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]   Go Down
  Print  
Author Topic: Transistor recommendation  (Read 2839 times)
OK8MP
Member

Posts: 3




Ignore
« on: January 29, 2012, 07:43:00 AM »

I have homebrew 160-10m QRPp transceiver with very clean class A output of 50 mW to 100 mW depending on band.  I would like to bring the power up to a very clean 20 to 30 Watts with linear amplifier circuit.  I have looked at RD15HVF1 based designs like HPSDR PennyWhistle and others as well as IRF510 designs such as HF Packer amp.  All have very poor IMD so I ruled them out. 

For the SS amp experts out there, what RF transistor would you recommend for something like this?  12 Volts is not a limitation, I can do 24 V or 50 V, the main requirement is low IMD for me.  Needs to be broad banded and cover at least 160 to 10 meters.  I was looking at the NXP MOSFETS like BLF175, etc..., but need some direction. It needs to be a device that is readily available too.  Any recommendations?

Logged
ZS5WC
Member

Posts: 410


WWW

Ignore
« Reply #1 on: January 29, 2012, 08:14:50 AM »

 Smiley

I used the old 2sc1969 (mitsubishi) transistors and they work well. In my single band rig they give me 25w in push pull with ease.
You could look at the K2 design (these transistors are used there as well..), and build a broadband amp from that.
Should give TX imd performance of -25dB at least at 12v.
Not wonderful I know, perhaps a valve (TUBE) will give a better TX imd.
From your power level you would need a buffer amp to give the required drive to the 2sc1969 pair.

I also am disappointed with the new mosfet TX imd--sad that all the new rigs on the market use them.

73 de William
ZS4L / ZS5WC
Logged
M0HCN
Member

Posts: 473




Ignore
« Reply #2 on: January 29, 2012, 09:16:24 AM »

Ok, so for a broadband stage you are going to want to run push pull (Makes the output filters easier), so we are looking for a pair of ~15-20W devices, and we probably want to do this at 24V or just possibly 50V.

Further we need to go from 50mW to ~25W, so in total we need around 27dB of gain, which is probably going to need two stages even with modern parts.

BLW50F springs to mind for the output devices, old school bipolars, can be run in class A (at the cost of much heat), but difficult (and expensive) to source these days, and like all bipolars would need a stiff bias supply.

BLF175 looks like a sane choice to me,  a push pull pair on a 50V rail will get you -35dB[1] or so IMD3 in class AB, and possibly a little more if you up the standing current a bit (heatsink dependent).
However a pair is really good for 50W - 60W with appropriate drain match, so may be overkill.
They are not quite man enough to make 20W push pull in full on class A, but if you have sufficient cooling you could probably run very high bias AB with good results.

Add some negative feedback (bifilar transformer between the drains in the manner of those old moto app notes), and make sure the DC injection ferrite is big enough and you might just hit -40dB[1] or so with as much as 17db of gain.

I would probably design the drain match for maybe 30 - 40W, then by running at 20W you will be well away from saturation (At the cost of rather low efficiency), the drain matching transformer could bear experimentation to optimise for distortion.

Regards, Dan.

[1] Note, I ***THINK*** the datasheet gives this number ref one tone (The wording is a bit weird), if so read as -41dB ref PEP.






Logged
OK8MP
Member

Posts: 3




Ignore
« Reply #3 on: January 29, 2012, 10:27:58 AM »

Thanks for the replays.

How about a single MRF150 in class A with a very good heatsink?  I am willing to waste the energy if I can get an extremely clean output.  I know I will have to proceed it with a driver stage too. 

I am hoping to target something in the -40 dB range.  Another requirement is that the transistor can be purchased easily and is currently available from places like Digikey, Farnell, or Mouser.

Thanks, MP.
« Last Edit: January 29, 2012, 10:31:14 AM by OK8MP » Logged
M0HCN
Member

Posts: 473




Ignore
« Reply #4 on: January 29, 2012, 10:46:08 AM »

Well I suppose you could, but the second harmonic will be a pain.

The big problem with a single anything is that the output filter becomes disproportionately more complex, because you need much more second harmonic attenuation, so the ability to use one filter for two bands becomes very much more marginal then it usually is. Granted class A will have lower harmonic output to start with, but still.

I would bet that the extra transistor is more then paid for by the simpler harmonic filtering.

Be a little careful as well, bias voltage temperature compensation is required to avoid the risk of thermal runaway in class A with mosfet stages (contrary to popular belief).

Regards, Dan.
Logged
G3RZP
Member

Posts: 4366




Ignore
« Reply #5 on: January 29, 2012, 12:26:01 PM »

For anything over 5 watts, you need 50 dB of harmonic and spurious attenuation at HF in Europe. (ERC REC 74-01)Below 5 watts, it's 43 + 10 log P dB.
Logged
KB1GMX
Member

Posts: 711




Ignore
« Reply #6 on: January 31, 2012, 12:34:02 PM »

The biggest issue is for a mere 50-100mW you need gain to get to 30W and that will not likely be  a
single stage.

