Forums :: Resources :: Features :: Photo Gallery :: Vintage Radio Shows :: Archives
Support This Site: Contributors :: Advertise


It is currently Feb Tue 25, 2020 9:54 am


All times are UTC [ DST ]





Post New Topic Post Reply  [ 29 posts ]  Go to page 1, 2  Next
Author Message
 Post subject: AM trans, audio distortion related to antenna matching?
PostPosted: Jan Thu 16, 2020 11:23 pm 
Member
User avatar

Joined: Jan Fri 02, 2009 11:32 pm
Posts: 641
Location: Equinunk PA 18417
Does it makes sense that I see & hear small, varying amounts of harmonic distortion from my section 15 AM transmitter as I make adjustments to the matching network? Why should that be so?

Let me explain.I am playing with an AM transmitter design I posted last year. After a full year away, my workbench area is finally intact again (better than before) and I have picked up from where I left off. Revised schematic posted below, not yet final, biggest change is doing away with feedback of demodulated signal, at least for now. It still sounds very good. This is a class E modulator. TTL square waves at the desired carrier frequency cause a mosfet to switch on and off. A DC current that has been modulated by AC (or, an audio signal with a constant DC bias, if you want to think of it that way) is switched on and off by the mosfet into a tuned inductor “tank” and antenna matching network. The tuned inductor resonates to reshape square waves into pseudo-sinewaves and, as part of a matching network, simultaneously helps match the low impedance transmitter to high antenna impedance. Some audio spectrum is reduced particularly higher audio frequencies. Perhaps inductor Q plays a role in this last observation, I think so, definitely can see it, there is more going on than I actually understand. I don’t know if distortion is caused somehow within the tank inductor, or antenna, or by loading of the mosfet, or what. I’ve been reading my arrl book, searched web but not finding it, maybe need better search terms.

This circuit’s radiated signal with a 10 foot straight wire is really quite good. Audio spectrum frequency linearity (of the demodulated radiated signal) is not perfect. High frequencies tend to be reduced (by how much appears related to tank inductor Q) but also there is added low-level audio distortion (usually single digit percentile) that baffles me. The problem is not before the mosfet. There is only VERY low audio distortion ahead of the mosfet (see graph 1). Very slight value changes of the matching network’s variable capacitor will cause measurable (and audible) changes of harmonic distortion while causing only slight changes of radiated power. I’ve also noticed that when tuning for peak radiated output, the peak may not necessarily result in lowest harmonic distortion. The optimum points are close but different. 2nd harmonic distortion increases in one direction, 3rd increases in the other direction. Graph 2 demonstrates this

Different inductors have different distortion characteristics. Graph 3 shows difference of harmonic distortions when the L2 tank inductor is changed from a 385uH mini inductor (typical low cost leaded part from TDK or Bourns) to my 395uH 7cm solenoid hand wound air core. Very nearly the same inductance, very different characteristics. The “touchiness” to find that exact narrow point of lowest distortion is not convenient and I doubt will be stable over time.

Antenna measurements are taken from the schematic’s “station monitor” output, which is demodulated rf tapped from the antenna connection. I can hear the same distortion effects on a nearby radio so I know distortion is not an artifact of my measurement methods.
Attachment:
File comment: Graph 1, very low distortion ahead of mosfet
G1 AM audio dist R9 sml.png
G1 AM audio dist R9 sml.png [ 58.5 KiB | Viewed 1148 times ]

Attachment:
File comment: Graph 2, distortion change from peak to off-peak. Solid line=2nd harmonic, dotted line=3rd harmonic
G2 tuning dist change from peak sml.jpg
G2 tuning dist change from peak sml.jpg [ 114.28 KiB | Viewed 1148 times ]

Attachment:
File comment: Graph 3, audio high freq reduction through inductors.
G3 AM audio freq compare 384mini and solenoid sml.jpg
G3 AM audio freq compare 384mini and solenoid sml.jpg [ 70.79 KiB | Viewed 1148 times ]

Attachment:
File comment: Graph 4, lowest distortion with air solenoid. Bumps caused by noise induced in antenna wire
AM audio dist solenoid 98-50 sml.jpg
AM audio dist solenoid 98-50 sml.jpg [ 110.54 KiB | Viewed 1145 times ]

