reliability of broadband high power amplifier

[Andrew77]

mrf6v1500n

Hello,

I'm designing a high power 100 W (min) , broadband (20-120 and 120-520) amplifier with #2 MRF6V2300N (LDMOS) in push-pull design.
The former amplifier was built with #2 MRF6V1500N (LDMOS) and has reliability issues, i.e. whenever I disconnect the load or the antenna one or both transistors break even with lower levels of power (50 W).
My goal now is to achieve a better reliability by using MRF6V2300N (300 W transistor instead of MRF6V1500N 150 W) biased in class AB @ 24 VDD. Suppose this amplifier is well matched and delivers an efficientyof 45-50% I wonder what else could I do to increase the reliability in case of high VSWR

thanks
Andrea

 

[vfone]

power amplifier site:edaboard.com

You can place an isolator at the output of the PA if the circuit needs to survive at 100W without antenna.
The spec mention 10:1 VSWR which is not impressive, but reducing the power and the Vdd from 50V to 24V will improve VSWR handling, due to less breakdown voltage.

 

[ChaoFan]

mrf6v2300n

So, why did you use it at 24v? from the datasheet, we see it can withstand 10:1 vswr at 300W output, which means it can suvive at 200W with load open. Anyway, reducing the vdd can improve the vswr handling. However, reducing the vdd can cause other issues, like unstable, oscillation at other freq, so that maybe a cause why you burn your parts.

 

[Andrew77]

hello,

I'm using 24 VDD because the amplifers specification, and I can't use a circulator because of the broadband (20 through 500 MHz). The amplifier is a multistage and the final transistors are #2 MRF6V2300NR1 in push-pull design (4:1 balun).
I think you are right, using 24 VDD instead of 50 VDD can cause other frequencies components; I'm just trying it and I see that there are harmonics: the 3rd is only 12 db below the carrier at the frequency carrier of 30 MHz at 100 W output. Do you think that harmonics must be well below the carrier, for example at least -30 dB? Or usually for this kind of amplifier it is normal to have higher harmonics?

Thanks

 

[vfone]

Reducing Vdd will not affect the stability if the input/output matching networks are tuned properly (as should be).

The impedances Z_source and Z_load from page 11 of the spec are changing with Vdd.
If you match in concordance the new impedances (at Vdd=24V) the device will work properly, with better VSWR handling compared to Vdd=50V.

Harmonics level is also dependable by the Idq bias current. More DC bias current, makes the amplifier more linear and generating less harmonics.
Intermodulation Distortion vs Bias (and indirectly Harmonics vs Bias) can check on the datasheet graph (page 5).

Harmonics at this power (without any LPF at the output) shall be better than -20dBc.

 

[ChaoFan]

I guess your matching circurt configuration is well compressing the 2nd harmonic. However your 3rd harmonic is much worse than 2nd harmonic. We can caculate the P1dB of that part at VDD=24v. Then we find it's not easy to have a better than -20dbc 3rd harmonic at 100W across the whole band, especially at the low freq. Because to maintain the broadband, you have to sacrifise p1dB and linearity. Anyway, I'm not sure you already tuned the parts at the best matching point. you have to spend some time on the tuning to see if there is any possibility to have a good 3rd harmonic. Good luck!

 

[Andrew77]

Up to date across the band 20-60 MHz the 2nd harmonic is -25 dBc about, the 3rd is about -12 dBc; better performances are at the end of the band.
Anyway I'm still trying to get better

 

[molloy]

I believe you cant do anything but maybe some hardware protections against the reflected power. Usually i found the nxp counterparts more reliable even if the 50V freescale are very good devices (at least the UHF parts) But be careful and test it with output filter and hardware protection. Keep in mind that these devices have a high gain, so it is very difficult to stabilize them in all load conditions