The AM Forum
March 28, 2024, 01:29:01 PM *
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: Homebrew 10x o-scope leads built into transmitter  (Read 6713 times)
0 Members and 1 Guest are viewing this topic.
ka1tdq
Member

Offline Offline

Posts: 1509


Red part turned in for a refund.


« on: August 03, 2017, 06:34:36 PM »

Calling Stu!  

I have 1000x probes that I built into my 75 meter class E rig.  There's a BNC on the chassis for each phase and a 10' piece of 50 ohm coax goes into each channel on the o-scope.  The BNC on each of the o-scope channels has a 50 ohm termination resistor by using a coaxial T. 

I hadn't looked at my drain waveforms in quite some time so I thought I'd have another look.  One phase was really noisy/bad an the other was fine.  I switched the coax cables around and the problem followed phase 'a'.  Nothing looks obviously wrong with the components for the 1000x probe on that side.

I connected the commercial 10x probes directly to each drain bus and both phases are fine.  

I'm thinking that RF is getting into that phase's probe somehow.  With a 1000x probe, you're dealing with such small voltages that anything could dirty it up.  I want to replace the probes with 10x components to get a stiffer current and higher sampling voltages.

Stu, you came up with the 1000x component values.  Could you throw out another schematic for 10x probes?

Jon
Logged

It’s not just values, it’s business.
AB2EZ
Member

Offline Offline

Posts: 1722


"Season's Greetings" looks okay to me...


« Reply #1 on: August 04, 2017, 01:26:55 PM »

Jon

I suggest that you first try again to fix the problem with the 1000x probe on the noisy phase.

Look for a loose ground connection on the shield side of the wiring leading from the capacitive voltage divider adjacent to the FET to the front panel BNC connector.

Does the "noisy" signal look like a noisy version of the clean signal on the other phase... or is it just noise?

Check all of the resistors associated with the voltage divider to verify that none of the them have "opened up", and that all of them are still properly connected at both ends.

Stu
 
Logged

Stewart ("Stu") Personick. Pictured: (from The New Yorker) "Season's Greetings" looks OK to me. Let's run it by the legal department
ka1tdq
Member

Offline Offline

Posts: 1509


Red part turned in for a refund.


« Reply #2 on: August 04, 2017, 06:08:15 PM »

Here's the waveforms.  I checked the resistor values and they're all good, and the same as the other side's readings.  Maybe one of the ceramic disc bypass capacitors fried at some point.  That's an easy check.  I can just replace them and see what I get.  I'll need to stop by Fry's and pick up a couple.

I also checked the values of my 50 ohm terminations at the scope and they're both good.

Jon


* IMG_1809-2.JPG (2110.16 KB, 4032x3024 - viewed 302 times.)
Logged

It’s not just values, it’s business.
AB2EZ
Member

Offline Offline

Posts: 1722


"Season's Greetings" looks okay to me...


« Reply #3 on: August 04, 2017, 07:49:16 PM »

Jon

Looking at the two waveforms, and noting that:

A) The blue scale is 100V/div and the yellow scale is 200V/div

B) The blue waveform has an average value above zero... which means there is a D.C. path for at least part of the signal

I still think that there is a problem with the direct path, via the coaxial cable shield, between the ground side of the voltage divider (where it attaches between the drain bus and ground) and the BNC connector on the front panel... or there is a problem with an open shield on one side of the coaxial cable between the front panel BNC connector and the scope. Something else is providing an indirect path for the ground side. This could be the chassis of the transmitter and/or the outside surface of the shield of the other phase's coaxial cable. The indirect path would add a lot of interference, reflections, etc.

Stu

Logged

Stewart ("Stu") Personick. Pictured: (from The New Yorker) "Season's Greetings" looks OK to me. Let's run it by the legal department
ka1tdq
Member

Offline Offline

Posts: 1509


Red part turned in for a refund.