Attachment:
File comment: schematic, if not readable use download link below
richfair AMtrans schematic rev0120 TEMP.JPG
richfair AMtrans schematic rev0120 TEMP.JPG [ 524.37 KiB | Viewed 1148 times ]

Attachment:
File comment: schematic working, NOT FINAL
richfair AMtrans schematic rev0120.pdf [818.51 KiB]
Downloaded 16 times

_________________
-Richard


Top
 Profile  
 
 Post subject: Re: AM trans, audio distortion related to antenna matching?
PostPosted: Jan Fri 17, 2020 1:56 am 
Member

Joined: May Thu 04, 2017 1:08 pm
Posts: 361
Location: Raleigh, NC area
Well this is interesting. I can't really comment on your issue... But I just recently purchased parts to build the design you and black85vette worked on in the "variation on the 35Z5 ECS oscillator transmitter." I didn't realize it was two generations back! I guess it will have to do; postage on parts is making a great sucking sound on my wallet!


Top
 Profile  
 
 Post subject: Re: AM trans, audio distortion related to antenna matching?
PostPosted: Jan Fri 17, 2020 4:51 am 
Member
User avatar

Joined: Jan Fri 02, 2009 11:32 pm
Posts: 641
Location: Equinunk PA 18417
Quote:
I didn't realize it was two generations back! I guess it will have to do; postage on parts is making a great sucking sound on my wallet!
That version works really well and with a low parts count. You'll be amazed! I do hope you'll keep us informed as it comes together.

link to that thread: https://antiqueradios.com/forums/viewtopic.php?f=12&t=335976

_________________
-Richard


Top
 Profile  
 
 Post subject: Re: AM trans, audio distortion related to antenna matching?
PostPosted: Jan Fri 17, 2020 5:24 am 
Member

Joined: Jan Thu 01, 1970 1:00 am
Posts: 2367
Location: Saskatoon
Yes, it is very likely that antenna matching (tuning actually) is affecting audio distortion. The electrically small antenna that is normal for a Part 15 transmitter will have an extremely narrow bandwidth. This means that you won't have flat response over the full channel width (±5 kHz) that you are transmitting on. The ends of the sidebands will be attenuated, and you will lose high frequencies. In addition, if the attenuation is unequal between sidebands, as would be the case if your antenna isn't tuned exactly to the carrier frequency*, then the unequal sidebands will also result in distortion when detected in a normal envelope detector. A receiver that uses a synchronous detector would likely avoid this type of distortion, but the kind of antique receivers that we would be dealing with wouldn't have that type of detector.

* In fact, the radiation of an electrically small antenna will increase with increasing frequency. So, if your antenna is tuned to the center of the channel to radiate maximum carrier, it's very likely that the antenna is radiating more upper sideband than lower sideband. Therefore, to minimize distortion you may have to tune the antenna slightly off peak (on the lower side).

I encountered this problem with my Part 15 AM stereo transmitter. To broadcast AM stereo, the proper balance between sidebands is critical to get proper a proper stereo signal. I found that I had to tune my antenna slightly off peak to get the best quality audio.


Top
 Profile  
 
 Post subject: Re: AM trans, audio distortion related to antenna matching?
PostPosted: Jan Fri 17, 2020 6:35 pm 
Member

Joined: May Thu 04, 2017 1:08 pm
Posts: 361
Location: Raleigh, NC area
BobWeaver wrote:
* In fact, the radiation of an electrically small antenna will increase with increasing frequency. So, if your antenna is tuned to the center of the channel to radiate maximum carrier, it's very likely that the antenna is radiating more upper sideband than lower sideband. Therefore, to minimize distortion you may have to tune the antenna slightly off peak (on the lower side).

I encountered this problem with my Part 15 AM stereo transmitter. To broadcast AM stereo, the proper balance between sidebands is critical to get proper a proper stereo signal. I found that I had to tune my antenna slightly off peak to get the best quality audio.


That's cool... So one should tune for the best sound, not the longest distance.


