You may be able to improve the gain and output power by shortening the length of the MOSFET Source lead, soldering it closer to the ground plain, especially on the higher frequency bands. At 7MHz, the design is capable of ~17.5dB Gain, ~55W at about 48% efficiency. Idq was set to 10mA/device for best two tone linearity. Thanks for sharing your home brew construction of the amplifier. 73, Mike, WA2EBY
Nice little project. I would have used a pair of slightly bigger devices... go for 70W but what you have there is a nice compromise between supply current and heatsink size etc
Hello Ron! Enjoyed watching your DIY project, Thank-you. I observed couple of points, 1) The binocular core names on schematics are marked wrongly, names just need to be swapped. 2) At 20 mA bias current, i could observe some crossover distortion, anyway you did increase the bias to 30mA. It probably would have fixed the distortion too. Thanks again! Best regards, bp (I am just a DC to RF enthusiast, not a ham)
Hi and thanks for your comments. I re-checked the schematic and found that the binocular core names are indeed correct. The smaller one on the input is the BN43-202 and the large one on the output is the BN43-3312.
Hello Ron, Glad to hear from you. Yes! you are indeed right. Sorry, my mistake. I probably made that mistake not realizing that the dimensions were actually given in inches for larger core and millimeter for smaller core on Amazon website. 73's Best regards, Bharath
That is what I call a really informative practical video. I have a few of those transistors. I intended to use them with a construction similar to the book 'building a transceiver' by ei9gq. I was impressed by the results from your construction, so I think I will try both. Your construction seems to use simpler transformers, so I will kick off with yours. I will be interested to see how your construction would work with more bias up to around the linear position for ssb. I will also try using a simple 50 ohm attenuator padder at the input and see what happens. Thanks for the video. Keith G0ODU
This is an answer on a filter question which is hidden as a reply below you either go to QRP labs and buy the filter from there or you use Motorola Application note 267 which is still available on the web. This covers several filter topologies including the Pi type filter which I believe is Net 3 in that note. Look it up and work it out. Zout for the amplifier (depends on mode - the above is oriented to CW mode but it will operate as a linear at lower power out. THe output impedance from the amp to match is approximately 2 * (Vcc - 2 ) squared / Pout where Pout is say 30 W for CW use and 16 W for SSB use using much higher bias approx 600 mA total quiescent current through the sum of the Drains. If you are designing a two PI filter you should chose sqrt ( output Z x 50) and match into that using the equations given in the application note. and using a filter Q of either two or three - your choice. Q is defined inn the application note. The RHS filter on the output can be designed using the tables in the Appn note as you know the RHS termination is fifty ohms. You need to have basic EE knowledge to do this ie you need to know how to go from impedance to L and back again and how to do the inductor turns calculations. Those are covered in the toroid tables of The Toroid king - search on line. Such filters can only be approximate as there is tolerance on the properties of the iron powder material. A size fifty toroid will be adequate for a 30 W rf filter. The capacitors should be either multilayer monolytic or Silver mica. The latter are preferred for low loss. The Toroid king sells them. If you want a seven element filter you have to extend the above description. THe application note shows how you can subtract Co of the respective Mosfet from the calculated filter C1 capacitor. This is not applicable for a fully isolated OP transformer as used here. It is applicable for class C choke type circuits which is a whole different matter. Basically, the Appn note assumes single ended RF construction so you have to show some initiative when applying it to a push pull RF amplifier. BTW you may wonder where I got the 16 W above from . Kees Talen used to offer a pretty sophisticated kit with a similar design and it used identical Active devices from MItsubishi. His considered max Pout with full Fcc Compliance was 16W out. This was on the basis of a 2 Pi filter or two section filter following the power amp and 14 V supply. If you take the voltage up to say 16 , which is outside the manufacturer's specification for these devices and if you over drive them (he used just 600mW in) whereas Ron uses one watt, you may well get of order 30 W out on CW but you may not be FCC compliant. The FCC requirements are considerably harsher for transmitters over 5 W out and some people don't allow for this. In addition of you wire the output transformer for 2T out you will have a nominal output Z of 72 ohms (running the amplifier for 16 W out nominal) and if you treat that as 50 R you get only three percent mismatch loss if you design the filters as 50 R in and 50R out. I think Diz specified 1:3T (I have his original schematic just to try to get a bit more efficiency out of the filter by running it as a higher voltage input device in the first two sections. Regards TE Fidler Nz.
Increasing power - a Japanese constructor made a similar amp years ago but called it a 40 W amp. IN fact it is about a 34 W amp for SSB use when fed on 14V supply. There was about twice as much ferrite in his core. He used a copper tube core but you can get a similar effect by using teflon insul coax cable with all the braid in contact. Ok so the only other major change was he connected one 100 R x 1W resistor to each gate , wound 6T of ECW around it and brought it out to a common point per pair of RD 16HHF1 mosfets, enabling two to operate as if one device. They were biased in common (the assumption is that the came out of the same production batch ). THe point of the inductor thus created was to stop ringing between the gates. The Softrock 50 which is an ultra reliable kit amp uses a three plus three arrangement to claim 50 W out for SSB - that is the Japanese claim. (The Japanese is likely getting probably more like 90 W PEP). It s food for thought when laying out such a board - allow for expansion. Such an amp would require about 1.8 W at the mosfet gates to achieve full output. TEFidler ...
Recently found an old BLF578 ldmos device that had failed. Measuring it i noted the failure on one side only. I threw it at a PCB along with a few variable Cs and a few inductors and wham ! 400W out with 1.5w in at 40v on the Drain, 80% efficiency. Based the circuit on the MRF174 data sheet. Great use of a device that was destined to be scrapped. Might end up on 6m as a ft8 PA. 73 from ZL.
