hundreds of watts per 2879 ???????????????????????

[dxhound]

If you take your example of 250 watts input power at 18 volts and 13.9 amps with 70% efficiency, I come up with 175 watts into the load and 75 watts of collector dissipation. Up the drive until the collector draws 25 amps and you have 450 watts input, 135 dissipation and 315 watts output. That is only 24 watts more output then I've ever seen from the device. The transistor was designed with enough collector dissipation so that it would not be the weak link in the chain. Going past 20 volts or 25 amps is what kills this part. This is what prevents you from ever reaching the output power that would be obtained at the 250 watt collector dissipation.

I read your posts in the other tread about building combining the two amps and using it on the other HF bands. Before you go through the work of doing that, I would test a single amplifier on each of the lower bands you intend to operate on to make sure it runs stable first. Even passing this test is not a 100% guarantee it will still be stable after you combine two but it's a great start. Keep the coax going to your combiners as short as possible and the same length.

At 94 watts each transistor they will be running comfortably and with the correct mic gain setting the IMD shouldn't be too bad. It is biased but as you probably know it's not the best bias circuit or amplifier for HF. If one amp works good I'd almost be tempted to leave it at that rather then come up with twice the DC amps to feed a pair and gain a half an "S" unit. There is also the chance of a self oscillation on some frequency that could wipe out transistors instantly too.

Oh yeah, the amps will be tested individually before continuing. I was thinking of taking them into work and doing a freq sweep with 5 watts output and doing both an input VSWR plot and gain plot per based on freq.

250 x.7 = 175 watts output 75 dissipation. Same setup but driven till 25 amps is reached you get 450 watts input 315 output and 135 watts disipation ok got it, Dont know what I screwed up when doin my math prior............oh wait hahaha never mind, i divide by 70 instead of multiplying by .7 hahahaha Wow, i need a vacation.

But what im still pondering is, the resistance at the devices output. Im sure the output transformer has at least 2 ohms of resistance, (I may be wrong). Take that into consideration and at 18v drawing 9 amps, the device would only then be producing 162 watts. This is where I am getting hung up. This is what the basis of my argument was.

Is this resistance what is accounted for already in the efficiency rating? The transformers secondary winding is what separates the transistor from the 52 ohm output correct?

 

[Shockwave]

Oh yeah, the amps will be tested individually before continuing. I was thinking of taking them into work and doing a freq sweep with 5 watts output and doing both an input VSWR plot and gain plot per based on freq.

250 x.7 = 175 watts output 75 dissipation. Same setup but driven till 25 amps is reached you get 450 watts input 315 output and 135 watts disipation ok got it, Dont know what I screwed up when doin my math prior............oh wait hahaha never mind, i divide by 70 instead of multiplying by .7 hahahaha Wow, i need a vacation.

But what im still pondering is, the resistance at the devices output. Im sure the output transformer has at least 2 ohms of resistance, (I may be wrong). Take that into consideration and at 18v drawing 9 amps, the device would only then be producing 162 watts. This is where I am getting hung up. This is what the basis of my argument was.

Is this resistance what is accounted for already in the efficiency rating? The transformers secondary winding is what separates the transistor from the 52 ohm output correct?

I see your question now. The collector load impedance of the transistor is what determines the DC current drain and this value changes with the drive level. Using ohms law we can see that in order to draw 25 amps at 18 volts the collector output impedance actually drops all the way down to 0.72 ohms! Almost looking like a short and why the output transformer primary is usually brass tubing to cope with the skin effect. It's now the transformers job to step that .72 ohms up to 50 ohms according to the turns ratio used.

Since the collector impedance changes with drive, current drain and power output, it is only possibly to obtain a perfect match at one particular value. That value is chosen based on the maximum power the amp is designed to make in order to obtain maximum efficiency and power transfer. Transformers are also extremely efficient.

 

[dxhound]

I see your question now. The collector load impedance of the transistor is what determines the DC current drain and this value changes with the drive level. Using ohms law we can see that in order to draw 25 amps at 18 volts the collector output impedance actually drops all the way down to 0.72 ohms! Almost looking like a short and why the output transformer primary is usually brass tubing to cope with the skin effect. It's now the transformers job to step that .72 ohms up to 50 ohms according to the turns ratio used.

Since the collector impedance changes with drive, current drain and power output, it is only possibly to obtain a perfect match at one particular value. That value is chosen based on the maximum power the amp is designed to make in order to obtain maximum efficiency and power transfer. Transformers are also extremely efficient.

ok, so the load the transformer represents is negligible in this application. Then the transformer matches the transistors "1" ohm to 50 ohm by varying the windings. Basically, if you had a transformer that matched 2ohm to 50 ohm, the highest output then would be limited to 162 watts. So really, the transformers efficiency is what determines the amps overall efficiency, and all the bias does is set the conduction angle for the transistor?
Am I understanding you right on this?

