vk3alk
Member
Offline
Posts: 253
|
|
« on: January 27, 2021, 10:05:53 PM » |
|
Just started a new thread... Maybe I should have started a new thread but been playing around with another 40M module... Basically the same as my 160/80M transmitters being 2 H Bridges modules into a combiner... I will be using this unit in my standby 40M transmitter thats on my bench.... Just experimenting with different output transformers and have settled on type 61 cores ... the ones on the left side with 2 cores... The transformer on the right uses type 43 cores.... The 61 cores run at appox 44 degrees C when at 150 watts carrier after about 45 minutes so with modulation they will be at their most efficient operating core temperature...... The hard work is now done as the combiner is easy to make....
Wayne
|
|
|
Logged
|
|
|
|
vk3alk
Member
Offline
Posts: 253
|
|
« Reply #1 on: January 27, 2021, 10:15:55 PM » |
|
Have installed on the right side new Transformer using the 61 cores... Tested each module and both are very similar in performance and good enough to use into a Magic T Combiner.... Unfortunately I seemed to have misplaced my Flange 100R resistor for use in the Combiner so have ordered another locally so shouldn't take long to arrive.... I wouldn't buy one from China as it would have a 10R resistor inside... The LPF is only a test tool and wouldn't survive long at 300 watts carrier with modulation... Each Module has a loading of 14R so both in parallel provide appox 7-8R to the modulator.... Wayne
|
|
|
Logged
|
|
|
|
vk3alk
Member
Offline
Posts: 253
|
|
« Reply #2 on: March 09, 2021, 08:46:59 PM » |
|
This Module in now finished although not fully tested with modulation at high power levels.... Produced 300 watts carrier without any alarms sounding...
The Analog CRO there is a Hung Chang 20Mhz from South Korea that gave up the ghost.... No Horizontal sweep. Has a Sawtooth waveform but doesn't get to the Horizontal amplifier stage so are now just looking for faulty components.... Are not all that skilled at repairing these things....
Anyway the 40M module will replace the one in my current transmitter at some stage....
Wayne
|
|
|
Logged
|
|
|
|
steve_qix
Contributing
Member
Offline
Posts: 2595
Bap!
|
|
« Reply #3 on: April 11, 2021, 10:48:23 AM » |
|
Wow, looks good ! Would love to see the schematic. Nice work.
|
|
|
Logged
|
|
|
|
vk3alk
Member
Offline
Posts: 253
|
|
« Reply #4 on: April 12, 2021, 11:38:50 PM » |
|
I think it looks good mainly because the PCB has been sprayed with lacquer... Have uploaded a circuit diagram.... The design is typical H Bridge using very few components really..... To be honest its all about what FETs you use....and not many work well .... most are a complete disaster and are sure people have given up using H Bridge because of this... But the IPP530N15 do work extremely well and you can expect efficiencies well into the 90s ... Their main drawback is the maximum voltage of 150 volts.... but have pushed modulation peaks well above without any problems at all... I am not an engineer so just experiment a bit.... On 160/80 I use current limiting resistors between the IXDD and those inductors but on 40M found they caused issues so removed them. Also on 160M and not so much on 80M the waveforms at the gates of the FETs are squarewave "ish" .... but on 40M there are no harmonics left just the fundamental so its like sinewave drive... Have found too using H Bridge can duplicate modules making them perfect for use with Combiners etc: Thats about it really.... Wayne
|
|
|
Logged
|
|
|
|
N4LTA
Contributing
Member
Offline
Posts: 1070
|
|
« Reply #5 on: April 13, 2021, 10:28:38 AM » |
|
Wayne
Looks Good
Do you have the details on the transformer? Thanks.
Just getting back into the shop and starting Nigel's PWM board that I layed out during the hiatus caused by the fire the fire.
Pat N4LTA
|
|
|
Logged
|
|
|
|
vk3alk
Member
Offline
Posts: 253
|
|
« Reply #6 on: April 13, 2021, 06:12:28 PM » |
|
Hi Pat
It has taken a while but at least now your getting back into the Shop....
The driver cores are 4T / 6T using FT50-75 cores....
The Output is 1T and 2T secondary using FB1020-61 cores ....
