replacement FET's for LARCAN power amps

[vk5zlx]

Hello to you all,
I see from the minuteman web site that MRF-151G's are a suggested
replacement for the SRF series of FET's that are fitted to the
LARCAN's, has any one actually tried them in a LARCAN power amp for
50MHz?

I ask this as 3 out of the 4 devices in my LARCAN module are faulty
(draw no idle current) and I would need to replace all 4 with a
similar device to maintain some sort of matching, so I am looking for
any real world experiences with substitute FET's.

Regards,
Peter, vk5pj

[Bob DeMattia]

Hi Peter,

The MRF-151G is identical to what's in the amp now.

The transistors in the amp are specially marked with

a Larcan part #.  This was told to us by a Larcan

technician.  We also independently confirmed this

with the current part supplier.

-Bob

"The board uses four SRF-3943-2 FET's for 1kW video/SSB or 600W CW/FM. According to Larcan, and confirmed by Freescale Semiconductor who now owns this Motorola business, the '3943 is functionally equivalent to a MRF151G. M/A-COM has the build rights and is currently producing these. This device is rated at up to 175MHz, 300W (150W per FET). The MRF151G is available from rfparts.com for $110 - $120. If you really want the original SRF part marking, you can buy one from Larcan for $200. The datasheet for the MRF151G is in the links section above. Eight 10A fuses protect each half of the four dual-FETs.  The fuse removal will also allow for optimal  biasing ofeach transistor for multi-channel wideband TV service. That adjustment is rarely required for ham operation due to the 5 KHz maximum modulation bandwidth typically used."

[Peter Sumner]

Thanks Bob,
I had read that onthe web site but wanted to make sure before I dive
off and try and buy some.

Peter, vk5pj

[Dave Olean]

Hi,
I remember reading that someone was experiencing a significant roll off
in gain as the frequency is lowered to 50.1 MHz for SSB activity. I just
converted my LARCAN amp for 50 MHz. I had a LO-HI unit that needed the LO-LO
mods. I checked the amp beforehand and saw 23.5 dB of gain by simply reading
the power on my Bird dummy load wattmeter, and measuring the input of my
signal generator used as a driver. It was reasonably flat across the high
band with a peajk at 92 MHz before it dropped off. After the conversion, I
am seeing barely 20 dB, but the gain up around 77 MHz is way up over 23 dB.
I tried running the amp with highe4r power. I used a MACOM broadband 4 watt
instrument amplifier and drove it with my HP8640B. I verified that the gain
roll off was there even at the higher drive levels. Some might not notice
this if they have no means to sweep the amplifier, and use a 50 MHz ham rig
instead. The problem I see is that the efficiency is not so great at 50.1
MHz. I can get plenty of output at 50 MHz. It's just that the gain and
efficiency is low.
I next started to tweak the amp. I played around with the input stage,
but that seemed to be good at 50 MHz. The output tuning was not good for 50
MHz. I ended up tweaking the value of C16 from 51 pf to 72-75 pf and saw the
gain go up. C13 seemed ok. I have not tried C14 and 15 yet. That may yield
an improvement. I also noted that no changes are made to the output
combiner. I wonder if that could be swept easily to see what is going on
there.
For the record, here are the standard changes I made in the conversion:
1. Remove the shorts/shunts on the output lines.
2 C1,2 to 15 pf
3. C3 to 63 pf
4. C4 to 200 pf
5. C13 to 82 pf
6. C14 & 15 to 180 pf
7. C16 to 51 pf
8. Remove C18 & 19
I ran a sweep with a nominal 50 watts output. Here is what I saw
FREQ PWR
50 22 watts
55 32 watts
60 35 watts
65 58 watts
70 56 watts
75 60 watts
80 61 watts
85 45 watts
Changed C16 from 51 to 75 pf
FREQ PWR
50 27 watts
55 32 watts
60 32 watts
65 46 watts
70 56 watts
75 67 watts
80 65 watts
85 40 watts

Increasing C16 really does improve things at 50.0 MHz. I just have to find
the other tweaks! I will keep playing with this as time permits. I wonder if
anyone has already solved this problem?

Dave K1WHS

[Bob DeMattia]

An unmodified LO-LO amp is designed to cover channels 2,3, and 4/54-72 MHz.

