HFU HF Underground
Technical Topics => The RF Workbench => Topic started by: OgreVorbis on November 12, 2018, 0104 UTC
-
As I am building the transmitter discussed in this thread, I decided to document the progress on my blog. If you're looking to build your own class D, check it out. There is a massive archive of PDFs and schematics in there as well as an in depth explanation. I am not as experienced as some here, but I always update any inaccuracies. I also approach the topic from a novice's perspective.
http://dosaidsoft.com/wp/category/radio/class-d/ (http://dosaidsoft.com/wp/category/radio/class-d/)
Original message:
I am trying to gain a better understanding of a class D transmitter, and while I do understand most of the concepts, I need a little clarification.
So I am probably going to start with this design: http://www.w1vd.com/40M375WclassDRev2.0.pdf (http://www.w1vd.com/40M375WclassDRev2.0.pdf)
Yes, I know it's old. And I think there might be some parts left out. It looks like there is an unlabeled resistor at the gate of each FET.
First thing I am not clear on is if the exciter needs to be biphasic in order to make a class D or E amp. You can see in the schematic I linked that it has phase 1 and 2 input and it says 12V, but how many amps does it need (on the input)? Can I build half of the amp and use a normal single output DDS module? I am not sure how to make a two phase exciter. I think class D is more broadband, so I want to go with that. I'm still not even exactly sure what the difference is between D and E despite them being described so often. I never find a clear difference.
So when I have the amp and exciter built, I think I'll start trying to drive it with a mod transformer instead of PWM just to start because it's easier. So how many ohms do I need to match right at the drain of the FETs? Is it 50? No, probably not because that's after the matching right? So if I use a generic audio amp, what transformer ratio do I need? I think I'll use a microwave oven transformer. I have one that has all coils removed, so it's just a bare core I can wrap around.
If I need a two phase exciter. Where can I find one? I do know about the classeradio.com guy, but he only sells the PWM and not the exciter. He didn't respond anyway. Probably because I'm not ham.
And, yes, Stretchyman, I know you're here too. Your transmitter is tempting, but I don't want to just buy it without understanding how it works because that gets annoying in the long run.
Would you mind explaining a little to me? Mostly on how your DDS module works (I assume you use DDS). Is it a two phase thing or is it essentially just a 9850 and a PIC on a board - like what you can get on eBay? I would really like to see the schematics and also if you can clarify how the exciter module works. If you'd like a small payment for your schematics, I'd be willing. I'd just buy the whole thing if you can tell me how to change it for more bandwidth.
If I was building it myself I'd have used Class D for more bandwidth, so it's not perfect from my standpoint. What about your amp makes it Class E and could it be changed to D? I've seen a class D 100W amp that was using two SiC devices and it works 3-8 MHz. What makes yours only 200 Khz range?
I know that's a lot of questions and I am being very critical of the design. I know it is better than most others out there given that it's using the newest GaN and drivers. I hope you understand I'm am trying to learn w/o just buying stuff and wiring it up. That's too easy for this hobby and it makes it boring.
-
First off, Class D and D/E or E are not the same. Class D/E or E amps are inherently narrow band affairs, because they operate correctly only over a small range of frequencies, usually less than 5%. Class D amplifiers operate over a wider range, but have a highly distorted output that must be filtered before being radiated. Class D amplifiers also become much less efficient and harder to drive beyond 3 MHz, which is why most designs beyond 80 meters are Class E or similar.
As to the W1VD stuff, his designs are similar to Class E. Note the resonant tank between the two halves of the amplifier. This tank network is responsible for the narrow band characteristic of the amplifier, but also allows it to output a much cleaner spectrum than a class D amplifier. The unmarked 'resistors' in the schematic between the gates and the fet drivers are actually striplines. I've found that driving the fets through a low ohm resistor, usually around 3.3 ohms, dampens the gate network and reduces overshoots. The drive current required will depend on what fets you want to use, and also what drivers. There are a lot of variables here, but at 40 meters, plan on 400mA per driver/fet combination when driving a fet with 1000pF input capacity with 15V. In practice this is low, 18-20V is more common with SiC devices, but with regular mosfets 12V is fine. I use IXDD614 drivers with C2M0160120D SiC's and seem to be getting around 93% efficiency when everything is happy, about 87% DC to RF efficiency with the PWM modulator. I also got rid of the 4:1 output balun in W1VD's design and chose to use a simple 1:1 ac coupled balun on the output to keep the current levels more reasonable, at the expense of higher required B+ and drain voltage ratings. I also borrowed the phase splitter circuit from a Nautel NX series transmitter to derive the push pull drive from a single ended source.
Are you sure you wouldn't rather just buy something to get started?
+-RH
-
Good to see the interest....
Would you mind explaining a little to me? Mostly on how your DDS module works (I assume you use DDS). Is it a two phase thing or is it essentially just a 9850 and a PIC on a board
Yes, just a PIC driving an AD9850 using the I/Q O/Ps driving SOIC FET drivers driving GaN FETs
if you can tell me how to change it for more bandwidth.
Er, no... it's class E and therefore narrow banded, 200KHz is loads anyway!
What about your amp makes it Class E and could it be changed to D? I've seen a class D 100W amp that was using two SiC devices and it works 3-8 MHz. What makes yours only 200 Khz range?
Same as above...Please point us towards the design you mention......
BTW, stay away from using some rewound transformer, PWM is simple, I can send you a PCB, 1 inch square, will mod ANY voltage and upto 10A, you'll just need to make a PWM filter, 2 caps and 2 toroids.
Simples....!
Str.
p.s. I'm not selling or offering the schematics, the design is simple as and the same as ALL the other class E designs out there, I just use modern devices, not those crappy 11N90's they keep wibbling on about, they'd rather use valves anyway, bless 'em!
-
For anyone interested - Information/Schematic Diagrams & more!
posted on my "archive.org" pages!
View Online-/ or Download
https://archive.org/download/Transmitters-Antennas/TX%20TRANSMITTERS/ (https://archive.org/download/Transmitters-Antennas/TX%20TRANSMITTERS/)
André
CoolAM Radio - Shortwave
-
Thanks for your responses.
OK, so I've made some progress reading some more Class D/E schematics I could find out there. I came to find that the FAT5 from shortwaveradio.co.uk has a huge PDF with a ton of info. None of the other sites/documents I could find were nearly as thorough.