Top
 Profile  
 
 Post subject: Re: AM trans, audio distortion related to antenna matching?
PostPosted: Jan Fri 17, 2020 7:59 pm 
Member

Joined: Jan Thu 01, 1970 1:00 am
Posts: 2367
Location: Saskatoon
For my mono transmitter, I've always tuned the antenna for peak power. I don't think I ever noticed any audible distortion when I did this, but I probably wasn't paying particular attention to it either.

If it sounds good, then don't worry about what your test equipment is telling you.


Top
 Profile  
 
 Post subject: Re: AM trans, audio distortion related to antenna matching?
PostPosted: Jan Fri 17, 2020 8:55 pm 
Member
User avatar

Joined: Jan Thu 01, 1970 1:00 am
Posts: 4259
I experienced similar with my Wenzel transmitter. I was frustrated that I had to accept less radiated signal to get near FM quality sound. IIRC, by reonfiguring the loading coil as an RF transformer (or autotransformer, I can't remember) the sound remained quality with almost the strongest radiation.


Top
 Profile  
 
 Post subject: Re: AM trans, audio distortion related to antenna matching?
PostPosted: Jan Sat 18, 2020 4:39 am 
Member
User avatar

Joined: Jan Fri 02, 2009 11:32 pm
Posts: 641
Location: Equinunk PA 18417
Well its good to know this isn't confined to only my transmitter!

Thanks very much for comments so far. Bob, your explanation about increased radiation resistance of sidebands, especially the lower sideband, makes a lot of sense but when I try to reason through to my tuning observations, I cannot connect. I’ve noticed while tuning the antenna with a vari cap, when I add capacitance to tune off-peak the 3rd harmonics increase while 2nd harmonics decrease. When tuned off-peak by reducing capacitance, the opposite occurs. I don’t understand how this happens. Here’s my logic. Let’s assume I am demodulating only with envelope detectors and therefore will only experience what happens in one half of the modulated rf. 2nd harmonic distortion is double the frequency of the fundamental, which suggests non-linearity happening on both positive and negative current swings and should therefore be present in both upper and lower sidebands equally. 3rd harmonic is triple the frequency, higher than 2nd, still present equally in both upper and lower sidebands but further away from the carrier center. Am I right so far? Now, if I add a tiny bit of capacitance to tune the antenna off-peak, the antenna is being tuned to a lower resonant frequency which should favor whatever is in the lower sideband. If I am detecting only the positive rf swings, I should detect higher 2nd and possibly higher 3rd harmonics. If I tune the cap in the opposite direction, to a higher frequency, I should again see both 2nd and 3rd harmonic increase even if the 3rd harmonic increase is a bit more (due to the antenna being more efficient at radiating higher frequencies). In other words, tuning in either direction should increase both harmonics. This is not what I observe. Perhaps my brain is too simplistic?

Another observation that is now puzzling is that when using my 395uH 7cm air core solenoid, tuning for lowest audible distortion finds the same point as peak carrier output. It is when I use a mini inductor (384uH or 220uF with different C value of course) that the peaks do not match. The solenoid reduces high frequencies more but nearly doubles the antenna rf radiation compared to the mini inductors, but why the difference in tuning peaks? Perhaps I should just accept all this as magic and move on to other problems.

I read a post not too long ago where the author commented section 15 transmitters, in not so many words, "I wouldn't waste my time on one that cannot modulate 100% with less than 1% distortion." I don’t perceive distortion that small when listening to real programming but I do hear it easily with sinetones. As a techy guy I would like to see if I can reduce it even if doing so doesn’t make my listening experience any better. I don’t have any am receiver that could possibly be considered high fidelity and yet, I can hear the sinetone distortion through some of them while tuning, the same as what I hear through my transmitter’s station monitor output.

_________________
-Richard


Top
 Profile  
 
 Post subject: Re: AM trans, audio distortion related to antenna matching?
PostPosted: Jan Sat 18, 2020 11:34 am 
Member

Joined: Jan Thu 01, 1970 1:00 am
Posts: 2367
Location: Saskatoon
richfair wrote:
Let’s assume I am demodulating only with envelope detectors and therefore will only experience what happens in one half of the modulated rf.

It is a common fallacy that one sideband corresponds to the positive half cycles of the carrier, and that the other sideband corresponds to the the negative half cycle of the carrier. This is absolutely not true. The positive and negative halves of the modulated carrier will be identical (except in sign) regardless of differences between sidebands. Unfortunately, the relationship between sidebands and the shape of the modulation envelope is far more complicated.