« Reply #4 on: August 04, 2017, 09:23:28 PM »

It's getting a little better.  All the shields ohm'd out fine everywhere.  As a test though, I did put BNC barrels in each coax line isolated from the transmitter chassis (electrically, it's just one piece of coax going to the scope for each channel).  The waveform did clean up quite a bit, but there's still some noise there. 

At this point, I'm thinking that running an open chassis 400 watt transmitter that's within 5 feet of a Chinese scope with homemade leads is bound to probably have some noise.

Jon


* IMG_1813.JPG (2491.63 KB, 4032x3024 - viewed 259 times.)
Logged

It’s not just values, it’s business.
AB2EZ
Member

Offline Offline

Posts: 1722


"Season's Greetings" looks okay to me...


« Reply #5 on: August 04, 2017, 09:46:40 PM »

Jon

Yes.... that's a big improvement.

One more thing to try is to disconnect both ends of the yellow channel coaxial cable between the transmitter and the scope (removing the yellow channel input to the scope)... and see how that affects the blue channel waveform.

Stu
Logged

Stewart ("Stu") Personick. Pictured: (from The New Yorker) "Season's Greetings" looks OK to me. Let's run it by the legal department
ka1tdq
Member

Offline Offline

Posts: 1509


Red part turned in for a refund.


« Reply #6 on: August 05, 2017, 01:00:25 AM »

I did manage to get a clean display. 

It was still noisy when I completely disconnected the yellow phase.  So then I started messing around with scope settings.  I went from "dc" coupling to "ac" coupling with the coax cables completely isolated from the transmitter chassis.  That's when I got the clean picture seen below. 

Once I added one channel to the chassis, the waveforms got a little modified (but not noisy).  I then added both channels to the chassis and the waveforms got even more distorted (but not noisy).

So it'll work fine.  I just need to wrap electrical tape around the BNC barrels so that they don't come in contact with the chassis.

I did manage to get a small sizzling fireball during testing though.  The coax from phase 'b' came in contact with the tank capacitor and made for an interesting show.  The second picture shows the burn mark on the capacitor.

Jon


* IMG_1817.JPG (2000.45 KB, 4032x3024 - viewed 279 times.)

* IMG_1819.JPG (1607 KB, 4032x3024 - viewed 297 times.)
Logged

It’s not just values, it’s business.
KD6VXI
Contributing
Member
*
Offline Offline

Posts: 2648


Making AM GREAT Again!


« Reply #7 on: August 05, 2017, 12:15:34 PM »

I'm thinking you need a common mode choke on the inputs to the scope bud.

Couple ferrite cores would go a long way to cleaning up noise on the shield.

--Shane
KD6VXI
Logged
ka1tdq
Member

Offline Offline

Posts: 1509


Red part turned in for a refund.


« Reply #8 on: August 05, 2017, 02:44:05 PM »

I'm not an expert in grounding and RFI, but I think the problem might be ground loops.  It's clean now, but the problem comes back when I use the BNC barrels mounted on the front of the transmitter chassis.  It's about an 18" run for each phase from there to the drain bus. 

I'm not too concerned about it now because it's working.  I just re-routed those coaxes to the back of the transmitter in the wiring bundle there.  But, a fix if I wanted to use the original BNC barrels might be to put a small ferrite bead over the very end of the coax on the drain side of things.  Maybe keep the RF off the shield to begin with.

To do that, I'd need to do soldering re-work over the drain bus.  From experience, if one glob of solder falls into that whole cluster of "stuff"... game over.  Completely re-do the MOSFET assembly.  I'm from Maine, but even that wouldn't be funny.

https://www.youtube.com/watch?v=feg6ICTaOGo

Jon
Logged

It’s not just values, it’s business.
AB2EZ
Member

Offline Offline

Posts: 1722


"Season's Greetings" looks okay to me...


« Reply #9 on: August 05, 2017, 03:18:22 PM »

Jon

I suggest that you place the ferrite beads at the point where each cable from each power divider attaches to the associated BNC connector. There should be plenty of room there.