So I've compiled some info and made a schematic of my own from the bits I could find.
It is linked at the bottom. Please let me know if you see any errors. Some things I'm not clear on:
The capacitors on the drains - are they correct?
Not sure why they used 8V regulators, wouldn't 12V be better and more standard? The datasheet on these drivers also shows a slight improvement in switching speed with more volts.
I want to keep it simple to test it, so I want to try and use the old fashioned mod transformer and audio amp. I should probably know this, but: What turns ratio would I need?
https://drive.google.com/open?id=1mbt60zDB50rt65chP51dezb1KN28foh5 (https://drive.google.com/open?id=1mbt60zDB50rt65chP51dezb1KN28foh5)
-
The topology to me seems odd, but then again I've been stuck in my own little world the last few years. How much power are you aiming for? One of the problems with class D is the ridiculous supply voltages required to get any kind of decent power. This is where (in my mind) CMCD/class E shines; the peak voltage across the drains is Pi times supply volts. To get the same voltage swing requires a 16:1 output transformer, or 4x more supply. I need around 79Vdc to achieve ~550W carrier. To achieve the same thing with class D you will need 250Vdc. If your output transformer is wound 1:16, only 62V is required.
+-RH
-
RH is right. It will however 'work'.
So build it and let us know how you get on.
Str.
-
So I've made some progress. I decided to just make a pro PCB instead of trying to make a dead bug. The PCBs from china are cheap enough anyway.
So here is what I made. It is based on the schematic I posted earlier, but I added spaces for 4 extra FETs. Let me know what you think or if you see any errors.
I created the board in Sprint Layout which I would highly recommend. CAD software is overly complicated for these types of designs imo (and I don't know how to use it :).
I'm curious to see how well this works compared to class E. I know it won't be quite as good, but I'm more interested in bandwidth than obtaining max power from each device.
https://drive.google.com/open?id=1p_pzBloYqnlwnSdilyqJNou5V1moF8OQ (https://drive.google.com/open?id=1p_pzBloYqnlwnSdilyqJNou5V1moF8OQ)
I decided to get two sets of mosfets: C2M0160120D and C2M0280120D
One has less capacitance, but lower current handling. As I understand it, the lower the capacitance, the higher the potential switching speed. I'm unsure of what is better though, more amps, or less capacitance. I find that the higher amp devices have more capacitance. So what is the higher priority? Amps could make up for the greater cap, but the less caps will make it faster. Not sure what's better.
-
PCB is missing a lot of holes!
Lower C devices are easier to drive but have higher on resistance.
Take your pick. The SiC devices are OK, however GaN are far superior but(5X)more $.
Below 5MHz everything is fairly easy but at 7MHz things aren't as efficient. Most of the designs will work upto 14MHz (with GaN) but losses are higher and generating good square wave drive problematic.
Good luck on your quest. However as mentioned D/E is the best method. You can cover 200-500KHz without retuning and that's enough to cover any band segment.
Str.
-
1 driver per fet will be required above 3 mhz in class D. Low RDSon is required for class D as well, I would stick to the lower R fets. I also don't like the thin and windy traced on the drains. This should be a large plane, not a trace. You will have a lot of ringing problems due to the stray inductance. Areas like this I also make sure there is no copper on the back side to create stray capacitance which will load the output unnecessarily. Same thing with the gate traces, large and fat equals lower inductance and better waveforms.
Little things like that make a BIG difference!
+-RH
-
1 driver per fet will be required above 3 mhz in class D. Low RDSon is required for class D as well, I would stick to the lower R fets. I also don't like the thin and windy traced on the drains. This should be a large plane, not a trace. You will have a lot of ringing problems due to the stray inductance. Areas like this I also make sure there is no copper on the back side to create stray capacitance which will load the output unnecessarily. Same thing with the gate traces, large and fat equals lower inductance and better waveforms.
Little things like that make a BIG difference!
+-RH
OK, that makes sense. I don't think I can fix the drain traces though. There's just not enough space on there. I think I'll attach them to a raised piece of copper clad board. I've seen this somewhere before. It shouldn't be a deal killer though, right? I mean at these frequencies, I wouldn't expect a trace like that to have any level of inductance that would cause problems. Most inductors at these frequencies use toroids, so a little loop shouldn't do much. That's just speculation though. I'm new to this and you do know more, so...
In terms of the 1 FET per driver. I did already know that, but then I saw this guy demoing his class D transmitter. It has 4 FETs and two drivers and makes 200W carrier on 40m. His 2 FET version made 100W carrier, so it doesn't seem like the drivers are limiting it. I'm also using the TC4452 and he is the lower amp TC4422.
https://www.youtube.com/watch?v=QnM2Uum-_Cs (https://www.youtube.com/watch?v=QnM2Uum-_Cs)
-
Something about this doesn't look right. He's using a MW slug on 40 meters (so much for calibration). Also, everywhere I've read said that using wire in output transformers is problematic above 80 meters due to circuit strays and coax should be used instead.
You would think a little L here and there wouldn't hurt, but it can create voltage spikes and transients that can kill your fets and other components. From scratch its taken me about three years to get to where my TX is now, and there were a lot of little details along the way that no one seemed to publish or know.
That's one of the reasons I just can't believe that the frankstein stuff over on the classeradio site actually works. None of this stuff is plug and play, and I have to agree with Stretchy, a lot of the secret sauce to this is the layout design.
Without seeing his design in action and looking at waveforms and such, its hard to know what is really going on. I suspect the efficiency is nothing to write home about, and who knows what kind of reliability your going to have.
+-RH
-
Hearing what you know, I'd be interested to see some pictures of your design. Do you have any?
If not, then maybe you could give me a general idea of the layout. Are you using PCBs? How do you make connections (other than short as possible)? etc. Or maybe just some tips or things that you did wrong during the process.
I'd really appreciate it, but I understand if you want to keep it a secret.
-
I have no problem releasing schemo's and pics, but this probably isn't the best venue for it, for reasons of liability.
Shoot me an email, and I'll show you what I'm working on.
+-RH
-
Only took me two years!
I had a lot of other designs to copy and a PCB layout to prove. That took a few goes too!
Like the man said, big fat traces.
Str.
p.s. ping me a mail toooo. Can show you the layout etc...
-
Only took me two years!
PFFT! as an RF engineer, you should have had it done in two weeks then!
