The AM Forum
March 28, 2024, 04:54:32 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] 2 ... 4   Go Down
  Print  
Author Topic: Zorch, again on my 8-FET 75 meter transmitter  (Read 78630 times)
0 Members and 1 Guest are viewing this topic.
ka1tdq
Member

Offline Offline

Posts: 1509


Red part turned in for a refund.


« on: February 10, 2018, 07:06:44 AM »

I was getting ready for the AM rally by actually turning on my 75 meter class E rig.  Everything worked fine.  I even ran the carrier back to 300 watts to be cautious since I previously blew a FET pair running it too hard.  I don't have overload protection in my transmitter.  Just meters for me to view parameters.

I also had prepared a whiz-bang antenna for the weekend.  I wanted to try a full wave loop at 75 meters since I had never done that before.  It barely fit in my backyard, but unfortunately most of the antenna was literally 6 feet off the ground.  As you'd expect, it didn't work very well.  I watched my signal on both web SDR receivers out this way, and even at 2am the signals were pretty lousy.  So, back to the original dipole.

I listen to my transmitted audio by sampling the RF output and feeding it directly to my headphones.  All sounded crystal clear on the loop antenna.  When I switched to the dipole, the audio had lots of hash in it.  Like a good ham, I kept transmitting to try to figure out what the problem was.  The only change was the antenna.  Well pop-bang-fire!  I blew the same FET pair on phase B as last time. 

I believe it's an RF problem getting into that phase.  That phase is physically located nearest the output of the transmitter and always ran slightly less clean that the other phase as seen on a scope.  You can see the two blown FETs in the picture. 

I've built the transmitter heat sink assembly such that I can rebuild it quicker than a bubba can swap out a transmission, but I'm looking for the cause of the problem. 

It could also be parasitics rising from local RF, and Steve, I know you mentioned the 300 ohm resistor trick... but what is that little resistance going to do for a device with super high input impedance? 

Jon 


* IMG_0233.jpg (2631.8 KB, 4032x3024 - viewed 1089 times.)
Logged

It’s not just values, it’s business.
KC2UDZ
Member

Offline Offline

Posts: 22


« Reply #1 on: February 10, 2018, 08:03:41 AM »

just my opinion get rid of the doorknobs, if you must we have used 5% silver micas 500v not the best but they will work. I would also try turning the output transformer 90 degrees from the tuning network. I cant see the tank coil but i would also keep that 90 degrees off the tuning network.


* E Rig Posting Pics 028.jpg (61.25 KB, 640x480 - viewed 941 times.)

* Class E Project 013.jpg (60.03 KB, 640x480 - viewed 933 times.)
Logged
ka1tdq
Member

Offline Offline

Posts: 1509


Red part turned in for a refund.


« Reply #2 on: February 10, 2018, 08:33:32 AM »

I just thought of this too:

The front chassis of the transmitter has BNC jacks to accept the output of the DDS VFO.  From there, it's twisted pair 20" long to the back of the transmitter and connects to the IXDD's.  It runs smack dab through the output tuning network.  Amazingly, it runs as clean as it does (especially phase 'A').

I just slow-boated a couple 24" BNC-to-SMA cables and I'll extend the coax run right up to the IXDD inputs.  I'm not a betting person, but...

Jon

And if I'm wrong, oh well.  I guess I'll spend another $10 in junk box parts for a couple FETs and a driver.
Logged

It’s not just values, it’s business.
KC2UDZ
Member

Offline Offline

Posts: 22


« Reply #3 on: February 10, 2018, 08:45:21 AM »

Again, Get rid of the Doorknobs thats old news.
Logged
ka1tdq
Member

Offline Offline

Posts: 1509


Red part turned in for a refund.


« Reply #4 on: February 10, 2018, 08:58:10 AM »

Right-o. I have some ATC’s on the way. That’s a quick fix.

Jon
Logged

It’s not just values, it’s business.
N1BCG
Member

Offline Offline

Posts: 836


« Reply #5 on: February 10, 2018, 09:38:48 AM »

The front chassis of the transmitter has BNC jacks to accept the output of the DDS VFO.  From there, it's twisted pair 20" long to the back of the transmitter and connects to the IXDD's.  It runs smack dab through the output tuning network.

