Page 1 of 1
Posted: Fri Sep 06, 2019 5:20 pm
Iv been looking around for different mosfets to try out. Every time I check the data sheet they say obsolete. So I don’t really want to design something based on an obsolete item.
Most of the round type SD series say obsolete, SD2931, SD2933, SD2943, etc.
Mostly all MRF series I see are gone.
Bipolar transistors are well gone. BLX series...... what’s left?
Only seems to be rectangle type fets now.
Posted: Fri Sep 06, 2019 8:06 pm
The ma/com mosfets are still in production (mrf171a , mrf173 ect.)
Posted: Sun Sep 08, 2019 10:19 pm
Many of the old bipolar types of RF power transistors are made now by a company called ASI or Advanced Semiconductor Inc. The BLX15 for example, is in their catalog, so you can still maintain or build equipment wtih some of the device types that were obsoleted by their original manufacturer some time ago.
But if you want to try out some brand new shiny mos-fets, recently there has been released the MRF101AN that can push 100W in a convenient TO220 package. There is a demo board for 88-108 MHz published with all the details in the manufacturer's website.... just in case...
Posted: Mon Sep 09, 2019 12:44 am
Have a look at NXP Semiconductor, they have a range of modern dual LDMOS devices.
https://www.nxp.com/products/rf/rf-powe ... F-INDUST-1
The ones I use are the MRFE6VP5300N (300W) and MRFE6VP5150N (150W), both only require 1W of drive, 1.5W max.
You can get a higher output impedance with a dual LDMOS package as opposed to a single ended devices, and therefore a much better broadband match.
Posted: Mon Sep 09, 2019 11:11 pm
I got three MRF101AN TO220 FETs this morning. I built my first PA this afternoon. 60V supply, 800mW input, 100W anywhere you want from 87.5 - 108 MHz. These are available from Farnell or Mouser for around £18 - pretty cheap for a really useful device!
Posted: Mon Sep 09, 2019 11:20 pm
Did you use the reference design from fig. 32/33 in the datasheet?
This one: https://www.nxp.com/docs/en/data-sheet/MRF101AN.pdf
Posted: Mon Sep 09, 2019 11:25 pm
It's very similar, but laid out to fit a heatsink box I have in stock. These are really good - cheap, easy to mount, efficient and pretty robust.
Posted: Tue Sep 10, 2019 12:00 am
Hopefully you built a better filter. 2f is only -32 dBc for that design or in the old money that your test equipment accepts, about 1.78 V RMS.
Posted: Tue Sep 10, 2019 12:28 am
On 98.4 MHz before the filter, the second was -44dBc and the third was -49 dBc. When the filter was added, the second was -71dBc and the third was in the grass. The filter is separated from the PA by a tinplate screen, and the coils in the filter are aligned to minimise mutual coupling. The capacitors in the filter are "AVX" branded chip caps.
The gain of these amplifiers is proportional to the supply voltage, which could prove useful if you need a very specific power output.
Posted: Tue Oct 01, 2019 5:59 pm
More on the MRF101N:
I managed to destroy one! I fried the gate when I dropped a screwdriver into the bias supply circuit. The output power just disappeared, and the PA stopped drawing much current. I realised what I'd done, and a quick replacement FET got the PA going again.
I think that I've got the things just about optimised.
The PA gives exactly 100 Watts output for 500mW input at 62V supply. The efficiency is ~78%, and the gain is substantially flat across the band: At 90MHz, it needs 510mW for 100W out, at 98.4 MHz it needs 495mW for the ton, and at 106.6MHz it needs 500mW exactly.
At 90 MHz, the second (after the filter) is -67dBc and the third is in the grass. At 98.4MHz the second is ~-71dBc, and at 106.6MHz the second is ~-73dBc. Obviously these numbers could be bettered with coax stub filters cut for the frequency in use (my favourite LPF!).
Interestingly, the bias is less touchy than with older FETs - obviously they die if you give them too many volts - but once they're into conduction, they're not really critical at all. The efficiency and the gain varies slightly, but they're really easy to set up.
These things are just £18 from Farnell, and are really easy to get going. There's not a huge amount of heat to get rid of (my 144 MHz version just uses the diecast box that it's in for the heatsink, since it's in intermittent use). My Band II exciter for these things is a half-frequency FET oscillator into a FET buffer driving a diode doubler, further FET buffer, monolithic bandpass filter, then a BFR96 pair for 500mW. The PLL is the three chip 74HC circuit that I often use. The whole thing fits into a moderately sized Hammond diecast box, with the exciter / PLL in a segmented tinplate box, and a separate tinplate box for the output filter. The power supply is a simple linear for the exciter and a cheap "telecoms" SMPSU for the 60+V supply. The whole thing is small and neat, and costs next to nothing to build.
I can't wait for these devices to come down in price. The Motorola rep I talked to yesterday suggested that they'd be around £5 in the fairly near future, since they have another new generation of wonder devices coming out shortly!
Posted: Thu Oct 03, 2019 11:47 am
This sounds very interesting Albert, I must give these devices a try. They're only £18.58 a piece on Digikey and could make for some cheap PA's.
https://www.digikey.co.uk/product-detai ... ND/9749131
Posted: Thu Oct 03, 2019 12:53 pm
I didn't realise Motorola were still in the RF transistors game, as far as I was aware they stopped making them a number of years ago and MA-COMM started manufacturing them instead??
Posted: Thu Oct 03, 2019 1:34 pm
it's made by NXP.
Posted: Thu Oct 03, 2019 3:33 pm
Ahh ok fair enough.
....That's a bit strange for a newly released transistor to have the old MRF (motorola radio frequency) part-number codes, expecially when it wasn't even made by Motorola.
Until today I just thought NXP made the old Philips devices. (BLF278 etc)