+-RH
-
Indeed! ;)
I have to work as well!
However most of the time was spent doing the layout and sorting the hardware. Even punched out cheeks for the PWM inductor former. That and piddling around with a new PWM design that didn't work as expected (sortid now).
Anyway all this seems a bit mute as have offered the RF generator with no takers.
I will (eventually) have the whole thing boxed up with SWR detector etc but will double the price.
I'm happy to sell a 100W (400W pep) quad GaN RF GENERATOR for $250. Has built in AD9850 DDS with 16 programmable channels. 60V max will yield 600W and will need forced cooling. Happier passively cooled at 48V giving 400W. So Mod tranny at 24V or PWM at 48V.
This works seriously well, reports to prove!!
Str.
-
That's one of the reasons I just can't believe that the frankstein stuff over on the classeradio site actually works. None of this stuff is plug and play, and I have to agree with Stretchy, a lot of the secret sauce to this is the layout design.
+-RH
Well the proof is in the operation of the transmitters for years on 160 and 75 meters putting out big power. Especially the 72 Fet one Steve WA1QIX runs. Go down to the AM window 3873 or 3885 and listen for a week and you will find out that the "frankstein stuff" really works. It is just amazing. 8)
-
To each their own, but 72 fets....really?
three MODERN devices will suffice just fine, and prove just as reliable.
+-RH
-
Yes I'm afraid they're somewhat behind the times, they're still Toob mad on there too.
As stated, designs on there are nearly 10 years old, nowt wrong with that but current technology has only been available for a year or so.
It's time to move on.
There's no point using bolt down FET drivers and FETs with high gate capacitance.
GaN devices easily driven with a SOIC sized driver all with 5V/8V.
AD9850 will directly drive SOIC FET drivers running at 8V which will switch GaN devices running at 150V.
That's what you should be using.
One chap on the AMFOAM (sic) website is doing just that. His name is Nigel I suggest reading his posts and looking at his designs as they are a tad more modern.
Str.
-
I will not make the jump to GaN until 1. the price comes down and 2. you can get decent power out of the devices. Nothing personal, but if it can't do 1 KW + modulation with two devices, I don't want it. Decent power requires beefy devices, and the GaN stuff just isn't there yet. SiC is a good compromise, as the devices can produce decent power, and are reasonably priced.
+-RH
-
True!
However vwith GaN you can drive them without using bolt down devices, I got fed up with drilling holes! That and the GaNs have a grounded tab so no washer needed.
All I wanted 400W at 48V.
I use SiC below 5MHz, GaN for 40m.
Str.
-
Your going to have to drill holes for the fets anyway...what's two more holes?
Bottom line; different design philosophy for different requirements. Yours and mine are not the same. I went the high voltage/lower current route, you did the opposite.
For my needs, my design makes more sense, and makes use of cheaper more readily available parts I can order from Digikey, Mouser, et al without reliance on one vendor. SiC is available from Cree, Toshiba, Rohm, etc so at this time obsolescence is not a concern.
I just wish you would stop beating the drum for GaN, as what makes sense for you often doesn't make sense to everyone else. There are folks out there who will be building tube linears and modulators until they'll have the make the tubes themselves. Likewise, your going to have folks like 'QIX that think the only way to build solid state AM is their way. I think in this case, he got lucky and wants to stick with what works, until he can no longer get the parts to build them.
I have an eye on higher power, and for the time being, SiC is the way to get there....until something better comes along, at a reasonable cost.
+-RH
-
Your going to have to drill holes for the fets anyway...what's two more holes?
Bottom line; different design philosophy for different requirements. Yours and mine are not the same. I went the high voltage/lower current route, you did the opposite.
For my needs, my design makes more sense, and makes use of cheaper more readily available parts I can order from Digikey, Mouser, et al without reliance on one vendor. SiC is available from Cree, Toshiba, Rohm, etc so at this time obsolescence is not a concern.
I just wish you would stop beating the drum for GaN, as what makes sense for you often doesn't make sense to everyone else. There are folks out there who will be building tube linears and modulators until they'll have the make the tubes themselves. Likewise, your going to have folks like 'QIX that think the only way to build solid state AM is their way. I think in this case, he got lucky and wants to stick with what works, until he can no longer get the parts to build them.
I have an eye on higher power, and for the time being, SiC is the way to get there....until something better comes along, at a reasonable cost.
+-RH
Very well started Redhat.
That's one of the reasons at this point I am not going into building a class E D H rig. If a really great new design comes along I may consider it. Until then I can go from 1MHz to 30 MHz with outstanding AM and SSB audio with no hassle in minutes with minimal tuning. One of the biggest limitations for me is class E D H is just AM only not SSB. SSB gets out better and can sound as good. But I still love AM too.
-
Must ask, what have you built that covers the whole of HF, All mode?
Pray tell....
Str.
-
Must ask, what have you built that covers the whole of HF, All mode?
Pray tell....
Str.
in my opinion, it does not really matter whether you build it from scratch or modify and re-purpose existing equipment.
what matters is the meaning of your purposes behind it and whether you desire a high fidelity AM broadcast that is clean and spurious harmonic free.
in my case, i use a Yaesu FT-757GX2 (low level IF AM modulation mixer) that i modified and re-purposed extensively and exclusively for full duty 50W, hi fidelity AM broadcasting operation, it is no longer considered an "amateur" radio after the extensive modifications to it, it is strictly a 0-50W AM transmitter only.
since this radio includes all the necessary components to operate a clean and spurious harmonic free transmission, i purchased it on the cheap in a non-operative condition which i easily repaired and then modified, since i already have the original paper service manual, its preferred by me over building one from scratch.
this type of re-purpose and modification can be done easily to any solid state "amateur" radio up to the manufacture year of 1990 (the older, the easier) but with extreme difficulty to post millennium models due to having 2-3 layer PCB's with micro-traces, tiny surface mounted components and "all in one" integrated circuits (IC's) making target modifications nearly impossible.
in conclusion, there are many solid state older "hybrid" non-working "amateur" transceivers available dirt cheap on several market places that can be repaired, then modified at less than the total cost of building one from scratch.
i took this approach simply because i dont have the time or patience to design, fabricate and build one to the desired specifications.
FYI - the FT-757GX2 modified TX audio frequency response is 5Hz to 20Khz as tested but i broadcast with a frequency response of 20Hz to 10Khz using "Stereotool" as the final sound/modulation processor.
