Call Search
     

New to Ham Radio?
My Profile

Community
Articles
Forums
News
Reviews
Friends Remembered
Strays
Survey Question

Operating
Contesting
DX Cluster Spots
Propagation

Resources
Calendar
Classifieds
Ham Exams
Ham Links
List Archives
News Articles
Product Reviews
QSL Managers

Site Info
eHam Help (FAQ)
Support the site
The eHam Team
Advertising Info
Vision Statement
About eHam.net

   Home   Help Search  
Pages: [1] 2 Next   Go Down
  Print  
Author Topic: RF power LDMOS transistors  (Read 9205 times)
KA4POL
Member

Posts: 1910




Ignore
« on: November 02, 2011, 02:29:39 AM »

There are some new RF transistors on the market for all those QRO projects up to 1.25 kW from 1.8 to 600 MHz. And for an ad about their ruggedness see: http://contact.freescale.com/content/RuggedVideo?t=el
Does anyone have any experience yet?
Logged
M0HCN
Member

Posts: 473




Ignore
« Reply #1 on: November 02, 2011, 04:50:47 AM »

I know of a couple of people experimenting, but the elephant in the room is, as ever, heat.
You need to read those data sheets carefully, and with your cynicism turned well up to get the real picture.

Absent water cooling you will have a real fight to get the heat away, to the point that a cluster of lower power devices and splitter/combiner networks is actually probably easier. 

That arc demo is pretty, but is close to the amplifier, so will appear as a mostly resistive mismatch, I would be interested to see the same thing on the far end of a hundred feet or so of low loss coax (The water cooled heatsink helps here as well).

Personally I would consider these to be 600 - 800W devices in the interests of reliability and being able to remove the heat.

The 1.2KW rating makes sense in something like an MRI machine, where high power low duty cycle pulses are what matters.

Regards, Dan.
Logged
KA4POL
Member

Posts: 1910




Ignore
« Reply #2 on: November 02, 2011, 05:40:16 AM »

I share your views. I downloaded the data sheet a while ago.

On this demo setup I am not sure what they did. It seems to be difficult to get the information. They claim VSWR<65:1 as a load mismatch. The total device dissipation is 1333 W which says it all. Would make a nice heater for the shack.
Logged
W8JI
Member

Posts: 9304


WWW

Ignore
« Reply #3 on: November 03, 2011, 05:51:27 AM »

I don't know why people can't or don't read data sheets.

The 1200 watt test is at 1200 watts PULSE with a 100uS width and 20% duty. The heat is equivalent to the heat from a 1200/5 = 240 watt output amplifier in normal modes.

I can run the same tests on a PAIR of MRF150 transistors and get almost 1200 watts output into the same dramatic short and open test loads.

All that stuff is really just creative misleading drama to get people's attention.

If we look we will note there are NO published SSB,  Morse code CW, or carrier ratings for that device. 
Logged
W6RMK
Member

Posts: 649




Ignore
« Reply #4 on: November 03, 2011, 06:50:11 AM »

I don't know why people can't or don't read data sheets.

If we look we will note there are NO published SSB,  Morse code CW, or carrier ratings for that device. 

Given the incredibly small market for such applications, I doubt anyone publishes Morse Code CW ratings in a datasheet for any semiconductor component, unless they're selling specifically to hams.<grin>

(although I confess I'd love to see an ad in the style of the old DAK ads. "These THUNDER POWER LDMOS devices, made custom for us, are rated at a FULL 50 WORDS PER MINUTE")
Logged
WB6BYU
Member

Posts: 13020




Ignore
« Reply #5 on: November 08, 2011, 12:38:35 PM »

But the on time is still critical for determining heat build-up:  clearly an hour on and 40 hours off
puts more thermal stress on the part than 10us on and 400us off, because the shorter the on
time, the more that the maximum heat build-up at the part looks like the average dissipation
rather than the peak.  The effect depends on the thermal mass of the junction, which can be
pretty small.  I can see a difference between a 200us and 400us burst length with some
wireless LAN parts using a constant 20% duty cycle, though the effects are small.
Logged
W8JI
Member

Posts: 9304


WWW

Ignore
« Reply #6 on: November 09, 2011, 05:32:27 PM »

By CW, I meant more traditional duty cycles useful for amateur communications, not pulsed "carrier" power. Most Hams use SSB and CW and carrier modes, very few use pulse transmission.

