Just finished up this project and after some careful planning I ended up with a really nice amplifier for medium wave. Power in is around 1 to 2 watt PEP input for around 35 watts PEP output. Since I am using a higher powered transmitter at the input I added a padding resistor stage to allow for the greater power.
I can get a perfectly clean 8 watts carrier out of it into a 50 ohm load. With full 100% modulation it puts out 35 watts PEP with no noticeable modulation or carrier distortion on the oscilloscope at my test frequency of 1610kHz.
The unit needs very little cooling. I added a small 12v computer fan to it but is running at half speed to keep quiet and it still is plenty enough to keep the components cool. I was thinking of just skipping on the fan, but better safe than sorry.
I have tested this for 30 minutes driving it with a full 35 watts CW and it takes it just fine drawing a bit more than 4 amps from the power supply. The schematic is a widely copied design that is found on the internet, but I modified to work with the FETs I used along with the other components.
Here it is at 8 watts driving a 13.8 volt car brake light as a dummy load. Actually the light was so bright I ended up blowing out the light right after I took the picture ;D..
Closeup..
Quick schematic I drew up using Dia software..
I still need to build a low pass filter. Even though the output is a perfect sine wave I am still able to pick up the first two harmonics around the house which makes me a little uncomfortable.
The MOSFETs were randomly chosen as I had a dozen or so of them out of an old junked car stereo amplifier. I'm sure IRF510s would have worked just as easily with the appropriate design changes.
Being around the antenna tuning unit I swear I actually felt dizzy. Holding a florescent light I was able to light the darn thing without it being connected to anything as far as 3 feet from the ATU and antenna itself Always amazes me how those first 5 to 10 watts really puts out the power!
MW is fun to play with on the perf boards... where things get a little scary at VHF
Cheers!
Actually I never had too many issues with building lower powered VHF amps on perf boards. There is a bit of technique to it though. I run thin copper wire around the outer square of the board to ground, then snake more copper wire between all the other wiring and connections under the perf board. That way it acts like a near field shield, much like it would if you were using a copper etched professional board. It doesn't look pretty but if you do it right the results can be as good as the more pro. designs and heck I am a cheapo so perf boards are my kind of thing. I find with VHF stuff it's all about shielding the hell out of all the components and have multiple ground points all around. Anything more than a few inches of free space should be hard grounded or surrounded by ground shielding. Now as far as UHF, forget it haha.
Lower HF frequencies are so nice for this though as you said. Long connections = no problem. As you can see I didn't even mount the FETs directly to the board, just ran wires to them, because at a few MHz it makes no noticeable difference. Also makes replacing a dead duck a lot easier.
I have been running this amp all morning over here and it's holding up perfectly. I don't have a clue as to the frequency range it can cover but I tested everything from 1000Kc to around 1900Kc and it is very broadband. Also able to handle complete mismatches and even no load connected at all without overheating. Bit of a power guzzler though. My 3amp power supply is being pushed far past its nominal rating.
edit: modified schematic, noticed an error in the drawing and added more subtext.
In all honesty I'm not sure. I think it's #22. I got it from a roll of wire I had in the junk bin. The toroids are from an old computer power supply that I removed the wire from and rewound with my own. Scraping a bit off the paint they all looked to be ferrite material or some mix of.
I ripped out the voltage regulating guts of my ratshack supply and replaced it with a simple bridge rectifier and three 8200uF caps. Provides 20 volts around 3 amps, but allows large pulses of current drain w/o taking too big of a toll on the transformer thanks to the high uF caps.
Reminds me of how people with these high powered car stereo subwoofers use gigantic capacitors to make bass notes not kill the battery voltage stability.
Anyways, with 20 volts I can get 10 watts carrier, 40 watts PEP with no carrier swing. Makes me wonder just how far I can push this linear amp. I know the ratings of the FPQ50N06 and the IRF510, but this thing seems like it could really push a lot more if I just had the voltage and current.
Hey I have extra FETs if I pop one so no big deal. 10 watts though still feels like a lot of power when you hear it sing through other house electronics that aren't even radios!
Got to be careful with them light bulb loads as they do not present a proper load to the TX. They are like wire wound heavily inductive resistors at RF.
Shouldn't really hear it through other devices even at 10 watts unless there is some spurious coming out of the amp too. But like the light bulbs, hearing the signal from non-radio devices at least lets you know the amp is indeed working!
Peace!
K-ROCKS RadioOne
ZeroPointRadio
AM Stereo 1670
FM Stereo 92.1
possible to replace those mosfets with typical cb final transistors? i want 80-100mW in and max 5 watts out
Not too sure what all would need to be changed but my guess is that with some basic modification it could be made to work with typical CB transistors, but don't expect anywhere near the power.
