HFU HF Underground
Technical Topics => The RF Workbench => Topic started by: redhat on September 18, 2016, 2022 UTC
-
I found this ages ago when researching what others were doing in the realm of class D on HF. A google search for '75 meter am transmitter paper pdf' will get you the original from the AM Phone message board.
This has some neat practical information for those of you who want to experiment with high power levels and high level modulation. His prototype will make 1500W PEP on 75 or 160 Meters (3.8 and 1.8 MHz respectively) at over 90% efficiency. It also covers some of the pitfalls of this method.
This is not a call to action, but some practical info from someone who has done it. With the drive system and fets in use, I doubt this would survive an excursion to 43 meters, but the same principles apply.
http://amfone.net/Amforum/index.php?action=dlattach;topic=30798.0;attach=31360 (http://amfone.net/Amforum/index.php?action=dlattach;topic=30798.0;attach=31360)
+-RH
-
Great RF design with a WTF modulation scheme.
Why on earth design a decent RF generator only to modulate it with the most outdated of methods?
Yes saw this ages ago and commented on AMFONE (that's where they all hang out!)
Maybe the author has never heard of PWM?
There's people on the same site still advocating the use of HEISING modulation, FFS it makes me want to bang my head against the wall (actually I'd rather bang there's!)
Anyway people, please don't bother with mod trannies for any Voltage over 12V as there is a better way.....
Oh Yes.
;)
-
Number one; your dealing with a crowd that values simplicity. I agree, PWM is the way to go, but there is something to be said for simplicity. My back says otherwise ;D
Number two; I was reminded of this when Antenna was mentioning all the trouble his friend had with modulating that kit rig, and some of the same problems I think this guy who wrote the paper discovered as well.
In this day and age, people are still building grenade clones, not because of their 'superior design and quality'...quite the opposite. There is a lot to be said for something that can be made and repaired from commonly available parts, and that is where the desire to build these things comes from. Its just another way to approach the problem.
*edit* upon further reflection, it could also be that the preference toward high level transformer modulation relates to nostalgia. To a lot of the older guys, unless it 'feels' solid with lots of iron, like big cars, it just doesn't feel right. Heavy iron may be reminiscent of old broadcast rigs. Just a thought...
All the best,
+-RH
-
Yes 'Nostalgia' !
Dont do that as I'm an Engineer at heart (albeit an 'Old' one!)
I rekon in 50 Years time folk will pick up PWM when the Earth's natural resources (i.e. IRON) run out!
Ah Well......
;)
BTW, rebranding my PWM, with Audio BPF and 6:1 compressor, 50V @ 10A MAX, 1 inch square!
-
I saw your comments on AMFone as well, glad to see you're not biased toward your PWM approach as being the only viable solution.... :-) Your statement about "better" ways to modulate is interesting. Better in what way? There are numerous ways to amplitude modulate a signal, all have advantages and disadvantages. I have built and used PWM, PDM, Outphasing, Class A series, and Heising type modulators. I am aware you think it is 'modern and best' to use PWM. Maybe so, maybe not....but the point of the article was not to make the best transmitter ever, as was clearly stated in the beginning.
If you didn't get your competitive back up about the modulator and read the intent of the article you would readily see, this was just an experimental approach to prove a couple of points:
1. Use of off-the-shelf hi-fi gear. Simple, cheap (like most of us hams! :-) )
2. Show that a simple, cheap off-the-shelf transformer can provide broadcast quality ( 3Hz to 30 kHz at -3dB frequency response ) using ONE CHEAP COMPONENT.
I had heard someone talking on the air who had the belief that tube AM is the only way to get good quality AM, and they emphatically made the statement "you can't modulate a solid state RF deck with a transformer" regardless. That narrow minded statement was partially what made me want to try it. I did. It worked. So what.
The bottom line is A) you're probably right PWM might be better by several measures (and of course YOURS would be so much better!) B) You missed the point. Criticism like this discourages sharing experimentation with wacky non-standard approaches. Yes Strandbeest isn't a very good car.
I too have a laundry list of critiques of that prototype. I later added a diode to ensure the modulation could not swing negative (D'OH!), an LED overmodulation indicator, and discovered much less audio power is needed (150W amplifier will do) and the transformer doesn't need to be nearly as big (300VA is sufficient).