Wayne
|
|
|
Logged
|
|
|
|
KD6VXI
Contributing
Member
Offline
Posts: 2679
Making AM GREAT Again!
|
|
« Reply #7 on: April 13, 2021, 09:16:36 PM » |
|
Wayne,
Is there a reason you used different mixes for input and output or was that just what you had on hand?
--Shane KD6VXI
|
|
|
Logged
|
|
|
|
vk3alk
Member
Offline
Posts: 253
|
|
« Reply #8 on: April 14, 2021, 12:30:59 AM » |
|
Hi Shane.... Will try not to rattle on too much .... It all came down to size really .... small PCB and heatsinking with minimal Ferrite... FT50 size cores looked good next to the TO-220 FETs but are limited to the number of turns ..... Type 43 worked OK but made the single IXDD work hard and developed heat .... So type 77 was tried and settled on type 75 material... The Output cores were going to run hot anyway... Using 4 X FB1020-43 worked fine but ran hot and after long overs probably would have required a fan particularly on a hot day ... To me it all looked ridiculous so tried FB1020-61 cores.... Using 2 cores they ran at about the same temperature and appeared to be more temperature stable .... I'm not confident about type 43 material and even though fan cooling helped was slightly concerned about the internal temperature and the curie point of the ferrite.... So settled on type 61 and well its curie point is so much higher also I think type 61 maximum efficiency temperature is around 50-70 degrees ( but could be wrong there not sure ) .... So thats really it Wayne
|
|
|
Logged
|
|
|
|
N4LTA
Contributing
Member
Offline
Posts: 1070
|
|
« Reply #9 on: April 14, 2021, 09:45:42 AM » |
|
I notice you have the output transformers inside forms made with PCB material and brass tubing. I assume this is a one turn winding and the wire is the two turn winding?
Thanks
Pat
|
|
|
Logged
|
|
|
|
vk3alk
Member
Offline
Posts: 253
|
|
« Reply #10 on: April 14, 2021, 07:20:40 PM » |
|
Cannot seem to find a photo of one....
But yes the primary one turn is made of cutout PCB on the ends with copper tubing through the cores.....
Wayne
|
|
|
Logged
|
|
|
|
KQ6F
Member
Offline
Posts: 124
|
|
« Reply #11 on: April 14, 2021, 07:40:23 PM » |
|
Wayne -
I notice that you are using a power combiner for summing the two module outputs. In the past I also tried that but found that simply connecting the two secondary windings in series and feeding directly to the LPF produced slightly better efficiency. I was using the type of combiner that involves a terminating resistor. I reasoned that the loss in that resistor plus a small amount of loss in the combiner core were the culprits.
Am wondering what type of combiner you're using.
73, Rod
|
|
|
Logged
|
|
|
|
vk3alk
Member
Offline
Posts: 253
|
|
« Reply #12 on: April 14, 2021, 10:09:17 PM » |
|
Hi Rod.... Forums are great with all these different ideas.... I did know that but for whatever reason never thought to give it a go..... Your 80M Class D/E TX uses an adder and so does Steves Class E modules... Will try it out with this module and post the results.... The Combiner I use is a Magic T from here : https://www.w8ji.com/combiner_and_splitters.htmHaven't detected any noticeable loss at all actually and my thought was MAGIC is a good name... Of cause both TX Modules have to be very very similar in performance otherwise an unbalance will occur and maybe an adder could be a better way.... I always check the 100R resistor to make sure its cold.... But then a Magic T does provide some protection and you would at least stay on air with one module etc: Thanks for the comments Rod... Wayne
|
|
|
Logged
|
|
|
|
ka1tdq
Member
Offline
Posts: 1506
Red part turned in for a refund.
|
|
« Reply #13 on: April 20, 2021, 09:47:40 AM » |
|
Wayne,
I find it amazing that one IXDD can drive four gates, at 40 meters no less. Out of curiosity, did you come up with that yourself or is that a well know trick that I've missed.
Jon
|
|
|
Logged
|
It’s not just values, it’s business.
|
|
|
vk3alk
Member
Offline
Posts: 253
|
|
« Reply #14 on: April 21, 2021, 05:51:16 PM » |
|
Hi Jon...
Only 2 FETs are on at a time.....
Not many FETs work well actually ..... the 11N90 would not be suitable at all ....
