Problems about 400MHz,4w,PA design

[l.j.zheng]

mrf1517 application notes

hi,everyone.

I got the job to design a PA at 400MHz, with P1dB>4W,the Vd<12.5V.
I have searched many many datasheets such as Minicircuits,triquint, as well as
freescale,philips,infeon and so on.But I can't find a good module or a discrete transistor.
Since the key requirement is that the required size of the PA <5cm.
but some mosfet such as MRF1518T,MRF1517,ne5511279a's typical application
circuits are all using transmision line to macth. because @400MHz,the λ is very long,so
the size cannot be small.

could any one introduce some methods or some transistors or modules to me ?
great thanks!
Best Regards!!

 

[Bouli]

high efficinecy 400mhz amplifer

You can use Philips transistor BLF1043 or BLF1822-10. This are almost the same transistor, but BLF1043 is in smaller case. I used BLF1822-10 to 10W 20-520 MHz power amplifier. I thing this are very good transistors.
You can heve big problems with heat using BLF1043 (because of small heat area), BLF1822-10 is better in this case, but is bigger. You have to choose something.
This transisors are 28V, and you need 12.5V transistor. You can use those transistor with 12,5V power supply, and you should achieve 4W output power with excellent eficiency. But you will get worst intermodulation distortion. So it depends what kind of signal you need to amplifier.
About space on PCboard. You don't have to use microstrip lines and stubs to match transistor. On this freq you can use air springs and SMD capacitors as weel as microstrips. I did this and it works.
Any other questions?

 

[l.j.zheng]

400mhz 5w pa

thanks a lot to Bouli!!

Although the PA's required size is so small, the heat problem is not big. because under the technical requiments, the Tx (PA working)time is quite short, you can take the Tx signal as pulses ! The intermodulation index is quite important (@P1dB-10dB<-40dBc), I wonder the BLF1822-10 could achieve this with the 12.5V supply .and could you give some methods to improve IP3, thanks again!!

 

[carrier]

400mhz power amplifier

I have BLF1822 but data for design is not complete.
you can use Mitubishi RF modulse from the selection guide in the site.
this module RA07M3340 is 7w from 330MHz to 400Mhz

 

[Bouli]

400mhz pa matching

In theory you can achieve the lowest intermodulation distortion workind in class A amplifier, but in class A you can get low efficiency, max 10%, so that causes big problems with power dissipation. BLF1822 and another LDMOS transistor for high power usually are prepared to work in AB class, thanks that you can get high efficiency and good linearity amplifier often even beeter then in class A. So my advice is: You can try to do class A amlifier with high Iq (about 500mA or more), or do AB class amplifier with Iq about 100mA or less. But you have to remember that the best intermodulation preformance you can achieve with precisely taked Iq. The problem is that Iq changes from one part to another, because of different Vth. So this can be a problem in mass production. Solution can be tunning.
About higering IP3, it depends on supply conditons and load and input impedance. You are limited about power supply so only you can do is making good matching network. For the begginign you can try do simple matchiing network as series L (about few nH0) and parallel capacitance (about few pF) in 50Ohm network. it should work.
Any other questions?

 

[l.j.zheng]

blf1822+data sheet

Thanks to carrier,I viewed the RA07H3340M datasheet of Mistubish.I found it good at nealy all aspects except its linearity. For the reason of modulation methods, the linearity is important for my PA.I wonder if you've used it, its linearity
can be achived well by tunning its IdQ and Vg for P1dB=4w output and good IP3 as refered above?

Thanks to Bouli. I'm a new comer for PA,your advice is quite helpful to me,I'll
try them with some mosfets. The Philips BLF1822's datasheet doesn't give the input
and output impence at 400MHz, since it says the Zin and Zout are variable with frequency.I think with BLF1822 I can't get good match at 400MHz,I'll try to find some other mosfets.

Thanks to you both again!

 

[Bouli]

5w pa 400mhz

Don't afraid looking for best Zin and Zl by yourself. It isn't as hard as it look like. It is good to use some trimmer capacitors and air coils to tunning this circuit.

Another thing I heaven't written befor about highering IP3 is Zl for second end third harmonic. It is also important. Not as for the main harmonic, but it is. When you find best Zin and Zl for main harmonic you can try to change Zl for second and third harmonic to find the best performace. Transistor don't "like to see" short or open for higher harmonics. It isn't easy to get optimal Zl for every harmonics, but nobady says that RF PA is easy.

Another thig you can try to do PUSH-PULL aplifier. But this is another story.

 

[molloy]

400mhz pa

If u need a good linearity use only LDMOS. Even if the P1dB of the Mitsubishi device is very high the linearity (as you can see from the datasheet) is not so good and anyway The LDMOS are very good: i got approximatively no fault from these devices. Pay attention in dissipating the devices (the surface of the heatsink and the carrier should be very polished). Maybe the real problem of the design are the dimensions.
regards

 

[springf2000]

rf power mosfet 520mhz 40dbm

this is a reference design:

1)use the ldmos MRF9045
2)use the DC/DC converter, rise the voltage to 28V
2)use the load pull to find the Zin and the Zout
3)then match it
4)then two AH312 as the driver
5)run it

 

[alatriste]

400mhz pa

I think you could use LDMOS from STMicroeletronics.
There are two transistor LDMOS could fit your design:
PD5508 with Vds12.5V and Pout 8 W ;
PD5503 with Vds 12.5 and Pout 3W;
furthemore there is a PD5403 with Vds 7V and Pout 3W.
This transistor have a very little package, suitable for your issues and good linearity;
I think , you could put your circuit in 5cm.
good luck

 

[l.j.zheng]

400mhz 10w pa

Great thanks to all.

To bouli, I wonder if I use a BLF1822 which P1dB is 10W under Vds 28V,could I achive P1dB of at least 4W under Vds 12V? I thinks the output power should
decrease 10log(28²/12²)≈7dB. So the output power under Vds 12V supply may
be 40dBm-7≈33dBm. is it right?

To Molly,I've noticed the question of what you said. The Mitsubish's Deviced aimed
for non-linearity use of FM, in fact , its 7W output isn't P1dB output, Its P1dB
is around 4W, So maybe I should do a test about its linearity.

To springf2000, I think the DC-DC will increase the size of module, could you introduce an excellnet DC-DC wiht high effiency and small size?

To alatriste,I'll see about it ,thanks for your advice.

Best regards!

 

[Bouli]

rf powewr amplifier with pd5503

Hallo
I have never seen the way of counting P1dB like you have written. So I can propose you to use "load line theory" to count output power. You can find some informatuion about it here:
http://scholar.lib.vt.edu/theses/available/etd-07152001-172453/unrestricted/Chap2.PDF
or
**broken link removed**
or
www.google.com

In few words. If you need to get 5W power from 24V supply you need to work with average current about 500mA. You can achive this using 50Ohm load (this data is not counted!!!). So if you decrease supply voltage twice (12V) you need to increase average current twice (1A) to get the same power. Another words you need to decrease twice load resistace (25Ohm). So you can achive the same output power from diferent voltage supply only by changing load resistance. Of course there is a limit of this:maximum voltage supply and maximum curent of used transistor.