-
A another great point.
-
A another great point.
I'm definitely missing out on something here!
Ah well....neh mind!
Refering to the title of the post I don't see what relevance repurposing a ham radio has at all?
Time and patience, Yes, plugging in a ham radio, anyone can do that.
We're talking about design, which you make clear you have absolutely no idea of.
I've spent years trying to get that right, I'm not there yet but will continue to try.
Str.
-
A another great point.
I'm definitely missing out on something here!
Ah well....neh mind!
Refering to the title of the post I don't see what relevance repurposing a ham radio has at all?
Time and patience, Yes, plugging in a ham radio, anyone can do that.
We're talking about design, which you make clear you have absolutely no idea of.
I've spent years trying to get that right, I'm not there yet but will continue to try.
Str.
You are getting confused in our old age my friend comming to the conclusion about what was said by "The Relay Station " not by "Radio Station"
I never said I have not built or designed anything . Next time before you make your comments spend the time needed to read first before you make your condescending comments.
-
Another great point then?
I think not.
Str.
-
Refering to the title of the post I don't see what relevance repurposing a ham radio has at all?
the relevance is about getting on the air and broadcasting, re-read my post.
Time and patience, Yes, plugging in a ham radio, anyone can do that.
again, you need to re-read my post, there is a bit more involved than that.
We're talking about design, which you make clear you have absolutely no idea of.
repairing, re-purposing and modifying existing equipment requires the thought of "design", does it not ?
-
Another great point then?
I think not.
Str.
in my opinion, it seems most people are against the idea of modifying existing equipment for the purposes of high fidelity AM broadcasting, like it is some sort of "sin" to discuss it.
i dont really see how it should be treated any differently, if you dont possess the knowledge to build one from scratch, how can you understand how to modify and re-purpose it then ?
-
Difference of philosophy I suppose, but I wouldn't call it a 'sin'. Modify to your hearts content, do what makes sense to you. I would rather start with something that is designed for full duty cycle (my FT857 sure isn't, heatsink is WAY too small). Most ham stuff Is designed with a price point in mind, as is all commercially available stuff. To that extent, a lot of corners are cut, and I do agree the older gear had a lot more meat to it than the modern flimsy electronics we know as ham gear. The heatsink alone in my prototype transmitter weighs more than most solid state ham transceivers I've seen.
+-RH
-
Thanks RH, sanity restored, somewhat!
Lets stick to Class D design, shall we...?
Str.
-
It doesn't make sense to try to modify something that has an antiquated and overly complicated design. Using the modern technologies of DDS and class E or D doesn't seem to be a part of commercial ham radios. Maybe I'm wrong because I don't own any ham gear, but that's what it looks like to me.
That aside, the PCBs for my project have shipped and the FETs are already here. I will keep this thread updated with the progress when I begin the build.
I also have two new questions:
1. I tested a modulation transformer on a small class D (not the one I'm building). I've encountered a problem with the transformer. It is a 1:1 115V toroidal transformer. I don't know the VA, but it weighs about 15 kgs. I've read of using toroidal transformers as modulation transformers before and it looks like there are hams doing this and that it works well. Now we get to the problem: How do they not overload the audio amplifier when the ohms drops so low at the low end of the audio spectrum? If I look at the resistance at 100 Hz, it is 1 ohm and at 1 KHz, it is 6 ohm. So how can I get any bass into my signal without overloading the audio amplifier. Do I just need to filter the bass a certain amount?
2. Can I stack ferrite cores on the output transformer? For example, if I have 8 FETs, can I have four cores on the bottom and four on the top and then run the wire through all of them similar to if they were lined up? In the designs I've looked at, it seems like there is always one core per FET. Is this for power handling reasons or is it because the matching needs to change for each additional FET? (ie could I get away with four cores on an 8 FET amp, given that the power level is not too high or is it for matching?)
-
1. I tested a modulation transformer on a small class D (not the one I'm building). I've encountered a problem with the transformer. It is a 1:1 115V toroidal transformer. I don't know the VA, but it weighs about 15 kgs. I've read of using toroidal transformers as modulation transformers before and it looks like there are hams doing this and that it works well. Now we get to the problem: How do they not overload the audio amplifier when the ohms drops so low at the low end of the audio spectrum? If I look at the resistance at 100 Hz, it is 1 ohm and at 1 KHz, it is 6 ohm. So how can I get any bass into my signal without overloading the audio amplifier. Do I just need to filter the bass a certain amount?
You may want to take a look at this paper as it covers some of the problems of using transformer modulation at the few hundred watt level.
http://amfone.net/Amforum/index.php?action=dlattach;topic=30798.0;attach=31360 (http://amfone.net/Amforum/index.php?action=dlattach;topic=30798.0;attach=31360)
2. Can I stack ferrite cores on the output transformer? For example, if I have 8 FETs, can I have four cores on the bottom and four on the top and then run the wire through all of them similar to if they were lined up? In the designs I've looked at, it seems like there is always one core per FET. Is this for power handling reasons or is it because the matching needs to change for each additional FET? (ie could I get away with four cores on an 8 FET amp, given that the power level is not too high or is it for matching?)
Does..not...compute. I would need to see a picture. The choice of core, how many turns, ect are design considerations made and influenced by power level and frequency mostly. Maybe your talking about the huge output transformer in the 'QIX design? I've found that making a large binoccular core out of four 1020 cable beads seems to be good enough until you get to the kilowatt level. In class D you will have more harmonic power to contend with and it would be best to make your output transformer a little overkill to make sure you've got enough headroom to handle the higher than normal harmonic currents.
+-RH
-
You may want to take a look at this paper as it covers some of the problems of using transformer modulation at the few hundred watt level.
http://amfone.net/Amforum/index.php?action=dlattach;topic=30798.0;attach=31360 (http://amfone.net/Amforum/index.php?action=dlattach;topic=30798.0;attach=31360)
+-RH
I've already read that article. It gave me most of the info I have now about the transformer, but I'm still not sure why the impedance is so low at low frequencies and how that can be rectified :P
I've discovered that the problem is not the actual transformer. If I connect an 8 ohm load to the transformer, I can see a pretty flat load on the other winding (not perfect, but no where near as bad as when it's connected to the TX). What determines the resistance here and why is it not flat? I tried changing the cap next to the RF choke, but it made no difference.