A LO-HI amp is designed to cover channels 5 & 6/76-88 MHz.  The mods outlined

in the article covert a LO-HI to a LO-LO, and thus rely on the assumption that

the amp is broadband enough to go down another 4 MHz.

So it makes sense that you could further adjust the component values to optimize

the LO-LO range down to 50 MHz.

Your results confirm what is stated in the article: the amps get around 23dB

at their intended frequency range, but go down to around 20dB at 6m.  If you

use an unmodified LO-HI amp, you'll see 15dB.

-Bob

To: larcan-6m-amplifiers@googlegroups.com

[Dave Olean]

Hello Bob and the LARCAN group,

    I think I made some headway with the converted LARCAN amplifier on 50.1 MHz.  I am not finished but am going in the right direction. I have made it much better than it originally was. I  looked at the output combiner by lifting one PA output lead at the 50 ohm point and looked through the two power combiners. I used a network analyzer. What I saw at 50 MHz was not too great. The power split was not 6 dB as it should be, and the return loss at that point was about 11 dB which is getting close to 2:1. I tried messing with the inductor L1 which is about 0.39uh. I tried larger and smaller coils here but saw little improvement. I did add chip caps at C4 and C5  (62 pf) and C1 and C8 (12 pf)  and improved the performance of the splitters at 50 MHz. Now I see 6 dB combiner loss and a return loss of 16 dB at 50 MHz, This is a big improvement!  (I started with 3.8 dB split loss and 11 dB return loss)  I misread the values of the parts and suspect that the added 62 pf is too much for C4 and C5.  I left it alone as the network analyzer seemed to like what I had!  So with these changes, I tried to spark it up and do a sweep. Here are the results:

50 MHz    21 watts output

55 MHz    25 watts output

60 MHz    25 watts output

65 MHz    42 watts output

70 MHz   45 watts output

75 MHz   59 watts output

80 MHz   70 watts output (This is the sharp peak that used to be at 92 MHz.)

I ran 4 watts into the amplifier, and saw 540 watts output on a CW carrier. Drain voltage and current was 44.2 volts at 34 amps for 21.3 dB gain and 35.94% efficiency. I need to get my 50 volt power supply running as a few more volts on the drain should help the efficiency I would guess.

    When I get this all working well, I will write it all up so it can be posted.  I did make BEFORE and AFTER plots of the combiner bandpass.  The combiner is now better than 16 dB return loss from 48 to 74 MHz. with my modifications. The original combiner was –14 dB (1.5:1) from 55 to 90 MHz.  I think this solid state amplifier is going to work out quite well.

 

73

Dave K1WHS

[Dave Olean]

Hello Bob,

    I got my 50 volt power supply running and hooked it up to the amplifier on 50.000 MHz.  I was using 44 volts for my testing.  With 4 watts input I saw 640 watts out key down. That is 22 dB gain, so I guess the combiner is working much better.  Yahoo! Now I need to do some metal work to button it all up, and rig up a current meter so I can monitor efficiency etc.

 

73

Dave K1WHS

 

From: Bob DeMattia

Sent: Monday, January 16, 2012 2:02 PM

To: larcan-6m-...@googlegroups.com

Subject: Re: 50 MHz Conversion Freq Response

 

[Peter Sumner]

Bob and Dave,
I think where the extra 3db goes to is in the output combiner, I
have done some sweeps of both modules I have and can easily see a 3db
roll off from 55 down to 50 MHz.

I have tried adding two turns of inductance on the input of each leg
of the combiner as I had guessed the strip line inductors were too
short and causing the roll off but that did not help.

So far I am struggling to get any sense out of the larcan modules but
with one of them now lacking capacitors and the second one with mostly
sick FET's I am between a rock and a hard place until the extra 180PF
caps show up from ATC in the usa (takes over a month to get caps from
them).

Dave will be interested to keep in touch, once I have a module with
all good FET's and the right caps I plan to see if I can lick this
efficiency problem.

In VK we have a few LARCAN tx's still in use on our CH0 (46.250 Video
& 51.750 Sound) so I am in contact with the support team who manage
one of those TX's to see if I can get the details of how thay have
been made to work at 46MHz :-) but they do not visit the site very
offten and all the books are at site.

