The AM Forum
December 10, 2024, 02:55:15 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   Go Down
  Print  
Author Topic: Waveforms for my single FET 40 meter CW rig  (Read 14873 times)
0 Members and 1 Guest are viewing this topic.
ka1tdq
Member

Offline Offline

Posts: 1506


Red part turned in for a refund.


« on: May 12, 2015, 09:20:24 PM »

Finally got my hands on a scope.

My 40 meter CW rig is driven by a UDB1008S Chinese DDS VFO.  The driver for the FET is an IXDD614 and the FET is a silicon carbide C2M0040120D.  Output power is 100 watts.

Picture 615 is the waveform at the gate with 5 volts/div and the VFO set for a 50% duty cycle.  This appears to be turning the FET on and off correctly.

The next two pictures are of the output of the DDS VFO at 1 volt/div.  Picture 616 is at 3.870 MHz and looks like a fairly decent square wave.  Picture 617 is at 7.290 MHz and looks almost like a triangle wave. 

I'm working on another transmitter using this same VFO (it's cheap), but I'm going to build it for 75 meters.  The waveform looks nicer at the lower frequency.

Jon
KA1TDQ


* IMG_0615.JPG (313.58 KB, 1280x960 - viewed 353 times.)

* IMG_0616.JPG (315.28 KB, 1280x960 - viewed 367 times.)

* IMG_0617.JPG (324.03 KB, 1280x960 - viewed 370 times.)
Logged

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

Posts: 2597


Bap!


WWW
« Reply #1 on: May 12, 2015, 10:33:21 PM »

Do you have any pictures of the drain waveform?

The gate waveform, if the zero volt point is the center of the scope, has the gate turned on way too long as compared to the off time.  The transmitter performance will improve quite a bit if the duty cycle is somewhat less than 50%.

But, the drain waveform usually tells the story !

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: 1506


Red part turned in for a refund.


« Reply #2 on: May 12, 2015, 11:21:02 PM »

I wasn't successful in capturing a picture of the drain or output waveform.  The scope I'm using I guess isn't up to the challenge.  I did vary the duty cycle down below and up past 50%.  The output dropped off both ways.  I get 100 watts output at that point and it drops off appreciably above and below that.

Jon
Logged

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

Offline Offline

Posts: 1506


Red part turned in for a refund.


« Reply #3 on: May 13, 2015, 09:43:22 AM »

I got a message about cleaning up the waveform by putting the VFO in sine wave mode and using a 74HC14 as a buffer.  There is space on the circuit board to put one.  I'll do that and see what the result is.

By the way, when I was capturing the picture of the gate waveform I was getting two results.  Most of the time I keyed down, I would get waves that were skinnier (had less of a duty cycle) than the one in the picture.  Sometimes when I keyed, this one would pop up.  I'm not sure why.  I don't know if it was a scope error or if sometimes I really am getting a different wave. 

Jon
KA1TDQ
Logged

It’s not just values, it’s business.
WD5JKO
Member

Offline Offline

Posts: 2002


WD5JKO


« Reply #4 on: May 13, 2015, 10:40:54 AM »


 As covered before on other threads, the probe must NOT be a 1X probe. Preferably a low capacitance X10 or X100 probe with a fast risetime (< 5 ns) and a VERY short ground lead touching the FET source lead. The probe compensation must be done with scope square wave test signal.

Jim
Wd5JKO
Logged
ka1tdq
Member

Offline Offline

Posts: 1506


Red part turned in for a refund.


« Reply #5 on: May 13, 2015, 11:53:36 AM »

It was a 1x probe...  I need to invest in a good scope/leads.

Jon
Logged

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

Offline Offline

Posts: 1711


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


« Reply #6 on: May 13, 2015, 06:32:10 PM »

Jon

You can add this circuit from drain-to-ground, and use your existing x1 scope probe. The connection of the scope will not disturb the behavior of the RF circuit.

1. Keep the leads short between the FET and this added circuit
2. Make sure the 33pF capacitor is rated for 1000V or more.
3. You can (if you wish) use 50ohm coax between the output of this circuit and the 50 ohm input of the scope (e.g. RG174U).
4. The 47kohm resistor should be a 1/2 watt carbon film resistor
5. The 1.5k ohm resistor can be an ordinary 1/2 watt or 1/4 watt carbon film resistor

This is essentially the circuit that would be inside of a 1000x scope probe. The 30pF and 1000pF capacitors form a 31:1 RF voltage divider. The 47kohm, and the 1.5kohm resistor form (when the scope is connected) a 32:1 DC voltage divider. The 1.5kohm resistor and the 50 ohm scope input form an additional 31:1 voltage divider.

Stu  
It was a 1x probe...  I need to invest in a good scope/leads.