I know this is probably hard to determine without knowing the actual circuit. Also, this is a preliminary question and I'm not even sure it will be a problem when I get the amp I'm building, but it could be because the designs are very similar.
-
Iron saturates at low frequencies, HPF your audio @100Hz.
I'd understand PWM if I were you, it's so simple you'll wonder why you ever used any other method.
My modulator probably weighs about 2oz.
The filter would weigh a bit but only 2-3lb.
You're building Class D RF so why not Class D audio too?
Str
-
How are you measuring the impedance? Stretchy is right, as frequency decreases, inductive reactance will go down, more so if the transformer is not tape wound with hypersil type steel.
+-RH
-
1. I tested a modulation transformer on a small class D (not the one I'm building). I've encountered a problem with the transformer. It is a 1:1 115V toroidal transformer. I don't know the VA, but it weighs about 15 kgs. I've read of using toroidal transformers as modulation transformers before and it looks like there are hams doing this and that it works well. Now we get to the problem: How do they not overload the audio amplifier when the ohms drops so low at the low end of the audio spectrum? If I look at the resistance at 100 Hz, it is 1 ohm and at 1 KHz, it is 6 ohm. So how can I get any bass into my signal without overloading the audio amplifier. Do I just need to filter the bass a certain amount?
You may want to take a look at this paper as it covers some of the problems of using transformer modulation at the few hundred watt level.
http://amfone.net/Amforum/index.php?action=dlattach;topic=30798.0;attach=31360 (http://amfone.net/Amforum/index.php?action=dlattach;topic=30798.0;attach=31360)
2. Can I stack ferrite cores on the output transformer? For example, if I have 8 FETs, can I have four cores on the bottom and four on the top and then run the wire through all of them similar to if they were lined up? In the designs I've looked at, it seems like there is always one core per FET. Is this for power handling reasons or is it because the matching needs to change for each additional FET? (ie could I get away with four cores on an 8 FET amp, given that the power level is not too high or is it for matching?)
Does..not...compute. I would need to see a picture. The choice of core, how many turns, ect are design considerations made and influenced by power level and frequency mostly. Maybe your talking about the huge output transformer in the 'QIX design? I've found that making a large binoccular core out of four 1020 cable beads seems to be good enough until you get to the kilowatt level. In class D you will have more harmonic power to contend with and it would be best to make your output transformer a little overkill to make sure you've got enough headroom to handle the higher than normal harmonic currents.
+-RH
Just stack em till you get the temp range you want. You're watching the temp of the core right?
-
Iron saturates at low frequencies, HPF your audio @100Hz.
I'd understand PWM if I were you, it's so simple you'll wonder why you ever used any other method.
My modulator probably weighs about 2oz.
The filter would weigh a bit but only 2-3lb.
You're building Class D RF so why not Class D audio too?
Str
I can understand that the iron core might be saturating, but this doesn't seem like the case. When an 8 ohm load is on the transformer, it's pretty flat down to 100 Hz, but NOT when it's connected to the transmitter.
Let me know when you have a PWM board ready and I'll order it. It's nice to make use of old things around the shop though and that's why I'm testing the transformer first as I have several of them laying around. BTW, my audio amp is a class D (still less efficient than PWM though).
How are you measuring the impedance? Stretchy is right, as frequency decreases, inductive reactance will go down, more so if the transformer is not tape wound with hypersil type steel.
+-RH
I am using an LCR meter that has resistance measurements at 100Hz, 120Hz, 1KHz and 10KHz. And I don't doubt the measurements because the amp shuts itself down unless the bass is equalized in the audio. When hooked up to an 8 ohm load, this is no longer the case. So the problem is only with the TX connected.
The HPF idea seems like not a bad one. I already tried that with an equalizer, but maybe I need more aggressive filtering like with a real HPF. That would probably work, but it seems like I'm dodging the real problem.
-
You could add some R in series with the transformer and see if that helps. I suppose you could add a capacitor as well to get a similar HPF effect, albeit not as sharp. Does the problem go away if you back off the modulation some?
+-RH
-
Cap on the I/P using the I/P Z and the Xc of the cap as an HPF, works for me.
Str.
-
OK, I'm back at this after a little break.
I've got my 8 SIC FET board mounted and populated with 4 x C2M0280120D and 2 x TC4452 drivers.
I am using a balun with 4 x FB 43 1020 (someone suggested the FB 61 1020 instead)
1 turn primary, 2 turns secondary
My output circuit is incredibly simple. It is just the balun right now. That's it :P. I've seen a guy pull off matching like this before, so I think I just need to change my core type. And yes, I'd prefer it this way because I want broadband as the highest priority.
So I did my first test today and this is what I got:
140W @ 40V, 7MHz, 6.8A
200W @ 40V, 3MHz, 7.1A
So not good efficiency, but I am happy that it works at all as this is my first build.
I have my drivers at 8V right now. The two changes that I plan to make are change the cores to type 61 and increase the driver voltage. The max driver voltage for the TC4452 is 18V. What do you think I should set it to? Do you have any other suggestions?
I also have some C2M0160120D in stock that I could try.
-
In order to get clean switching, you really need to hit the gates of SiC fets with 15V or greater. 18-20V seems about optimal, any less and RDSon comes up pretty quick and will deteriorate your efficiency and cause excessive heating.
Also, are your power measurements before or after the lowpass filter? Without a filter, power is likely to read substantially higher due to large amounts of harmonics, which on conventional class D is going to be quite high.
Not sure what your output transformer looks like, but hopefully the primary is made out of large-ish copper tuning to reduce Q and therefore excessive ringing due to circuit strays. I would try and do something similar to what is found on solid state VHF PA's where some semirigid line is bent into U shapes and the shields soldered together. This will also help maintain impedances and keep ringing down.
+-RH
-
In order to get clean switching, you really need to hit the gates of SiC fets with 15V or greater. 18-20V seems about optimal, any less and RDSon comes up pretty quick and will deteriorate your efficiency and cause excessive heating.
Also, are your power measurements before or after the lowpass filter? Without a filter, power is likely to read substantially higher due to large amounts of harmonics, which on conventional class D is going to be quite high.
Not sure what your output transformer looks like, but hopefully the primary is made out of large-ish copper tuning to reduce Q and therefore excessive ringing due to circuit strays. I would try and do something similar to what is found on solid state VHF PA's where some semirigid line is bent into U shapes and the shields soldered together. This will also help maintain impedances and keep ringing down.