I also suggest that you use large ferrite beads... one, each, of the type that you use for making the RF transformers. Even better, if you have enough slack, loop each cable two times through its bead... to get 4x the inductance that will be produced by a single turn.

I also suggest that... if you re-attach the BNC connectors so as to make contact to chassis ground... that you place one of the beads (2 turns) over each of the cables running between the chassis and the scope.

Stu
Logged

Stewart ("Stu") Personick. Pictured: (from The New Yorker) "Season's Greetings" looks OK to me. Let's run it by the legal department
KA2DZT
Member

Offline Offline

Posts: 2192


« Reply #10 on: August 06, 2017, 12:00:54 AM »

Could the problem have anything to do with poor electrical contact between the two parts of the BNC connectors??  Maybe even something to do with the two halves of the BNC connectors being made from different metal or different plating.  Some are silver plated others are nickel plated.
Logged
ka1tdq
Member

Offline Offline

Posts: 1509


Red part turned in for a refund.


« Reply #11 on: August 06, 2017, 01:00:02 AM »

It's possible.  The BNC connectors that go from the chassis to the drains are from a Goodwill 2nd hand store.  I found the jumper cable there and cut it in half in order to have two lengths.  Someone else put those BNC connectors on and who knows how good/bad they are.  They do ohm out ok though.

The BNC barrels are Chinese bulk specials.  I'm using the same barrels as the chassis mounted ones for the working setup.  It all works there, so I'm falling back to some RFI/ground loops when mounted in the chassis.

I did throw on some clamp-on ferrite chokes that I had in the junk box.  It's just for added protection since it's already working.  I could connect it back up to the chassis with the ferrite chokes to see if it helps, but I actually like the coaxes coming out the back now.  Feng-shui goes up in this case.

Whenever I get around to building a 1kw class E rig, I'm going to mount the BNC's on an insulating piece of material to avoid this problem.  Especially since with a 1kw rig, there's four drain busses to look at.  I have most of the parts for the RF deck sitting in a box.  I'm thinking that it would be a nice father/son project in some years.  If Wyatt ends up taking an interest in all this stuff, I can be like, "Hey, I know something we can build..."

Maybe submit it as a 4th grade science project?

Jon


Logged

It’s not just values, it’s business.
KD6VXI
Contributing
Member
*
Offline Offline

Posts: 2648


Making AM GREAT Again!


« Reply #12 on: August 06, 2017, 12:17:56 PM »

If you don't have multiple passes through those clamp ons, your pretty much wasting time.  Impedance goes up with the square of the turns ratio.  5 to 6 turns would be Bueno.  One turn, makes us feel good.

Good luck on the science project.  I tried building a working TX with my son.  Ended up in a fight with the school board that culminated in the science fair being closed for that year.  Seriously. A vinegar volcano would have been OK.  Since they didn't understand the theory it was no good.

Those transmitters have thousands of volts (we where going pwm class e)....!!!

Rediculous.

--Shane
KD6VXI
Logged
ka1tdq
Member

Offline Offline

Posts: 1509


Red part turned in for a refund.


« Reply #13 on: August 06, 2017, 01:06:42 PM »

Science fairs are highly political.  I remember in grade 6 winning the science fair, but having to share the spotlight with a teacher's kid.  The student of the month was based on the winner.  That month, we had two students of the month.  Maybe your class E transmitter was going to disrupt the coronation of the expected winner.

And by the way, volcanos are overrated.  You get much more attention if you have a cardboard face with the tube behind the mouth.  I called it the "puke tube."  (grade 4).

Jon

Logged

It’s not just values, it’s business.
DMOD
AC0OB - A Place where Thermionic Emitters Rule!
Contributing
Member
*
Offline Offline

Posts: 1768


« Reply #14 on: August 06, 2017, 09:24:58 PM »

TO show you how bad leads can affect measurements, I had designed a new transmitter and on the Spectrum Analyzer it showed first and second harmonics 10 and 20 dB below the carrier, even though I had harmonic traps on the output.

