Wayne's 40M H-Bridge

[w9jsw]

This Class E forum needs some new activity!

Now that I have a working board layout for Rod's PWM, I am going to give this circuit a go.

I have a few questions -

1. I am going to drive it with a 2-phase board that has a pair of IXDD614 drivers. So don't need the input inductors to create a phase difference. Is it still wise to keep them in the design?
2. Can this design be tweaked to be able to go 160/80/40? Or should I plan to go like the super senior and have separate 160/80 and 80/40 implementations?
3. Will try those 150V MosFETS. Target voltage is 96V with a carrier of around 48V. This is based on an already existing supply I have on hand. Will I be able to achieve 200W or better? Will I blow the FETs? I am also going to try some 16N50s that I have on hand.
4. When tying this together will I have to sequence the PWM with the driver board? i.e bring up the pwm, then bring up the driver and do the reverse when going back to rx, or can the drive and pwm be brought up together. Reason is I have a 2 stage sequencer/TR board (already in hand) for the antenna and the boards. Having to sequence the driver with the pwm will require a new 3 step sequencer.

My schematic is what I have come up with so far. I previously used those SiC devices and got 375W of carrier with 96Vcarrier at 89% efficient. Impedance was 20.6R which is too high for the PWM, I fear. Also, it would not do 40M.

John

[w9jsw]

Got an answer on the sequencing. A 2 stage sequencer process will work.

RX -> PTT -> ANT relay/Drive enable -> wait -> MOD enable = TX

TX -> drop PTT -> drop Mod -> wait for RF to drain off -> drop drive and ANT Relay (will take a few ms to settle) = RX

[vk3alk]

Hi John...

Will try to answer your questions.
I am not an engineer and the only qualifications I have is an Amateur Radio licence like you...  

1.   Your change in design using IXDDs will not work...
      If you look at the design using transformers you notice 2 things .... there is a transformer across the gate source of each
      FET...and how they are arranged enables the middle of the bridge to float in respect to ground....
      You could use IXDDs but would need 4 of them and the circuit would be far more complicated and is just not worth it...
      Transformer drive is simple ...low cost and there are no adjustments etc: although the drive requires a duty cycle of 50:50..

2.   The circuit will work 160/80/40 as a multibander but only with the FETs in the design...
      Can I leave this for the moment as will promise to post later with this build...

3.    Only use the FETs as detailed...there is another one but would stick to the IPPs.
       Ok on your Power Supply of appox 96 volts....
       Best to use the standard Mark to Space ratio of 60:40 so using appox 100 volts on the Modulator the carrier voltage would
       be 40 volts...
       You need headroom for your positive peaks  

4.     No sequencing is needed....just one relay enabling the pulse trains in the PWM and TX and others like Antenna relay muting
        etc:

40M is a little more involved so maybe starting on 160/80M would be a better way to go....
I myself would build up one of those PWMs and test them hard before giving them to anyone as all problems would maybe come back to you...
Build one TX module for say 80M and makeup a little transmitter....with only 40 volts though expect no more than 150 watts at most....
Post your progress on the Forum....


Wayne

[w9jsw]

Thanks Wayne,

I am anxious to know more about your multibander.

I have a few ideas on how to tweak my current prototype output transformer as well. After being immersed in transformer math, I have come up with a 7T:19T on a T225-2B as my next try. Hope to get the current up higher at a lower voltage to get the output impedance below 10 ohms. We shall see.

John

[w9jsw]

I see that Nigel ran his PWM to drive the PA at 13 ohms of source impedance. That is not far off of what I have now. I was going to run this at 93V carrier and 180V unmodulated. With what I have learned, 75V and 180V are better goals.

I am going to target 13 ohms as that is what Nigel's schematic shows with inductor values already determined for the PWM filter.

Now I am getting 90% efficiency with 93V to get 375W out. I am using a 10T:13T on a T225-2 core shunted by 1200pf.

Going to try 2 different cores next. For one I will use a T225-2B core with a 7T:14T which with an impedance ratio of 3.84 should give me the proper source impedance of 13 ohms. The other try will be the current single core T225-2 using the same 10T on the primary and 19T on the secondary. Hopefully the need for shunt capacitance will go away when I get the proper impedance ratio on the OT.

We shall see.

[vk3alk]

Hi John..

Sounds like your having some success there....
Your circuit doesn't show a DC blocking capacitor in the floating middle part ... appox 1uF poly as it gets quite warm or maybe thats an error etc:
Also how come you have all those components like a 1N4148 diode and 22R / 4.7K resistors and the .047uF cap across the secondary of your transformers...
Try taking them out and see how you go ...
How much current is your IXDD drawing ?
What voltage are you applying to the IXDD ?
Have you made the duty cycle at the gates of the FETs 50:50 ?