+-RH
OK, thanks! It is really invaluable having your help, no one around me in my day to day could answer, so I appreciate it.
I am measuring after the lowpass filter. It won't fully work on the 3MHz though which might be contributing to the higher number there.
The wire I am using I believe is 12 AWG silicone stranded wire. It is not a coax. I do have some RG-402 though.
What type of ferrite cores do you use? I suspect my main problem is the core type.
Another thing I am wondering for the future: I have four 1020 cores arranged in a binocular form. My board has space for eight fets even though now I am using only four. When I add the additional ferrite cores I won't have space, so can I stack them so I have four on the bottom and four on top. And then loop the wire through all of them (not sure if I would need to cross over to the opposite top side). I have never seen it done this way.
-
I've tried both 43 and 61 material ferrites, not much difference I could note, although with your higher harmonic power I could see some difference in this application, 43 may be better.
I would try something similar to this, wound with coax, shield grounded only on the load end http://www.communication-concepts.com/rf2000-transformer/ (http://www.communication-concepts.com/rf2000-transformer/)
+-RH
-
Yep You need 12V min on the gates, 15V is optimum for SiC I have found. O/P tran wise check the FAT5 design, I copied that and works a treat. Stack a pair or T200-2's and wind LH CW and the RH CCW, if that makes sense?
Check amfone for the same design of o/p tran.
http://amfone.net/Amforum/index.php?action=dlattach;topic=42504.0;attach=54688;image
Str.
-
OK, so I've increased the voltage from 8V up to 18V. Here is what I am getting now:
3MHz 13.8V 4.0A 50W H=5.2W 90.6%
7MHz 13.8V 3.4A 32W H=14.9W 68.2%
3MHz 28.1V 7.5A 200W H=10.7W 94.9%
7MHz 28.1V 6.67A 140W H=47.4W 74.7%
3MHz 48V 14A 640W H=32W 95.2%
7MHz 48V 12.7A 410W H=198W 67.4%
I also briefly tested 5MHz, but my power supply is rated at 600W, so I was really pushing it and didn't get proper measurements.
It made 600W at 5MHz, so clearly it doesn't drop off linearly. It is also possible that my 640W reading it limited by the power supply. I have a 1500W supply on the way.
I have yet to change the core material and that is my next trial. It is amazing how much better it is operating with the 18V.
Despite the success, I still haven't met my goal of 7MHz. Any ideas of other changes I could make? What is usually the weakest link here?
-
How are you introducing 'dead time' into the phases of the RF drive? At 7MHz, you will need about 40% duty cycle on each phase to prevent cross conduction (shoot through). This could be one of the reasons for the relatively poor efficiency on 7 MHz.
+-RH
-
How are you introducing 'dead time' into the phases of the RF drive? At 7MHz, you will need about 40% duty cycle on each phase to prevent cross conduction (shoot through). This could be one of the reasons for the relatively poor efficiency on 7 MHz.
+-RH
I think it is 40% 60%, but I am using a pre-made DDS module (made for this purpose). I checked it on the scope a while back to just see what voltage it was putting out, but I didn't look at duty cycle. I really need to check it again to verify. I am also going to check the outputs of the TC4452 to make sure they're staying square.
So I am having correspondence with a knowledgeable guy via email. He tells me that I should use 20V (in which case I'd have to upgrade to a different driver and change my PCB). Are any of you using the TC4452 or TC44xx on 40 meters? Could the voltage or the driver itself be the problem? My suspicion is the guy is just being nitpicky about the 20V and it doesn't really matter. (As I said earlier, I am using 18V now.)
In addition, I tried changing the cores to type 61 1020. I seem to be getting an erroneous reading from my bird meter now. It says I'm putting out 800W instead of 600W with the old core at 3 MHz even though my power supply is only drawing 700W. At 7MHz the power and the efficiency did not change.
-
20V won't make any difference, don't waste your time. As stated I've built the same and 15V is fine.
Str.
-
I'm using IXDN614 drivers. The waveforms looked better with these drivers than the '4452's. 15-20V should be fine, I think your bigger problem is the overlap in the PA due to 50% duty cycle drive.
I've also had problems with some RF slugs indicating greater than 100% efficiency when input/output power is calculated. This is more noticeable with higher power slugs when measuring at less than 75% full scale, as the linearity deteriorates at the lower half of the scale.
+-RH
-
Measure power with a calibrated scope probe and some sums (math). I use a 100:1 probe with a BNC T piece, works a treat. Those drivers are rather 'old school' BTW. NCP81074A are way faster. Please read up on the design I pointed towards on amfone. That's the current best one I know.
Str.
-
Maybe OT but could you modulate a class D amp by controlling the duty cycle of the drive?
Does that exist, if so does it have a name?
-
It is theoretically possible, however, getting conventional devices to switch that fast (100x carrier frequency to achieve 99% modulation) is almost impossible.
+-RH
-
Alright, well I popped two fets, but I learned some things. . .
So I set the 40% 60% duty cycle. It improved a bit (the RF watts went up to 460W from 410W and the efficiency went to 80% up from 74%), so I changed it a bit more and then they popped immediately. I wasn't able to even read the amps. I have to change the duty cycle for each wave individually. Is having them not exactly the same or overlapping a huge problem? I got a new high power current limiting supply, so I'll use that in the future. I am not 100% sure that it was 40 60, but it should be pretty close. My scope doesn't have any fancy stuff. I just had to read the ns/division and do the calculation. I am surprised how little of a change beyond the 40 60 did it.
Anyway, here is the data I gathered:
The DDS exciter is making a decent looking square wave. It has 7ns of rise time regardless of frequency (from 3 to 8 MHz it stays at 7ns). And . . . the drivers have a 20ns rise time (again - regardless of frequency). The fall time on the drivers, does however, seem to get even slightly worse with increased frequency (goes up to like 25ns). So I think the drivers are the culprit. I didn't do the exact math, but I'm thinking 20ns is probably too much for 7 MHz. Is this correct?
If so, then I'm on to making a new PCB. I am going to use the drivers stretchy suggested and put them as close as possible to the fets. I gave the datasheet a quick look and the pinout looks confusing. I am not sure how to hook them up. I addition to the new drivers I am also going to have a custom crystal on the main PCB with an inverter to generate the second wave. It's either going to be the SN74AHCT14N or SN74HC132N for the inverter.