Regards,
Peter, vk5pj

[Dave Olean]

Hello Peter,
Well my tweaking has produced 2 dB of extra gain at 50.000 MHz. With my
50 volt ASTEC PS I am seeing 640 watts out with 4 watts of drive. I have a
few more tweaks, but I am happy with my results so far. I will post exactly
what I did when I get a chance. So far I added 62 pf to the 47 pf capacitors
in the final splitter (C4, C5), and I added 12 pf to the first splitter,
(C1, C8). I played with the inductor, but saw little change. I also changed
C16 in the output side of the FET from 51 pf to 75 pf. I made some plots of
the combiner and did see major problems there as far as 50 MHz goes. My amp
still has extra gain up higher in frequency after all my tweaks.

73
Dave K1WHS

-----Original Message-----
From: Peter Sumner

Sent: Wednesday, January 18, 2012 7:18 AM
To: larcan-6m-...@googlegroups.com

Subject: Re: 50 MHz Conversion Freq Response

[Bob DeMattia]

Hi Dave:
I'd be interested in knowing the efficiency of the amp at 50, 52, 54 MHz,

and out of curiosity, up at the 70+ MHz power peak.

-Bob

On Wed, Jan 18, 2012 at 8:17 AM, Dave Olean <k1...@metrocast.net> wrote:

Hello Peter,
  Well my tweaking has produced 2 dB of extra gain at 50.000 MHz. With my 50 volt  ASTEC PS I am seeing 640 watts out with 4 watts of drive. I have a few more tweaks, but I am happy with my results so far.  I will post exactly what I did when I get a chance. So far I added 62 pf to the 47 pf capacitors in the final splitter (C4, C5), and I added 12 pf to the first splitter, (C1, C8). I played with the inductor, but saw little change.  I also changed C16 in the output side of the FET from 51 pf to 75 pf.  I made some plots of the combiner and did see major problems there as far as 50 MHz goes. My amp still has extra gain up higher in frequency after all my tweaks.

73

Dave K1WHS

-----Original Message----- From: Peter Sumner

Sent: Wednesday, January 18, 2012 7:18 AM

To: larcan-6m-amplifiers@googlegroups.com

[Dave Olean]

Well Bob, I tried to check things for you at 70 MHz and with my modifications, I fried one of the transistors.   At 50 to 54 it worked just fine. I saw these results with 4 watt of drive:

50 MHz   640 watts    43 amps 51 volts  36.9%

52 MHz   630 watts   36.5 amps 51 volts  33.9%

54 MHz  600 watts    38 amps   51 volts   31%

70 MHz   Blue flame on one FET  Blown fuse.

 

So now I have a blown amplifier. I hate it when that happens. I guess I need to scrounge up another device.  Any junkers laying around?   I suspect it was the change of 51 pf to 75 pf in the output stage. (C16)  That must have been very bad at 70 MHz.  I think the combiner will work OK at 70, but I have no desire to run anything there.  I built a metal frame for sealing up the components on the bottom. I need to cut out a bottom plate with some 0.063” aluminum and then I am done except for the keying line.

 

73

Dave K1WHS

[Bob DeMattia]

Oops.  

We used all the spares we had fixing amps before we sold them, so we don't

have anymore.  Somebody has them available on eBay right now for $98 (inc shipping).

http://www.ebay.com/itm/1PCS-MRF151G-MRF-151-RF-Power-Amplifier-Transistor-m-/320815038315?pt=LH_DefaultDomain_0&hash=item4ab210ff6b#ht_3355wt_1164

It looks like you improved the coupling at the lower end.  I only mentioned

the 70 MHz because one of your earlier posts indicated a peak up there.  Maybe

this was before you changed the components.  It looks like you've got it

optimized at the low end now.  If you'd like me to post your write up on our

website, let me know.

On

[Dave Olean]

Bob,

Ok on the LDMOS. I’ll get one ordered. I think I will leave well enough alone, and stick to the 50 MHz end of the scale. It is working a lot better down at the low end now.

 

Dave K1WHS

 

From: Bob DeMattia

Sent: Wednesday, January 18, 2012 2:28 PM

To: larcan-6m-...@googlegroups.com

Subject: Re: 50 MHz Conversion Freq Response

 

[Dave Calhoun]

Hi Bob,

 

    I'm currently using a TS-2000 for an exciter so all I have is 6 meters. The gain drops off at the bottom of the band but I have not checked efficiency. That seems difficult because it's easy to exceed 1 kw by running into gain compression.