Top
 Profile  
 
 Post subject: Re: AM trans, audio distortion related to antenna matching?
PostPosted: Jan Sat 18, 2020 5:04 pm 
Member
User avatar

Joined: Jan Fri 02, 2009 11:32 pm
Posts: 641
Location: Equinunk PA 18417
You're right, the phrase you quoted does not demonstrate that I do understand that upper sidebands are not upper audio frequencies. I think it is fair to say that higher audio frequencies are further away (in either direction) from the carrier frequency. I can grasp and math proves that upper and lower sidebands are present equally in both pos and neg halves of the modulated carrier. * A simple diode placed in either direction can demodulate equally well. This is central to my previous post's questions and I highly appreciate your effort to explain it to me, it helps a lot, but I still am no closer to understanding what is observed.

A related circumstance is when one tunes their am receiver slightly off-frequency of an am broadcast. I have no experience with synchronized detectors but with every radio I have ever used, off-tuning in either direction emphasizes the upper frequency audio content of the broadcast (as well as modern digital data buzzing also included there). The sidebands are more audible regardless of the tuning direction. Bringing the discussion back to my observations, I know that all harmonic distortion (in fact all audio content) is contained only in the sidebands, both upper and lower. Logic therefore dictates that both 2nd and 3rd harmonic distortions would be affected approximately equally when tuning the transmitter antenna off-peak in either direction. That is not what is happening, so I am left to look for a different cause.

Let me try this theory. My transmitter's mosfet is operating as a switch, not operating as a linear device. It is either fully on or fully off. It is generating a huge amount of harmonics but none related to the audio program, instead related to the switching carrier frequency. Perhaps those harmonics add and subtract in magical ways (meaning: in ways I don't understand) during the phase and magnitude transformations being made to them within the tank+antenna. These transformations happen in real time. So, can I conclude that when tuning the antenna I am also changing the nature of these magical transformations and thus of sideband contents?

*One explanation I read uses "phasors" to help describe the math behind phase relationships of upper and lower sidebands (relative to the carrier), and how those relationships add and subtract to define the instantaneous magnitude of a modulated carrier. An unmodulated carrier has no sidebands, phasors have zero values, and the result is a "flat" carrier.

_________________
-Richard


Top
 Profile  
 
 Post subject: Re: AM trans, audio distortion related to antenna matching?
PostPosted: Jan Sat 18, 2020 6:37 pm 
Member
User avatar

Joined: Jan Thu 01, 1970 1:00 am
Posts: 4259
There may be another place to look. Sometime back, Bob posted the results of his study of LC ratios for part 15 loading coils, which concluded that the most L and the least C produced the strongest radiation from a ten foot antenna.

IIRC, too much C has increasing "attenuating" effect the lower the frequency.

We're pushing the limits of what my brain can understand here, but might it be with so much C (220pF), that your lower frequency sideband is practically not equal to the upper one? Or something along these lines?

If that were true, then both operating at a higher frequency and using all L and no C might improve audio quality as well as radiation.


Top
 Profile  
 
 Post subject: Re: AM trans, audio distortion related to antenna matching?
PostPosted: Jan Sat 18, 2020 7:58 pm 
Member

Joined: Jan Thu 01, 1970 1:00 am
Posts: 2367
Location: Saskatoon
That's an interesting thought. What I found was that the RF current was split between the variable capacitor and the antenna itself, because they are essentially in parallel (at least they were for the antenna configuration that I was using). It should be straightforward to put together a spreadsheet to see how the L/C ratio in the antenna tuner affects the balance between sidebands.

In reply to Richard's last post, I've been trying to think of some simple way to relate sideband imbalance to modulation envelope shape. The closest thing that I can think of is vestigial sideband modulation. This is where part of one sideband is deliberately eliminated during broadcast. I tried looking up the term "vestigial sideband" and get a bunch of hits, but every one of them is extremely math intensive. However, one of the sites I found, mentioned that demodulation of vestigial sideband requires techniques similar to DSBSC (double sideband suppresses carrier). In other words, you have to use synchronous detection. Now, if you look at the envelope of a DSBSC signal, you'll see that the positive and negative sides of the envelope cross over each other, which will definitely introduce distortion. However, this distortion should be mostly 2nd harmonic because it amounts to a doubling of the modulation frequency. So, if the unbalanced sidebands are creating an effect similar to DSBSC, then that explains the 2nd harmonic distortion. As for the 3rd harmonic distortion, I'll have to think a bit more on that.


