New Modulator - 4kW peak output you can hold in your hand!!!

Welcome and feel free to ask questions about class E, mosfet's, and AM.
You must be registered to post.

New Modulator - 4kW peak output you can hold in your hand!!!

Postby steve_qix » Mon Feb 06, 2006 6:01 pm

Hi!

The modulator shown in the picture is part of the "Class E-Z" transmitter I've been working on. This transmitter is a two band (80 and 160 meter) transmitter featuring a broadband, untuned driver stage flat from 100kHz to 4mHz. The driver also works on 40 meters, however this needs much more experimentation before making any pronouncements in this area. The only RF tuned circuit in the unit is in the RF output stage. The picture, below, is of the modulator.

Image

The modulator is a Pulse Width design, and includes a 4 pole torroidal filter, keeping size and weight down. The switching frequency is 160kHz, and the output filter corner is set at around 15kHz. The modulator also includes a 6 pole anti-aliasing (input) filter. Due the nature of pulse width modulators, and the high efficiencies involved, the power supply requirements are significantly reduced as compared to other designs. The transmitter, as shown, uses a 135v 3.5 ampere power supply.
The picture below shows the entire, operational transmitter.

Image

The transmitter runs 45 volts at around 9 amperes - about 400 watts of power. The modulator can deliver up to 135 volts at 30 amperes giving a positive modulation capability of 200% positive!! The modulator is ultra clean, and the frequency response is flat from dc to around 5500hz, at which point the input filter starts attenuating the modulating signal. A DC coupled negative peak limiter is part of the design. The builder can optionally choose any filter corner frequency within the audio range.

My goal is to publish the entire transmitter design in QST, and to this end, I have chosen "conservative" filter corner frequencies.

This has certainly been an interesting project!!!

All feedback welcome and accepted :-)

Talk later and Regards,

Steve
Visit the class E web site at:

http://www.classeradio.com
User avatar
steve_qix
Class E Guru
 
Posts: 434
Joined: Fri Jan 31, 2003 11:03 am
Location: Townsend MA

Postby W2INR » Tue Feb 07, 2006 7:05 am

CAn't wait for the doc

Type A here

Dying

Help
G

The INR
W2INR
Member
 
Posts: 25
Joined: Fri Apr 18, 2003 12:03 pm
Location: Syracuse New York

Postby cableman » Tue Feb 07, 2006 9:35 am

Steve,

Don't go to work anymore!
Don't sleep!
Only work on the document!!

I am a fool and want a foolproof design :lol:
cableman
Member
 
Posts: 57
Joined: Tue Jan 04, 2005 8:40 am
Location: Netherlands

Postby W1DAN » Tue Feb 07, 2006 6:24 pm

Hi Steve:

Looks great and sounds great on the air!!!

a couple of questions:

1) did ya just get rid of two of the four PDM output transistors, or use better transistors in the output section? Any different diodes?

2) is this bi-phase or single phase PDM?

3) did you notice any non-linearity on the output filter because of the Toroid core at the input? Wanna pass along the core info and values?

Some thoughts:

I have been playing with PDM output filters by looking at the calculators by KE1GF and data by WA1GFZ here, and by experimentation. I finally was able to really attenuate my 120kc switching images extremely well by using 3 air coils and 3 caps as well as a single coils/cap as a 120kc notch. My audio is flat to about 8kc. I am hoping to get all this down to one LPF coil and cap and a single notch. The notch is really good, but the phase shift starts early.

My other thought is to do bi-phase PDM.

On your low level multi-pole anti-ailiasing filter...I really like the Twin-Tee notch filter. It provides a good steep drop at your corner frequency. Maybe follow that with a single LPF.

BTW N9NEO's full bridge switching supply/modulator works very well.

Good stuff!!!

Thanks!

73,
Dan
W1DAN
W1DAN
Member
 
Posts: 58
Joined: Wed Oct 15, 2003 11:29 am

Postby frank carcia » Tue Feb 07, 2006 7:52 pm

yes I would like to know the flux density of those inductors running at such a high power. What is the material. I had to use Ui=60 iron on the viking 2 and still air wound inductors sound better. The cores I used were much bigger 3" od 1 " ID 1 " tall
frank carcia
Senior Contributer
 
Posts: 765
Joined: Fri Jan 31, 2003 10:48 am
Location: Enfield Ct. / Niantic Ct.

Postby steve_qix » Tue Feb 07, 2006 9:04 pm

frank carcia wrote:yes I would like to know the flux density of those inductors running at such a high power. What is the material. I had to use Ui=60 iron on the viking 2 and still air wound inductors sound better. The cores I used were much bigger 3" od 1 " ID 1 " tall


Hi Frank,

I'm using the CWS-Bytemark part # CH777060 High Flux core (two of them in the filter). U=60.