If the twisted pair connects to the BNC jack then that's the problem. The jack is an unbalanced component, twisted pair is balanced. You have a 20" antenna in a high RF field.

Same thing as grounding one lead of an open wire line feeder.

Use coax.
Logged
ka1tdq
Member

Offline Offline

Posts: 1509


Red part turned in for a refund.


« Reply #6 on: February 10, 2018, 10:21:51 AM »

That’s what I was thinking, and I’m amazed it worked as long as it did. Basically like you said, it’s an oscillator power level signal trying to get through a 400 watt carrier field. I’ll make both changes and see what happens.

Jon
Logged

It’s not just values, it’s business.
steve_qix
Contributing
Member
*
Offline Offline

Posts: 2599


Bap!


WWW
« Reply #7 on: February 10, 2018, 01:13:38 PM »

I don't know how you're getting the signal from the frequency source to the drivers, but you _must_ put a 300 ohm resistor directly in series between the driver input pin and everything else.  Then, the "feedline" connecting from the frequency source must be terminated in its characteristic impedance.

Shielded cable is the absolute best.  The cable can be quite long.  I've run 15 feet of cable between the VFO output (12V signal) and the drivers with no issue - this is with thin coax.

The problem you had does sound like a parasitic.

The signal from the frequency source should be substantial - 12V, 0V referenced (AKA, the signal goes from 0V when off, to 12V when on.  This improves noise immunity.
Logged

High Power, Broadcast Audio and Low Cost?  Check out the class E web site at: http://www.classeradio.org
ka1tdq
Member

Offline Offline

Posts: 1509


Red part turned in for a refund.


« Reply #8 on: February 10, 2018, 01:34:35 PM »

My existing configuration is too complicated to describe without eyes glazing over, other than to say it's wrong.  It worked, but it's wrong.  I'll run 50 ohm coax directly to the drivers and use the 300 ohm resistor inline. 

I was using a sine wave but not near 12volts.  I think it was around 7.5 volts pk/pk.  I can't go any higher because the duty cycle will get too high coming out of the drivers.

Anyway, I think this will greatly improve things.  Thanks!

Jon
Logged

It’s not just values, it’s business.
steve_qix
Contributing
Member
*
Offline Offline

Posts: 2599


Bap!


WWW
« Reply #9 on: February 10, 2018, 02:42:31 PM »

Ok, I see what's going on.  Terminate the coax in 50 ohms.  Analog drive to the driver ICs is unpredictable, and I don't know how the drivers will react to any sort of negative going input signal.  This type of operation may be undefined (you'd have to look at the documentation), or it may be that the signal is dc-shifted up due to clamper action within the driver input.
Logged

High Power, Broadcast Audio and Low Cost?  Check out the class E web site at: http://www.classeradio.org
ka1tdq
Member

Offline Offline

Posts: 1509


Red part turned in for a refund.


« Reply #10 on: February 10, 2018, 02:53:37 PM »

Apparently something like that is going on because it does work. I could easily throw a buffer inline to square it up, but I don’t think it’s needed.

Jon
Logged

It’s not just values, it’s business.
steve_qix
Contributing
Member
*
Offline Offline

Posts: 2599


Bap!


WWW
« Reply #11 on: February 10, 2018, 04:14:28 PM »

Apparently something like that is going on because it does work. I could easily throw a buffer inline to square it up, but I don’t think it’s needed.

Jon

It definitely will work.  It's just that the behavior is not defined.  The series resistors and the terminating resistors at the end of the feeding coax may clear up the parasitics completely.

Let us know !
Logged

High Power, Broadcast Audio and Low Cost?  Check out the class E web site at: http://www.classeradio.org
ka1tdq
Member

Offline Offline

Posts: 1509


Red part turned in for a refund.


« Reply #12 on: February 10, 2018, 06:52:05 PM »

I've got coax coming, ATC capacitors, and 300 ohm resistors.  I'll adjust things one at a time until I get a clean wave.

And to be clear, it will run "dirty."  I monitor the audio directly and when things are fine, the audio is clear.  Only when I get hash in the audio do things go to hell.  It'll be very simple for me to determine what the fix is (after I replace 2 blown FETs and a driver).