Very clean construction technique. I like very much the method to create pads inside a ground plane. Also, it will be easy to change the transistors when burned by a SWR accident, as we all do ;-) Thanks, F1SOC
Great video Ron; really enjoyed the step-by-step construction and test walk-thru. For isolated pads I bought a punch-tool from Harbor-Freight, filed off the punch center-tips and use it to punch-out small round pads out of scrap PCB that then I super-glue down to the main PCB.. The reason I adopted this was that I could make tiny island-pads that are easy to fit into tight spaces and therefore really able to shrink the size of my overall circuits...
This is very nice. Certainly I have learned some things that will be useful. I have not watched a deadbug circuit put together. Even simple things for me are new. Thanks!
Мудрость древних, навесной монтаж над полностью покрытым медью двухсторонним текстолитом, усилитель класса AB, трансформаторы на марганец -цинковых сердечниках из сетевого фильтра, комутация на реле со схемой детектирования сигнала на диодах
Thank you for all the explanations. This was really helpful in understanding the circuits needed and how everything ties in. Now I'm not so scared to try to make my own. 73!
I think that ieland cutter is very beautiful, unfortunately we cannot buy them here in the Netherlands! Healthy and Friendly Greetings from the Netherlands! Rob
The digital wattmeter is adapted from an article by W8TEE and AC8GY titled “Build an Inexpensive 150 Watt Dummy Load with Wattmeter” published in the Nov 2018 issue of QST magazine.
Very nice build and explanation video. Glad to see another homebrew project using real RF Mosfets like the RD series instead of IRF switching Mosfets. Excellent work brother. 73. KC3MOP
I built my first successful RF Amps ever with these RD series Mitsubishi MOSFET Transistors they are awesome and require very low drive to make WATTS even my BK Precision Arbitrary Waveform Generator will drive them!
very good video : thaks for sharing your experience. Your explanation is very accurate and it will so simple for me to buildup one. The point for me is the output filter since I do not have special instruments available here at my station: could you please post the data to build a power low pass filter for 80, 40 , 20 and 10 meter? Thanks for all.
you either go to QRP labs and buy the filter from there or you use Motorola Application note 267 which is still available on the web. This covers several filter topologies including the Pi type filter which I believe is Net 3 in that note. Look it up and work it out. Zout for the amplifier (depends on mode - the above is oriented to CW mode but it will operate as a linear at lower power out. THe output impedance from the amp to match is approximately 2 * (Vcc - 2 ) squared / Pout where Pout is say 30 W for CW use and 16 W for SSB use using much higher bias approx 600 mA total quiescent current through the sum of the Drains. If you are designing a two PI filter you should chose sqrt ( OUtput Z x 50) and match into that using the equations given in the application note. and using a filter Q of either two or three - your choice. Q is defined inn the application note. The RHS filter on the output can be designed using the tables in the Appn note as you know the RHS termination is fifty ohms. You need to have basic EE knowledge to do this ie you need to know how to go from impedance to L and back again and how to do the inductor turns calculations. Those are covered in the toroid tables of The Toroid king - search on line. Such filters can only be approximate as there is tolerance on the properties of the iron powder material. A size fifty toroid will be adequate for a 30 W rf filter. The capacitors should be either multilayer monolytic or Silver mica. The latter are preferred for low loss. The Toroid king sells them. If you want a seven element filter you have to extend the above description. THe application note shows how you can subtract Co of the respective Mosfet from the calculated filter C1 capacitor. This is not applicable for a fully isolated OP transformer as used here. It is applicable for class C choke type circuits which is a whole different matter. Basically, the Appn note assumes single ended RF construction so you have to show some initiative when applying it to a push pull RF amplifier. BTW you may wonder where I got the 16 W above from . Kees Talen used to offer a pretty sophisticated kit with a similar design. His considered max Pout with full Fcc Compliance was 16W out. This was on the basis of a 2 Pi filter or two section filter following the power amp and 14 V supply. If you take the voltage up to say `16 , which is outside the manufacturer's speck for these devices and if you over drive them (he used just 600mW in) whereas Ron uses one watt, you may well get of order 30 W out on CW but you may not be Fcc compliant. The FCC requirements are considerably hasher for transmitters over 5 W out and some people don't allow for this. In addition of you wire the output transformer for 2T out you will have a nominal output Z of 72 ohms (running the amplifier for 16 W out nominal) and if you treat that as 50 R you get only three percent mismatch loss if you design the filters as 50 R in and 50R out. I think Diz specified 1:3T (I have his original schematic just to try to get a bit more efficiency out of the filter by running it as a higher voltage input device. Regards TE Fidler Nz.
@@ambroribo6079 If you go to the Toroid King Site you will find in his kits that he sells LP Filter PCBs If you want just two frequencies then two off DPDT relays (use contacts in parallel ) will allow switching between them. Don't try it with just one relay - there will be too much signal bypassing occurring. The single relay approach has been used in cheap Indian kits and it turns out to be a bit of a disaster technically. TEF
Hi Ron, I am seeking an RF amplifier that can input 250mw and boost to 3 watts+. Frequency range is 470-480mhz. Do you have any recommendations? Is this something you could make? Thanks!
You might try one of the Toshiba or Mitsubishi RF power modules sold by RF Parts. One they currently have that might meet your needs is the Toshiba S-AU99H. Good luck
Excellent video. Great for a beginner like me. Would this amp work on top band? 160m? If not, what would need to be changed to make it work at lower frequencies? Thanks!
It should work there but you need to choose your ferrite material appropriately for the transformers and capacitors may need to be added across the primary windings of the input and output transformers
Hi, thank you!!! Four questions: - are you using the braided outer conductor for the coax for that turn in the first core? - if so, why not just use a wire? - for the second core, why 3 turns and 6 turns, why not just 1 and 2 ? - are those special "rf" relays? thank you!
