HFU HF Underground
Technical Topics => The RF Workbench => Topic started by: digitalmod on November 27, 2017, 1348 UTC
-
I am a retire engineer and yes very frustrated with the CB-Ham old man crap so now what's next? Most AM stations are exclusively using FET class E transmitters these days. I find very few designs for such as a Pirate high quality transmitter. In fact, so far I have located only FETS and DDS chips which after all, without a lab and shop equipped system including a scope, rather make making my own rubbish at this time. But perhaps somebody has designed a pirate class E AM transmitter. They are very high quality with wonderful baseband audio potential. Just, look as I have, they are no where, as scare as hen's teeth.
Jay
-
I've had a 14 watt class E transmitter for nearly 20 years, and it was used when I got it. The problem is the guys who build them come and go. It's a labor of love for them, it's not like they're making a pile of money from it.
There are people who will make them to your specs, including one who has his 10 watt rigs for sale on the site, but be prepared to lay out a few bucks for the high powered ones.
I'd ask our guy on this board if he's up to building one with that kind of power. He backs up what he sells and will handhold you through getting it on the air.
-
Digitalmod:
If you're just looking to get on the air with a Class E AM shortwave rig, the easiest and cheapest route would be to buy one of Stretchy's Lulu transmitters. He's running a Xmas sale on them at the moment for HFU members:
https://www.hfunderground.com/board/index.php/topic,38921.0.html (https://www.hfunderground.com/board/index.php/topic,38921.0.html)
However, if you want to have some fun and learn something in the process, you should try building your own "Lulu". Later you can move onto higher powered rigs if that's your fancy.
For starters, join the homebrewpirateradio Yahoo Group, where you will find schematics and photos of all sorts of homebrewed transmitters, plus a willing support group to answer your questions and help you build your rig.
Here are a couple of links with more info on the Lulu for you:
http://www.qsl.net/lu8eha/lab_ref2.html (http://www.qsl.net/lu8eha/lab_ref2.html)
https://hfunderground.com/blog/?p=72 (https://hfunderground.com/blog/?p=72)
There is a solution out there if you want to take the plunge with a higher powered Class E transmitter, pulse-width modulation and DDS/VFO in kit form, however you really should have an oscilloscope to simplify the Class E tuning procedure and some prior transmitter building experience:
http://www.shortwaveradio.co.uk/projects.htm (http://www.shortwaveradio.co.uk/projects.htm)
Welcome to the dark side!
-
Cheers Chaps!
I'll wait for a PM!
Have what you want already!
4 X GaN or SiC, PCB's.
Youll have a bit of work to do but much less than building it from Scratch.
Youll need a O'scope, NO excuse, cheap enuf new these days (Siglent et al)
A very large (KW) dummy load and a high power PSU 50V min 10A, more volts is better mind around 150V max.
Str.
-
Agreed, without a scope, don't bother with class E. It is necessary to observe the waveforms for tuning.
I'm working on something similar to W1VD's CMCD stuff (basically push-pull class E). Right out of the box, I haven't seen anything documentation wise that fits my needs. The AM Ham guys seem to be reluctant to put much out there. It's 2017, why are they building RF decks to get 375 watts and need 8 fets...ludacris. With the 2 1200V cree SiC's my testing shows good results to at least 600W carrier. With better cooling, 1 KW seems obtainable.
Once I get something stable, I'll post the schematics and board layouts. It's going to be a while though. A companion stereo exciter will be included.
+-RH
-
Of course...and MW.
One of the problems I noted when building these things is getting proper drive to the PA stage. The PA I'm working on has the necessary circuitry built in, all you need to do is provide TTL drive from an exciter The stereo exciter will have a means to do this from a quadrature DDS source which can be disciplined to whatever you can afford (cesium fountain clock anyone, lol). C-QUAM will be included, Kahn may be as well if I can build precise enough phase shift networks.
Not for the faint of heart, everything will be surface mount with the exception of the power components.
+-RH
-
Still waiting for PM.....
;)
-
Yes, waiting on PM. Indeed. I have the little one for now and I like POWER, but.. that's asking for shark attacks if you are on FCC radar screen and some idiot files a jolly complaint. Such people I tire of greatly. FCC acts only on complaints!
I was think in terms of a push pull class E affair and computed a nice in range out put impedance easily matched with the very cheap (yes 125 dollars US) Hi Fi high power amps. Plenty of ways to bridge stereo output for more Juice if needed or maybe a less than perfect impedance match.