Jon
Logged

It’s not just values, it’s business.
ka1tdq
Member

Offline Offline

Posts: 1509


Red part turned in for a refund.


« Reply #13 on: February 11, 2018, 03:35:41 AM »

I did some experimenting with my VFO and discovered that I did have it set without an offset, meaning that half the wave was a negative going pulse.  I set the position for maximum and now everything is above zero DC going out.

Also, my current transmitter configuration has resistors totaling somewhere around 500 ohms at the driver input.  I did that back in the day so as not to drag down the VFO.  That was a mistake.  

I tested again with just the VFO and an o-scope and terminated the VFO with a 50 ohm load.  For about a 43%-ish duty cycle, it's going to take between 11 and 12 volts output out of the VFO.  No problem - it'll do that.  So, after I re-terminate the drivers with 50 ohm inputs and feed it with coax, things should behave better with a 'stiffer' driving current that's more immune to RFI.

Jon
Logged

It’s not just values, it’s business.
steve_qix
Contributing
Member
*
Offline Offline

Posts: 2599


Bap!


WWW
« Reply #14 on: February 11, 2018, 07:30:55 AM »

sounds good!  Just do make sure that there is a resistor between the driver input pin and everything else.  500 ohms is also just fine on 75 meters.  It's a non-critical value.

Regards, Steve
Logged

High Power, Broadcast Audio and Low Cost?  Check out the class E web site at: http://www.classeradio.org
ka1tdq
Member

Offline Offline

Posts: 1509


Red part turned in for a refund.


« Reply #15 on: February 11, 2018, 08:08:33 AM »

Ok, thank you.

Jon
Logged

It’s not just values, it’s business.
GW0FZY
Member

Offline Offline

Posts: 28


« Reply #16 on: February 15, 2018, 03:33:10 PM »

Steve,
I'm building your 400w 80m transmitter; I note on the schematic on the class e site you don't have those 300 Ohm resistors shown between the input of the driver FET. I have been following the circuit pretty much to the letter and have not put them in,but as yet I'm not in a position to test as there's still work to be done on the RF deck.

Justin
Logged
K6IC
Contributing
Member
*
Offline Offline

Posts: 745


« Reply #17 on: February 15, 2018, 04:46:48 PM »

Hi Justin,

From what was stated by QIX Steve  in the   ...   Starting 24 FET ...   Thread,   there is an UPDATED schematic that does show these resistors:
http://classeradio.com/24_fet.htm

Here is the Link to the other discussion:
http://amfone.net/Amforum/index.php?topic=43529.0

FWIW,   GL,   Vic
Logged
GW0FZY
Member

Offline Offline

Posts: 28


« Reply #18 on: February 16, 2018, 04:46:14 AM »

Cheers Vic, that explains it. Will also put them in
Logged
ka1tdq
Member

Offline Offline

Posts: 1509


Red part turned in for a refund.


« Reply #19 on: February 16, 2018, 06:23:38 PM »

The mood finally hit me to start tearing into this transmitter.  I mounted two SMA connectors near each phase's driver input.  There was already a tapped 4-40 hole on each side of the heat sink for them.  It's a sign.  It was meant to be.  Ordained from God.  Angels descended and sang as I tapped them in.

I still need to replace the blown side of phase B and add the 50 ohm resistors to the SMA connectors.  AND, that slow boat from China is still chugging its way over here with my two coaxial jumpers.

Jon


* IMG_0274.jpg (2692.02 KB, 4032x3024 - viewed 993 times.)
Logged

It’s not just values, it’s business.
ka1tdq
Member

Offline Offline

Posts: 1509


Red part turned in for a refund.


« Reply #20 on: February 18, 2018, 12:22:20 AM »

Surgery is complete.  Testing tomorrow.

(just so I said it... stuff is laying on the output network only for the gate drive testing.  Full output testing will be done a little more prudently)

Jon


* IMG_0338.jpg (2748.89 KB, 4032x3024 - viewed 946 times.)
Logged

It’s not just values, it’s business.
ka1tdq
Member

Offline Offline

Posts: 1509


Red part turned in for a refund.


