"Pill" questions...

[Shockwave]

Now consider that every mono band, multi transistor HF amplifier you've ever seen was wasting half of the transistors useful output power in push pull, broadband configuration. Twice the output can be achieved combining single ended stages with resonant, LC input and output sections. Since this is RF, the output coil fills in the missing half of the waveform just like a tube amp.

Linear operation with class AB bias is easily achieved with single ended RF amps. Just look at the final stage in any SSB CB radio. They are all single ended and run linear in SSB. After 30 years of competition amplifiers being built, you would think some of the builders would have taken notice by now...

 

[kopcicle]

http://www.seboldt.net/k0jd/mot-an762.pdf
http://www.om3bc.com/datasheets/mrf421.pdf

"The MRF421 with a specified power output of 100 W PEP
or CW is the largest of the three RF devices. The maximum
dissipation limit is 290 Watts, which means that the
continuous collector current could go as high as 21.3 A at
13.6 V operated into any load. The data sheet specifies 20 A;
this is actually limited by the current carrying capability of
the internal bonding wires. The values given are valid at a
25°C mount temperature."

Are you suggesting that the (admittedly dated) reference to the MRF421 as a 100 watt device is capable of in excess of 180 watts as a single ended device?

you would think some of the builders would have taken notice by now...

Helge did.

https://cache.nxp.com/docs/en/application-note/AN749.pdf

So there are the combiners.

Now back to the MRF421. Even in a single frequency, tuned circuit, with adequate cooling I have never seen in excess of 110 watts for any length of time. I lived with this sand state for over two years. It's one of the few BJT's I actually know anything significant about. Even if you eliminate the broadband push pull losses you still have the losses in the combiner(s). TANSTAAFL. Any gains in single ended implementation (actually theoretically possible) result in minimal percentage gains in efficiency at the expense of bandwidth necessitated by higher "Q" . These gains in no way equal half of the transistors useful output power.

 

[Shockwave]

It's not that single ended provides twice as much output per transistor. It's that Single ended gets the full output from each transistor simultaneously while a push-pull pair never gets more than the full output from one transistor at a time. If you know that one MRF421 can support 20 amps of collector current, measure the current drawn by a push-pull pair. The circuit will not support a single amp more with two transistors in push-pull. Using a pair of single ended transistors with combiners and splitters will allow the full 40 amps to be drawn. Doubling the input and output power as compared to a pair in push-pull.

 

[kopcicle]

.

 

[Shockwave]

http://www.seboldt.net/k0jd/mot-an762.pdf
http://www.om3bc.com/datasheets/mrf421.pdf

"The MRF421 with a specified power output of 100 W PEP
or CW is the largest of the three RF devices. The maximum
dissipation limit is 290 Watts, which means that the
continuous collector current could go as high as 21.3 A at
13.6 V operated into any load. The data sheet specifies 20 A;
this is actually limited by the current carrying capability of
the internal bonding wires. The values given are valid at a
25°C mount temperature."

Are you suggesting that the (admittedly dated) reference to the MRF421 as a 100 watt device is capable of in excess of 180 watts as a single ended device?



Helge did.

https://cache.nxp.com/docs/en/application-note/AN749.pdf

So there are the combiners.

Now back to the MRF421. Even in a single frequency, tuned circuit, with adequate cooling I have never seen in excess of 110 watts for any length of time. I lived with this sand state for over two years. It's one of the few BJT's I actually know anything significant about. Even if you eliminate the broadband push pull losses you still have the losses in the combiner(s). TANSTAAFL. Any gains in single ended implementation (actually theoretically possible) result in minimal percentage gains in efficiency at the expense of bandwidth necessitated by higher "Q" . These gains in no way equal half of the transistors useful output power.

The link you provided for broadband combiners covers a lot of useful information. Things like that 4 port combiner would surprise people who have never seen a 8 transistor amplifier with only a single RF splitter and one output combiner. Something that has been done on high end RF generators for decades. My area of interest concerned the combining of single ended amplifier stages. Something I have never seen on HF before doing it myself. Although, I have seen it on VHF in some higher power 2 way communications equipment and that's what gave me the idea to apply the same method on HF. Two transistors combined in a single ended configuration provide twice the power handling capacity and output as two in push-pull.

 

[kopcicle]

Two transistors combined in a single ended configuration provide twice the power handling capacity and output as two in push-pull.

umm, no.

 

[Shockwave]

umm, no.

Please provide a technical reason why you disagree? You will quickly change your mind on this topic as soon as you measure the current drawn by any broadband, push-pull amplifier. No pair will ever show an input current that exceeds the rating of a single device even though there are two. I've made this claim several times and have been met with skepticism but no one even tries to measure that input current.

Single ended, resonant LC circuits easily sustain the input current and match the output power of a push-pull pair, using one transistor. Nothing stops two from being combined so long as phase and magnitude are balanced in each stage. Since this pair operates in phase, they both conduct simultaneously, doubling input and output power capability.

 

[Shockwave]

"The MRF421 with a specified power output of 100 W PEP
or CW is the largest of the three RF devices. The maximum
dissipation limit is 290 Watts, which means that the
continuous collector current could go as high as 21.3 A at
13.6 V operated into any load. The data sheet specifies 20 A;
this is actually limited by the current carrying capability of
the internal bonding wires. The values given are valid at a
25°C mount temperature."