:-)   
Logged
KA4POL
Member

Posts: 1910




Ignore
« Reply #7 on: January 14, 2012, 09:00:59 AM »

Well, not exactly homebrew, but the Italians have read the data and put the transistors to work: http://www.italab.it/prodotti_uk.php?cat=4&scat=2
Logged
KB1GMX
Member

Posts: 711




Ignore
« Reply #8 on: January 14, 2012, 04:38:28 PM »

Look at the MRFE6VP61K25H..

1250W CW with an efficiency of about 70%  meaning 30% of the _input_ power (1785W) is heat.
If you ca get rid of the 530W of heat your in business..  However there are cases of mismatch that
will increase that to near full output power and that takes a lot of metal to absorb it.  The amp they
show is built on a copper block looks to be about an inch or more thick.

What they demonstrated is typically bond wire failure from the FET die to the leads, At around 40A
you need seriously strong conductors to carry the current or they just melt like fuses.

FYI: pulse rating are fairly close to typical SSB when peak to average of 30%.  Like tubes, running semiconductors at or above ratings is a game of dealing with heat.  Sweep tubes would melt if
full bore tuned into a bad load as well, commercial tubes had bigger plates and all to handle the
extra heat.  Semis are the same game.

Now there are other little things to observe.  Working with a BLF278 amp at 2M I've fried output caps
at 350w, didn't kill the transistor, but finding ATC caps that didn't fry at the RF current presented by
that power was a challenge.   At 350W out the heat load is 300W presented in an area an inch square
so using .25" copper plate under the device to spread the heat out to the aluminum heatsink
is critical.   

So building a 1250W  amp is a matter of managing a 2000W energy source that can be RF power
or heat.  Also at 50V that 2000W is 40A, all the conductors and such have to stand a lot of DC
amps and RF amps that takes a lot of copper.

FYI: it's how people can build linear amps using the IRF510 a 23W package with rotten thermal
characteristics to get two tone 40-60W PushPull without melting (at 28V).  It's not a great device
but it proves with a little care (and a big thick heat sink) taken power can be had.  Same applies
for the bigger devices.

To do high power for ham use the devices have to have both the peak power and current rating,
sufficient dissipation, and excellent thermal transfer characteristics.  The latter is most important.
The rest is care in design especially with output networks as some get very high loss at the low or
high edges of the at 10M or 160M.  That impacts efficiency and contributes to the heat issues.

I'd look at a pair of those monsters running at a conservative 750W (lower than max voltage and
current) and combining them for 1500w and they would be far harder to destroy.  This has been done
with lower power parts like the MRF150.

Allison


Logged
M0HCN
Member

Posts: 473




Ignore
« Reply #9 on: January 15, 2012, 03:00:15 PM »

Don't forget that that 70% drain efficiency is probably running the device well into saturation and is far from the worst case for a mode like SSB.

Linear class AB (which is where most amps operate for SSB service) will have a drain efficiency more like 50% at 100% output, dropping rapidly with the envelope power.

600 - 700W per device might be doable for SSB service, but personally I think that pairs of SD2933 or 43  and a couple of power splitter/combiner networks looks like a better bet (Easier to cool, and should work out about the same price).
4 modules, doing say 400 - 500W each would give a nicely conservative design for 1.5KW output, with graceful fallback if a module faulted.

Apart from anything else, a lower power stage is going to be easier to get to work well at both top band and 6M (Physically smaller magnetics, so electrically shorter windings at 6M), quite apart from making the development blowups less spectacular.
Adding 6M to a all HF band amp is actually really a bridge too far in terms of the engineering compromises, designing a drain match for 160 - 10M is very much easier then 160 - 6M.

Got to love watching a chip cap unsolder itself from the board when you get the current distribution wrong......

IIRC AVX have some ceramic parts similar to the ATC products that may be worth a look in larger sizes.

Regards, Dan.
Logged
W8JI
Member

Posts: 9304


WWW

Ignore
« Reply #10 on: January 15, 2012, 03:08:19 PM »

Well, not exactly homebrew, but the Italians have read the data and put the transistors to work: http://www.italab.it/prodotti_uk.php?cat=4&scat=2

That appears to be FM, not linear service.
Logged
WB8VLC
Member

Posts: 116




Ignore
« Reply #11 on: January 15, 2012, 03:42:23 PM »

Some of the Freescale hams down in Tempe/Phoenix have the 600w and 1 Kw parts operating from 1.8 to 54 MHz along with a 1Kw application board running on 2 meter and 220 MHz SSB and FM.

There are a couple of HF amp dev kits that were made available for either the 300, 600 and 1 Kw parts in this family to present and ex Motorola/freescale employees to evaluate over this frequency range.

I have a 300/600 watt dev kit which works very well over the entire 1.8 to 54 MHz range with the proper heat sink of course, %eff is around 60 to 65 % over such a wide band.

