The AM Forum

HoooKay.

With the idea of getting more folks into building their own rigs, and actually using them on the air, I will issue this challenge.  Design a simple, inexpensive, easy to build AM rig that will be suitable for a first time builder, with minimal experience. Post schematics and parts lists etc here. Board members can review, offer suggestions etc.  Any one wanting to build one will have a great starting point (hopefully) and resource right off the bat.   There's so much experience on this board this should be cake!

Design criteria -

1- LOW COST - Using Solid State, Catalog (DIGIKEY or MOUSER) parts.
2- SIMPLICITY- Bare minimum parts count, Point to point, Squashed Bug, or Manhattan style construction.
3- LOW POWER - from 5-25 watts out put.
4- VFO Coverage of at least the General Phone portion of 75-80 meters.
5- EASY-to-BUILD - in sections or stages. (this way someone new can build a piece, test it, and move on when it works)
7- PERFORMANCE - Output should be reasonably clean and stable. Modulation should be 90-100%.

Here's What I'll be drawing up and building:

Starts with a Clapp Oscillator for a VFO, built around an MPF102 JFET, feeding two 2N3904's in a buffer, varicap tunning using a pair of 1N400X type diodes, about as vanilla as it gets, but simple and proven effective.

This will feed an IRF-510 based RF amplifier, which I think can be drain modulated using a filament type power transformer or perhaps a 12V power transformer (thinking easy to get radio shack type parts).

Audio will be op-amp based preamp, into a higher powered LM-380 variant chip, possibly feeding a couple more IRF510s as modulators or...   

instead of using the IRF 510s to make an amp, do like they did with the retro 75 and use a 10 watt or so audio amp IC, then feed the transformer with that.

If you want to make it more simple, get rid of the modulation transformer and use a class A series modulator.  The design just "falls out", and the audio will be much better.  Also, the positive peaks are limited only by the power supply voltage.  Negative peak limiter is also very easy with a series modulator.

If 25 watts output is really desired, a larger MOSFET should be used.

But, such a rig would be fairly easy to build.  My son Michael built a class E transmitter when he was 12 years old that would put out up to 5 watts max.  Used a class A series modulator and a voltage tripler off of a fairly hefty wall wart.

Here ya go.  Already designed and laid out for you

http://mysite.verizon.net/sdp2/id11.html

Al

Al

Nice job on that xmtr.   Sure the heat sink is big enough????

Fred

That's Stu's site - AB2EZ

I'm using a single IRFP260. Get's me up to 40 W using an old converted 13.8 power supply as an DC modulator. The 260 can provide more but I was limited to the power supply.
The FET has a dedicated digital drive chip which does a good job even under heavy modulation.

Frits

Frits,

I would be interested in seeing what scheme you used to modulate the supply?  Is it a switcher or a linear?   Using the supply as the modulator makes the most sense to me. 

Joe, W3GMS

Hello Joe,
It's an old linear power supply. I just pulled the old electronics and build in a pretty
simple modulator circuit (see attached image..my version is slightly different though)
Very simple analog design, but sound pretty good!

Frits,

Looks good.  Just essentially modulating the reference voltage.  I bet it sounds just fine.  A very similar scheme works for getting audio onto the screen grid of the final tube.  The beauty is, you can dial in the amount of audio you want while keeping the resting or steady state voltage the same.

I bet it sounds good  as well.

Joe, W3GMS   

If anybody is interested , last year I have build a 2W AM rig. This guy is a "standalone" with build in VFO but still using digital drive and series modulation.
Here's a download link to the schematics:
http://files.myopera.com/Frister/projects/80MPW.zip

Thanks to guys like Steve WA1QIX , running the Class E website with tons of info  8)

Frits

Fritz,
Great little PW rig!  Nice job.
Joe, W3GMS

Frits, that lil rig is da bomb.

;D Thanks guys. At some point in the QRP net last season, Timtron called me the channelmaster  8)
seriously, good luck on building your transmitter Ed. I'm curious to see what you cook up.
I first started by trying to modulate in a earlier stage with 2 transistor in push pull , and they were driving the Gate of a IRF510. I had a hard time getting clean audio out of it, at same point is was half way decent, but still not satisfying. That's when I ended up with the digital drive and modulating the Drain of the FET.
A whole handful of components made place for a single chip and no more worries about linearity of earlier drive stages.

Frits

Ed, having been there and done that with voltage tuned oscillators, I highly recommend you use a variable capacitor instead of varicap tuning.  There are many, many potential pitfalls with varicap tuning that are easy to avoid if you use a good variable capacitor.   FMing, unwanted RF feedback causing FMing and drift, FMing due to dirty spots in the tuning potentiometer, non-linear tuning, drift from minute changes in Vref, drift and low Q (causing sluggish oscillation) due to 1N4xxx diodes being pressed into varicap service, phase noise, and I could go on. 