The good that came out of that work though is a way to quickly, cheaply, slap together a quality transmitter in a weekend using a minimum of readily available parts. Also the H-Bridge RF Deck was another separate experiment, while it was successful it was unnecessarily complicated, and not easily scalable in frequency, e.g. not the best approach. Glad to hear you liked that though, thanks.
I hope this answers your WTF comment.
Regards,
Mike
Great RF design with a WTF modulation scheme.
Why on earth design a decent RF generator only to modulate it with the most outdated of methods?
Yes saw this ages ago and commented on AMFONE (that's where they all hang out!)
Maybe the author has never heard of PWM?
There's people on the same site still advocating the use of HEISING modulation, FFS it makes me want to bang my head against the wall (actually I'd rather bang there's!)
Anyway people, please don't bother with mod trannies for any Voltage over 12V as there is a better way.....
Oh Yes.
;)
-
Better as in more modern, cheaper, smaller more efficient, more linear and incredible simple as PWM and driver combined in a single package with virtually lossless FETs makes it for me! That and a KW modulator with aircored inductors will weigh very little in comparison.
Str.
-
PWM is hardly a modern concept, been around since the 50s (I was there!) guaranteed not cheaper than a $30 transformer, it probably is smaller but you've not sent any specifications, it is not more efficient than a passive transformer, and probably not more efficient than the Crown XLS1000 + passive transformer, I doubt it is any more linear than the Crown but again you've not provided any specs. Virtually lossless FETs? That's funny! No such thing exists. All FETs including Silicon Carbide and GaNFETs have on-resistance.
What is the input and output voltage range of your modulator?
What range of modulation impedance can it drive?
What is the frequency response?
What is the Efficiency vs load? Is it duty cycle limited?
What is the linearity or IMD vs load?
What is the switching frequency leakage vs. current drawn?
What are the limitations of its operation? Have you done any critical analysis of your design?
If you have actually characterized these parameters please send along a link, I would like to see the data. I might be interested in buying one if A) you can tell me what it is in technical terms, and B) if I could get a sense that you were trustworthy. Being on a forum I am starting with neither :-)
Mike
Better as in more modern, cheaper, smaller more efficient, more linear and incredible simple as PWM and driver combined in a single package with virtually lossless FETs makes it for me! That and a KW modulator with aircored inductors will weigh very little in comparison.
Str.
-
I'm going to get some popcorn and watch this from a safe distance ;D
+-RH
-
Hey Redhat, sorry to dissappoint. I concede I have been trolled.
I am done here. 73
-
I'm sorry to see you go. I am glad you made this project available to all of us, and I certainly learned a lot from it when I was starting out building my own stuff.
Please don't be a stranger, and I hope you won't let this experience taint your opinion of the rest of us. We're all here to learn, some folks forget that.
+-RH
-
Oh... did I just read Outphasing !!! I thought I was the only one. Redhat please pass the popcorn.
tAC
-
A lot of those guys on 75 like to hear that initial big "Chonk!" in their monitors that comes with high power and Heising modulation. A certain ham in Maine would likely throw himself off his collection of school buses of spare parts if he had to do it another way.
-
Apologies for not answering sooner!!
Questions, questions, OK.....
PWM is hardly a modern concept, been around since the 50s (I was there!) guaranteed not cheaper than a $30 transformer, it probably is smaller.
Yes, much smaller about and inch square + the humongous filter tho'! that and you forgot the cost of the amp ($1000?)
What is the input and output voltage range of your modulator? ANY upto whatever the FETs will take so 12V-500V
What range of modulation impedance can it drive? ANY, filter designed to match.
What is the frequency response? DC to Infinity, limited by filter ONLY so usually use active filter on I/P to cut over 7KHz.
What is the Efficiency vs load? Is it duty cycle limited? 95%+ I2R as per, so with 90mR FETs (4 FETs = 22.5mR, u can do the sums..)
What is the linearity or IMD vs load? No idea...
What is the switching frequency leakage vs. current drawn? Depends on filter again, more poles for high power and less clock leak.
What are the limitations of its operation? Have you done any critical analysis of your design? None, er..No.
