Making a small HF ssb amplifier a broadband one?
The input core is a ferrite one (broadband) but the output core is an iron powder and resonated (single band).
I wonder, if I replace the output core to a ferrite one as well, without resonating it, won't it be made into a broadband amp?
soemone has done this, with discrete transistors and pig nose ferrites broadband
Just remove the tuning capacitors from T2, or at least reduce them so the resonant frequency is higher than the top of your required bandwidth. Note that the 'Q' will go to almost zero so the amount of power you get out will drop dramatically. You trade the efficiency of a resonant circuit for the bandwidth.
Brian.
As far as I remember, I have tested it replacing the tuned circuit (Iron powder) with a ferrite (43 material) and I could not get significant power out of it without the resonant circuit. Maybe that is why. The case of low Q is also valid if using ferrite (broadband) instead of iron powder?
I also wonder why an input transformer phase splitter is needed? Can't this be made simpler by using a transistor phase splitter instead?
http://www.hobbyprojects.com/junction_transistors/phase_splitter.html
Some other means of biasing the transistors might be needed though, so it might not worth the effort.
The transformer material is best chosen to suit the frequencies you are using but consider that ferrite is actually iron powder anyway!
You can use a transistor phase splitter but it probably wont work as well as the transformer. The reason is that you need to drive the output transistors from a low impedance to ensure they will work at high frequency. The internal capacitances and stored charge has to be 'overcome' by the drive signal and to do that you need to provide significant base current. Using the splitter makes it's emitter and collector resistances a restriction on available current, if you drop them down to a low enough value, maybe <50 Ohms each, the splitter stage will draw more current than the output stage!
Brian.
Thanks Brian,
A final question, what will be the difference if I use TCA671 (Ic 200mA) instead of the CA3083 (Ic 100mA)?
Would I be able to get more power output?
Do I need to do any bias change for this or any other change?
I doubt you would get more power by simply changing the device to one with higher rating. You might be able to increase the supply voltage slightly to allow more current to flow through it.
The bias is set by Q1 on the basis it should match the other transistors and is obviously in close thermal contact with them. Changing to a different transistor array would in theory not need a bias change as all the transistors should still be equal. It might be worth experimenting with a capacitor (~10nF) from the center tap on T1 to ground (so it is across Q1) to keep the drive impedance as low as possible.
Brian.
But if I change the 2.2k to a lower value, won't I be able to get more current (which produces more output power?) at the same voltage?
Also, having a device which is capable of more current means thet it is heated less for the same amount of output power (or not?). Or is capable of tolerating more vswr than the less powerful one, without heat sinking.
Increasing the available bias current just sinks more down Q1 which is wired as a 0.6V stabilizer. Even if you provided more bias current to the other transistors it would only result in more DC being drawn (=hotter!) but not necessarily more RF out.
Power is V*I regardless of the device rating so the dissipation is the same whatever the device is. The difference in current rating is more to do with the construction of the transistor junction. Unfortunately, the larger junction needed to carry additional current also has higher capacitance which may be counter-productive at high frequencies. I'm afraid to get more out you have to scale everything up, including the level of drive signal and provision to remove heat.
Brian.
Hm... the TCA671 are specified for amplifiers to 30MHz, although I think the datasheet refers to the single transistors, not them connected in parallel.
For more power output, do you believe adding more transistors in parallel would worth it?
For example instead of having two parallel pairs, to have four parallel pairs.
But which transistors should be matched in that case, Q4 to Q2 or Q4 to Q5? I.e. the push-pull ones or the parallel ones?
This would require another array and you lose the advantage of the thermally coupled diode of course.
Adding more in parallel might work but at the end of the day you still have a broadband amplifer producing maybe 50mW of RF. If instead you used a single small RF power transistor you would get considerably more power out from a simpler circuit.
The transistors inside the TCA671/CA3083 are designed for small signal amplification rather than running at high current and when the spec says "amplifiers up to 30MHz" I don't think it means they will manage that at full rated current. Typically they are used for amplifying very tiny voltages where the close matching of the transistors can be utilized in low-offset input stages.
Brian.
There are a few special problems with broad banding an HF power amplifier.
Its a lot easier to get something working on one band only.
The first concerns impedance matching the source into your transistors over a wide bandwidth. Your best bet will probably be a broadband transmission line transformer of the Guanella or Ruthroff type. If the input impedance matching is not suitably efficient, your power gain will be dismal...
Next problem is the power gain of the transistors themselves will vary hugely over say a 10:1 frequency range. And that usually leads to instability problems at the low end, and often rather poor power gain at the top end.
Commercial designs often use negative feedback to both improve linearity and flatten the gain curve.
Matching the transistors to the load is just as difficult at the output as at the input. Its sometimes helpful to use a higher supply voltage and less current.
Just because it must run from a nominal 12v supply, does not mean you cannot use a switching power supply to supply dc to your RF power amplifier at a much more suitable (and fully regulated ! ) supply voltage.
I have never actually tried to build anything like this myself, I only know what I have read, and techniques I have seen used in commercial designs, and the explanations in service and repair manuals of if why it was done that way.
Actually copying an existing "known to work well" commercial design might save you a lot of trouble, even if the correct specific semiconductors are relatively expensive, which they probably will be.
Thank you!
I have not seen any such small amplifier being broadband. All the designs are 1 to 10w, not 1-1000mw.
What if I use discrete transistors do I really need the emitter resistors if I closely match them for beta?
In PA design to get broadband capability (BW > 3 octaves) more important than the bandwidth of the output match is the bandwidth of the input (interstage) matching network.
With the PA schematic attached I was able to get 10W (+/- 10%) from 3MHz to 30MHz.
Many thanks for this design!
What is the Input power requirement?
What is the current drawn by the amplifier?
How did you manage to wind 4T bifilar and 10T single wire onto the output toroid? Maybe you actually wound two separate windings for the primaries end secondary?