Driving the FETS in PP mode clocks 180 degree pulses from f out x 2 driver.
I have never tried to build such a beast, but it would be fun regardless.
:'(
-
Do these class d/e 10-20 watt txmitters play nice with an ab1 type linear amplifier?
rgg
-
Why bother with linears when you can have a TX that is a lot more efficient and will have lower distortion?
Did I mention lighter too?
Yes, they should play nice provided your amp provides a flat and non-reactive load.
+-RH
-
To Gaga: Yes, but here it the issue, ten watts not a lot of juice. So say you had a big Grand daddy Amp.. it might not drive it without yet another intermediate.
Ten watts I think isn't going to raise much smoke, but if you be in USA know FCC sharks signals that are LOUD and come on air at predictable times. Its a shame isn't it and such a waste of tax money when we are in such debt we shall never be free.
But, Let Free Radio reign I say. 250 watts would be great, I'd love such a rig, but its a fairly big gun and makes lots of noise. Yes, enough for 20 db sn about everywhere on east coast.
But as a music producer and former engineer, I dig the idea of lets play IT 8)
-
That's true. My old amp needed about 100 watts PEP to get full power.
+-RH
-
:D
Streachyman: Sounds like you are onto the "goods". Do you have a working prototype that I could see on Utube.
I could supply a power source and yes, even a modulator if needed.
You mention a scope, and that is in the works for this new year.
Thanks
-
Yes, have a 4 X GaN design finally up and running at 150-250W carrier @ 48V. I'll use 2 X 48V supplies to give 1 KW PEP.
I've ordered 2 X 10A 48V PSU's so when they come I'll wire it all up and do a vid.
Modualtor wise it's a PWM system I've been using for a while but I'll use some SiC FETs on the O/P.
I've had the RF side on a PCB since March 17 but have had 'issues' but have managed to resolve them at last and now have a stable system.
It's got an onboard DDS so easy to prog 32 X 5KHz channels, however 160KHz is somewhat wide for class E and Power and Eff% will drop without retuning and tuning caps are large, expensive and no longer manufactured! so best to stick to 16Ch and a 80KHz range.
Did I mention a 'Scope?
You simply cant do this sort of work without one, simple as...
Str.
-
8)I used to work in a University EE lab, but alas, I now need rebuild my lab and construction space. Something to do in new year.
BTW mate.. the lil box is working Great. I had a tug with ordering BNC to PL 259, but its of course a no brainer.
Will look for your prototype. :-X
-
Sweet, thanks for the 'Flowers' !
Taming the high power version is keeping me busy!
Str.
-
::) yes, when you tame it will be Sweet I am sure. With regard to 15 watt Class E I am getting hum and high residual audio noise.
It would appear the audio chip has an Automatic Gain Control? If that be the case, is it possible to disarm it because as is the effect does increase distortion and non linearity.
I monitored the output on both an SDR and a FT900. The hum or noise rises if there is no audio present. So, I assume that is a gain effect in the modulator that I wish I could disable.
Since I posted top, I discovered an issue with power supply, a ground loop with a 850mv ac gradient. WOW, now that could put a hum on most transmissions! Ordered NEW 50 amp supply, my old one, an Astron 35 amp has been with me since about 1985 !
-
If your not monitoring directly from an RF sample and attenuator from the output of the transmitter (i.e. relying on your receive antenna to pick up some signal leakage) you will get excessive hum and noise. Check the output with a scope. If its clean, your monitoring system is suspect.
+-RH
-
Thanks 4 suggestion. I thought initially it was the SDR, then when I monitored with "other" radio still got that bassy noise, it just expands when no audio input, like filling up a balloon. :-X
A bite of post script here. Was using an Astron power supply. Suspected a ground loop. Took a DMM and yes, for some reason, 890mv of ac between the true ground and the DC ground. That's what the skunk in works is. So, solution a switching 14 volt supply that will not involve any ground loops to utility grounds.
I still am unsure why I see 830mv ac , high between power DC ground and utility ground. You can see we can't in electronics make any immediate judgements until tests are run.
This whole MESS would also place AC on audio inputs .
-
Nearfield RFI can cause all sort of problems in the lab, which is the main reason I only use transformer based couplers to sample the RF. Open air sampling is unreliable because the RF induces itself into the power line and reradiates mixing products of the carrier frequency and harmonics of 60 hz. My house is so old the electrical system has no ground anyway, so I need to be careful working on anything as the signal ground is usually not at the same potential.