Top
 Profile  
 
 Post subject: Re: AM trans, audio distortion related to antenna matching?
PostPosted: Jan Sun 19, 2020 3:22 am 
Member
User avatar

Joined: Jun Sun 22, 2008 2:03 am
Posts: 1451
Location: Burke, VA 22015
Class E modulator is not easy to tune. It is designed for a specific load resistance. When the load resistance changes, the modulator needs to be re-tuned.

Once the feed inductor L1 and drain-source capacitor C14 are set, the series capacitor C13 and inductor L1 are chosen such that the drain-source voltage goes to zero just before the switch closes to minimize switch losses. The Sokal's equations can be used to compute the values for these components. L1 is chosen such that its reactance at the carrier frequency is from 5 to 10 times the load resistance - that's the loaded Q factor. For C13, my rule of thumb is to choose it to have the same reactance as L1 and slowly increase one or the other until the desired waveform is obtained.

The example below shows a design for 1MHz carrier and 50-Ohm load.
Attachment:
E_Mod_2.jpg
E_Mod_2.jpg [ 54.42 KiB | Viewed 920 times ]

Steady-state voltage waveforms across switch (green) and load (blue). Note that the voltage across the switch looks almost like a half rectified sine wave.
Attachment:
E_WF2.jpg
E_WF2.jpg [ 236.06 KiB | Viewed 920 times ]

FFT of voltage across the load. 2nd harmonic is almost 30dB below fundamental.
Attachment:
E_FR2.jpg
E_FR2.jpg [ 174.71 KiB | Viewed 920 times ]

When the load is increase to 500 Ohm, the harmonics increase and thus more waveform distortion. 2nd harmonic is now only 15dB below fundamental. The voltage across the switch is no longer zero before the switch closes - more switch loses.
Attachment:
E_WF2b.jpg
E_WF2b.jpg [ 204.58 KiB | Viewed 920 times ]

Attachment:
E_FR2b.jpg
E_FR2b.jpg [ 194.94 KiB | Viewed 920 times ]

When a vertical antenna is attached to output instead of a dummy load, I can't decide the best way to do it without detuning the modulator but one is shown below. An LC matching network is used.
Attachment:
E_Mod_6b.jpg
E_Mod_6b.jpg [ 78.46 KiB | Viewed 920 times ]

Waveforms (switch voltage (green) and current (blue) and antenna voltage) and FFT for final circuit. 3rd harmonic at antenna is over 40dB below fundamental.
Attachment:
E_WF6a.jpg
E_WF6a.jpg [ 255.21 KiB | Viewed 920 times ]

Attachment:
E_FR6b.jpg
E_FR6b.jpg [ 172.99 KiB | Viewed 920 times ]

_________________
Binh


Last edited by bb.odin on Jan Sun 19, 2020 4:04 am, edited 5 times in total.

Top
 Profile  
 
 Post subject: Re: AM trans, audio distortion related to antenna matching?
PostPosted: Jan Sun 19, 2020 3:43 am 
Member
User avatar

Joined: Jun Sun 22, 2008 2:03 am
Posts: 1451
Location: Burke, VA 22015
Example that shows L2 is slightly below the optimal value for efficiency but the output second harmonic is 5 dB lower (35 dB below fundamental).
Attachment:
E_FR2c.jpg
E_FR2c.jpg [ 400.43 KiB | Viewed 915 times ]

_________________
Binh


Top
 Profile  
 
 Post subject: Re: AM trans, audio distortion related to antenna matching?
PostPosted: Jan Sun 19, 2020 2:21 pm 
Member

Joined: May Thu 04, 2017 1:08 pm
Posts: 361
Location: Raleigh, NC area
Hi Binh, nice to see the LTSpice Master drop in!

For those of us struggling to understand the posts from you, Rich, and Bob, is the green trace in the latest post showing switch current? I can see how the left-right assymmetry contributed to the 2nd harmonic distortion component. Should this waveform ideally be a rectified sine?