Two RD70HHF in push pull should get to 70+W and will be clean at 12V is the bias is around
50-100MA per device. But you only get 13=14DB gain (100mW in 20W out).

There is a design published in QST using 4 of RD16HHF, 2ea in parallel in a push pull amp.
that was either 2010 or 2011 home brew challenge.  It was 50W for 1-2W in.

Either way push pull is the best topology for bandwidth and ease of filtering. Also if you do
push pull than class A operation is not needed and likely not desired as AB1 is clean.  With 30W
class A system you going to be running at least 60W of heat.

The 28 Volt devices will be cleaner and give better gain and the 50V parts do better still.
If you take the 50V 100W devices and run them at 40V you will get an amp that is near
unbreakable yet clean.  The typical 50V device as push pull should yield around
20-22DB gain for HF.  For 100mW that is around 80-100W.

Something like a RF16HHF driving a pair of RD70HHF will get you to 100W at 12V
but not as clean.


Same problem for bipolar devices with the added problem of getting a broad band match.
They typically have lower gains in the 11 to 15db range for 12V devices.  the problem is
that devices like the 2SC1969 are getting hard to find (other than counterfit).

You do not say if the source QRPp is CW or SSB.  Usually CW has far less problems with IMD
in the amplifier and running class A is not needed.

In the end your pitting low drive power, device availability and wide band operation.  You will
have to pay attention to input and output transformer design and ferrite used for best operation
at the edges (160 and 10M).


Allison
Logged
OK8MP
Member

Posts: 3




Ignore
« Reply #7 on: January 31, 2012, 12:48:52 PM »

The biggest issue is for a mere 50-100mW you need gain to get to 30W and that will not likely be  a
single stage.

Two RD70HHF in push pull should get to 70+W and will be clean at 12V is the bias is around
50-100MA per device. But you only get 13=14DB gain (100mW in 20W out).

There is a design published in QST using 4 of RD16HHF, 2ea in parallel in a push pull amp.
that was either 2010 or 2011 home brew challenge.  It was 50W for 1-2W in.

Either way push pull is the best topology for bandwidth and ease of filtering. Also if you do
push pull than class A operation is not needed and likely not desired as AB1 is clean.  With 30W
class A system you going to be running at least 60W of heat.

The 28 Volt devices will be cleaner and give better gain and the 50V parts do better still.
If you take the 50V 100W devices and run them at 40V you will get an amp that is near
unbreakable yet clean.  The typical 50V device as push pull should yield around
20-22DB gain for HF.  For 100mW that is around 80-100W.

Something like a RF16HHF driving a pair of RD70HHF will get you to 100W at 12V
but not as clean.


Same problem for bipolar devices with the added problem of getting a broad band match.
They typically have lower gains in the 11 to 15db range for 12V devices.  the problem is
that devices like the 2SC1969 are getting hard to find (other than counterfit).

You do not say if the source QRPp is CW or SSB.  Usually CW has far less problems with IMD
in the amplifier and running class A is not needed.

In the end your pitting low drive power, device availability and wide band operation.  You will
have to pay attention to input and output transformer design and ferrite used for best operation
at the edges (160 and 10M).


Allison

I am now convinced that push pull is the way to go for this.  My QRPp transceiver is both CW and SSB.  I am thinking of going with something like the BLF175 in AB1 push pull now since I can get these devices easily.  I have not picked out the transistors for the first stage yet.

Here is a link I found that uses the RD70HHF as well as the RF16HHF devices http://neon.skydan.in.ua/PA.php with a OPA2674 driving them.  IMD is -30 to -36 range at 50W and almost -40 at 30W supposedly.

I wonder if the CPU coolers can be adapted for a heatsink.  They have nice copper mounting area.  http://www.coolermaster-usa.com/upload/product/3049/gallery/thumb/5.jpg?10807823

Thanks, MP
Logged
AD6KA
Member

Posts: 2232




Ignore
« Reply #8 on: January 31, 2012, 02:15:07 PM »

Quote
I wonder if the CPU coolers can be adapted for a heatsink.  They have nice copper mounting area.  http://www.coolermaster-usa.com/upload/product/3049/gallery/thumb/5.jpg?10807823
Sure, lots of folks use them on small amps.
Just make sure to electrically insulate the PA transistors
with mica or Kapton pads and plastic shoulder
washers. NTE offers a mica pad and plastic shoulder
washer set, I forget the part number.

All Electronics has a very good selection of CPU cooler
heat sink/fan units at very reasonable prices.
One example:
http://www.allelectronics.com/make-a-store/item/CF-395/CPU-COOLER/SOCKET-7-370/1.html
My written catalog from All Electronics
shows more selections than the online catalog for
some reason....
Good luck with your project.
73, Ken  AD6KA
Logged
Pages: [1]   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!