Yes it’s braid but wire is ok too at this power level. I used the higher number of turns on the other transformer to keep the inductance high enough so it would work well at 80 meters. Those are just plain old junk box relays. Thanks… Ron
@@rontaylor9026 oh okay Appreciate reply do you have any guide how to make one for These frequencies? i want atleast a boost of 10-20 watts current have 5 watts output of Walkie talkie
Hi Ron, really enjoyed the build - thank you. I've been looking for a reliable source (non Far East) for the RD16HHF1. Kits & Parts don't list on their site. Do you have any suggestions for a supplier? Thanks. Neville, New Zealand.
@@rontaylor9026 Hi Ron. Thanks for the info. I've placed an order with RF Parts. I've ordered enough for your project and another project I like the look of ... and a couple of spares :).
NM You can also get a similar part from Mini Kits in SA ie Adelaide - Tim Fidler. THe chinese supplier, UTSource is normally reliable as in not selling dud or repackaged parts. Basically if you are buying out of PRC you should be paying 4 to five USD per part . If you pay a lot less you will be buying junk. I have bought parts from Polida in PRC with good results (leaded J310s ) but not this particular part. Tim Fidler WRE NzL. My take - RF Parts US has very high min order value and he charges top dollar too for what he sells. However he does not sell unknown parts from second tier suppliers.
@@F_Tim1961 Thanks Tim from Whangarei. I've added all the sources you mention to my list of places to check for future purchases. I must say I was very happy with my RF Parts transaction. 7 transistors including shipping etc NZD$90 and they arrived within a week of ordering.
@@nevmarr Fair enough I think you would have paid around the same out of ADL. - a quote from his site "Mitsubishi RD16HHF1 RF MosFET Small Signal N Ch 12.5v 16W 30MHz,.. Priced from AU $8.60." Bear in mind that this amplifier you are building is a NON Linear at the power levels given. It is a 16 W RMS SSB amplifier only. The bias for SSB use should be around 500mA per device ie 8W each before they even produce any RF out. The video does mention CW only. The unit would benefit from a different material of OP transformer if intended for 10m Use. Specifically type 61 material. The type 43 material was what DIZ says he used. (he calls it the MiniPIG Amplifier) However Diz says a lot of things .He calls it a 20 W linear which is practically impossible at 14V supply. The amplifier as shown will not produce 37 W out (non linear) when operated within Mitsubishi Specs for the devices. However , It might produce about 65/45 x 20 W = 28.8 W out based in typical efficiencies Class C to B ratio and run on a 24 Vcc supply, which is above Mitsi max for infinite bad VSWR resistance by about 9 V. This is based on real word results for a Ukr pre built amplifier and comparison with data from Kees Talen (USA) who used to sell a kit with a similar topology.
Hello Matias. Actually, I set each transistor for a quiescent drain current of 30 mA. They aren't un-even. With bias starting at zero, there is some idle current. Write that down. Then, with no input RF, key the amplifier on and adjust one of the bias pots slowly until the total current rises by 30 mA. Then adjust the other bias pot until the current rises ANOTHER 30 mA. The new total will be 30 + 30 + unbiased starting idle current. I hope this makes sense. 73 ... Ron
It's certainly possible, but it may take a total re-design. The IRF types have the drain lead on the tab of the device so they have to be isolated from ground. They also exhibit different impedances, so the turns ratios of the transformers may need to be modified. And if for example you use IRF-510s, they will not make as much power as the ones I used when running at 12 volts, and will need to be run at 18-20 volts. Finally, I believe the devices I used are much less susceptible to damage due to high SWR than are the IRF devices. Hope that helps ...Ron
I built it for use with a 40 meter CW transceiver and yes it works well. It does exhibit gain reduction with increase in frequency but should be usable at least from 80 meters to 17 meters.
You could do something similar but you'll need to make changes. The FETs should be changed to the RD15HVF1 instead and the input and output transformers will probably need to be altered. Of course on those frequencies you can drop the relays as TR switching is not going to be necessary
Possible error - the 2n3906 PNP looks strange - should not the Emitter be mirror imaged on the schematic ?? No I see - all good - it is very hard to read on my screen.. that's my excuse on the closer pin of the 78 L it looks like the word OUT but it is very faint. TEF
Actually it is correct Tim. The 12v enters at the emitter and appears at the collector when the PNP transistor turns on and supplies input to the 5 volt regulator. The 5 volts out supplies bias to the transistors on transmit. That same 12 volts at the emitter is present on one end of both the relay coils. The other side of those relay coils is returned to ground by the NPN transistor to energize them on transmit as well.
@@rontaylor9026 I thought you might take up the business of the "missing" finessing capacitor ACROSS the primary which I mention below. ?As say H Granberg and many many others talk about it and it does increase power transfer but the specification depends on things like the parasit inductance of the primary of the OPT (and the secondary for that matter ) . I don't have the RF gear to measure that. The proof of the pudding is the cct will work without it but it is more efficient with it. 2. The circuit has not option for a ground -= Run PTT circuit. The most beautiful way to implement this is a BS 250 P channel Mosfet switching type. this has to be biased on with say 5 V via 1K resistor. When the cold side of the resistor is pulled low , and the gate goes low the BS 250 conducts ( up tot 700mA) That will then fire up the relays by pulling current through the drain of the Mosfet and the relay coils in //. 3. You need to make it very clear that when you use a MLCC type capacitor it is "accept not substitutes" . For example the input 0.01uF caps are going to have to pass 1 W RF each and higher power as a peak. They can be MLCC or SMica but if other subs are made the results may be ugly. New constructors may not realise this. 4. Theoretical calcs show that if you have two off To220 close to the ctr of a HSink then you are FAR better off thermally than splaying out as you did. It seems counterintuitive but it is true and is supported by quality H Transfer theory books. So if each of the RD16HHs is rotated then the board layout becomes a lot more beautiful and in fact thermally superior.