You can make a transformer sampler by winding 20-40 turns on a torroid of convenient size such as a T106-2 and passing the center conductor of your coax through it. The winding should then be terminated by a known resistance, I use 10 ohm, 15 watt film resistors here. It can all be built into a small diecast aluminum box with two N connectors for the RF and a BNC for the sample.
20 turns should yield around 24dBm @ 100W, 40 turns around 18dBm, still too hot to feed directly into an SDR, but fine for most real spectrum analyzers.
+-RH
-
Didn't notice the earlier posts....Complaints about my TX!
ALWAYS a shitty PSU.
My Tx's are absolutely fine, NO RF Feedback either.
Hmphh....!
OK, more fun the in th' Lab.
I've got a 4 SiC FET system working well but have a 4 GaN system working even better.
God knows how much power it can put out but with 200W @ 48V it's stone cold.
I've got to build up my PWM system now to modulate it and check it's linearity.
If anybody would like a built up RF generator ONLY, please let me know, will just need bolting to a heatsink (doesn't need isolating tabs) has a built in DDS for 16Ch.
Just let me know what frequencies you want prog'd.
This has the very latest FET drivers and GaN FETs.
PCB is 5"X4", needs 7min-12Vmax for drivers and DDS and 2X48V 10A supplies for 200W Carrier 800W pep
Str.
-
I've noticed this even when testing my QRPp 22 meter beacons, powered by a 9 volt battery. Lots of 60 Hz (and harmonics) on the monitoring SDR.
Nearfield RFI can cause all sort of problems in the lab, which is the main reason I only use transformer based couplers to sample the RF. Open air sampling is unreliable because the RF induces itself into the power line and reradiates mixing products of the carrier frequency and harmonics of 60 hz. My house is so old the electrical system has no ground anyway, so I need to be careful working on anything as the signal ground is usually not at the same potential.
You can make a transformer sampler by winding 20-40 turns on a torroid of convenient size such as a T106-2 and passing the center conductor of your coax through it. The winding should then be terminated by a known resistance, I use 10 ohm, 15 watt film resistors here. It can all be built into a small diecast aluminum box with two N connectors for the RF and a BNC for the sample.
20 turns should yield around 24dBm @ 100W, 40 turns around 18dBm, still too hot to feed directly into an SDR, but fine for most real spectrum analyzers.
+-RH
-
Dearest Streachy,
Your wee box is powering several north American freedom outputs and one that I know of is in North lands.
I didn't want to insult you because the wee box in my lab is doing well. The passband is FM grade. The Output is as specified 15 watts and the so called frequency drift I checked yesterday in my growing electronic lab. The crystal is essentially inside a wee phone booth closed and when initial power , it (the crystal) heats a bit and I tracked about 250 hertz down as the wee box comes up to oven temp. I am amazed some of the people here have accused the wee box of instability. These people know NOT the olden days of VFO driven am rigs.
Every crystal is temp sensitive. I used to work in my youthful days at a BC band transmitter operator Gates Pre WW2 10 kw and it had an oven heater for crystal that NEVER was turned off. I think the approximate 200 or so hertz is indeed nominal and anyone who is listening in the mode of the wee box AM will not have to track anything. I mean 200 hertz drift. Not bad Old Boy. btw, After relative thermal equilibrium, the modulation is applied. I did detect a wee amount of FM, BUT.. I was still able to get reasonable reception with lower or upper sideband. In SSB mode, which is NOT the mode the box is intended. Of course PLL and other complexities would help, BUT the old VFO days of ham radio AM, many of these vintages exist. One can attempt to use sideband and it is not amazing the carrier wobbles! This was pretty much expected before PLL and DDS and so on. The complexities of keep a VHF signal on frequency even in the sixties presented issue. Not until phase locked loop and integrated circuits did RF transmitters become rock stable.
Well, I thought I would give the wee box a good review, but its also wee power. We await a real transmitter , one that we can really launch with.
I do not as yet own a RF spectrum analyzer , but in any design lab that is absolute must. Spurious signals can arouse all sort of nasty. I checked in band and out to +/- 1 mhz. It was definitely clean with 15 watts. There are sensitive military and aircraft channels in 43 meter band, beware. In a practical sense, no pirate will use a ship always at anchor. That's just asking for trouble, making it easy for the storm troopers. DDS allows stability and ability to haul up anchor and move! :'(
-
Thanks 4 suggestion. I thought initially it was the SDR, then when I monitored with "other" radio still got that bassy noise, it just expands when no audio input, like filling up a balloon. :-X
A bite of post script here. Was using an Astron power supply. Suspected a ground loop. Took a DMM and yes, for some reason, 890mv of ac between the true ground and the DC ground. That's what the skunk in works is. So, solution a switching 14 volt supply that will not involve any ground loops to utility grounds.