Jon


* Scope voltage divider.jpg (26.42 KB, 960x720 - viewed 316 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: 1506


Red part turned in for a refund.


« Reply #7 on: May 13, 2015, 07:07:16 PM »

Hey Stu

Thanks again!  I'll build that onto the PC board using a BNC connector.  That way I can always test the drain waveform easily by just connecting a BNC jumper cable.  I think I'll be able to... I need to look at it when I go home and see if there's room on the board at that location.

But first, I do want to try driving the IXDD614 with the 74HC14 chip.  It should be a nice, clean square wave.  I'll change the output of the VFO to a sine wave and run that into the 74HC14 chip.  To change the duty cycle at that point, I'll have to mess with the amplitude of the sine wave and DC offset.  Hopefully I'll still be able to get the 100 watts out like I do now with the square wave.

It's nice to operate a homebrew CW rig with frequency accuracy to several decimal places, at 100 watts and not sound like a train whistle... (beeyouuuu, beeyouuuuu).    Smiley

Jon
Logged

It’s not just values, it’s business.
WD5JKO
Member

Offline Offline

Posts: 2002


WD5JKO


« Reply #8 on: May 14, 2015, 07:49:52 AM »


  Stu,

   That voltage divider is interesting. This could be used as a RF probe feeding a spectrum analyzer instead of a scope. I might need to add an AC coupling capacitor however. I have been fighting some 3 KW commercial RF generator problems where the DC voltage feeding the output RF Fet's is not adequately filtered to achieve AM modulation spec of -50dbc

   Jon,
     
     Some scopes have a selectable input impedance of 1 meg, or 50 ohms. If the scope you use does not have a 50 ohm input option, you can add a 51 ohm resistor to that 1000:1 probe.

Jim
Wd5JKO
Logged
ka1tdq
Member

Offline Offline

Posts: 1506


Red part turned in for a refund.


« Reply #9 on: May 14, 2015, 12:31:38 PM »

There is room on the circuit board to add a BNC connector, so I just ordered the components. The scope I'm using just has 1meg ohm inputs, so I'll use the 50 ohm resistor in parallel with the output.

And actually, I'll look at the drain on the scope before I re-engineer the rig right now.

Jon
Logged

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

Offline Offline

Posts: 1506


Red part turned in for a refund.


« Reply #10 on: May 15, 2015, 11:11:58 PM »

Ok, I did end up changing the drive to the IXYS chip from just the DDS VFO to now a 74HC chip in between the VFO and the IXYS.  The VFO is still set for square wave output.  Good news is, it worked...

I don't have the o-scope at home right now to verify, but I think it cleaned up the drive a little bit.  Maximum power output is now at 45% duty cycle.  Power output went up a little bit as well, depending on the frequency.  I believe the discrepancy along the spectrum is due to the funky output filter I'm using.

I'm getting 112 watts output at 7.117 MHz and the lowest power output is 98 watts between 7.001 and 7.060 MHz.  So, I'm calling it 100 watts on average.

As you can see from the picture, I've also mounted a BNC connector in there.  I'm waiting for the parts to come in from DigiKey so I can finish the drain o-scope interface circuit. 

Jon
KA1TDQ


* IMG_0630.JPG (2250.46 KB, 3264x2448 - viewed 384 times.)
Logged

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

Offline Offline

Posts: 1506


Red part turned in for a refund.


« Reply #11 on: May 16, 2015, 05:24:18 AM »

...and actually, I had to go back to my original configuration.  Luckily I have a pretty big heat sink because something was causing the FET to get really hot! 

I have two LED's on the transmitter.  One monitors the 48 volts drain and the other samples the RF output and lights up when I key-down.  The one which monitors 48 volts would actually glow brighter when the output of the VFO was on, even though I was not key-down and there was no power applied to the IXYS gate driver.  It's almost like something wasn't jiving with my switching 48 volt power supply and the RF, which was only going up to the 74HC chip. 

Anyway, I'm back to my original configuration and I'm still getting 100 to 110 watts across the CW portion of the band.  The FET runs just slightly warm; not hot at all. 

My parts should be in tomorrow for the 1000x probe circuit.  I'll finish that and get drain o-scope readings on Monday.

Jon


* IMG_0631.JPG (2276.18 KB, 3264x2448 - viewed 348 times.)
Logged

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

Posts: 1806


« Reply #12 on: May 16, 2015, 02:01:20 PM »

Hi Jon,

I heard you on 40m the other night and while conditions were not great, your audio was very good.

Were you using this transmitter?

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
ka1tdq
Member

Offline Offline

Posts: 1506


Red part turned in for a refund.


« Reply #13 on: May 16, 2015, 02:39:02 PM »

Thanks Phil! That's the ultimate compliment for AM.