When I find things are going well I reduce the voltage to the IXDD until the output power starts to drop and at that point nip it up a bit...
You might find that voltage to be say 10 volts and the current drawn much lower.
Makes the IXDD life a little easier...

I am surprised with the turns ratio of 1:3 but maybe thats Ok using those CREE devices...

[w9jsw]

I collaborated with a colleague on this design. We added the 22R to get the input SWR closer to 1:1 when driving it from a SDR transceiver. I recall that the diode and resistor were present to protect the more sensitive gate of the SiC device. The diode is a negative surge suppression diode. The resistor is used to pull the gate to ground during the lack of gate drive. The barrier diode protects the device overall. It is used to suppress conduction of the body diode on the die. It is an 8A SiC Schottky diode.

I am going to use a 2 phase driver on the next iteration of the board so that I can have full control of the width of the gate pulse on both phases. Using a single drive will not allow this.

The 3:1 turns ratio is to get low gate inductance with higher drive as well as control the gate voltage. The secondary is made with a loop of tinned braid. I takes 4W of power to drive this prototype to 375W. This would be well and good except for the risk of shoot-thru. A different colleague thinks I may not be getting enough gate drive on all of the fets. I have a new 4 channel scope so hope to check that out better this week.

Attached is the original 1.0 version of the design which matches the prototype. 1.2 is where I am at now on my thinking.

More to come.

[w9jsw]

I don't think that decoupling cap near the OT is necessary. I have been following this attached ap note and some others.

[vk3alk]

Hi John...

I see now where your getting information from....
Will print it out and have a read .... seems interesting ....

This is just my opinion .....
I feel you have turned a simple circuit into something rather complex....
The whole idea of h bridge is its simplicity....
Many FETs just do not work well in H Bridge mode and are better left for Push Pull designs...
I wouldn't use Sinewave drive .... use digital drive ...
To vary the Duty Cycle is simple .... your VFO into an inverter with a little bit of DC bias will do the job very well ...
Have uploaded two photos ....
The test board I finished a while ago now and I switch the Flip Flop in / out for different functions etc:
But you can follow things along... crystal oscillator going into the 7404 inverter and that little trimpot there adds DC bias to its input changing the duty cycle etc:
When you are successful with your TX module you can makeup something a little better etc:
Even Steves variable Duty Cycle circuit could be used....
This is all off single drive ....

The other photo is of two of the four waveforms without Drain voltage.... you should be able to get something like that..
Ohhh this is on 80M though ... not 40M...
On 40M there are no harmonics left only the fundamental which is sinewave but adjusting the DC is the same...
When Drain voltage is applied the waveform will distort somewhat but should be similar...
When you apply digital drive you just adjust for 50:50 duty cycle thats all....

But this is what Amateur Radio is all about trying different things and we all have our own ways of doing things...  


Wayne

[w9jsw]

Digital drive - yes. Don W9BHI has built a board that I will use to drive this. It combines his VFO with QIX's circuit to manage the duty cycle. Should work FB.

Following the article, the extra components are to lessen shoot-thru due to overshoots/undershoots. However, I now have the board laid out so that I can omit the extra components and take a look at the waveforms to see if they are necessary. If so, I can add them.

Also building the input transformers per the article.

Will be trying multiple FETs, starting with the CREE ones first. Also going to try 16N50s that i have in hand and any that you can suggest that can run at 180v modulated, 75V carrier.

Going to use the tank circuit from the article. I like the idea that I could perhaps omit a tuner using that approach on a well-behaved dipole. My dipoles are not always 50R across the bands. A tuner for 375W carrier and 1500W PEP would be expensive, even if homebrewed due to vac cap costs.

Board is 100mm x100mm so fits nicely on my current heat sinks. I prefer under board mounts. Will have to see if the driver ICs cause any new heating. Prototype got barely warm under extended testing.  All SMT parts are 1210 or larger so super easy to solder.

John

[w9jsw]

These are the boards that are due to arrive any day. First build will be with IPP530N15 running at 40V carrier. Have 4 of them I bought when still available. This will be my baseline build.

I have built a newly designed W9BHI driver board that uses a DDS to generate the 2 phase drive.

Second build will be with Transphorm devices, most likely TP65H070G4PS. Will be targeting 80M first and if the wave forms look good, will give 40M a try.

I laid out the board to also use SIC devices but may not go that way. Schematic shows an expensive SiC device that I may try eventually.

[w9jsw]

DDS Driver board works! Going to need an enclosure...but hey, no clip-leads!