@@rontaylor9026 Another question - do you have any feel for why Diz used the FT 50-61 for the phasing choke. Just about every other design of this type for the RD 16HH uses a type 43 ring core for the P Choke, to drive up the inductance . I know the original design was for 28 Mhz but you are using it at much lower frequencies. I have always wondered too why Diz used type 43 material for the output transformer - given the frequency. It has occurred to me that the material types were transposed because he never went into serious production with the design, where he error, if it was, would have been caught. Type 43 of course on the OPT would be fine for 3.5 and 7 Mhz. Just curious TE Fidler There's no point in asking Diz about this as he no longer has any interest in the design.
@@F_Tim1961 Hi Tim. Thanks for addressing the filter question. I missed that. You may be right about the swapped ferrite types. I’ll have to play with that someday. I was happy with the output on the lower bands where I tend to use it so didn’t try any efficiency modifications. Ron
@@rontaylor9026 The single mode impedance of the phasing choke needs to be about 5 x the impedance of the final RHS resistive load as reflected through the OPT to the primary. This is because a closely coupled Ph. choke doubles Vcc as seen by the PA mosfet which is the same as the reflected load being halved as seen by the phasing choke. So the Real load as seen in parallel with the single mode Z of the Ph Choke is in that case on tenth of the Ph. choke's 5 x Z . Now a FT 50-61 with 8 t is a bout a 4.2 uH inductor (if it is slow wound to cover 330 deg say it will be lower and by perhaps 30 percent). That 4.2 at 10 Mhz comes in at 240j ohms. If the amp is ouputting 16 W at supply 14 V the output Z is about 18 R. so we want at least 90 j as the single mode Z of the Ph Choke. That means the Ph choke as built is only Just acceptable at 3.5 Mhz in terms of the current that will just be shunted and not appear in the load. As soon as the power out increases to 25 W or 30W then it is not acceptable. The choke winding table I used assumes that the Cu is in intimate contact with the core and for a bifilar wound core, this is not strictly true so full L will not be achieved. The OPT core of type 43 is correct for the lower bands but is a lousy choice at 14 Mhz and above. Tables on page apprx 90 of the old Fairrite cat 14 show this very clearly. TE Fidler
Thanks for this, I learnt a lot and the individual filters is a great idea.
You may be able to improve the gain and output power by shortening the length of the MOSFET Source lead, soldering it closer to the ground plain, especially on the higher frequency bands. At 7MHz, the design is capable of ~17.5dB Gain, ~55W at about 48% efficiency. Idq was set to 10mA/device for best two tone linearity. Thanks for sharing your home brew construction of the amplifier. 73, Mike, WA2EBY
Solid good old RF construction technique!
vy 73, HB9BTL
Great video! Thanks for making the extra effort to document the construction and share it.
Nice little project. I would have used a pair of slightly bigger devices... go for 70W but what you have there is a nice compromise between supply current and heatsink size etc
Hello Ron! Enjoyed watching your DIY project, Thank-you.
I observed couple of points, 1) The binocular core names on schematics are marked wrongly, names just need to be swapped.
2) At 20 mA bias current, i could observe some crossover distortion, anyway you did increase the bias to 30mA. It probably would have fixed the distortion too.
Thanks again!
Best regards,
bp
(I am just a DC to RF enthusiast, not a ham)
Hi and thanks for your comments. I re-checked the schematic and found that the binocular core names are indeed correct. The smaller one on the input is the BN43-202 and the large one on the output is the BN43-3312.
Hello Ron, Glad to hear from you. Yes! you are indeed right. Sorry, my mistake. I probably made that mistake not realizing that the dimensions were actually given in inches for larger core and millimeter for smaller core on Amazon website. 73's
Best regards,
Bharath
That is what I call a really informative practical video. I have a few of those transistors. I intended to use them with a construction similar to the book 'building a transceiver' by ei9gq. I was impressed by the results from your construction, so I think I will try both. Your construction seems to use simpler transformers, so I will kick off with yours. I will be interested to see how your construction would work with more bias up to around the linear position for ssb. I will also try using a simple 50 ohm attenuator padder at the input and see what happens. Thanks for the video. Keith G0ODU
Good video, I have worked on several tube amplifiers but never tried my hand at a solid state one.
This is an answer on a filter question which is hidden as a reply below
you either go to QRP labs and buy the filter from there or you use Motorola Application note 267 which is still available on the web. This covers several filter topologies including the Pi type filter which I believe is Net 3 in that note. Look it up and work it out.
Zout for the amplifier (depends on mode - the above is oriented to CW mode but it will operate as a linear at lower power out.
THe output impedance from the amp to match is approximately 2 * (Vcc - 2 ) squared / Pout where Pout is say 30 W for CW use and 16 W for SSB use using much higher bias approx 600 mA total quiescent current through the sum of the Drains. If you are designing a two PI filter you should chose sqrt ( output Z x 50) and match into that using the equations given in the application note. and using a filter Q of either two or three - your choice. Q is defined inn the application note. The RHS filter on the output can be designed using the tables in the Appn note as you know the RHS termination is fifty ohms. You need to have basic EE knowledge to do this ie you need to know how to go from impedance to L and back again and how to do the inductor turns calculations.