 

What I will do is rig the thing up for a 2 tone ssb test and see what it looks like on the S.A.

Then I can figure efficiency at a given grunge level.

 

I can also light up some driver modules and use my HP 8640B to check the modules properly. Time would have to permit.

 

Here's what one amp is doing:

7 watts drive steady carrier as read on the TS-2000

 

53.999 MHZ  996 Watts

53.0     MHZ  956 Watts

52.5     MHZ  937 Watts

52.0                921

51.5                903

51.0                892

50.5                868

50.0                840

 

Here's pictures of my test.

73,

Dave W2VW.

[Lance Collister, W7GJ]

Hi Dave,

I have been sending out a number of Harris Platinum I Channel 2 TV amps to folks, and
put a lot of info on my web page that may be of some interest to people trying to
combine amplifiers:

http://www.bigskyspaces.com/w7gj/HarrisAmps.htm

The combiner and splitter shown on my web page are working quite well with the pair
of Harris amps here. I realize the rest of it is really irrelevant to the challenges
you have been facing with the Larcan amps, because these amps were basically made to
run on 6m to begin with and need no modification to any components. With these amps,
you just need to add cooling fans, coaxial relays, a 50 VDC power supply and a low
pass filter.

GL and VY 73, Lance

On 1/17/2012 10:51 PM, Dave Olean wrote:
> Hello Bob,
> I got my 50 volt power supply running and hooked it up to the amplifier on 50.000
> MHz. I was using 44 volts for my testing. With 4 watts input I saw 640 watts out key
> down. That is 22 dB gain, so I guess the combiner is working much better. Yahoo! Now
> I need to do some metal work to button it all up, and rig up a current meter so I can
> monitor efficiency etc.
> 73
> Dave K1WHS

> *From:* Bob DeMattia <mailto:b...@demattia.net>
> *Sent:* Monday, January 16, 2012 2:02 PM
> *To:* larcan-6m-...@googlegroups.com
> <mailto:larcan-6m-...@googlegroups.com>
> *Subject:* Re: 50 MHz Conversion Freq Response

> To: larcan-6m-amplifiers@__googlegroups.com
> <mailto:larcan-6m-...@googlegroups.com>

> Subject: Re: replacement FET's for LARCAN power amps
>
> Thanks Bob,
> I had read that onthe web site but wanted to make sure before I dive
> off and try and buy some.
>
> Peter, vk5pj
>
> On Mon, Jan 9, 2012 at 12:57 PM, Bob DeMattia <b...@demattia.net
> <mailto:b...@demattia.net>> wrote:
>
> Hi Peter,
> The MRF-151G is identical to what's in the amp now.
> The transistors in the amp are specially marked with
> a Larcan part #. This was told to us by a Larcan
> technician. We also independently confirmed this
> with the current part supplier.
>
> -Bob
>
> "The board uses four SRF-3943-2 FET's for 1kW video/SSB or 600W CW/FM.
> According to Larcan, and confirmed by Freescale Semiconductor who now owns
> this Motorola business, the '3943 is functionally equivalent to a MRF151G.
> M/A-COM has the build rights and is currently producing these. This device
> is rated at up to 175MHz, 300W (150W per FET). The MRF151G is available from

> rfparts.com <http://rfparts.com> for $110 - $120. If you really want the

> original SRF part
> marking, you can buy one from Larcan for $200. The datasheet for the MRF151G
> is in the links section above. Eight 10A fuses protect each half of the four
> dual-FETs. The fuse removal will also allow for optimal biasing
> ofeach transistor for multi-channel wideband TV service. That adjustment is
> rarely required for ham operation due to the 5 KHz maximum modulation
> bandwidth typically used."
>
>
>
>
>
>

--
Lance Collister, W7GJ
(ex WA3GPL, WA1JXN, WA1JXN/C6A, ZF2OC/ZF8, E51SIX, 3D2LR, 5W0GJ)
P.O. Box 73
Frenchtown, MT 59834-0073
USA
TEL: (406) 626-5728
QTH: DN27ub
URL: http://www.bigskyspaces.com/w7gj
Windows Messenger: W7...@hotmail.com
Skype: lanceW7GJ
2m DXCC #11/6m DXCC #815

Interested in 6m EME? Ask me about subscribing to the Magic Band EME
email group, or just fill in the request box at the bottom of my web
page (above)!