-
How are you introducing 'dead time' into the phases of the RF drive? At 7MHz, you will need about 40% duty cycle on each phase to prevent cross conduction (shoot through). This could be one of the reasons for the relatively poor efficiency on 7 MHz.
+-RH
Does that mean 40% on, 60% off, or the opposite? And it's not relative to each other, it is just on time vs off time for each phase individually?
I want to make sure I did this right. I have a feeling I did it backwards.
-
40% on, 60% off. The idea is that the drivers are better at turning the fets on than off, due to miller and reverse transfer capacity in the mosfets. As such, the duty cycle will be stretched, and if not compensated for, cross conduction occurs in which for a very short time, there is a dead short across the PA voltage source which destroys the transistors. The way around it is to introduce dead time. I use the circuit out of a Nautel NX-50 PA module to accomplish this. The idea is that two drive signals are compared, one normal, and one delayed by a fixed amount. The two signals are compared, one by NOR, and one by AND. This produces two signals with a fixed amount of dead time, which prevents cross conduction. Some experimentation will be necessary to find out what amount of dead time produces best efficiency, but 10-20nS should work. I believe I used the DS1100z-40 in this application and its working well.
The reason the NCP driver has two output pins is to allow you to tailer the gate resistor values for optimum rise and fall times, usually using a smaller resistor for the low side.
Whenever checking your input drive waveforms, be sure to look at them with a dual channel scope so you can verify that overlap is not occurring (phase A CH1, phase B CH2).
(https://i.imgur.com/rBnENSU.gif)
+-RH
-
Using the NCP drivers which are super fast BTW, I've never bothered with any dead time adjustment and driven the gates directly from both O/P pins, DC coupled with NO RES.
Mind you I have been using the GaN fets, again super fast, but previously used SiC and they were fine only having to use 15V rather than 6V as the driving Vcc.
Again I'd point you toward the article on AMFONE, there's even better FET drivers now with an RF isolated barrier (what next!) but the NCP jobbies are fine.
Str.
-
Using the NCP drivers which are super fast BTW, I've never bothered with any dead time adjustment and driven the gates directly from both O/P pins, DC coupled with NO RES.
Mind you I have been using the GaN fets, again super fast, but previously used SiC and they were fine only having to use 15V rather than 6V as the driving Vcc.
Again I'd point you toward the article on AMFONE, there's even better FET drivers now with an RF isolated barrier (what next!) but the NCP jobbies are fine.
Str.
I looked at that amfone link, but it lead me to an image. I managed to get the thread from the URL, but I didn't find anything in the thread discussing a better driver than the NCP. What is the name of it?
On my new board I need to make sure that I have enough space for my output transformer. Can I use only 4 cores for 8 fets? Would I just need to alter the windings to compensate or will it overheat or not match? If not, I asked before, but can I stack the cores on top of each other? This should be the last question I have before I can begin designing. It will really change my layout depending on the answer to this.
-
Can't find the exact post but look for, VE3ELQ.
The NCP A ones are fine anyway.
Regards stacking the cores I gave up with the binoculars and stack 2 T200-2's, they be fine.
Str.
-
I would leave a small amount of R in series with the gates for two reasons;
1. it helps dampen any tendency of the network and other strays to ring (Better drive waveforms).
2. should a transistor failure occur, it prevents the output of the driver from going complete short circuit. I have lost a few drivers due to transistor failure, No more once a little R was added. I'm using three 10 ohm 1206 resistor in parallel to make a wider low L resistor, and also spread the dissipation.
+-RH
-
Leaving space on a PCB is no bad thing!
I must admit I've lost 1 or 2 drivers in the past when thing went a bit west and the O/P's went S/C.
BTW I've tried to find any mention of the alternative drivers but it would appear that the post has been removed.
Stick to the NCP's.
Do read up on ALL the post on AMFONE concentrating on the posts by Nigel.
Personally I'd stick to the toroidal O/P tran and make sure the windings are opposed as shown in the design on there.
By careful choice of series L/C components (low Q?) you end up with a kinda class D/E thing where the efficiency is around 93-95% and the B/W around 1MHz.
A good compromise I think....
Str.
-
I just had a thought. Is there a reason why there is always PWM to the drain of the mosfets? Why not apply PWM into the drivers instead and eliminate a separate PWM board. Why is it not done this way?
-
Per my previous (two posts back) reply, device limitations in regard to switching speed is the main reason. To get 99% modulation depth, the switching devices would have to be capable of switching at 100X carrier frequency, hence at 7 MHz, the devices would have to produce something resembling a square wave at 700 MHz.
There are ways around this problem. Harris series combines a bunch (48?) PA modules, arranged in binary order, big steps, and little steps. Their drive signals are switched on and off according to how much instantaneous power is required for the envelope power as dictated by the incoming audio. Nautel at this point is conventional PWM with DSP based AM/AM and AM/PM correction, similar to the smaller DAX series transmitters from Harris/Gates/Who ever they are this week.
My attitude is, if it were possible and financially viable, it would already be in a commercial product.
+-RH
-
I'm on my way with the updated design and I have a few questions.
I have this audio amplifier: https://www.parts-express.com/wondom-aa-ab31242-1x600w-class-d-audio-amplifier-board-(t-amp-technology)--320-3346 (https://www.parts-express.com/wondom-aa-ab31242-1x600w-class-d-audio-amplifier-board-(t-amp-technology)--320-3346)
It is a class D design and I was originally going to use it with a transformer, but I think I may be able to use it directly. What do you think?
Here is the datasheet for the IC it is using: https://www.ti.com/lit/ds/symlink/tas5630.pdf (https://www.ti.com/lit/ds/symlink/tas5630.pdf)
So I have to check if the ground is isolated, right?
Then I need to apply whatever voltage I need for the carrier at the ground?
Is this how it should be done?
My other question is about the inverter circuit. I don't have too much experience with logic circuits, but I know I probably need one of these: SN74HC132N, SN74AHCT14N, or CD74HCU04E
Which should I use? I am going to use it along with a crystal oscillator to generate the inverted wave.
-
The amp may work direct without a transformer if the protection circuitry is primitive enough to ignore DC imbalance on the driver pairs, something that unfortunately can only be found through trial and error. Output power will be square law proportionate to input voltage, so I would be sure that the limitation input voltage of the modulator (40V) is enough to achieve the PEP you are aiming for.