Those are covered in the toroid tables of The Toroid king - search on line. Such filters can only be approximate as there is tolerance on the properties of the iron powder material. A size fifty toroid will be adequate for a 30 W rf filter. The capacitors should be either multilayer monolytic or Silver mica. The latter are preferred for low loss. The Toroid king sells them. If you want a seven element filter you have to extend the above description.
THe application note shows how you can subtract Co of the respective Mosfet from the calculated filter C1 capacitor. This is not applicable for a fully isolated OP transformer as used here. It is applicable for class C choke type circuits which is a whole different matter. Basically, the Appn note assumes single ended RF construction so you have to show some initiative when applying it to a push pull RF amplifier.
BTW you may wonder where I got the 16 W above from . Kees Talen used to offer a pretty sophisticated kit with a similar design and it used identical Active devices from MItsubishi. His considered max Pout with full Fcc Compliance was 16W out. This was on the basis of a 2 Pi filter or two section filter following the power amp and 14 V supply. If you take the voltage up to say 16 , which is outside the manufacturer's specification for these devices and if you over drive them (he used just 600mW in) whereas Ron uses one watt, you may well get of order 30 W out on CW but you may not be FCC compliant. The FCC requirements are considerably harsher for transmitters over 5 W out and some people don't allow for this.
In addition of you wire the output transformer for 2T out you will have a nominal output Z of 72 ohms (running the amplifier for 16 W out nominal) and if you treat that as 50 R you get only three percent mismatch loss if you design the filters as 50 R in and 50R out. I think Diz specified 1:3T (I have his original schematic just to try to get a bit more efficiency out of the filter by running it as a higher voltage input device in the first two sections.
Regards TE Fidler Nz.
Increasing power - a Japanese constructor made a similar amp years ago but called it a 40 W amp. IN fact it is about a 34 W amp for SSB use when fed on 14V supply. There was about twice as much ferrite in his core. He used a copper tube core but you can get a similar effect by using teflon insul coax cable with all the braid in contact.
Ok so the only other major change was he connected one 100 R x 1W resistor to each gate , wound 6T of ECW around it and brought it out to a common point per pair of RD 16HHF1 mosfets, enabling two to operate as if one device. They were biased in common (the assumption is that the came out of the same production batch ). THe point of the inductor thus created was to stop ringing between the gates.
The Softrock 50 which is an ultra reliable kit amp uses a three plus three arrangement to claim 50 W out for SSB - that is the Japanese claim. (The Japanese is likely getting probably more like 90 W PEP).
It s food for thought when laying out such a board - allow for expansion. Such an amp would require about 1.8 W at the mosfet gates to achieve full output.
TEFidler ...
Recently found an old BLF578 ldmos device that had failed. Measuring it i noted the failure on one side only. I threw it at a PCB along with a few variable Cs and a few inductors and wham ! 400W out with 1.5w in at 40v on the Drain, 80% efficiency. Based the circuit on the MRF174 data sheet. Great use of a device that was destined to be scrapped. Might end up on 6m as a ft8 PA. 73 from ZL.
Very clean construction technique. I like very much the method to create pads inside a ground plane.
Also, it will be easy to change the transistors when burned by a SWR accident, as we all do ;-)
Thanks,
F1SOC
That's true, but these transistors are supposed to be able to take infinite SWR without harm. I haven't tested that though.
@@rontaylor9026 Hi Ron, thanks for the video, infinit swr is a property of the transistor? Or of the construction itself?
Nice build Ron. Not enough home brewing like the good old days. I sure miss Ham Radio magazine.
Excellent video Ron!! Thank you, your English is so clear, speed is rigth. I really appreciate your video !!!!
Thank you for you video, hope you could make more video like this for example HF transmitters, receivers, etc.
Great video Ron; really enjoyed the step-by-step construction and test walk-thru. For isolated pads I bought a punch-tool from Harbor-Freight, filed off the punch center-tips and use it to punch-out small round pads out of scrap PCB that then I super-glue down to the main PCB.. The reason I adopted this was that I could make tiny island-pads that are easy to fit into tight spaces and therefore really able to shrink the size of my overall circuits...
Excellent work.
Thank you for your generosity in sharing your knowledge.
Respects from Argentina.
Thanks Nestor. Glad you enjoyed it
Excellent job Ron! I have a few of those Mitsubishi transistors somewhere. I might try to make this amp... 73 VK4QP.
The real ham radio spirit 👍🤩
Excellent work!
Thank you for your generosity in sharing your knowledge!
Hi Ron, really great video... Thanks for the clarity of explanation and the effort you put for this. 73 de Vic, I2GPT
This is very nice. Certainly I have learned some things that will be useful. I have not watched a deadbug circuit put together. Even simple things for me are new. Thanks!
Ron, put more videos, they are very instructive. 73 of PS8AJR BRAZIL.
Мудрость древних, навесной монтаж над полностью покрытым медью двухсторонним текстолитом, усилитель класса AB, трансформаторы на марганец -цинковых сердечниках из сетевого фильтра, комутация на реле со схемой детектирования сигнала на диодах
Very interesting brother. If I was able to build something like this, I doubt it would look that clean. Keep up the good work.
Thanks Bro. You can do it. Just jump in and swim! .. 73
thank you sir
Excellent video - I shall definitely be having a go at this. Thanks.
Have fun! Hope it works for you.
Thank you for all the explanations. This was really helpful in understanding the circuits needed and how everything ties in. Now I'm not so scared to try to make my own. 73!
Thanks for the comments Cam. Glad it was useful. And, yes, the only thing to be scared of is burning your finger with the soldering iron..73
Great video and explanation!
Your "digital readout wattmeter" seems interesting, do you have information about it?
Beautiful circuit.
Excellent video. Thanks for sharing.