[Dave Olean]

Lance,
Where did you find the amps and how much are you selling them for? I
guess that is a good way to earn some DX-pedition funds etc. The Larcan
amps are pretty good. They will do 1000 watts PEP, but the size is a bit
awkward, long and flat. I will need to build a filter for the Larcan. My 5
pole design did not really kill the second harmonic too much, maybe 20 dB.
I'll have to look at this Larcan when I am done, and see what the second
harmonic is. Making filters is fun! I want to get some solid state
instant on amps so I do not need the diesel generator. I am making a 600
watt 144 amp as well. I would be interested in a Harris amp as I believe
they are much more efficient than the Larcan units. I am seeing 37%.

Dave K1WHS
I ran into an old girl friend from Jr High and High school recently. She
also went to Middlebury 1963-67.

[Lance Collister, W7GJ]

Hi Dave,

I started collecting these great Harris Platinum I Channel 2 TV XMTRS in 2010, so I
could send them to "barefoot" 6m operators in rare DXCC overseas. Since I started, I
have acquired quite a few more and have been making them available to NA hams too.
The nice thing about the Harris amps is that they are already enclosed, are very well
protected, and don't need any conversions or modifications. You do have to add
cooling fans, external relays, a power supply and a low pass filter. Also, without
the proprietary Harris connectors, it is necessary to rig up a way to connect the RF,
DC and PTT line to the back of the amp. There already are some great technical links
discussing the amps, and I added them to the web page I developed to assist people in
getting these units on the air on 6m:

http://www.bigskyspaces.com/w7gj/HarrisAmps.htm

My goal was to make these quality amps very affordable, especially for overseas hams.
I have shipped out about 20 of these units so far, but still have not covered my
costs yet. A number of them have been shipped in pairs to people interested in
combining them to create a really rugged 6m amp, and I try to show how to do this on
my website by explaining how I did it here. The combiner really works slick!

You will notice that I incorporated your famous low pass filter into the combiner,
but I also run the output through the other two filters that I had already in line
for my tube amp. Your contribution to 6m with that great filter is appreciated by
many of us 6m operators ;-) If you need more attenuation, one thing you might want
to consider is adding an open quarter wave stub at 100.300 to the output of your amp
to further attenuate the second harmonic. I have had good success using such stubs
on my tube amps, but I don't know how the increased SWR will be tolerated by these
solid state amps. The Harris might need to have the SWR shutdown pot tweaked a bit.
I will try testing a quarter wave stub with one of the Harris amps as soon as I get
a chance and will let you know how it seems to work here.

The Harris amps also are not all that efficient either, due to the ultra linear
nature of the units. I suppose you could adjust the bias to make them more efficient
in certain modes such as JT65 or CW or even SSB. But it is great to have something
that is very clean... and with two of them, I have more power in reserve than I need
and they don't even get warm at 1500w output on JT65a MODE.

GL and VY 73, Lance

[Dave Olean]

Hello Dick, Peter, and LARCAN crew,
I spent some time working on getting the LARCAN amp to provide a better
account of itself on 50.0 MHz. I have some preliminary stuff written up and
sent it along to Bob to post on these LARCAN pages. Basically I attacked the
input and output combiners to make them work better at 50 MHz. I changed
capacitance values there as well as a changing few values in the power
output side of the individual FETs. At low drive levels I am seeing
efficiencies that are in the high 30's and that gets better when I press the
pedal to the metal. (over 50%)I really need a better setup to see my
ACCURATE power output on peaks. (At the moment I am using a 1000 watt slug
rated for 100-250 MHz) I substituted a Yaesu FT-625RD in place of my 4 watt
instrument amplifier and was amazed at what came out. I kept the drive down
to under 10 watts and looked at the output on my scope. The scope has great
bandwidth so 50 MHz is easy! At the point of clipping I could peg the 1000
watt slug pretty convincingly. I can drive this amp now to beyond the
current rating for the transistors at the 50 volt level, so you do not want
to hold the key down. I used up all the current my ASTEC supply could
deliver. For reliability, I am going to limit the current to about 36 amps.
Next step is to retrieve my peak reading LP-100A wattmeter from my remote
hamshack and measure the peak power available along with the drive power
being used. I calibrated the LP-100A with a high power Bird 30 dB pad and an
HP 435 power meter. It is within about 1.5% at all power levels. Rough
testing with Bird slugs indicates about 21.5 dB gain at 50.0 MHz, and 22 dB
at 54 MHz. I will update the notes when I get better figures. I really like
this amplifier now!
73
Dave K1WHS