If you have actually characterized these parameters please send along a link, I would like to see the data. I might be interested in buying one if A) you can tell me what it is in technical terms, and B) if I could get a sense that you were trustworthy. Being on a forum I am starting with neither :-)
Looks like you're perfectly capable of building your own.
You're on AMFONE and there's a few designs on there, however I'd stick to using the UCC35702 as it's very linear, has ripple cancellation and a few other nice features, use this to drive any of the new half bridge FET drivers and some SiC FETs.
Whole design will fit in the palm of your hand, don't forget the chunky filter tho'!
Str
p.s. Please don't leave :(
-
With that circuit you need isolation because of the line operated power supply. An off the shelf transformer is a good choice. If I was going to build something like that to pirate with I would go with Bluetooth and Class A series.
-
Line isolation is not a big problem if you have transformer coupled output and drive signals. Above the 1KW (5KW PEP) or so level, this makes more sense than stringing together more switchers, both from a weight and cost standpoint. I see a lot of the high power Greek pirates doing this, and at the levels they are now pushing (12KW) the transformer required would be quite massive. It all gets rather complicated with regard to metering, in this case you would need analog photocouplers with the associated circuitry to get voltage and current samples to the outside world, isolated from the mains. Watch your signal grounds!
+-RH
-
I see no need to measure voltage in a circuit like that. It's going to be about 1.4 X the RMS line voltage. I can see wanting to know the current when first testing it. After that temperature, forward and reverse power would be all I care about. Search for w1209 on ebay. Those are cheap enough to have one for every FET. youtube w1209 to.
-
When playing with solid state, good metering is essential. PA Volts and PA Amps measurement will allow you to do indirect power calculations, and when compared to output metering provides a troubleshooting aid.
If you want to stay in the dark thats fine, but your going to have a lot of trouble trying to fix it when something isn't right.
+-RH
-
Not Class D, that's like Switch Mode Power Supplys. You don't troubleshoot SMPSs with the power on. You just replace all the bad components and power it up. If you miss any then you do it again. All you need is a Huntron Tracker.
-
Some measurements you have to do with the power on. Gate drive waveforms can tell you a lot about the health of a mosfet and the load downstream. Lousy drive makes for lousy efficiency, and eventually burnout. Drain waveforms will tell you about parasitics and whether the output network is properly tuned.
Better information makes for better troubleshooting. Case in point, I just got my prototype transmitter running last night. The thing had survived a 3 hour excursion to 1200 watts carrier on 4070 KHz in the back yard a few months back. Following that I retuned it to 43M but never tested it at power. The next day I drove to a site and setup. When I flicked the HV switch, the PA failed, as indicated by a driver overcurrent condition. I brought it home and replaced all the silicon, brought it up to 250W with modulation, and it ran for about two minutes and failed again. Over the last few months of armchair troubleshooting, I theorized that some of the passive components must have been overstressed and developed partial breakdown problems that was taking out my fets. I again replaced all the silicon, and all the passive parts in the RF power path (DC blocking caps and resonator caps). In the process I decided to change out the output balun core and go back to material 61. I quickly discovered that I could not unwind the balun because the supposed teflon jacket on the coax had melted and several of the adjacent turns had punched through, as evidenced by soot marks here and there.
I still don't know if the passive components failed, but I changed them anyway. In light of the smoking gun, it seems they were fine. At this point its not worth the trouble to find out, especially since all Cree SiC devices across the board are out of stock until October and I have a limited stock of replacement devices.
BTW in any commercial transmitter, PAV, PAI, and FWD/REV power at a minimum are measured. This monitors not necessarily the peak voltage across the output devices, but the average DC input to the PA stage following the modulator. PAV and PAI are multiplied together, then multiplied against an assumed efficiency factor to determine power output. This is called indirect power measurement. Direct power measurement is carried out with a calibrated watt meter, or using calorimetry.
+-RH
-
I know all about how the commercial guys do it. Worked for a place that made and installed UHF TV transmitters. Is it me or is management all right-wing dick-heads in those places? Anyway, I'm only talking about hobby/pirate stuff.
BTW When I first heard about Cree I thought it would be a good place to invest.
-
I'm only talking about hobby/pirate stuff.
Why should we expect any less? I guess it boils down to operator philosophy, but I would rather carry something into the field that was designed with commercial 24/7 operation in mind. There is a better chance it will work when I flip the switch.
+-RH