I still am unsure why I see 830mv ac , high between power DC ground and utility ground. You can see we can't in electronics make any immediate judgements until tests are run.
This whole MESS would also place AC on audio inputs .
Interesting I had an Astron power supply of the same vintage that became prone to ground loops as it aged. It started emitting a mid-range hum that you could definitely hear in the transmitter audio. It became a doorstop after a couple of days of trying to fix it. I got a new one that is still kicking.
-
Dearest Streachy,
Your wee box is powering several north American freedom outputs and one that I know of is in North lands.
I didn't want to insult you because the wee box in my lab is doing well. The passband is FM grade. The Output is as specified 15 watts and the so called frequency drift I checked yesterday in my growing electronic lab. The crystal is essentially inside a wee phone booth closed and when initial power , it (the crystal) heats a bit and I tracked about 250 hertz down as the wee box comes up to oven temp. I am amazed some of the people here have accused the wee box of drift. No so, and every crystal is temp sensitive. I used to work in my youthful days at a BC band transmitter operator Gates Pre WW2 10 kw and it had an oven heater for crystal that NEVER was turned off. I think the approximate 200 or so hertz is indeed nominal and anyone who is listening in the mode of the wee box AM will not have to track anything. I mean 200 hertz drift. Not bad Old Boy.
The drift is indeed thermal, it's pretty easy to see this if you plot out the carrier frequency over time. I've done this a few times and posted the results, it has the characteristic curve that looks sort of like a decaying exponential. Lots of drift at first, then things settle down. After about 20 minutes give or take, it has (mostly) stabilized. I typically see a 200 to 300 Hz drift when listening to these rigs. Since we're dealing with AM and not SSB, that's not a major issue. Worst case tweak the tuning frequency if you're running a tight IF bandwidth due to a weak signal. If the filter is open, you won't even notice it.
In the past I've plotted carriers of commercial AM stations, this post has a plot the 1490 kHz graveyard channel: https://www.hfunderground.com/board/index.php/topic,23649.0.html
You can see dozens of stations there. Many have interesting sawtooth style patterns, likely due to either thermal regulation or possibly even the PLL circuitry in the oscillator. Those patterns and frequency offsets can actually be used to ID MW stations by their carrier characteristics. Note the rather tight regulation, nothing is changing by more than a few Hz. Of course this level of frequency regulation isn't necessary for what operators are doing, just an example of what is possible. Some pirate stations do indeed have tight frequency regulation, as I've noticed when listening to them and watching the carrier frequency.
I have no firsthand experience with one of these 43m transmitters, but from the photos I've seen, the crystal is inside the box, and as digitalmod notes, it heats up as the transmitter runs. The IC chip used as an oscillator (I think LuLu transmitters tend to use a 74HC240?) also heats up, and stray capacitance is going to change, effecting the loading on the crystal, and the operating frequency. Plus lots of lower order effects we don't even think about. In a former life I worked in Industrial Process Control, specifically the use of beta/gamma/x-ray radiation to measure the thickness of metals, plastics, etc while being made. We used various radiation sensors, all of which had temperature sensitivity issues, so I've had some experience with these kinds of issues. We'd use an assortment of methods to keep things relatively thermally stable. Did I mention our sensor was sometimes a foot or two away from orange hot steel in a rolling mill? ;D
Some other transmitters of this style (Radio Animal's Grenade is generally considered to be the first popular such rig) had a socket for the crystal, which was mounted on the outside. I think Channel Z's transmitters are this way as well. I'm sure this was initially done so operators could easily switch out the crystal to change frequency (plus these were the giant WWII surplus FT-243 crystals). But the added advantage is that the crystal is now outside of the transmitter box. So it does not tend to heat up as much, resulting in significantly less drift. The choice of crystal is going to have some effect, lower tempco crystals are available, but for a price.