-----Original Message-----
From: Dick Hanson
Sent: Wednesday, January 18, 2012 8:24 AM
To: Dave Olean
Subject: RE: 50 MHz Conversion Freq Response

Way to go, Dave!
Will look forward to reading more later.

73
Dick, K5AND

> -----Original Message-----
> From: larcan-6m-...@googlegroups.com [mailto:larcan-6m-
> ampli...@googlegroups.com] On Behalf Of Dave Olean
> Sent: Wednesday, January 18, 2012 7:18 AM
> To: larcan-6m-...@googlegroups.com
> Subject: Re: 50 MHz Conversion Freq Response
>

[Dave Olean]

Hi Peter,
I sent the article to Bob DeMattia. Bob had said he would post the
article over the weekend. I am sure it will show up pretty soon. I have had
my amplifier sitting here unconverted, for a long time. My problem is that I
have very little free time, and usually only in our winter months between
Jan 1 and about March 15th. That is when I do all of my homebrewing and
such. I had heard stories that the Larcan amp was a dog at 50.0 MHz with
rather poor efficiency, but I know that the MRF151G is a pretty hot
performer. Being the cheap ham that I am, I wanted to turn this amp into a
nicer unit. I think it is now a better deal than the Harris ones, as those
transistors are now Unobtanium, I believe. I like the Larcan circuit as the
transformer losses are very low. The efficiency is way up now. I think I
saw 38 or 39 amps when it was putting out 1200 watts peak. That is 60% or
better.
What I provided to Bob was a first draft and I need to clean it up and
polish it a bit. Some of the plots need to be cleaned up etc. My final
results were taken with my K3 low power exciter. I retrieved my calibrated
LP-100A peak reading wattmeter and here is what I saw:
50 MHz 7.34 watts PEP 1200 watts out
54 MHz 7.5 watts PEP 1350+ watts out
these numbers are off the top of my head. My notes are around here
somewhere,but I can't put my fingers on them. I remember figuring 21.9 dB
gain at 50.0 and 22.93 dB at 54.0 MHz. This is much better than the original
numbers. There still is some roll off in the combiners/ splitters, but my
mods improved the matching on the input and the output by a large amount.
Now the amp really stands up and takes notice when you drive it. I whistled
it up to almost 60 amps on my 150 amp ammeter and 50 mv shunt. Who knows
what the real peak current was, but the amp pegged a 1000 watt slug
violently. That amounts to 15 amps per dual FET and I think that may be
too much so I have not tried that trick again. My little K3 at 7+ watts
drives it in the range of the high 30's for amperage and peak powers are
about 1200 which is 4 X the 300 watts that the MRF151G is rated for. I am
leaving it there. I hate buying new FETs

Dave K1WHS

-----Original Message-----
From: Peter Sumner
Sent: Saturday, February 11, 2012 7:04 PM
To: k1...@metrocast.net
Subject: Re: 50 MHz Conversion Freq Response

Hi Dave,
if you cold FWD that email with your write up I would appreciate it.
I am amazed that it has taken this long for any of the arcan owners to
click that the performance of the amps is so poor at 50MHz. From very
early on here I could see it was not very good and thought it must be
my modules as no one else was saying anything else.

I am waiting for some more ATC caps to come for my second larcan PA,
so I can complete the basic conversion then I will merge your finding
and then do some freq sweeps with the spec analyser here. A 3dB roll
off is big hurdle and means lots more drain current to get a
worthwhile output so I am keen to get that as low as I can.

my progress is documented here:
http://www.users.on.net/~pedroj/larcans/index.html

Regards,
Peter, vk5pj

[Kevin Hobbs]

Hi Dave and Others

 

Sounds like your amp now performs the same as my unmodified lo-lo.

