Ben

Both seem to use the same output device

The device is  designed for wideband data modes at 400mW. This requires high linearity, hence the typical single tone P1db of 34dBm (min is +30)

Efficiency at 26dBm is not great, probably better in CW mode but you've got to get quite a lot of heat out of a small area.

The data sheet from Skyworks does not give any thermal info and it may be that key down CW at 2W will overheat the junction.

In amateur use the TX duty cycle helps us and   it may be able to run at 2W  out  for typical QSOs, if on a good heat sink.  The 40dB gain module  will need to be bolted to one. We also don't often have high ambient temperature that may push the limits

So, with the usual caveats, if the devices are genuine and full spec and the board has been properly designed (thermally)  2W PEP is not  impossible without damage.

It is interesting that the second one  states 1W for the same device, perhaps being a little more conservative.

None of the vendors are saying how long the device will last at  the P1dB level key down...

Even at 1W these represent reasonable value for money... if they work as advertised.  If they dont, send em back!

Mark GM4ISM

5G-6GHz One-Way Microwave RF Power Amplifier Module 40DB SBB5089+SE5004

( 114749899192 )

SE5004 1W Microwave Power Amplifier RF Power Amplifier 5.15GHz-5.85GHz 30DBm

( 114556927546 )

Did anyone try these? Both use the SE5004 which seems rated to 34dbm max, 

I'm guessing the first unit rated at 2W is 2 Chinese Watts. The second unit might just do it.

Probably get 1W out of first but it has more gain so smaller i/p needed I guess.

 Ben.

Virus-free. www.avg.com

Ben

160 C junction temperature  is pretty standard for this technology, the data that is missing is the thermal resistance, junction to case measured in C/W, and power added efficiency.  Without it you cannot calculate the permitted dissipation of the device and the expected output power WRT the junction temperature. Yes it can manage a P1dB of 2W  but maybe only with a 20% duty cycle for thermal considerations.

If the value  of thermal resistance is say 15C/W, then if you dissipate 5W in the junction, the junction rises to 75 degrees above the case,  that would be manageable as the heatsink would have to keep the case at a max of 85C

if the value is 30 then the same 5W dissipation would put the junction at 170C even if you kept the case anchored at 20C with a fantastic heat sink.

we can guess from the figures.. at 400mW the data sheets states that the max case temperature is 85C  and that this is at 5V and 600mA  That is 3W in, 400mW out   2.6W dissipation  ie about 14% efficiency

If you assume that  at these values the case temperature is allowed to approach the max of 85C the the junction will also be OK below 160 C, some 75C above case.  75/2.6 is  28.8 C/W thermal resistance.

Lets say the efficiency reaches 30% at the P1dB point of 2w out,  that would require a junction dissipation of 4.4W

With that thermal resistance which is a constant, the junction would be 127C above the case. Thus the case must not rise above 33C

Thats quite low, requiring a very good heatsink and likely a low duty cycle

These calculations make a few assumptions and may be quite a bit off, I may even have  made a slip on the calculator  but they look about right to me and support the supposition that  1W or maybe a bit more output, with good design and typical amateur duty cycles, is probably feasible.

There are a lot of devices out there that look good for really high output power at the P1dB point but being designed (thermally) for a lower mean power  means they just cant realise that P1dB for a useful time.  Not all data sheets tell you the whole story :(

Mark GM4ISM

The data sheet from Skyworks does not give any thermal info and it may be that key down CW at 2W will overheat the junction.

160deg C Mark.

. if they work as advertised.  If they dont, send em back!

Mark GM4ISM

I doubt worth trying to send back to China. Still you get either a nice little project box or a nice heatsink depending on which you buy, hi. 

Ben,

Virus-free. www.avg.com

“ The SKY66293-21 is an older part.  I’d recommend the SKY66318-11 for this frequency range, however, the target output power of our small cell Pas is +27dBm ~ +28dBm with back off of 8dB.

 For all the SKY662xx series, we ran all of our thermal analysis at case temp = 85C or 100C (depending on the part number) at the rated target output power of +27dBm ~ +28dBm.  Since the target application is LTE modulated or 5G NR, we don’t expect the PAs to continually transmit at output powers higher than +27dBm ~ +28dBm with respect to thermals "

So as Mark says the datasheet doesn't tell you the whole story.

Dave G8KHU

Found one on ebay, £26 from China or £30 from Amazon. 

I blew up two of those, so treat them *very* gently. They usually have  a pair of the same 5mm square devices that everything uses.  In the end, I fixed a slab of aluminium over the top of the chips to press them down to a bigger heatsink and help remove more heat from the package. I also scrapped the noisy switchmode regulator and used a two-stage linear regulator

Neil G4DBN