Moving the clip-on-probe showed the amplitude of the first and second harmonics would change in amplitude.

Replaced the whole lead assembly. No more first and second harmonics.


Phil - AC0OB
Logged

Charlie Eppes: Dad would be so happy if we married a doctor.
Don Eppes: Yeah, well, Dad would be happy if I married someone with a pulse.NUMB3RS   Smiley
AB2EZ
Member

Offline Offline

Posts: 1722


"Season's Greetings" looks okay to me...


« Reply #15 on: August 09, 2017, 10:38:21 AM »

Jon

I was thinking about the issue you were/are having... and it occurred to me that there is an additional implementation detail that might be helpful.

See the attached schematic

Note that the voltage 1000:1 divider actually consists of a 31:1 capacitive voltage divider feeding into a 31:1 resistive voltage divider (the resistive voltage divider consists of a 1500 ohm series resistor feeding a 50 ohm coaxial cable with a 50 ohm termination).

I have added a vertical dashed line to the schematic, to emphasize this.

In my implementation of my 4-FET class E transmitter... I located the 1500 ohm series resistor directly at BNC connector. The BNC connector and the 1500 ohm series resistor were (in my implementation) located on a small angle bracket attached to the outside of the chassis... and a short piece of RG-174 coaxial cable, running from the capacitive voltage divider (on one side of the dashed line) to the 1500 ohm series resistor, exited through a 1/8" hole in the chassis (actually drilled through the heat sink). The hole in the chassis was located close to the drain bus... so the short length of coaxial cable was, except for about 5 cm of its total length, isolated from the magnetic and electric fields inside the transmitter.

That way, the center conductor of the short piece of coaxial cable, running between the capacitive voltage divider and the 1500 ohm series resistor +  BNC connector, had 31x more voltage to ground... compared to the alternative of locating the 1500 ohm series resistor directly at the 1000pF capacitor... and was well isolated from the fields inside the transmitter. 

This short coaxial cable was not looking into a 50 ohm resistive termination at either end... but since it was short (less than 25cm long) it didn't cause any significant problems with reflections.

Stu


* Slide2.JPG (40.68 KB, 960x720 - viewed 262 times.)
Logged

Stewart ("Stu") Personick. Pictured: (from The New Yorker) "Season's Greetings" looks OK to me. Let's run it by the legal department
ka1tdq
Member

Offline Offline

Posts: 1509


Red part turned in for a refund.


« Reply #16 on: August 09, 2017, 09:56:24 PM »

Ah, centimeters... forces me to think and do a quick conversion/guess-timation about how long that really is.  I'm still in lbs, inches, barns mode. (actually, barns may be universal; they didn't tell me).

Yeah, transmitter layout and planning for all this stuff is key.  I like building open frame stuff for the 'wow' factor rather than showing somebody a square, aluminum box.  And, my probes were an afterthought.  The transmitter was already built and the coax run to the only faceplate on the rig was 18" away from the drains (and, conveniently parallel to and just inches away from the tank components).

I'm amazed that my setup works as well as it does.  I don't have RF in my audio or any other RFI problems really to speak of after adding a few ferrite cores here and there.  I did switch to a mechanical sequencer though.  My automated one did have some RFI problems, but I did cure them with ferrite cores.  I decided to switch to a mechanical setup since other station functions don't need all events.  For example, when I'm on CW I only need event 1.  Class E 75 meters needs all 3 events, and 40 meters class E only needs 2 events.

The next rig I build will be a KW carrier version.  That's the logical next step.  Everything else is boring.  But, I will give this rig lots of thought about layout and planning for everything (including probe construction).  With 4 drain busses, it's critical to monitor the waveforms.  

Thanks for the tips on how you did it.  I'll try to keep it short and shielded next time.

Jon  

*PS>  I just Googled it.  Barns is a metric unit, so it is universal.  I can't imagine how complicated some of the formulas might be if they used English measurements.  
Logged

It’s not just values, it’s business.
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 0.121 seconds with 19 queries.