« Reply #21 on: February 18, 2018, 05:26:29 AM »

Ok, I think I may have found the problem.  

Firstly, the transmitter is working once again.  I spoke into the microphone at 300 watts carrier and the TX audio in my headphones was clear.  

When I dekey the drain relay, I'm drawing an arc.  I can hear the arc as well as see it on my drain DC meter.  The meter is located at the transmitter post-relay, and it jumps about when I hear the sizzling.  I was experimenting with increasing the duty cycle of the gate drive and found that as I increased the duty cycle, the arc got worse.  The longest duration lasted several seconds, so long that I had to jump up and run over to the drain power supply and turn it off.  

I had inserted across the relay an RC combination to mitigate this when I noticed the arc before, but it doesn't totally work.

I guess I could just put in a vacuum relay.  That's probably the simplest and best solution.

Jon


*** And I just realized what I'm doing wrong.  For this new setup, I'm not cutting off drive to the gates and the FETs are always 'on'.  There's a low resistance path to ground on the opposite side of the relay from the DC supply.  It continues to try like hell to pass current.  I'm going to install a sequential relay to cut off the 12 volts DC to the IXDD's and see if that helps.
Logged

It’s not just values, it’s business.
steve_qix
Contributing
Member
*
Offline Offline

Posts: 2599


Bap!


WWW
« Reply #22 on: February 18, 2018, 12:58:36 PM »

*DON'T* cut off the drain drive until AFTER the DC input to the RF amplifier has fallen to zero.

This is a sequencing problem, and it is absolutely necessary to solve it or there will be failures.

So, when going from transmit to receive, the modulator needs to be put into a state where it generates no output first.  The drive stays on.  Then some number of milliseconds after that (maybe 200ms or so - give the modulator plenty of time for the output to fall to 0), stop the drive to the gates of the RF amplifier.

I don't remember (or maybe I didn't ever know) what you are using for a modulator.  If you're using a PWM type of modulator, stop the pulse train.  That will reduce the modulator output to 0V real fast.

If you're using something else such as an analog modulator, you could use a couple of IRFP260Ns in parallel as an electronic switch.  The devices are put into saturation when you're operating, so the loss is negligible, and put into cutoff when you're in receive.

There are different ways to handle it, but the bottom line is - the drive STAYS ON until the DC fed to the RF amplifier is at 0V.

Regards,  Steve
Logged

High Power, Broadcast Audio and Low Cost?  Check out the class E web site at: http://www.classeradio.org
ka1tdq
Member

Offline Offline

Posts: 1509


Red part turned in for a refund.


« Reply #23 on: February 18, 2018, 01:37:41 PM »

I'm using a Heising modulator for this transmitter.  The audio amp is a commercial unit which provides 8 (or 4, I forget) ohms to the Heising circuit. 

Everything is sequenced using a mechanical sequencer.  Timing is dependent upon hand motion.  Ha, ha. It goes IXDD 12 VDC and antenna relay, drain voltage, then modulator cut-in.  It goes in reverse order from TX to RX.

Sequencing is there, but during my testing phase this time, I forgot that I kept the drive on all the time.  That gives the drain power supply ample incentive to draw an arc and keep it going.  I've now put the IXDD power supply in the sequence. 

I'll test everything again later on today when I won't propagate (if I get back from going out in time).  If not, I'll just blast over a QSO.  Just kidding, I won't.

Jon
Logged

It’s not just values, it’s business.
ka1tdq
Member

Offline Offline

Posts: 1509


Red part turned in for a refund.


« Reply #24 on: February 18, 2018, 05:21:34 PM »

The transmitter is functional again.  The arcing is gone with the gate driver power supply sequenced and the audio is clean in the headphones.  In the picture is the gate waveforms for a little over 300 watts carrier.  This was transmitting into a dummy load, and it's not tuned exactly as it should be, but it's close.

Next is to try the dipole again.  I need to do some work on my tuner before I do that test, but at least it's generating RF again (rather than smoke and fire).

Jon



* IMG_0339.jpg (2036.67 KB, 4032x3024 - viewed 865 times.)
Logged

It’s not just values, it’s business.
Pages: [1] 2 ... 4   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 0.064 seconds with 19 queries.