Don W9BHI used a AD9850 with a PIC controller to feed the circuit design that Steve WA1QIX uses (with permission) from his sine wave to 2 phase board.

Sweet!

[w9jsw]

Have the new board built. Trying to decide if I should use a T225-2 toroid wound at 1:1 for the output transformer or use a broadband one like Wayne uses. I have some FB43-1020 ferrites so that would be easy. I think Wayne used a 1T:2T ratio. I could also try a QIX one also.

I plan to add a QIX style LC match after the OT if necessary to get a good efficiency.

What say ye?

Think I'll cruise around the internet to see what others have used for a CMCD layout.

[vk3alk]

Hi John...

Was surprised that you bought some IPPs so they are now liquid gold and have to be treated with care....
There is a couple of things that I have not mentioned....
Because we were talking about 40M I didn't include Ballast Resistors... as I found that on that band the input waveform was better without them... maybe more about that later....
On 160M and 80M its best to include them just to make life for the IXDDs a little more comfortable....and the input waveforms seem better with them in circuit ( I use 8.2R )...
They do get quite warm so that means there doing what they should in protecting the driver IC...etc:
I know you ordered some nice looking PCBs but when you are experimenting its good to have a board that can add or take away components easily etc:
Have uploaded 2 photos...
I like mucking around too but my components are different from what yours are so the results will differ…
For the Driver cores I used FT50-75 with 4T Primary and 6T secondary….
The extra inductance aids with 40M much more….
The Output cores are FT114-61….
I do have the artwork for the PCB if you want and can email it to you but you have to use a laser printer and photo quality paper ….iron it to the PCB …. soak the  paper off and then etch it out etc:..


Wayne

[w9jsw]

Thanks for the reply, Wayne. I wound my first attempt with FB-43-202 on a 3T:3T with #24 teflon wire. Will see what the waveforms look like with those first. I have some FT50-43 cores can also try. May have some 77s also in the stash. I tend to buy extras.

On your OT, 7 turns bifilar? What diameter wire? Looks like 16ga enameled? I have some of those FT114 cores but not sure of the mix. Will check.

I am going to try out a transmission line transformer using 1T:2T and FB-61-1020 cores with no class E style network. Will feed a LPF. At 1:4 impedance ratio should get 12.5R of PA output impedance.

Fair-rite's number for these FB-61-1020 cores is 2661102002. Took me a while to find it. Mouser stocks these cores and has more reasonable shipping than ordering direct from Amidon. Will use RG-142.

I may also try out a binocular core that W1VD uses with transmission line. Fair-rite 2861010002. I will have to use smaller RG-174 on this so not as confident it will handle the power.

Parts on the way - stay tuned.

Some eye candy attached...

[vk3alk]

This might help in winding the OT.......

One of the windings ( 6T ) is the primary and the other two are connected in series making 12T to the output...
For a 10R load the ratio is 6T to 14T....

http://amfone.net/Amforum/index.php?topic=45787.msg328440#msg328440

[w9jsw]

With respect to the LPF, what size toroids are suggested for a 350W carrier 1600W PEP transmitter? I bought some T106-2s but now thinking those will be insufficient.

[vk3alk]

They probably are being strengthed a bit but try them out first anyway.....

Are you hoping for that output with the IPPs ?


Wayne

[w9jsw]

No, the IPPs are to establish a baseline. If I had more of them I could combine 2 boards to achieve 350W.

I will probably have to use different devices at a higher carrier voltage to get to 350w with one board. The IPP 150V limit is the main issue. I got an easy 350W from the Cree devices but could not get 40M to hit the efficiency I want. Perhaps better control of the input waveforms will yield better results. Using a 1:4 impedance ratio to get a 12.5R PA impedance, I need to hit 70V at 5.6A to get 392W input. 90% efficiency would yield 352W. 70V/40%=175V with modulation - too much for the IPPs.

If I play with a 6T:14T OT like yours, 9R of PA impedance, I could drop the Vcarrier to 60V. That would get me into the 350W range. Vmod would need to then be around 150V. Still too much for the IPPs.

What do you think about leaving some longer tails on the secondary and getting maybe 6T:18T or would the transformer efficiency fall away.

Lots of things to try out...

John

[vk3alk]

Hi John...

You could always order some IRFB4019PBF from Infineon.....they will work OK ....
Increasing the turns ratio of the OT probably will not increase efficiency and you will reach a point of low loading where the efficiency drops greatly....
Haven't seen any of your gate waveforms yet... maybe when your testing next you could post some...

Wayne

[w9jsw]

Waiting on Ferrites to show up.