 

My measurements are below. Setup used consisted of a commercial Sorensen 60V / 50A PS, Power Masters on both the Input and Output, so I am confident the numbers are very accurate. Drive was a string of dits about 15WPM. Idle current with bias off was about 8A.

 

Larcan	VE3KH Measurements Nov 24, 2011
Pin (w)	Pout (w)	Vin (V)	Iin (A)	PS (W)	Eff	Pgain (dB)
8	1100	48	33	1584	69%	21.4

 

I am now just integrating a 2KW HP Blade Server PS along with metering and cooling into an HP Rack Mount chassis … pictures to come.

 

73, Kevin

VE3KH

[Dave Olean]

Hello Kevin,

That sounds pretty neat.  I have been running my amplifier at 500 ma per side. I think the gain will go up if you increase the quiescent drain current.   Mine barely had 20 dB gain at 50.0 MHz after the conversion to a LO-LO. at the 500 ma  Idq level ( 4 amps total)  I was concerned that the combiners present a poor match to the MRF151Gs at 50.0 MHz.  The output combiner looks like 2:1 or worse at 50 MHz.  It was much better from 55 up.

 

Dave K1WHS

[Kevin Hobbs]

Hi Dave

 

You may be right … but, I am pretty happy with the performance as-is … I trust that the MFG did it right … so I am not inclined to touch anything!!!

 

73 KH

[Dave Olean]

Peter,
I suspect that many users are running the LARCAN amps higher in the
band for FM where the gain and efficiency is much better. A few friends had
noted that the LO-LO amp was somewhat of a DOG at 50 MHz SSB with
efficiencies in the 25% range. They either sold them or converted them to
144 MHz. I knew I would have to tweak it, but I knew that the MRF151Gs were
very nice FETs and capable of producing 300 watts plus each. I think I like
the Larcan now better than the Harris Platinum amps as it is now more
efficient with better gain and with less loss in the impedance transformers.
The Harris amp also has FETs that are made with "Unobtainium". The MRF151Gs
are still being made at least for now! I was toying with getting a Harris
after reading about them. They are nice amps too.
I started converting my amp shortly after you first posted to the users
group about poor gain etc. As far as I could see, no one had dissected the
amp to find out what needed changing. It was a good exercise at getting me
more familiar with solid state amplifiers. Good luck with your conversion.
I will update things as I get more info.

Performance again with an accurate (1.5%) peak reading wattmeter
(LP-100A) :

50.0 MHz 7.34 watts input PEP ~1200 watts output PEP SSB 163.5X
22.1 dB
54.0 MHz 7.5 watts input PEP ~1350 watts output PEP SSB 180.0X
22.5 dB
Efficiency is up well over 60% at these power levels. I am not sure how
accurate my current shunt is, but I was seeing under 30% when I started. I
sure would like to know what the maximum current is per FET. I am keeping it
under 9 amps now (36 amps total) to be safe. It will put out much more power
with more drive and more drain current! Maybe I am reading the MRF151G data
shweet incorrectly. It says 9 amps at 50 volts in FIGURE 5, (Safe Operating
Area) but now I am thinking that is the per side value and 18 amps may be
the correct maximum number for the two fet sections. After all 9 amps times
50 volts is 450 watts. They spec 50% efficiency at 175 MHz. It would not
meet the 300 watts output at 175 MHz. This amp will put out much more power
if the drain current exceeds 36 amps total. Maybe some expert could point
me in the right direction.

Dave K1WHS

-----Original Message-----
From: Peter Sumner
Sent: Saturday, February 11, 2012 7:04 PM
To: k1...@metrocast.net

Subject: Re: 50 MHz Conversion Freq Response

Hi Dave,
if you cold FWD that email with your write up I would appreciate it.
I am amazed that it has taken this long for any of the arcan owners to
click that the performance of the amps is so poor at 50MHz. From very
early on here I could see it was not very good and thought it must be
my modules as no one else was saying anything else.