The core is $15.00 (expensive), but it sure works well in the application. The audio appears to be unaffected over the range. I've done the old triangle wave test. No noticable effects. The cores do not get warm, even after extended use.

The rig is single phase PWM - 160kHz switching frequency.

So far, so good :-)

Regards,

Steve
Visit the class E web site at:

http://www.classeradio.com
User avatar
steve_qix
Class E Guru
 
Posts: 434
Joined: Fri Jan 31, 2003 11:03 am
Location: Townsend MA

Postby W1DAN » Wed Feb 08, 2006 9:30 am

Hi Steve and Frank:

Enjoyed listening to Ya'll on 160M last night. Got some my answers there! But..are your transistors and diodes beefier or did ya just not use all 4?

When I play more with my filter, I am thinking that I will keep my first inductor air wound and the others will probably go toroid.

I heard ya say something about ripple correction with the UCC25701 chip. Bon N9NEO asked this question on the board a while ago and was wondering if anyone used a small PS filter cap and nulled the ripple in the PDM chip. It reminded me of the hum balance pot in my Dynaco amp years ago...

If I ever get my PDM filter down to be simple, I may look also into negative feedback around the modulator..if the phase shifts fo not get me!

73
Dan
W1DAN
W1DAN
Member
 
Posts: 58
Joined: Wed Oct 15, 2003 11:29 am

Postby ke1gf » Fri Feb 10, 2006 4:32 pm

Steve, I started building my new rig, I found that the undervoltage protection on the IXD(D,H)4?? chips is very sensitive, bypass capacitors are super ultra critical... When I placed my scope at a slower time/div I found that I had dead spots in my drive... Paralleling several high quality capacitors of 3 decades in value solved my problems, I didn't test it at 8MHz yet! This would be a good thing to mention in your article, don't cheap out on thoes caps. :wink:
User avatar
ke1gf
Young Brainiac
 
Posts: 229
Joined: Mon Feb 03, 2003 11:38 am

Postby frank carcia » Fri Feb 10, 2006 4:45 pm

Bill I had exactly the same problem with the 160 meter rig when I went to 4423s. I also found them fighting each other when a bunch were in parallel. I even had 10 ohm series resistors on each driver. I had 2 chips 8 drivers in parallel. one or 2 parts would get very hot. I ended up removing 2 devices so 2 chips or 4 drivers are in parallel. The remaining unused parts will be used to put the second final on line someday.
This was even a problem after i added all kinds of bypass caps of different values. fc
frank carcia
Senior Contributer
 
Posts: 765
Joined: Fri Jan 31, 2003 10:48 am
Location: Enfield Ct. / Niantic Ct.

Postby ke1gf » Fri Feb 10, 2006 6:11 pm

Frank I'm using IXDH430CI IXYS T0-220-5 drivers, about $6/piece from all american. One driver chip per FQA11N90. I just tested 40M, looked good, needs some polishing. I built a little 4mW drive, 75M 100W carrier E-rig today. I'm using it for the driver on my 20 pill rig at the moment, but I'm getting ready to build up my 160 deck. I decided that I wasn't going to do multiband. I can always build a multi-band deck later if I want to. I was really stumped when the chips kept cutting out, but I read the data-sheet again and checked the ripple on the Vdd and found the problem quick. Very cool 2 less driver chips per fet, 2 less (N-P)chans per FQA11N90, two less TTL chips. Drive's is good on 40. Not bad for an afternoon eh?

--Bill 'GF
User avatar
ke1gf
Young Brainiac
 
Posts: 229
Joined: Mon Feb 03, 2003 11:38 am

Postby frank carcia » Sat Feb 11, 2006 3:27 pm

Bill,
very good. These new drivers will make it easy to get both finals working in the 160 meter rig. Maybe we could pool our money and buy a bunch of the best ones for the price. I prefer single band finals myself. This way if you take it apart or blow it up you can still operate. Also changing bands is a PITA.

The dropout thing is weird because at fast scope speeds all you see is a bit of fuzz. I went looking for something fast before I found the low frequency amplitude change. My predriver is a proto board with a ground plane and it had a pile of bypass caps. I added a bunch more of different values and just couldn't get rid of it. One of the parallel chips would also run a lot warmer to the point of almost hot. I suspect it has something to do with the inductance between outputs setting up a low frequency oscillation. The VCC bypass caps were right next to the pin. I even had parts with the same date code.
That is good you have a part that drives a final I need to think about a driver for the 75 meter rig. Band was so dead this winter I stuck to 160.
frank carcia
Senior Contributer
 
Posts: 765
Joined: Fri Jan 31, 2003 10:48 am
Location: Enfield Ct. / Niantic Ct.


Return to Class E Discussion

Who is online

Users browsing this forum: No registered users and 1 guest

cron