Better is to use as high-Q an inductor, variable cap, and any other caps in the tank as possible.  Use as light a coupling as you can to the tank and still maintain reliable oscillation.

A decent quicky oscillator is a NE602.  It includes a voltage regulator, a buffer, and balanced outputs.  Needs one transistor to amplify the outputs to drive a TTL load.  The nice thing about this is most of the components are on the chip and therefore rigidly mounted.  Using SMD caps and a robust tank assembly and you can hit it with a hammer and it won't budge.

Hi Frits,

Just curious why C4 and C5 are in series, or was C4 supposed to go to ground?


Thanks

Phil - AC0OB

Went with a Air-variable cap in an MPF102 based armstrong oscillator (tickler type) very easy to build. Feeds a couple of 2N3904s (I think, the dang things seem to shrink more each year) as a buffer.  Works good on the breadboard setup, and is pretty stable for not having been soldered together yet.   Going to get a perf board and assemble it permanently.

Next will be the IRF-510 PA.  Should be good for 5 watts or so.   

Hi Phil,
Yeah, that wasn't really an up to date schematic. The first thing that I changed on it , was dropping C5 and
C4 is just an audio coupling capacitor. I made a few other changes, like dropping R1 all together.

If attached a more update version of the modulator, which all so includes the final stage.
(the rest of the files are in the zip file in one of my previous posts)

Frits

I am thinking a a simple one transistor mosfet oscillator in an Armstrong tuned Drain circuit with a link coupled output into a two section PI filter. A big mosfet with 28VDC on it should make some noise on 75M. Might take some fooling with the biasing and tapping of the coil but it should put up a big signal with a serious Mosfet like one from an old ATX power supply.
Hook this up to your 3055 modulator.

Mike WU2D

Modulated oscillators are great if you wanna piss off everyone else on the band...

be just like a PW solid state version of timtron's SBE rig.

Mike -  I like the modulated self-excited oscillator idea (I built a tube QRP version for the Net) but one needs a well regulated power supply and there are other issues to tame before it's ready for routine use. They are sensitive to layout and mechanical rigidity (I heard of one you could modulate by yelling at it!) so heavy buss wire and good solid constrution throughout are required.  Also, they can be tough to tune on frequency due to hand/body capacity. 
 
Still, it can be done, and done well, if you want to work a bit at it.  Personally, I think HLRs example is optimised for a different set of design criteria, there are others who run them using 811 tubes and you'd not notice unless you ran your BFO and heard a bit of "wobulation" or if you were running a flex type rig and were watching the screen.

I'd go for it.  With low power on 80 meters, you're not going to bug too many folks while you work out the issues.

i have to study this thread very carefully. in the mean time, i just want to say that ALL my previous posts have been oriented towards building a rig with similar design specs as Ed's first post. I AM the newbie builder that wants to construct a flea power rig from scratch as an educational project. :)

i'm dead in the water on my IRF510 amp but i have i have several working oscillators and small signal amps for 80m and 40m. i really need to troubleshoot my IRF510 design and figure out how to get it producing a clean signal and then learn how to modulate it.

i look forward to digging through and closely studying this thread.

THANK YOU.

I have worked Bob, W2ZM many times on 160 with his old buzzard Heising modulated Hartley oscillator.  As long as he kept the audio level down a bit it was amazingly stable and sounded quite good. 

Joe, W3GMS 

Hi Solomon,
I recommend trying my setup , to drive the Gate of the IRF510 with a single chip.
Before I came to that conclusion , I've played around with driving the IRF510 the analog way with 2 transistors in a so called "push pull" configuration. No problem getting power out of the 510. In the schematic you'll see my attempt to modulate the push pull transistors, but I never got it sounding OK. The trick is to make sure everything is perfectly linear. And dealing with bias potmeters can be tricky to.

Anyway, maybe you can use the idea of the the 2 push pull transistors to drive the gate of the 510.
*Please note that this schematic is without any low pass output filtering*

Frits

Solomon try this IRF 510 based amp. a Class C QRP amp for CW use.  It should be something you could Drain (source?) modulate.  Google "MINIBOOTS QRP AMP" for details.  Schematics below. 

I built one for my 40m QRP CW rig and it works FB with one watt or so input providing about 10 watts out.

Also has T/R switching built in as well.  Very neat and compact design, just needs a good power supply.

thanks frits and ed!

i have a semi-operational pushpull irf510 amp i have been working on in the last thread. i really dont want to give up on it and go back to a single ended design if i can make it work.

is there anything different about how one would modulate a push/pull pair versus a single fet amp stage?