I am waiting for some more ATC caps to come for my second larcan PA,
so I can complete the basic conversion then I will merge your finding
and then do some freq sweeps with the spec analyser here. A 3dB roll
off is big hurdle and means lots more drain current to get a
worthwhile output so I am keen to get that as low as I can.

my progress is documented here:
http://www.users.on.net/~pedroj/larcans/index.html

Regards,
Peter, vk5pj

On Fri, Feb 10, 2012 at 12:51 AM, Dave Olean <k1...@metrocast.net> wrote:

[Jeff DePolo]

> I sure would like to know what the maximum current is per FET.
> I am keeping it
> under 9 amps now (36 amps total) to be safe.

That should be very safe, assuming the matching is correct. If you plan on
running the amplifiers at reduced power, lowering the drain voltage rather
than reducing drive is preferred, as is reduces the liklihood of a
catastrophic failure in the event of a fault (such as high VSWR).

> It will put out
> much more power
> with more drive and more drain current! Maybe I am reading
> the MRF151G data
> shweet incorrectly. It says 9 amps at 50 volts in FIGURE 5,
> (Safe Operating
> Area) but now I am thinking that is the per side value and 18
> amps may be
> the correct maximum number for the two fet sections.

Yes, that's per side, but it is the absolute maximum. From a dissipation
standpiont, I'd consider 50% of the spec'ed total device dissipation as
being a practical maximum when using traditional (mechanical clamp)
mounting/heatsinking as in the instant case, assuming all other limits are
met with an adequate safety margin. The TDD for these parts is 500 watts,
so half of that is 250 watts, or 1000 watts for four devices. 1200 watts
TPO at 60% efficiency equals 800 watts dissipation, so you've got a bit of
headroom at the rated output (combining losses notwithstanding).

My experience with MRF-151G's, BLF-278's, SD2932's et al in non-amateur
amplifiers is that they are very reliable as long as you treat them nice.
They are extremely unforgiving of high VSWR and overdrive. They are quite
unlike good ol' bipolar RF transistors as are commonly found in ham "brick"
amplifiers and two-way radios. Good protection is mandatory if you want
them to survive. The stock Larcan protection circuitry, while adequate when
these PA's are being combined with some inherent reflected protection
afforded by the combiner, may be inadequate in a typical ham installation.
For one thing, the stock circuitry doesn't totally shut down the device when
a fault it is sensed, it only reduces bias. Killing the drain voltage
completely, or applying a negative bias to the gate, would be preferred, and
definately recommended if using the module standalone. Rigs with TPO
overshoot at initial key-up are to be avoided (and I'd venture to guess that
more than half of the ham rigs out there overshoot). Just some warnings
from someone who has replaced dozens, if not over a hundred, LDMOS FET's in
broadcast transmitters over the years...

--- Jeff WN3A

[Dave Olean]

Thanks for the info Jeff,
I can see that the Larcan protection is not very good. I really want a
power FET switch on the 48 volts. I am looking for a design. I know that the
Harris Platinum amp has that. Still, I get the impression that I can run a
bit more current than the 36 or 38 amps I am running now. The thing that I
like is that my low power K3 at under 8 watts, provides just about the
correct drive at full power. At full power it is hard for it to overshoot!!
I should try to experiment with negative cutoff bias for protection.
Removing the bias voltage doe not kill the RF power entirely. I have a way
to go before this amp is ready for "Prime Time".

Dave K1WHS

-----Original Message-----
From: Jeff DePolo
Sent: Wednesday, February 15, 2012 10:48 AM
To: larcan-6m-...@googlegroups.com
Subject: RE: 50 MHz Conversion Freq Response

[Jeff DePolo]

If you're using one of the surplus blade server power supplies, it has an
enable pin -- switch it on/off as your means for killing drain voltage
(assuming that the amplifier is the only thing being powered by the supply).
Otherwise, a high-current FET switch or a brute force relay/contactor,
either mechanical or solid state, can do he job (if solid-state, make sure
it's DC-DC switching). Check out some of IXYS' "GigaMos" MOSFETs - for
under $20 they'll switch 100+ amps.

--- Jeff WN3A

> -----Original Message-----
> From: larcan-6m-...@googlegroups.com
> [mailto:larcan-6m-...@googlegroups.com] On Behalf Of Dave Olean
> Sent: Wednesday, February 15, 2012 12:57 PM
> To: larcan-6m-...@googlegroups.com