The mosfets I used are biased just slightly above cutoff point to run in Class AB. With the bias voltage adjusted properly it will reproduce a sinewave with very little distortion, thus it works great in linear operation which is needed when using low level modulation like I did. However keep in mind that using low level modulation requires an amplifier that will do 4 times the power output of your carrier signal, so this amp as is will give a 100mw modulated carrier a carrier output of around 8-10 watts, or 35-40 watts 100% modulated.
My guess is that CB transistors would get quite hot operating at power levels near that. I know most CB transistors run about 3.5 to 4 watts carrier and take a beating when used in full service, so you will be pushing their limits if running them in push pull at 8 to 10 watts carrier.
I would still recommend using mosfets. They seem to be more durable, easier to build with, and quite common in old car stereo amplifiers and even the IRF510 which is quite cheap and would probably work fine in place of the random FQP50N06s. Also with CB transistors you may have to work harder at preventing spurious oscillations, especially if you plan on using them for my original purpose of working on mediumwave frequencies. If I were you I would give it a shot if you have some spare CB transistors laying around, who knows maybe it will work quite well.
Oh yeah I forgot about the impedances also. That would be a headache since mosfets expect a low impedance output which is why transformers are used to convert the low impedance to a higher one for the coax connection. You may need to increase the impedance on the primary of T5. All those values would have to be found through experimentation.
And you'll likely have to put 2 x 4:1 transformers on the input, as bipolars have a very low input impedance.
MOSFETS are voltage-input devices in parallel with high capacitance while bipolars are current-input devices parallel with just a few pF at SW and practically neglible at MW.
And you'll likely have to put 2 x 4:1 transformers on the input, as bipolars have a very low input impedance.
MOSFETS are voltage-input devices in parallel with high capacitance while bipolars are current-input devices parallel with just a few pF at SW and practically neglible at MW.
Cheers!
I must admit I have had instances where I have placed bipolar transistors in place of mosfets in push pull or half biased single stage linear test projects, and even the other way around and happened to have the test part work, though usually not as great as the original. Of course it's only a matter of bias, load, and input section right?
I'm not suggesting that either can replace one another, but with a bit of modification it can happen. The issue comes in with impedance on various transformers and basic characteristics of the initial amplification/switching devices themselves.
Put it this way, I have been around a breadboard and out of boredom plugged in a CB style transistor (hoping for fireworks because of getting angry with circuits not working as planned) in place of a MOSFET, and sometimes they have worked! Though requiring a lot different bias voltage and usually causing a lot more heat and less space for thermal breakdown. Of course with proper impedance matching on their in/out that would probably make them workable. I think though the efficiency comes into play.
The main issue I see here is the initial question. Can he make an amplifier providing 10 watts carrier with 40 watts PEP using nothing more than CB PA transistors. I just don't see it happening. I have been there done that, and unless he has some bulletproof transistors made from quality donor CB radios it's not gonna happen. Most of the CB radios I have scrapped come with the most basic junk <30mhz transistors in existence. There are exceptions with Uniden models and some of the Cobras that included SSB.
We are talking 3-4 watt transistors at full power. Pushing them push pull at 10 watts can happen, but the stability factor is iffy imho.
I'd love to see the original poster do this, and I do believe it can be done if proper procedures are in place like RFI beads, metal enclosure, choked power wires, and running the amp knowing that it produces a clean trapezoid pattern while modulated.
--------------------
Now to the original poster... There is no reason not to try. My recommendation would be to at least try. I have been told by people just like me in the past that it won't work because of A or B or C. The real heart of building ones own setup is experimenting. You will fry a few transistors trust me. Maybe even drive away listening to your station and once 5 miles out suddenly your signal dies and you race back home only to find a home studio full of burnt capacitor smoke.
My biggest recommendation is to simply experiment and find what works best. Pirate radio is exciting to the listeners who stumble upon a mystery signal, but the real ride comes to the one making the station. There is nothing like putting on a show, and not just playing music and producing a show, but making the very items that allow that show to be listened to
I have been there done that, and unless he has some bulletproof transistors made from quality donor CB radios it's not gonna happen.
I've tried this before with MRF475 even.... what happens is SURE! You'll get the power easily... but at what IMD? A spectrum analyser will make you want to barf!
Now running an MRF454 at 1/4 its rating and you'll come close to the spectrum of a MOSFET.
There is reasons the industry is doing what it's doing
i have an 80mW 50 ohm panasonic cquam signal generator for medium wave. i'm looking to boost it for use in carrier current operation with my tcu30. I own an SA, 200m quad scope, DMM, MW FIM and have boxes of parts but no cash to work with.
i live in an apartment so 15.219 is not feasible only carrier current. others tests reveal that 2 watts output seems optimal for most c-c setups anything more overloads front ends of local radios
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
