HF rf amplifier circuit without balun core transformer

[Dom13c]

For the very beginning, you can certainly use some of those classic transistors; the IRF510 is seen still quite often in RF amps anyway; but as a word of caution: the pricy RF transistors have their prices because of the favorable rf parameters, which you normally don't see on the first look; one big point are the intermodulation properties of a transistor, where a real rf component beats out the shit off the IRF510 and friends.

It's an ever repeating issue among radio amateurs - on one side you want to construct for a bargain, but in the interest of everybody on the frequency, you have to take things like imd, harmonic suppression and phase noise into consideration - even though many amateurs give a literal shit about it.

 Actually it is damn true. I just am trying to put together the whole thing as cheap as possible. So trying to avoid those rf buddies

[Dom13c]

Worth noting that the design referenced is very similar to that of an awful of of CB 'burners' and that you're going to need to tinker with it to get a FET working in it (it can be done, I managed to get >35W @ 24V on 28MHz from a single FQP13N10 in a similar design before it grenaded).

If you can track down an Eleflow RF transistor or even the EKL ERF2030, again, it'll make your life simpler as the venerable BD139 will not get you 10W output, but, they're 'special' RF transistors and cost way more.

 Actually , it's the cost of those rf buddies that sucks me. Otherwise they are way too better than the BD139 .

[CJay]

Agree, they're expensive, they have a load of advantages but you pay for them, the prices make my teeth itch, there must be decent bipolar parts out there somewhere, I don't think (for instance) Eleflow are manufacturing, I think they're rebadging, I could of course be wrong...

[xaxaxa]

I don't know where you guys have been looking but "proper" RF transistors aren't expensive...

MRFE6S9125N: $5 on taobao, 125W and operates from HF to VHF
MRF5S9070N: $1 on taobao, $5 on aliexpress, 70W from HF to VHF
MRFE6S9160: $12 on aliexpress, 160W from HF to VHF
and many many more...

These days a 2.4GHz 160W transistor costs $10 there isn't much reason to be using IRF510 or vintage bipolar parts...

[CJay]

I don't know where you guys have been looking but "proper" RF transistors aren't expensive...

MRFE6S9125N: $5 on taobao, 125W and operates from HF to VHF
MRF5S9070N: $1 on taobao, $5 on aliexpress, 70W from HF to VHF
MRFE6S9160: $12 on aliexpress, 160W from HF to VHF
and many many more...

These days a 2.4GHz 160W transistor costs $10 there isn't much reason to be using IRF510 or vintage bipolar parts...

I agree, you get a *lot* of transistor for the money, UHF cellphone station FETs are bargains these days but they're nowhere near as easy to work with as the parts we were discussing, the package is a PITA to work with as it's designed to be soldered down and there's a good amount of heat needed to do that, cell phone parts usually need 28V or higher which isn't common in a shack and they take some taming to keep them well behaved or you end up with a 70W UHF oscillator.

They're also considerably more powerful than the OP needs and would be hard to get working well without ferrite impedance matching transformers.

For those prices though, personally, I will order some just to have on the shelf so I can experiment at some future time.

[bd139]

Have to agree. One of my constraints is time and it’s very easy to whip out a mismatch smoked TO220 compared to proper RF transistors. Every PA design I have done has exploded at least twice. I thought I sucked but no this is apparently normal. Also the cheap switching FETs can be had for pocket money with free next day delivery here from reliable suppliers. $8 for 10x Vishay IRF510’s from RS. No slow boat, no risk of fakes or out of spec devices.

Personally I’m trying to stay under 10W as that leads to manageable projects as well.

[CJay]

Have to agree. One of my constraints is time and it’s very easy to whip out a mismatch smoked TO220 compared to proper RF transistors. Every PA design I have done has exploded at least twice. I thought I sucked but no this is apparently normal. Also the cheap switching FETs can be had for pocket money with free next day delivery here from reliable suppliers. $8 for 10x Vishay IRF510’s from RS. No slow boat, no risk of fakes or out of spec devices.

Personally I’m trying to stay under 10W as that leads to manageable projects as well.

The switching FETs have that disadvantage though, they aren't designed for linear operation so you're balancing on a knife edge where one slip will cut you, 'proper' RF parts are happier over a wide range of linear operation, swings n roundabouts, you pay extra for that and a datasheet that gives you that information whereas IRF, FQP etc switching FETs are best guess but dirt cheap so you can fry a few while you experiment (and I have, I've gone through a couple of dozen FQP13N10 devices while experimenting)

There's a £12 70W MOSFET amplifier on it's way to me from China (ordered last week) to tinker with as well

It's all fun and I'm trying to sort out a TaoBao account to get my hands on 20-30 of those dirt cheap 70W FETs so I can play some more.

Maybe once I've tamed those I might have a go with the set of VRF2933 FETs I've been hanging onto for the last three or four years...

[bd139]

Yes they are tetchy. I've got a hole in the wallpaper where a bit of MOSFET pinged off. I was being lazy as my current limited bench supply was being used to power the rig and connected my radio SMPS to it which promptly chucked 22A through it before it exploded.