 

[Shockwave]

ok, so the load the transformer represents is negligible in this application. Then the transformer matches the transistors "1" ohm to 50 ohm by varying the windings. Basically, if you had a transformer that matched 2ohm to 50 ohm, the highest output then would be limited to 162 watts. So really, the transformers efficiency is what determines the amps overall efficiency, and all the bias does is set the conduction angle for the transistor?
Am I understanding you right on this?

The transformer doesn't present the load, it's just changing the value of impedance between the collector and 50 ohm output based on it's fixed turns ratio between primary and secondary. If the transistors collector impedance was at 2 ohms at 18 volts and the transformer was correctly matching that to the 50 ohm load, 162 watts sounds right.

Now if you take that same 2 ohm example and drive the amp down to 1 ohm at the collector, the output impedance would drop to 25 ohms with the same transformer. The output impedance of a transformer will change proportionally with the input impedance.

It is the transformers loss that is nearly negligible. I haven't actually measured the loss of the transformer itself but Tom mentioned a 10% loss with the transformer and a harmonic filter combined. I think the filter could have more loss then the transformer. Bias plays the bigger part in the efficiency of the amplifier by controlling the conduction angle. The longer the duration the conduction angle is, the lower the efficiency.

 

[dxhound]

To anyone intrested, go to qrz and read this thread, it was mentioned earlier in this thread as well. VERY good info there!!!!
http://forums.qrz.com/showthread.php?t=248732&page=6

 

[BOOTY MONSTER]

LOL , even the pork butts found it an interesting topic .
too bad the mods/admins at world wide pirates of dx have no faith in their members to be able to discuss it .
at least they seem to be able to post watergates without hurting each others feeling .
maybe they should change their name to ......

World Wide Pirates Of Watergates .

 

[Shockwave]

I'm actually impressed that we came up with nearly identical results in both forums. We agree that the 2SC2879 can easily produce 250 watts output without burning up. It can do it for way more then a few seconds. Even without any harmonic filtering, well over 90% (closer to 95%) of the power is on the fundamental frequency. There is virtually no harmonic content past the 3rd order. And that forum implied you can't get accurate information from this one.

 

[Moleculo]

Quote from other thread:

I had a feeling where I would find that thread.
If these guys would get out and read the volumes of information on solid state linear amplifier design they would change their tune, and maybe improve their audio quality.

So it seems to me after W8JI tested, documented, and validated that it is indeed possible to do this, that "these guys" did their homework after all. So which is it...the "volumes of information" are lacking or perhaps the "volumes of information" was indeed read and some just understand it better than others?

This topic has actually been discussed here a long time ago and got the same type of response over there. It's kind've nice that Tom (W8JI) actually tested it and documented it this time around.

 

[dxhound]

So why did the fatboy with 2 final 2sc2879's at around 350 watts output show harmonics well into the 800mhz range????

 

[Shockwave]

So why did the fatboy with 2 final 2sc2879's at around 350 watts output show harmonics well into the 800mhz range????

That's a question I don't have an answer to because it's nearly impossible for a HF amp to generate harmonics all the way up to the 29th order or 800 MHz. The power level contained in harmonics is reduced exponentially as they extend further from the fundamental. The HF matching transformers used would be so highly inductive at 800 MHz they would attenuate virtually all of the signal even if the amplifier was excited with an 800 MHz signal at reasonable drive power. I would think the internal capacitance of the 2SC2879 would certainly destroy all gain the device could have at 800 MHz also.

I'm curious myself as to what could cause harmonics to extend any where near that. If you asked me if it were possible, I would have said "no way". That wouldn't explain how you could see it on your spectrum analyzer though and because of that it warrants further investigation. I'm not a member of QRZ yet so I can't post there but I would suggest you ask W8JI for an explanation on this one. He's one of the brighter bulbs in these forums and I'd be curious to know if he can come up with a logical possibility that I've overlooked.

 

[dxhound]

I asked, so we will see soon his answer. Hey shockwave, Did you see his comments about harmonics and watt meters? At least I was right about somthing.......

 

[Shockwave]

Yes I did see that but it's important to know it relates to the quality of the meter too. Take a Dosey and feed it with twice the rated frequency and it will show much higher readings. That's not going to happen on the Bird 43. It will read higher but only by about 10%.

 

[BOOTY MONSTER]

so how much power is a 2879 actually good for on a normal running vehicles voltage , assuming the electrical system has enough amperage to power it also ?

 

[Shockwave]

About 175 watts on 12.5 volts. IMD will still be suffering at this power level.

 

[dxhound]

Yes I did see that but it's important to know it relates to the quality of the meter too. Take a Dosey and feed it with twice the rated frequency and it will show much higher readings. That's not going to happen on the Bird 43. It will read higher but only by about 10%.

Ill buy that for a dollar

converted from redneck: I agree