I dislike how LTSpice picks that dark blue as the second color in list of waveform color defaults. I can't see it against the black without turning up the screen brightness all the way.


Top
 Profile  
 
 Post subject: Re: AM trans, audio distortion related to antenna matching?
PostPosted: Jan Sun 19, 2020 3:50 pm 
Member
User avatar

Joined: Jan Fri 02, 2009 11:32 pm
Posts: 641
Location: Equinunk PA 18417
Binh, thank you for this. I have much to digest. As I drifted off to sleep last night I began to think I need to go back to the crazy waveforms present at the mosfet-inductor junction, that are strongly affected by minute tweaks of antenna tuning. Your posts demonstrate (more scientifically) what I was beginning to wonder about (unscientifically). No doubt, this is a large of my problem.

_________________
-Richard


Top
 Profile  
 
 Post subject: Re: AM trans, audio distortion related to antenna matching?
PostPosted: Jan Sun 19, 2020 9:08 pm 
Member
User avatar

Joined: Jun Sun 22, 2008 2:03 am
Posts: 1451
Location: Burke, VA 22015
Hi Dave,
The trace in the latest post shows the voltage across the switch. When L2 is slightly decreased, the frequency of the damped resonance in the output loop is higher than the switching frequency. You can see that half of the resonance cycle completes and the second half already starts before the switch closes. That explains the ugly waveform. If the damped resonance frequency is the same as the switching frequency, the voltage across the switch will look like a half rectified sine wave.

Hi Richard, I hope these posts help.

_________________
Binh


Top
 Profile  
 
 Post subject: Re: AM trans, audio distortion related to antenna matching?
PostPosted: Jan Mon 20, 2020 12:09 am 
Member

Joined: May Thu 04, 2017 1:08 pm
Posts: 361
Location: Raleigh, NC area
Oh! I get it. Cool, thanks!


Top
 Profile  
 
 Post subject: Re: AM trans, audio distortion related to antenna matching?
PostPosted: Jan Mon 20, 2020 12:50 am 
Member
User avatar

Joined: Jun Sun 22, 2008 2:03 am
Posts: 1451
Location: Burke, VA 22015
DaveInNC wrote:
Oh! I get it. Cool, thanks!

You're welcome, Dave.

Hi Richard,

I'd like to point out that the value of C13, the capacitor that resonates with L2 is way too high for two reasons (see the value used in my schematic diagram).

1) The choosen value of C13 doesn't resonate with L2 at the switching frequency.

2) One may think that C13 is used to block DC and forms a high-pass filter with L2. I'd like to think that the capacitor is used as an envelope tracker. The voltage across the switch is one sided. The average envelope in C13 is used to subtract from the switch voltage to create a two-sided voltage. To minimize distortion, C13 must faithfully track the audio envelope. I fell into this pitfall before. Check its value to ensure the high-pass cutoff is about twice the highest audio frequency (e.g. ~40 kHz). This condition is automatically satisfied when the first condition is satisfied.

_________________
Binh


Top
 Profile  
 
 Post subject: Re: AM trans, audio distortion related to antenna matching?
PostPosted: Jan Mon 20, 2020 3:07 am 
Member
User avatar

Joined: Jan Fri 02, 2009 11:32 pm
Posts: 641
Location: Equinunk PA 18417
There is much here for me Binh, you have shone a light from the darkness once again. I never thought at all about C13 needing to resonate, only to block dc. As far as matching the antenna, I may wish to keep variability since my antenna (and others who may benefit) will not be permanent and some adjustment is desired.

Thank you once again, I'll post back here after ingesting this.

BTW, those waveforms are close to what I've been accustomed to seeing except rarely as nice as images WF2 or WF6. The 720p cap across the mosfet was chosen by trial and error to give the best waveforms across the am band, using the (too large) C13 and other components in my schematic.

_________________
-Richard


Top
 Profile  
 
Post New Topic Post Reply  [ 29 posts ]  Go to page 1, 2  Next

All times are UTC [ DST ]


Who is online

Users browsing this forum: No registered users and 4 guests



Search for:
Jump to:  






























-->


Privacy Policy :: Powered by phpBB