To avoid any possibility of cross conduction (caused by logic gate propagation delays), I would use a 74LS86 configured as a push pull converter to get your drive.
Stretchy, where are these links you keep talking about?
+-RH
-
I love the simplicity of rc coupling compared to buffers or xformers and so on, but sometimes the added complexity of xformers and buffers expiates many sins in the most effective if not efficient manner.
-
The links have gone but the original design work is by VE3ELQ, so just check for posts by him.
Modulator wise I tried one of those Amps and drove a MT , if caught fire!
Not sure what you're up to with the logic inverter?
Pray tell?
Str.
-
(https://i.imgur.com/BI4dIdp.gif)
+-RH
-
OV,
Here is Nigel's (VE3ELQ) response to my question re driver chips;
"The best RF deck FET drivers I have tried by a wide margin are the NCP81074A. With a Tr/Tf of 4 ns, matched delays, and 10A of source/sink they are both super fast and capable of driving higher gate C FETs. No problem with the 150pf SIC FETs at 7.3 mhz or the newer GaN FETs. They are inexpensive and small so I recommend 1 driver per FET up nice and close to keep the gate lead as short as practicable. They work great.
https://www.onsemi.com/pub/Collateral/NCP81074-D.PDF"
Do you have any spare boards to sell although I believe you'll need to re-do them should you decide to use the above drivers. So cheap now I'd gladly contribute to any updated board.
I do like the look of your 8 device boards.
Transmitter Man
-
Yep, thems the ones! They drive the GaN FETs with ease and are OK (get a bit warm) for the SiC devices. Makes for a greatly simplified design.
Cardinal CPP Osc - NCP Driver- FET O/P- LPF.
-
OV,
Here is Nigel's (VE3ELQ) response to my question re driver chips;
"The best RF deck FET drivers I have tried by a wide margin are the NCP81074A. With a Tr/Tf of 4 ns, matched delays, and 10A of source/sink they are both super fast and capable of driving higher gate C FETs. No problem with the 150pf SIC FETs at 7.3 mhz or the newer GaN FETs. They are inexpensive and small so I recommend 1 driver per FET up nice and close to keep the gate lead as short as practicable. They work great.
https://www.onsemi.com/pub/Collateral/NCP81074-D.PDF"
Do you have any spare boards to sell although I believe you'll need to re-do them should you decide to use the above drivers. So cheap now I'd gladly contribute to any updated board.
I do like the look of your 8 device boards.
Transmitter Man
My new updated board should be here for experimentation in a few weeks.
RH: I tried that logic circuit that you posted. I'm assuming this is to prevent cross conduction. Well, I tried it with the 74LS86 and a 74HC04E (didn't have the 7406) and it doesn't seem any different than my simple circuit with 74AHCT14N, so I stuck with that for this design. It does cross for a couple ns at a really low level, so I think it's OK.
(http://oi68.tinypic.com/25sagm1.jpg)
-
Do you have any spare boards to sell although I believe you'll need to re-do them should you decide to use the above drivers. So cheap now I'd gladly contribute to any updated board.
I forgot to answer the question. Yes, I do have 4 more of the old PCBs left if you want one. I don't have time to sell a parts kit right now though. Maybe when I finish development and I'm at a good stopping point.
I've been looking to sell them anyway. They work perfectly up to 5 MHz, but no higher. This new design you see above should do 10 MHz easily, I hope. . . I can also easily swap the MOSFET pins to make it work with GaN devices in the same package. If it works, then I will also sell a GaN version of the board. The cheapest GaN to work with it costs about $20 each and the SiC are only $3.50.
-
See last two posts.
I realized after my design that it looks like the NCP drivers have an inverter built-in. The truth table in the datasheet makes it look like it won't work for this purpose though, but I don't know why else they would have such a feature. Is this to use in a push-pull amplifier with one square wave input? If so, then I don't even need the inverter :P
-
OV,
I missed the fact the previous board only works up to 5Mhz as I would much prefer the higher frequency board :-(
I will wait :-)
Transmitter Man
-
See last two posts.
I realized after my design that it looks like the NCP drivers have an inverter built-in. The truth table in the datasheet makes it look like it won't work for this purpose though, but I don't know why else they would have such a feature. Is this to use in a push-pull amplifier with one square wave input? If so, then I don't even need the inverter :P
To drive a High side and Low side Devices I guess? Or tie them together for single FET. Just use a dual Inverter from the O/P of your Osc. Sine in and square out, inverted and double inverted to drive either side.
Str.
-
To drive a High side and Low side Devices I guess? Or tie them together for single FET. Just use a dual Inverter from the O/P of your Osc. Sine in and square out, inverted and double inverted to drive either side.
Str.
Sorry, I don't understand the term "high/low side device". Does this mean the devices on one side of a push-pull amplifier (like what I'm building)? So then I don't need an inverter. Just use the NCPs?
-
Yes! High and low side being terms used for PWM boost voltage conversion.
https://engineering.purdue.edu/Courses/ECE433/exp5_5th~6thweek_.pdf
Similar methods if not slightly higher frequencies in use here!
The design is already there, just build the same one!
http://amfone.net/Amforum/index.php?topic=42262.0
Str.
-
Be careful when relying on the gate driver to act as the inverter! If your clock source gets loaded down or slightly shifts duty cycle then your amplifier will be imbalanced.
I ran into this issue at 7Mhz whereupon the gates were being driven 35/65 instead of 50/50.
That breaks things after a while.
My advice, use an external inverter or better yet, a flip-flop driven at 2x frequency, those are good at maintaining an even split. For added protection in case your clock stops, make diode level shifters with decay resistors that prevent the drive signal from remaining high.
See last two posts.
I realized after my design that it looks like the NCP drivers have an inverter built-in. The truth table in the datasheet makes it look like it won't work for this purpose though, but I don't know why else they would have such a feature. Is this to use in a push-pull amplifier with one square wave input? If so, then I don't even need the inverter :P
-
I have resumed work on the project. It is very near completion. I just need a good heatsink for the FETs and I'm searching for one now. I want a cooling aggregate type.
I still don't know what to do to the transformer when I add another pair of FETs.
Do I change nothing?
Do I need to add another winding on the secondary?
Do I need to add more than four cores for eight FETs?