Not really, in both cases you can modulate the final (IRF510) in a "high level" modulation fashion. In my earlier design I was trying a "low level" scheme, which produces the AM in an earlier stage. Which certainly can be done, but the amplifier stages after that need to perfectly linear to reproduce the AM.
Guess, I took the easy way out  ;D all though when choosing the "high level", way you have to build a modulator that can deal with the higher currents.

Frits

Apply the audio signal to the input voltage of the Push-pull stage, you could use a 12 or 24 volt power transformer (yes it would work, simple but not hifi) as a modulation transformer and pump in audio ;)  Remember you'll need a half as much power in audio as in RF final. 

worked up a Voltage Regulator for the VFO using an LM317 adjustable regulator, for 9volts inot the VFO.   I have the VFO feeding the IRF-510 fet.  0.2 volts input into the FET and I am getting ~2 volts out of the FET (all peak-peak on the scope).  The FET input voltage is 15 volts, Gate bias was adjusted with a pot to be around 8 volts or so to get maximum stable output.  This looks like a class A amp to me (device is always ON, output varies with 360 degrees of signal).   

Methinks I need a driver device? do I need to boost the VFO output somewhat to get more OOMPf out of the FET?   I have a monster 2N350x mounted in the metal can with mounting tabs on it.  If the final needs 5v on the Gate for full on, do I need to kick up the input?

I agree. I would like to have available 1 W max to drive the 510. You can pad it back as required, improves the load isolation for the VFO. So a driver would be nice. Say a couple of 5109 units in //.  I have a boatload of them and placing them in parallel with small emitter R's (similar to ballast R's in power devices) works fine. Say 4.7 ohms.

I also have some NTE 197s which look promising.  and for some reason a boatload of 2n6111... probably could work something up with one of them I suppose.

i think that if you use the 12 or 24v transformer you may want to use something like modified heising, that way the transformer doesn't become saturated by the dc on the one side.

yep, that would work quite well.  and components to handle these voltages won't break the bank either.

Excellent challenge:  I got my General class this year and I am working on a station now.  I will be interested in the replies to this challenge.

Question: why limited to a VFO?  A VXO circuit with a switch and two crystals would possibly cover the whole 80 meter AM portion, be easier to adjust to get working and be more stable as well.

Also has any one ever attempted to modulate an LM317 regulator (like they described in CQ Aug 2002)  I am wondering if it could provide modulated voltage for an FET or Transistor final stage for a transmitter.

I like the old CMOS 4049 with all sections paralleled up. You can run them with 12V on them and they will swing hard and drive the FET.

 http://www.ebay.com/itm/5-x-CD4049-4049-Hex-Buffer-Converters-IC-TEXAS-/220710920552?pt=LH_DefaultDomain_0&hash=item3363656168#ht_1352wt_1164

Actually, if you can make a VXO cover the phone portion of the band and keep it simple/low cost go for it!   

I admit the LM317 idea is interesting...  would that be a really simple form of series modulation?

Yes series modulation is what I am thinking.

Schematic of a very simple VFO, Buffer, and Final Amp.   The FET in the amp is a IRF-510. 

Not much power out, needs to be optimized a bit, but it is SIMPLE and would work OK.

You need a 50-100 uF bypass cap at the B+ feed of your final if you're gonna modulate that, otherwise the audio will get into the rest of the circuit.

I'd also suggest a L/C tuning network between the VFO and the driver - it'll help get rid of harmonics and make the thing easier to drive.

hey frits. would you recommend this series modulator setup for a an AB push/pull irf510 amp. i know you already recommended to me a low level modulation circuit driving the gates of the IRF510s but i am really interested in learning about high level modulation.

Hi Solomon. Well, I wasn't very succesfull with the "low level" modulation scheme and thats why I went for "high level" modulation.
Doing it the high level way, you don't have to worry much about how linear the amplifier stages are. The downside is that if you modulate the final, you have to deal with modulating higher current.
Take a look at the attachment of my latest RF Deck and you realize how simple it is to drive a FET.
For that particulair RF Deck , I'm using an old converted 13.8 Power supply as a high level modulator and I'm driving the deck with a DDS VFO that I've build in the past.

If you like,there are more pictures and info on my website / blog :
http://my.opera.com/Frister/blog/

hi frits. i've got a (almost entirely) functioning 35W irf510 AB amp. i just need to get it modulated. :)

is this the modulation circuit you are currently using for the amp you have attached in your last email? if so, how does it sound? can it reach 100% modulation?

by the way. your construction photos look really nice! great job on the board layout.

thanks,
sol

Hi Sol,
Thanks, yeah I like the old "dead bug" construction method. And for RF work its great to have a chunk of grounded copper clad board  8)
That schematic is a little bit old and I've made a few changes since, but that
is indeed the basic design ( Attached is a newer design). In my case , there are
two 2N3055's to split the current. The modulator has no problem reaching 100% as long as the carrier voltage is set correct and the power supply can deliver the current needed.
Congratulations on the push pull AB amp! did you use two IRF510's?