I think that ieland cutter is very beautiful,
unfortunately we cannot buy them here in the Netherlands!
Healthy and Friendly Greetings from the Netherlands!
Rob
Great amp! And a great presentation! Tks es 73!
Thank you very much. Very interesting material and enough to get me out of my analysis paralysis and building stuff! 73, KC1PAN.
Awesome, instructive, good info
Gracias por compartir su conocimiento muy bien explicado.
saludos desde CD. Juárez Chihuahua MÉXICO!.
Nice job. 73’s KG4LAH
Awesome video
Ron very nice thank you!
Great video sir! Full of information.
Thank you a lot for sharing this Knowledge
i enjoyed watching, thank you for the effort and uploading the video, much appreciated.
Great video 👍
Thanks ❗
Thank you so much for sharing this knowledge
Glad it was helpful!
The digital wattmeter is adapted from an article by W8TEE and AC8GY titled “Build an Inexpensive 150 Watt Dummy Load with Wattmeter” published in the Nov 2018 issue of QST magazine.
Should i use this in usb 11 meter band?how much power on 27mhz?
Thank you so much for that info
@@nicolalamotta9554if it is legal in ur country
What happens if you try to use on higher frequency? Still work just less output? Distortion? Nothing? Im curious. Thanks for the video
very nice! 73
Thanks Ron Nice work
Nice job, Ron. I didn't know you had any videos out. This just showed up in my recommended list! 73 de Randy WB0SMX
saludos desde sevilla (spain) , good work my friend .
Is they the XYL playing the piano in the background? Thanks for the video Ron. +/-~2years later.
Must be her. I didn’t really notice it before you mentioned this.
Very nice build and explanation video. Glad to see another homebrew project using real RF Mosfets like the RD series instead of IRF switching Mosfets. Excellent work brother. 73. KC3MOP
Thanks for the comment. Sorry I was slow responding. Glad you enjoyed it
I built my first successful RF Amps ever with these RD series Mitsubishi MOSFET Transistors they are awesome and require very low drive to make WATTS even my BK Precision Arbitrary Waveform Generator will drive them!
Do you have a link to that rotary saw with spring loaded tip? I could use one of those.
this video worth more thumbs up
Thanks Mike. Glad you liked it.
Thanks for making this!
What a nice vidéo !!
Why are RF amplifiers so simple compared to audio ones ?
They are to an extent. They have their specific nuances
Pushpull Linear 👍
very good video : thaks for sharing your experience.
Your explanation is very accurate and it will so simple for me to buildup one. The point for me is the output filter since I do not have special instruments available here at my station: could you please post the data to build a power low pass filter for 80, 40 , 20 and 10 meter?
Thanks for all.
you either go to QRP labs and buy the filter from there or you use Motorola Application note 267 which is still available on the web. This covers several filter topologies including the Pi type filter which I believe is Net 3 in that note. Look it up and work it out.
Zout for the amplifier (depends on mode - the above is oriented to CW mode but it will operate as a linear at lower power out.
THe output impedance from the amp to match is approximately 2 * (Vcc - 2 ) squared / Pout where Pout is say 30 W for CW use and 16 W for SSB use using much higher bias approx 600 mA total quiescent current through the sum of the Drains. If you are designing a two PI filter you should chose sqrt ( OUtput Z x 50) and match into that using the equations given in the application note. and using a filter Q of either two or three - your choice. Q is defined inn the application note. The RHS filter on the output can be designed using the tables in the Appn note as you know the RHS termination is fifty ohms. You need to have basic EE knowledge to do this ie you need to know how to go from impedance to L and back again and how to do the inductor turns calculations.
Those are covered in the toroid tables of The Toroid king - search on line. Such filters can only be approximate as there is tolerance on the properties of the iron powder material. A size fifty toroid will be adequate for a 30 W rf filter. The capacitors should be either multilayer monolytic or Silver mica. The latter are preferred for low loss. The Toroid king sells them. If you want a seven element filter you have to extend the above description.
THe application note shows how you can subtract Co of the respective Mosfet from the calculated filter C1 capacitor. This is not applicable for a fully isolated OP transformer as used here. It is applicable for class C choke type circuits which is a whole different matter. Basically, the Appn note assumes single ended RF construction so you have to show some initiative when applying it to a push pull RF amplifier.
BTW you may wonder where I got the 16 W above from . Kees Talen used to offer a pretty sophisticated kit with a similar design. His considered max Pout with full Fcc Compliance was 16W out. This was on the basis of a 2 Pi filter or two section filter following the power amp and 14 V supply. If you take the voltage up to say `16 , which is outside the manufacturer's speck for these devices and if you over drive them (he used just 600mW in) whereas Ron uses one watt, you may well get of order 30 W out on CW but you may not be Fcc compliant. The FCC requirements are considerably hasher for transmitters over 5 W out and some people don't allow for this.
In addition of you wire the output transformer for 2T out you will have a nominal output Z of 72 ohms (running the amplifier for 16 W out nominal) and if you treat that as 50 R you get only three percent mismatch loss if you design the filters as 50 R in and 50R out. I think Diz specified 1:3T (I have his original schematic just to try to get a bit more efficiency out of the filter by running it as a higher voltage input device.
Regards TE Fidler Nz.
@@F_Tim1961 very useful response: thanks.
@@ambroribo6079 If you go to the Toroid King Site you will find in his kits that he sells LP Filter PCBs If you want just two frequencies then two off DPDT relays (use contacts in parallel ) will allow switching between them. Don't try it with just one relay - there will be too much signal bypassing occurring. The single relay approach has been used in cheap Indian kits and it turns out to be a bit of a disaster technically. TEF
nice work . EP4TAT
Voice reminds me of Clint Eastwood.