The only reason I choose to evaluate the 300/600 watt parts and not the 1kw device was due to my 50 volt power supply limitations, other wise I'd have gone with the 1 kw dev kit.

These amps do get hot running FM mode but so far no dead devices, as with all LDMOS parts the major killer is overdrive and these devices can take a bit more than the typical 3 dB overdrive before they will fail; however the short and open circuit operation with proper drive power is pretty amazing with these parts.

I also have one of the 300/600 watt kits for 144 to 225 MHz use, it's only a small change to use either device in this circuit.

All in all these new parts are considerably more rugged than the 2nd and 3rd generation, HV3 and HV4, LDMOS devices that I used to design with when I was at Motorola back in the late 1990's, and yes Freescale still consists of some of the original design engineers/team members that designed the first LDMOS devices back in the 1990's.

And a good number of them are hams.


Mike
Logged
N3QE
Member

Posts: 2078




Ignore
« Reply #12 on: January 16, 2012, 05:49:04 AM »

Unless a ham station is in a 100% duty cycle mode like say long RTTY... would the pulse rating of these transistors be comparable in cycle to traditional ICAS ratings of tubes?

I mean, I know when I'm running in a contest in CW, that my actual duty key down duty cycle is at most 20 or 25%.

Or is the time scale of typical CW key down too long for the assumptions of pulse operation? (Not the thermal mass of a 3-500Z!)
Logged
WB8VLC
Member

Posts: 116




Ignore
« Reply #13 on: January 21, 2012, 10:25:26 AM »

I use the 300 watt part on FM without any problems, remember as W8JI mentions heat is the 2nd biggest killer while overdrive is number 1.

Some parts are more tolerant of heat while most can't survive a 3 dB overdrive and some not even 1.5 dB to 2 dB.

When Warren B. at freescale designed these new rugged HV6 LDMOS parts his goal was to make them as rugged as possible but with the understanding that you needed to adhere to proper drive level and heatsinking of the parts, then you would be able to survive a 65:1 vswr test running CW power.

But increase drive by 2 to 3 dB and run the same test at a 5:1 or 10:1 and the part could start complaining.

When we were designing the MRF275 part back in 1997 I had the pleasure of testing several 6 meter amplifiers operating that were mounted to the heatsink of an old 70 watt motorola HT land mobile amplifier, if I set Vdd at 24 to 26 volts,(they are 28 volt parts) and the drive around 1 dB under the maximum rating the part survived FM operation at ~ 160 watts output into a 2.5 1 antenna.

If I increased drive over 1 dB of max suggested and ran the amp into the same antenna I would start to get failures.

 I saw the same thing with the engineering samples which were used in the Yaesu FT100 radio lines except they were almost equal for input overdrive and heat failures.

I still have a stick full of engineering devices used on the FT100/D radios and there ruggedness is not adequate enough to obtain decent power out that I am presently using a custom made Point 9 technologies MOSFET in my FT100D V/U power amp.

Best recommendation is to #1:  run Vdd slightly under the rating (46 volts works good with these newer devices rated for 50 Volts), #2: keep the input Drive at or below the recommended rating and use an input overdrive shutoff circuit, and #3 supply adequate heatsinking with a thermal shutoff and none of these ridiculous relay controlled shutoff circuits will work, they must be fast switching circuits.


Mike
Logged
KB1GMX
Member

Posts: 711




Ignore
« Reply #14 on: January 24, 2012, 05:25:08 AM »

The efficiency of class AB push pull LDMOS amps tends to be quite high.  However the measurement is at
full(not saturated) power and running at lower power it can be quite low due to standing bias current.

I just finished a BLF278 amp for 2M and it's about 72% at 330W out and 50V.  For safety it will be
running at 48V maybe less.  to give a perspective saturated power the devices do 400W but that's at
1db compression but the efficiency is near 77%.  That same amp at 100W output is still good but you
start to appreciate that .25A of standing bias and all the support for the amp are part of the total
efficiency.  Also that is with a good 50 ohm load and ideal conditions.  Its important to measure it as
it also indicates that the output matching is working correctly.  I've seen amps where the efficiency
was poor and the output matching was running very hot.

In the end it's about heat and managing it.  Also it's components that will work at 50V DC and tens of
Amps DC and possibly more at RF.


Allison



Logged
Pages: [1] 2 Next   Go Up
  Print  
 
Jump to:  

Powered by MySQL Powered by PHP Powered by SMF 1.1.11 | SMF © 2006-2009, Simple Machines LLC Valid XHTML 1.0! Valid CSS!