Wayne
|
|
|
Logged
|
|
|
|
ka1tdq
Member
Offline
Posts: 1506
Red part turned in for a refund.
|
|
« Reply #15 on: April 22, 2021, 10:28:46 PM » |
|
oh
|
|
|
Logged
|
It’s not just values, it’s business.
|
|
|
vk3alk
Member
Offline
Posts: 253
|
|
« Reply #16 on: April 22, 2021, 10:50:11 PM » |
|
Hi Jon... H Bridges have been around for quite a while now and are used frequently in Broadcast Transmitters and stuff... This link explains things a bit and although it controls a motor the same principal applies ... https://blog.digilentinc.com/what-is-an-h-bridge/#:~:text=An%20H%2Dbridge%20is%20built,of%20current%20to%20a%20load.&text=If%20you%20close%20switch%201,with%20S2%20and%20S3%20closed. Just another thing too... The Modulators I use and Nigel ... and also what Pat is building are known as Half Bridges and it's important that correct switching is applied just like a Full Bridge otherwise BANG.. The delay in switching one FET off and the other on is called Dead Time and lasts for a few nS and in effect the circuit is not in control and causes distortion although low... The amount of deadtime created is controlled by resistors leading into the Gates of the FETs if you look at circuits etc: You can also use controlled Shoot through which is what I use in my PWMs ..... This means that both Top and Bottom FETs are slightly ON and conducting and the power is absorbed or controlled by a resistor.... Sorry I sound like a school teacher Just thought you might like to know Wayne
|
|
|
Logged
|
|
|
|
M0VRF
Member
Offline
Posts: 175
|
|
« Reply #17 on: April 23, 2021, 03:20:18 PM » |
|
I've decided to use the somewhat more modern Si8234 and can dispense with those gate res forever!
Waiting on some PCBs from CH...
Dc-Dc for powering the drivers (4xNCP81074), relay driver (to isolate DC) enabled from a PIC controller and some 'Modern ish' FETs too!
JB
|
|
|
Logged
|
|
|
|
vk3alk
Member
Offline
Posts: 253
|
|
« Reply #18 on: April 23, 2021, 07:31:44 PM » |
|
Hi John... You do sound rather modern - ish to me ..... only kidding of cause I do use controlled shoot through so definitely no dead time at all but doesn't make all that much difference I suppose... You will be pleased to know that are going to try those NCP driver things... Have all the parts to build Nigels 20M TX and will post the results as I progress along .... Will use 4 of those GAN FETs and 1 driver per FET might be best and see how it goes and aim for 200 Watts of carrier..... Wayne
|
|
|
Logged
|
|
|
|
M0VRF
Member
Offline
Posts: 175
|
|
« Reply #19 on: April 24, 2021, 02:58:49 AM » |
|
Hi Wayne, Yes same as the design I use. 4 GaN for 40m or 4 SiC for lower frequencies. I only want 100W carrier @24V.
Have sold 10+ now to US broadcasters as backup Tx's and a few to Aus on 5MHz.
Only built 25 and not really worth the effort commercially but has been an interesting experience!
The modulator is rather neat and hopefully the final version!
|
|
|
Logged
|
|
|
|
M0VRF
Member
Offline
Posts: 175
|
|
« Reply #20 on: May 13, 2021, 01:49:27 PM » |
|
Made a few adjustments to the layout and have built 10.
If anyone would like one to play with, please get in touch.
I'd like to see how they perform.
Regards
JohnB.
|
|
|
Logged
|
|
|
|
N4LTA
Contributing
Member
Offline
Posts: 1070
|
|
« Reply #21 on: July 15, 2021, 12:16:42 PM » |
|
Wayne
What voltage did you run these modules at?
Thanks
Pat N4LTA
|
|
|
Logged
|
|
|
|
vk3alk
Member
Offline
Posts: 253
|
|
« Reply #22 on: July 15, 2021, 05:51:39 PM » |
|
Hi Pat...
The load was appox 14R so around 40-45 volts will produce something from 100-150 watts... That's using just one module though ....
Wayne
|
|
|
Logged
|
|
|
|
N4LTA
Contributing
Member
Offline
Posts: 1070
|
|
« Reply #23 on: July 15, 2021, 06:39:07 PM » |
|
Thanks Wayne
i made a board a few months ago and am just getting around to building it. I may use it on 160 or 80. I didn't leave a place for gate swamping resistors. I probably ought to add them. 4.7 or 10 ohms should work I think.
Pat
|
|
|
Logged
|
|
|
|
|