Hi Ron, I am seeking an RF amplifier that can input 250mw and boost to 3 watts+. Frequency range is 470-480mhz. Do you have any recommendations? Is this something you could make? Thanks!
You might try one of the Toshiba or Mitsubishi RF power modules sold by RF Parts. One they currently have that might meet your needs is the Toshiba S-AU99H. Good luck
@@rontaylor9026 Thanks for this! I am actually looking for a complete unit. 12-15v input. BNC or SMA preferred. Thoughts?
@@Voiceguitar Sorry. I don't have any recommendations for that... 73 .. Ron
Ron, Where did you get the pad cutter?
I got that pad cutter at Harbor Freight. Smaller ones can be purchased from QRPme
Excellent video. Great for a beginner like me. Would this amp work on top band? 160m? If not, what would need to be changed to make it work at lower frequencies? Thanks!
It should work there but you need to choose your ferrite material appropriately for the transformers and capacitors may need to be added across the primary windings of the input and output transformers
new sub from new zealand
Hi, thank you!!! Four questions:
- are you using the braided outer conductor for the coax for that turn in the first core?
- if so, why not just use a wire?
- for the second core, why 3 turns and 6 turns, why not just 1 and 2 ?
- are those special "rf" relays?
thank you!
Yes it’s braid but wire is ok too at this power level. I used the higher number of turns on the other transformer to keep the inductance high enough so it would work well at 80 meters. Those are just plain old junk box relays. Thanks… Ron
@@rontaylor9026 thank you!!!!
Good video , But your audio is so quiet it's barely audible .
That's because the amplifier was not built yet.
88-108mhz
Para que feceucn es ese amplificado/ For to frecuency is this amplifier?
This amplifier was intended for use on Amateur Radio frequencies in the HF band. Primarily in the lower part of the HF Band from 3.5 to 14.350 MHz
Would you like a printed enclosure or are you putting in in something metal?
Planning to use it in a larger project. Thanks for the offer
Ron, thanks for the video, good luck to you! Greetings from Russia!!!
Can it be used with VHF UHF ?
frequency 144mhz and 400mhz Plus?
No, the gain rolls off after HF.
@@rontaylor9026 oh okay Appreciate reply
do you have any guide how to make one for These frequencies?
i want atleast a boost of 10-20 watts
current have 5 watts output of Walkie talkie
Apakah itu Untuk Frekuensi 40M Master 🙏
Hi Ron, really enjoyed the build - thank you. I've been looking for a reliable source (non Far East) for the RD16HHF1. Kits & Parts don't list on their site. Do you have any suggestions for a supplier? Thanks. Neville, New Zealand.
Hi Neville. Those transistors came from RF Parts here in the U.S. Good luck. Thanks for the nice comment ...
@@rontaylor9026 Hi Ron. Thanks for the info. I've placed an order with RF Parts. I've ordered enough for your project and another project I like the look of ... and a couple of spares :).
NM You can also get a similar part from Mini Kits in SA ie Adelaide - Tim Fidler. THe chinese supplier, UTSource is normally reliable as in not selling dud or repackaged parts. Basically if you are buying out of PRC you should be paying 4 to five USD per part . If you pay a lot less you will be buying junk. I have bought parts from Polida in PRC with good results (leaded J310s ) but not this particular part. Tim Fidler WRE NzL. My take - RF Parts US has very high min order value and he charges top dollar too for what he sells. However he does not sell unknown parts from second tier suppliers.
@@F_Tim1961 Thanks Tim from Whangarei. I've added all the sources you mention to my list of places to check for future purchases. I must say I was very happy with my RF Parts transaction. 7 transistors including shipping etc NZD$90 and they arrived within a week of ordering.
@@nevmarr Fair enough I think you would have paid around the same out of ADL. - a quote from his site "Mitsubishi RD16HHF1 RF MosFET Small Signal N Ch 12.5v 16W 30MHz,..
Priced from AU $8.60."
Bear in mind that this amplifier you are building is a NON Linear at the power levels given. It is a 16 W RMS SSB amplifier only. The bias for SSB use should be around 500mA per device ie 8W each before they even produce any RF out. The video does mention CW only. The unit would benefit from a different material of OP transformer if intended for 10m Use. Specifically type 61 material. The type 43 material was what DIZ says he used. (he calls it the MiniPIG Amplifier) However Diz says a lot of things .He calls it a 20 W linear which is practically impossible at 14V supply.
The amplifier as shown will not produce 37 W out (non linear) when operated within Mitsubishi Specs for the devices. However , It might produce about 65/45 x 20 W = 28.8 W out based in typical efficiencies Class C to B ratio and run on a 24 Vcc supply, which is above Mitsi max for infinite bad VSWR resistance by about 9 V. This is based on real word results for a Ukr pre built amplifier and comparison with data from Kees Talen (USA) who used to sell a kit with a similar topology.
Would this work on lorawan devices
No, this is designed for HF frequency bands, for Amateur radio only. Sorry
Very nice and clear instructions. I'll build this amplifier later this year. Thanks de OZ1BXM Lars
hola veo que a un transistor le dio 80ma y al otro 20ma total 100ma quedaron desparejos
Hello Matias. Actually, I set each transistor for a quiescent drain current of 30 mA. They aren't un-even. With bias starting at zero, there is some idle current. Write that down. Then, with no input RF, key the amplifier on and adjust one of the bias pots slowly until the total current rises by 30 mA. Then adjust the other bias pot until the current rises ANOTHER 30 mA. The new total will be 30 + 30 + unbiased starting idle current. I hope this makes sense. 73 ... Ron
Hi! Can I replace the mosfets with IRF types?