A PLL inside one of these rigs is still going to drift, because it's got a crystal lurking somewhere in the circuit, as a reference frequency. Of course you could always use an external GPS derived 10 MHz reference if you want nearly absolute frequency stability, but that's overkill. I don't see anyone doing that. Well, OK, I see redhat and one or two other ops doing that :)
Another thing you can see when you peer closely at the carrier of some of these rigs is FMing. From the schematics I've seen, these LuLu style transmitters use a 74HC240 logic IC as the oscillator.The slightest change in the voltage applied to the IC affects the operating frequency. I use a 74HC02 in my little 22m circuit, so I've had first hand experience with that. As the transmitter is keyed with CW (even on other gates in the IC) there's a slight shift in the operating frequency. The only way to prevent this would be to have a separate IC dedicated to the oscillator, and take great care to regulate the voltage. And even after doing that, you also need to deal with ground bounce/shifting issues. A ground plane on the PCB is not always a solution here, you might need a separate ground plane/system for the oscillator IC, with local regulation. It's not a trivial issue.
If you watch the carrier while the station is on the air, you'll observe two things. When there's no modulation, the carrier is narrow, as you'd expect. When there's modulation, it becomes extremely wide and variable, due to the carrier frequency moving around with the audio modulation. Likely because the power supply voltage and/or ground is moving around with the modulation. Dedicated voltage regulation and careful treatment of the ground of the oscillator and its connection to the rest of the circuit is required to reduce or eliminate this. As a practical matter I generally don't notice significant audio distortion due to this. Generally the issues I observe as a listener are either over or under modulation, but that's a station setup issue, not a flaw with the transmitter. IIRC, the Grenade had a rather clever built in compressor/limiter that involved an LED and photoresistor, to act as a sort of AGC. As a result, it was much more tolerant of input audio level settings.
None of this is meant as a criticism of any particular transmitter, only observations from the point of a listener and frequency measurement nerd.
-
Trouble is everyone is listening with their eyes, if they just used their ears!
Future versions will use AD9850 or AD9833.
I've noticed a little FMing and it's actually mainly due to the inductor I use to pull the Xtal as it picks up some RF from the output.
I HEAR you on the under/over mod. One op just had it turned up full as he thought it adjusted the power. I have informed folk that all material is best prerecorded and limited to a correct level. I've trialled the ssm comp chip but as it's meant for mics has way too much gain and hard to tame espec with RF present!
It's a basic TX, gets the job done.
Str.
-
To my dearest Christ,
Yes, all good observations. I worked in tech side of broadcasting until age 26 and realized I was stuck, went back to school.. that's a long boring story. Well, may I state FCC was always in over drive about AM radio until.. dada.. Ronald Reagan and his distain for Feds. But, when deReg came my station dumped its engineering staff. Kept the one that brown nosed the most as a part timer. Again long story.
Up shot: YES as late as 1975, no kidding, FCC had every AM station get a certified frequency check once a month! My station was on low end 550khz.. like duh.. how off frequency could it be, man? There was a old dude in Cambridge Crystals who would attempt to phase out other stations on 550 and get a frequency check against some unknown PRIMARY standard traceable to National Bureau Of Standards! Today Bolder has cesium atomic clocks that lose less than a second in a million years!!
Now to me 1975 was not that long ago, but look how far electronics and frequency stability has come.
-
Dearest pigmeat:
Well, what can be said. The old 35 amp Astron weathered it for years.. but when I began to investigate sources of NOISE.. yes there was the sad truth. Now I have a 50 amp switcher and the audio mysteriously ..(not really) became clear. Yes, the power line 120 volt wiring can make a good antenna.. for such pick up. It makes things confusing as to why audio and RF are getting so distorted. I kept the astron as a battery charge and a back up. All good things come to their end.. :-[
-
Another thing you can see when you peer closely at the carrier of some of these rigs is FMing....
(snip)
...If you watch the carrier while the station is on the air, you'll observe two things. When there's no modulation, the carrier is narrow, as you'd expect. When there's modulation, it becomes extremely wide and variable, due to the carrier frequency moving around with the audio modulation.
I do a bit of listening to Dutch and Greek MW pirates for entertainment.
Most of them have stable transmitters (stable enough) without a hint of major issues. Some are horrible. Some clearly have some sort of steady-ish low-frequency AM or FM on them, making the carrier wide. Some have some pretty serious FM issues. If my receiver's synchronous detector comes out of lock every time your transmitter exceeds 0% modulation (i.e., anything other than a dead carrier), then you might just have a problem with your transmitter.
-
If my receiver's synchronous detector comes out of lock every time your transmitter exceeds 0% modulation (i.e., anything other than a dead carrier), then you might just have a problem with your transmitter.
Hah yes. I don't think I've had issues with the sync detector coming out of lock, but I will have to listen carefully for any low frequency rumble. Although it's probably difficult to hear any if there's enough modulation to cause the FMing.