No, I wasn't using this rig. This is my CW transmitter. The one I was using also uses a single FET, but is analog driven with a ricebox and Heising modulated. I think I was talking with N5AEA from Texas.  Nice to hear that I'm hitting way out in Iowa!

Jon
Logged

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

Offline Offline

Posts: 1506


Red part turned in for a refund.


« Reply #14 on: May 18, 2015, 08:52:55 PM »

Well, it's not much to look at but here it is.  This is the waveform for the drain using a 1000x probe circuit.  The Y axis is .1 volts/div.  

And after taking better readings at the gate, it appears that it is only getting a 1 volt peak square wave.  That's not nearly enough to turn it on/off.  I looked at the output of the VFO and it is miniscule.  I'll need a buffer to amplify the VFO and then put it into the IXYS.  I think the 100 watt reading was erroneous. 

Jon


* IMG_0646-2.JPG (1892.93 KB, 3264x2448 - viewed 389 times.)
Logged

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

Offline Offline

Posts: 1506


Red part turned in for a refund.


« Reply #15 on: May 19, 2015, 08:57:37 AM »

Lo and behold... I happened to check the probe this morning and it has a 1x / 10x switch and it was set to 10x.  So, the gate is getting around 10 volts.  Phew!

I did make another change and tried lowering the duty cycle to 25% and again I'm getting 100 watts output.  I'll use the 10x probe to look at the waveform after the filter.

Jon
Logged

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

Offline Offline

Posts: 1711


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


« Reply #16 on: May 19, 2015, 11:01:49 AM »

Jon

The signal to noise ratio of the drain voltage waveform measurement is too low.

Depending upon where the noise is coming from, you may get a better estimate of the drain voltage waveform as follows:

Leave the existing voltage divider in place (including the existing 50 ohm termination of the output to ground). Verify that all of the resistors are carbon resistors, or carbon film resistors, or metal film resistors... not wirewound resistors.

Use your scope probe (initially set to 10:1) to measure the drain voltage waveform between: a) the point where the two capacitors and the two resistors are all joined together, and b)ground. The voltage waveform between this point and ground should be .032 x the actual drain voltage waveform.

Make sure the scope input is set to "DC" ... so that you don't remove the DC component of the drain voltage waveform..

After confirming that the peak voltage is not too large, you can switch the probe to 1:1.

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: 1506


Red part turned in for a refund.


« Reply #17 on: May 19, 2015, 11:30:51 AM »

Ok. I can do that tonight.

If output power is actually 100 watts, can I directly put the 10x probe on the antenna jack? If my Ohm's law is correct, that should be a max of 50 volts.

Jon
Logged

It’s not just values, it’s business.
VE3ELQ
Contributing
Member
*
Offline Offline

Posts: 258


« Reply #18 on: May 19, 2015, 11:49:33 AM »


If output power is actually 100 watts, can I directly put the 10x probe on the antenna jack? If my Ohm's law is correct, that should be a max of 50 volts.

Jon

Yes OK on antenna jack. If load is 50ohms non reactive this is probably the best way to measure power out.

No
Power is E sqr / R. So E(RMS) = Sqr Rt PxR.    E peak = E rms X 1.414
100W x 50Ohms = 5000  sqr rt = 70.71RMS x 1.414 = 99.98 Volts peak RF.

Be mindful of the scope input voltage rating. With a 1:1 probe you can easily blow out the input amp.
73s Nigel
Logged

A "Tube Free" zone.
w1vtp
Member

Offline Offline

Posts: 2632



« Reply #19 on: May 19, 2015, 12:31:15 PM »


If output power is actually 100 watts, can I directly put the 10x probe on the antenna jack? If my Ohm's law is correct, that should be a max of 50 volts.

Jon

Yes OK on antenna jack. If load is 50ohms non reactive this is probably the best way to measure power out.

No
Power is E sqr / R. So E(RMS) = Sqr Rt PxR.    E peak = E rms X 1.414
100W x 50Ohms = 5000  sqr rt = 70.71RMS x 1.414 = 99.98 Volts peak RF.

Be mindful of the scope input voltage rating. With a 1:1 probe you can easily blow out the input amp.
73s Nigel

Wouldn't you have to do the  calculations for a PTP value to determine what the scope input is going to see (both sides of the RF voltage - plus and minus)  Good point about the caution regarding the 1:1 probe.  With the right kind of resistors you could make up a "calculated" sampler that takes off the RF output and feeds into the 50 ohm input of the scope.