I might have a go with one of the Chinese kits. At that price it's no major loss if it turns out to be a turd.

[Wirehead]

There are some beefy ones, which given proper cooling and a good power supply, will actually keep going until the insides splutter all over the place.. this was at 300W CW Carrier on 80m band...  - simple switcher mosfet

Shorted out and sustained about 50A ..  No smoke, just molten guts

[CJay]

Yeah, we used to have to chisel MOSFETs (IRFP450 I think) off heatsinks in Compaq 386S desktop power supplies, they spewed a jet of molten guts out through the tab, insulator and welded themselves to the heatsinks quite effectively, often leaving a hole right *through* the heatsink so even when you'd drilled out the weird rivets you couldn't get the damn things to move.

I seem to remember it was usually the IRFP in the primary that was closest to the switching transformer...

oops, the PSU seen in this video:

[Dom13c]

I don't know where you guys have been looking but "proper" RF transistors aren't expensive...

MRFE6S9125N: $5 on taobao, 125W and operates from HF to VHF
MRF5S9070N: $1 on taobao, $5 on aliexpress, 70W from HF to VHF
MRFE6S9160: $12 on aliexpress, 160W from HF to VHF
and many many more...

These days a 2.4GHz 160W transistor costs $10 there isn't much reason to be using IRF510 or vintage bipolar parts...

I agree, you get a *lot* of transistor for the money, UHF cellphone station FETs are bargains these days but they're nowhere near as easy to work with as the parts we were discussing, the package is a PITA to work with as it's designed to be soldered down and there's a good amount of heat needed to do that, cell phone parts usually need 28V or higher which isn't common in a shack and they take some taming to keep them well behaved or you end up with a 70W UHF oscillator.

They're also considerably more powerful than the OP needs and would be hard to get working well without ferrite impedance matching transformers.

For those prices though, personally, I will order some just to have on the shelf so I can experiment at some future time.

Well, I do agree with you.  When it comes to buy rf transistors online, the problem of getting fake parts really sucks. And those matching transformers connected in circuit without proper measurements are nothing but an arrow thrown in the dark.

[CJay]

I think you're overstating how difficult it is to make an impedance matching transformer, they're not difficult, especially for HF, the ratio is determined by the number of turns of wire, the impedance can be calculated fairly easily if you have the spec of the core, if you don't have the spec of the core then you can work it out with some wire, a calculator and an LCR meter, one of the cheap ATMega ones is good enough.

[David Hess]

It also is not critical in the respect that you are not matching the output impedance of the power amplifier to the line or load; doing so would result in lower efficiency.  This is why broadband power amplifiers get away with transformer turns ratios limited by construction.  The turns ratio and supply voltage determine output power assuming that the output devices can support it.  The design works backwards to the output devices.

[Dom13c]

I think you're overstating how difficult it is to make an impedance matching transformer, they're not difficult, especially for HF, the ratio is determined by the number of turns of wire, the impedance can be calculated fairly easily if you have the spec of the core, if you don't have the spec of the core then you can work it out with some wire, a calculator and an LCR meter, one of the cheap ATMega ones is good enough.

okay, I'll try it out the real problem arises when you dont have a lcr meter . the same is with me. It seems like I have to invest some bucks to buy one of them. cause the hobby I'm in definately needs one of them

[Wolfgang]

Hi,

if you want something transformerless you could go for a complementary amplifier.
The web has some info from Chris Trask.

http://home.earthlink.net/~christrask/Complementary%20Push-Pull%20Amplifiers.pdf

I made one using 2N2905 and 2219 transistors, like that:

https://electronicprojectsforfun.wordpress.com/rf-module-gallery/the-amplifier-module-gallery/a-complementary-hf-push-pull-amplifier/

The frequency range is ca. 50MHz (-3dB) and you could extract 100mW output with a good linearity (38dBc). IP3 is 30dBm.
Its linear and foolproof against output open, short or mismatch.
 

[David Hess]

http://home.earthlink.net/~christrask/Complementary%20Push-Pull%20Amplifiers.pdf

The more things change, the more they stay the same.  That design is very similar to simple high impedance active probes.  Eventually they ended up with a diamond buffer output stage.

The problem is that it is difficult to extract power from emitter follower outputs because their low impedance increase the effects of parasitic inductance.  This is why higher power amplifiers also use higher voltages to allow higher impedance into the output transformer or network.

I made one using 2N2905 and 2219 transistors, like that:

https://electronicprojectsforfun.wordpress.com/rf-module-gallery/the-amplifier-module-gallery/a-complementary-hf-push-pull-amplifier/

That design is very similar to the high voltage z-axis amplifier on a CRT oscilloscope.

It does not help us here but very broadband push-pull power amplifiers for cable distribution use a push-pull stage driving a balun core transformer except that unlike a similar HF amplifier, the transistors are configured in common-base mode instead of common emitter.  They made special transistors in the same packages with the base and emitter leads swapped just for these.

One alternative I might consider is using multiple surface mount RF transistors in parallel in a more traditional circuit.  It is too bad all of the inexpensive RF power transistors are gone.