Have you ever seen a push-pull pair consuming more than the maximum 20 amp rating of a single device? That's not even a possibility with each transistor only conducting half of the time. While one transistor works, the other is asleep.

Are you suggesting that the (admittedly dated) reference to the MRF421 as a 100 watt device is capable of in excess of 180 watts as a single ended device?

Look back at the threat titled "Hundreds of watts from a 2SC2879" and you'll see where that is the case using a very similar transistor. After much skepticism, W8JI finally confirmed what many of us already knew. One of those transistors in a single ended, resonant circuit quite easily makes over 200 watts without ever exceeding its maximum collector current or dissipation.

Now back to the MRF421. Even in a single frequency, tuned circuit, with adequate cooling I have never seen in excess of 110 watts for any length of time. I lived with this sand state for over two years.

Is this the maximum sustained output you could get from one in a single ended circuit with a resonant input and output circuit, or half of what you've seen from a push-pull pair?

 

[AudioShockwav]

"Hundreds of watts from a 2SC2879" and you'll see where that is the case using a very similar transistor. Aftermuch skepticism, W8JI finallyconfirmed what many of us already knew."

100% correct.
This was done with a single original Toshiba 2sc2879 .
The thread both here and over on QRZ also proved that the 2879 was capable of much more than the 120 watts that the transistor spec sheet stated that was published by Toshiba.
Another reason I say that Spec sheets are not the final word on many transistors.
It is worth reading both threads.
I wish Tom had posted the test jig he used.
73
Jeff

 

[Shockwave]

The 2879 was an impressive transistor when it came out. It was while building current limiting protection circuits for these transistors that I learned something very revealing. The current limit to keep a single ended transistor within its Safe Operating Area was exactly the same as with a pair in push-pull.

Both the single ended and push-pull circuits had to remain below 25 amps or the transistors burn open. That told me the pair in push-pull, could not support any more input current than one device. Therefore, the DC input power and RF output power of the pair in push-pull, cannot reach levels beyond a single ended, resonant circuit.

 

[Tallman]

I have seen one case where "Stacking" transistors actually provided more output than a push/pull configuration and it was not used in a radio configuration.
It was used as a speed control for a DC motor.(FETs)
The maximums listed by Transistor manufacturers are just that. They are not holding back higher performance just to play with your emotions.
The energy lost in a combiner is not as much as you would think. If the transistors are properly matched and biased properly the loss would be negligible. There would be some power lost in DC resistance of the wire.
One thing I have not seen in this thread is the "De rating Curve." It figures heavily into the Junction Temperature" maximum. Heatsinks are forever changing temperature and the higher the temp the less power it can handle.
Also with the negative temperature coefficient the hotter it gets the more power it conducts and you end up in a thermal runaway. Lots of showy magic smoke coming out real quick.

 

[AudioShockwav]

The Toshiba spec sheet rates the 2879 at 120 watts.
I can attest to the fact the a single 2879 will do double that without a doubt.
The " test jig" that Toshiba used to produce that number has been called poor by a well known RF engineer that has been designing and building amplifiers for years.
This same engineer bolted a single 2879 to a large heat sink using a soft copper gasket in a single ended class c amplifier circut and at 5 MHz this single transistor pulled 30 amps at 18 volts and produced almost 300 watts output regardless of what the spec sheet says.
For those that will say that it is all harmonics his testing showed that Less th a 10% of the signal contributed to the total output.
His call sign is W8JI
Have a look here:
https://forums.qrz.com/index.php?threads/2sc2879-producing-250-watts-hahaha.248732/page-6
The Toshiba spec sheet is not the final word on these transistors.
I have run these same devices at 20 volts more that once and they survived with no magic smoke released.
It is an extremely rugged transistor that exceeds anything ever posted in a spec sheet.
You would be surprised at how much abuse they could take.
It is a shame Toshiba stopped making them and the clones that are on the market today will never equal the original design.
73
Jeff

 

[Tallman]

http://html.alldatasheet.com/html-pdf/30139/TOSHIBA/2SC2879/246/1/2SC2879.html
The data sheet say that it's power output is 100 watts pep. Which has nothing to do with the collector dissipation. The package design for the transistor could be capable of 250 watts of power dissipation, that does not mean the transistor can do that and remain functional. I noticed some one compared them to top fuel dragster engines.which are rebuilt after every single run. That's a pretty good analogy. You might get 125 watts pep, but that is not an RMS dead key carrier.
There was a figure that was used by shady marketing groups that exaggerated power output claims and it was called IPP watts. That stood for Instantaneous Peak Power. Reading those figures you would think you had the most powerful boom box in the world.

 

[AudioShockwav]

Sigh.
There is no such thing as RMS watts when measuring RF power.
I will stick to my statement.
1, Manufacture spec sheets are not the final word in device output.
2, The 2879 is more than capable of producing in excess of what the manufacture has stated in there specification sheets and survive just fine if feed, driven and cooled properly.
And I will add that this also applies to other RF transistors as well.
If you have read the entire thread and don't understand what was posted then there is nothing more that I can offer.
Have a good evening.

73
Jeff

 

[Shockwave]

The law of thermodynamics still applies. The input power to two push-pull transistors is split 50-50, minus circuit losses in the passive components.

73

How does that work when one is an open while its driven into the negative region? Does the base impedance remain low enough to absorb half of the drive? Measuring the DC input power to a push pull pair will prove that both never contribute to the output power simultaneously but I'm still trying to learn about what takes place with the drive power during the negative alternation where one transistor is cutoff.