Here's a sampler design spreadsheet I cobbled up a while back.  The idea is to use a TEE with one resistor in series with the sampled line going into a 50 ohm load (as found in your scope or an external 50 ohm feed thru scope termination).  It gives one all the values needed to design the sampler.  Good non-reactive resistor is needed to avoid stray capacitance problems etc.  The series resistor needs to be put into a small project box to avoid shock etc.  This isn't a good primary way to measure one's power butat least you would have an idea if you are having problems - flattoping etc

Al


Logged
AB2EZ
Member

Offline Offline

Posts: 1711


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


« Reply #20 on: May 19, 2015, 01:15:25 PM »

Al

The voltage divider you have designed is okay for low frequencies (e.g. audio), but not very accurate for RF. The parasitic capacitance across the larger resistor may have a lower impedance than the nominal resistance of the larger resistor.

For example:

Let's say the larger resistor has a value of 5000 ohms (for a 100:1 voltage divider).

If we want to look at a 7MHz periodic waveform, including the first 4 harmonics... then the reactance of the parasitic capacitance across the 5000 ohm resistor, at 35MHz, has to be at least -j5000 ohms. I.e. we don't want the parasitic capacitance to enhance the 4th harmonic output of the voltage divider by more than 3dB, or shift the phase of the 4th harmonic by more than 45 degrees.

This implies that the value of the parasitic capacitance across the 5000 ohm resistor has to be less than 1/[2pi x 35,000,000 Hz x 5000 ohms] = 0.91pF.

By using the design I posted (also posted below), the 33pF capacitor's capacitance will "swamp" out the capacitance of the 47kohm resistor... while the resistors will provide the required voltage divider for DC and low frequencies.

In any case, keep in mind that the larger value resistor must have a sufficient wattage rating (and also be non-inductive to avoid resonances).

Stu


* Scope voltage divider.jpg (26.42 KB, 960x720 - viewed 337 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: 1506


Red part turned in for a refund.


« Reply #21 on: May 19, 2015, 09:00:04 PM »

Things were still pretty noisy.  I'm not sure if it's the scope, leads, transmitter or what.  I can't seem to get a repeatable waveform anywhere.  I'd like to get an accurate reading of power output. 

I do know that when this rig was reading 35 watts a couple months ago and I hooked it into my linear, I would get 500 watts output.  Now that output is reading 100, I wanted to see what the linear would go to (carefully).

BUT, in trying to get a good waveform tonight, I blew up my shiny new SiC mosfet.  Drain to source is zero ohms.  I'm going to go back to the FQA11N90's for experimentation... they're cheaper.

Jon
Logged

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

Posts: 1806


« Reply #22 on: May 19, 2015, 10:46:16 PM »

Hi Jon,

If you need to delay phase B wrt to Phase A, AND set the timing pulse widths individually, below is one such implementation.


Phil - AC0OB

* Timing Circuit for AMPHONE.pdf (26.64 KB - downloaded 163 times.)
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
ka1tdq
Member

Offline Offline

Posts: 1506


Red part turned in for a refund.


« Reply #23 on: May 20, 2015, 09:39:19 AM »

Thanks for the schematic.  It looks simple enough to build.  I'd need a better o-scope to make the fine tuning. 

I'm going to throw a new FET on and call it good.  I brought it into work some time ago and threw it on the service monitor.  2nd harmonics are almost non-existent, so it's clean. 

Jon
Logged

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

Offline Offline

Posts: 1711


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


« Reply #24 on: May 20, 2015, 01:50:39 PM »

Jon

If you are getting lots of noise on the scope display when you measure the gate waveform (which you didn't have when you measured the gate waveform before), then I would check to see if there is a loose ground in the scope probe. I.e. it is essential that the ground clip on the scope probe is making good contact (via the probe's cable) with the grounded shell of the BNC connector on the scope.

Do you see lot's of noise/spurious signal on the scope display when you run the transmitter, but the scope probe is not plugged into BNC connector of the scope? [If yes, the would suggest that RF is getting into the scope via the power line or some ground path that leads to the scope.]

If not:

Do you see lot's of noise/spurious signal on the scope display when you run the transmitter, but just connect the probe's ground clip to the probe's tip? (I.e. short out the probe at the probe end, leave the probe's cable extended so that is it near the transmitter...but do not connect the ground clip or to probe tip to anything on the transmitter). [If yes, this would suggest that RF is getting into the cable of the scope probe]

If not:

Do you see lot's of noise/spurious signal on the scope when you run the transmitter, and touch the probe's tip to the same point on the transmitter's ground plane where the probe's ground clip is attached? (I.e. same as above, but both the probe's tip and the probe's ground clip are touching the ground plane of the transmitter).[If yes, this would suggest that the scope is responding to a common mode signal between the scope's BNC input signal(s) (i.e. tip and ground) and the scope's electronics. I.e. the scope's input preamplifier is not producing sufficient common mode rejection, and the RF signal between the transmitter's ground plane and the scope's ground is not being sufficiently rejected.]

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
Pages: [1] 2   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 0.067 seconds with 18 queries.