up. I've designed a fair number of filters and find orange drop perform very well for low frequency stuff but you need a better cap in parallel with it to deal with high frequency crud.
The only poly film cap I have ever seen that works good at RF was a custom and I can't buy them. Orange drops not even close.



So, what type of caps should we use?

I know there's been other posts on the subject, but........
User avatar
kf1z
Senior Contributer
 
Posts: 558
Joined: Sat Mar 18, 2006 7:07 am
Location: Ely, NV.

Postby kf1z » Sun Jun 17, 2007 9:28 pm

Well, built up the window comparator circuit today, using 2 TL3016 ICs.

Seems stable from 50 khz to about 14 mhz, then starts to act up...

I am positive that it is layout.... nice dead bug style... :D


At any rate, I can adjust the duty cycle of the vfo/driver from

54% ON / 46% OFF... to 20% ON / 80% OFF


I have it pinned down from power supply, VFO, right through to a pair of IXDD414 driver ICs...
Maybe just a couple more bypass caps to be sure.
(And I suppose a shielded cover would be in order....)

Now to run a couple cables to the drivers in the transmitter, and see what happens!
User avatar
kf1z
Senior Contributer
 
Posts: 558
Joined: Sat Mar 18, 2006 7:07 am
Location: Ely, NV.

Postby frank carcia » Mon Jun 18, 2007 7:33 am

Those comparators are very fast so layout will be critical.
frank carcia
Senior Contributer
 
Posts: 765
Joined: Fri Jan 31, 2003 10:48 am
Location: Enfield Ct. / Niantic Ct.

PreviousNext

Return to Your Photos and Articles

Who is online

Users browsing this forum: No registered users and 1 guest

cron