It's certainly possible, but it may take a total re-design. The IRF types have the drain lead on the tab of the device so they have to be isolated from ground. They also exhibit different impedances, so the turns ratios of the transformers may need to be modified. And if for example you use IRF-510s, they will not make as much power as the ones I used when running at 12 volts, and will need to be run at 18-20 volts. Finally, I believe the devices I used are much less susceptible to damage due to high SWR than are the IRF devices. Hope that helps ...Ron
@@rontaylor9026 Thanks a lot!
What about irf520 device and what voltage would they need?
What is rf frequency range does it work?
I built it for use with a 40 meter CW transceiver and yes it works well. It does exhibit gain reduction with increase in frequency but should be usable at least from 80 meters to 17 meters.
I Want fm 88-108mhz
Thanks Ron for the explained Rf project.
Can it work on frequency 80mhz to 120mhz?
You could do something similar but you'll need to make changes. The FETs should be changed to the RD15HVF1 instead and the input and output transformers will probably need to be altered. Of course on those frequencies you can drop the relays as TR switching is not going to be necessary
Would this work with mrf477's?
No. MRF477 is a NPN BJT transistor. The RD16HHF1s that I used are N-channel MOSFETs.
@@rontaylor9026 Ah, kk, ty 👍
Possible error - the 2n3906 PNP looks strange - should not the Emitter be mirror imaged on the schematic ?? No I see - all good - it is very hard to read on my screen.. that's my excuse on the closer pin of the 78 L it looks like the word OUT but it is very faint. TEF
Actually it is correct Tim. The 12v enters at the emitter and appears at the collector when the PNP transistor turns on and supplies input to the 5 volt regulator. The 5 volts out supplies bias to the transistors on transmit. That same 12 volts at the emitter is present on one end of both the relay coils. The other side of those relay coils is returned to ground by the NPN transistor to energize them on transmit as well.
@@rontaylor9026 I thought you might take up the business of the "missing" finessing capacitor ACROSS the primary which I mention below. ?As say H Granberg and many many others talk about it and it does increase power transfer but the specification depends on things like the parasit inductance of the primary of the OPT (and the secondary for that matter ) . I don't have the RF gear to measure that. The proof of the pudding is the cct will work without it but it is more efficient with it.
2. The circuit has not option for a ground -= Run PTT circuit. The most beautiful way to implement this is a BS 250 P channel Mosfet switching type. this has to be biased on with say 5 V via 1K resistor. When the cold side of the resistor is pulled low , and the gate goes low the BS 250 conducts ( up tot 700mA) That will then fire up the relays by pulling current through the drain of the Mosfet and the relay coils in //.
3. You need to make it very clear that when you use a MLCC type capacitor it is "accept not substitutes" . For example the input 0.01uF caps are going to have to pass 1 W RF each and higher power as a peak. They can be MLCC or SMica but if other subs are made the results may be ugly. New constructors may not realise this.
4. Theoretical calcs show that if you have two off To220 close to the ctr of a HSink then you are FAR better off thermally than splaying out as you did. It seems counterintuitive but it is true and is supported by quality H Transfer theory books. So if each of the RD16HHs is rotated then the board layout becomes a lot more beautiful and in fact thermally superior.
@@rontaylor9026 Another question - do you have any feel for why Diz used the FT 50-61 for the phasing choke. Just about every other design of this type for the RD 16HH uses a type 43 ring core for the P Choke, to drive up the inductance . I know the original design was for 28 Mhz but you are using it at much lower frequencies. I have always wondered too why Diz used type 43 material for the output transformer - given the frequency. It has occurred to me that the material types were transposed because he never went into serious production with the design, where he error, if it was, would have been caught. Type 43 of course on the OPT would be fine for 3.5 and 7 Mhz. Just curious TE Fidler There's no point in asking Diz about this as he no longer has any interest in the design.
@@F_Tim1961 Hi Tim. Thanks for addressing the filter question. I missed that. You may be right about the swapped ferrite types. I’ll have to play with that someday. I was happy with the output on the lower bands where I tend to use it so didn’t try any efficiency modifications.
Ron
@@rontaylor9026 The single mode impedance of the phasing choke needs to be about 5 x the impedance of the final RHS resistive load as reflected through the OPT to the primary. This is because a closely coupled Ph. choke doubles Vcc as seen by the PA mosfet which is the same as the reflected load being halved as seen by the phasing choke. So the Real load as seen in parallel with the single mode Z of the Ph Choke is in that case on tenth of the Ph. choke's 5 x Z . Now a FT 50-61 with 8 t is a bout a 4.2 uH inductor (if it is slow wound to cover 330 deg say it will be lower and by perhaps 30 percent). That 4.2 at 10 Mhz comes in at 240j ohms. If the amp is ouputting 16 W at supply 14 V the output Z is about 18 R. so we want at least 90 j as the single mode Z of the Ph Choke. That means the Ph choke as built is only Just acceptable at 3.5 Mhz in terms of the current that will just be shunted and not appear in the load. As soon as the power out increases to 25 W or 30W then it is not acceptable. The choke winding table I used assumes that the Cu is in intimate contact with the core and for a bifilar wound core, this is not strictly true so full L will not be achieved. The OPT core of type 43 is correct for the lower bands but is a lousy choice at 14 Mhz and above. Tables on page apprx 90 of the old Fairrite cat 14 show this very clearly. TE Fidler
เปลี่ยนเป็นภาษาไทยได้ใหม ไทยแลน
Thanks. 73s kb9jzk
You may be a good RF designer and engineer-please work on a audio PA so we can hear what your talking about
I’m sure it’s a good video but I can’t hear you volume all the way up so yeah your audio is way too low
Turn on closed captions
@@haxboi5492 sorry phone and headset. And eyes no 20yo anymore