I think you may have trouble VXOing a crystal at 3.8Mhz very far  - anything further than a couple Khz will be difficult. That has been my experience anyway.

I have been recently experimenting with a NE602 as a VFO and am inpressed with the quick start up stability and overall stability.

I used a T68-6 and a small C0G 390pf ceramic cap and got a very stable 3.2 Mhz VFO which I am using for a High Fidelity 3.7 Mhz AM receiver.

The receiver sounds very nice with a discrete transistor 2 watt audio  amp driving a decent quality 8" speaker. I am using a 455Khz IF with Mouser IF transformers and a high fi detector followed by an audio filter. It is modeled after Robert Batey's high fidelity AM tuner for the BC band. It was made with the intend of being a monitor for the local net on 3715 Khz. Right now a HF 15 variable cap moves it about 60 Khz.
My plan it to put it in a box with a 6" good quality speaker and keep it tuned to the net frequency. Right now the bandwidth is about 10Khz or a little less. I may add a ceramic filter that is switchable to narrow it down when needed.

The audio uses a complimentary pair BF140 and BF139 and a 2N3904 and 2N3906 and (I think) sounds much cleaner than the IC audio amps.

very cool. thanks for the updated schematic. indeed i am using two irf510s. i'm attaching a photo of the board. you inspired me and i rebuilt it yesterday to try to clean it up a bunch.

i need to build a 30dBm gain amp stage to drive my amp (i have been testing the amp with a little 1w qrp transmitter that i dont want to use in the long run). i have some 80m and 40m hartley oscillators already assembled and buffered/amplified to 0dBm. would you recommend a transistor to use for 30dBm gain operation and what class should i run it in?

after i do all that i will be putting together your series modulator circuit and i will have an entire homemade AM transmitter!

Sol,
I've attached a schematic of simple drive based on a design originally by W7EL.
It will get your oscillator up to about 1 to 5 W , which I think will be enough to drive
your amplifier. Now the L network in the schematic is designed for 3.9 Mhz , for any other
frequency just plug the details into you favorite (online) calculator like,
http://home.sandiego.edu/~ekim/e194rfs01/jwmatcher/matcher2.html (http://home.sandiego.edu/~ekim/e194rfs01/jwmatcher/matcher2.html) or
http://www.changpuak.ch/electronics/calc_18.php (http://www.changpuak.ch/electronics/calc_18.php)
..You might have to experiment a little bit with the values.

Nice work on your amp, I hope you will attach all the leads to the FETS?  ;D

oh wow. thanks for the schematic frits. i think i might have everything i need for that driver in my scrap bins. :) do you think i could substitute a J310 for the VN-10?

the FETs in that photo are disconnected because they are burnt out! i am waiting for new ones to arrive in the mail from china. it might be a week or more  before i can get the rebuilt amp running. :(

I don't think the J310 will work. But you can certainly give it a try. And for
your parts, may I recommend http://www.mouser.com (http://www.mouser.com)
The usually have a decent selection and reasonable shipping

hi frits. i assembled your driver example. i was hoping i could get away with parts from my junk box. i put the circuit together using a J310 but im not sure if it is working properly.

what kind of output should i expect with 0dBm input?

my output is 1.7V Pk-Pk with a dc offset of about 6.7V. the waveform is heavily clipped as well. i'm not doing any filtering.

Do you use the matching L-Network Sol? ,there's no DC offset if you do and the L-Network is necessary to put a 10 ohm load on the IRF510 while using 50 Ohm on the output terminal.
I'm not sure about the input requirements, but in my QRP transmitter , I get about 3Vpp from my VFO / Frequency Divider section presented to the VN-10.
(At that point probably 10 dBm)
The VN-10 is a small signal amplifier MOSFET and the J310 is a JFET which probably doesn't work as a replacement.
In my setup , I was able to achieve about 8 W output on 3.9 Mhz

i did not include the matching network because i also did not use the IRF510 stage. i was planning to use your circuit to drive my irf510 AB amp.

i didnt realize the VN-10 is a mosfet not a jfet. i ordered some from mouser just now (along with the parts i need for your modulator circuit) and will hold off on messing with this circuit until i get the proper components.

hi frits,

in your single irf510 design, what value do you use for the source impedance when calculating the L-Network values?
edit: i just noticed in your schematic it says 10ohm to 50ohm.

i tried putting the driver together in spice but it is not working properly. it looks like the problem might be with the bjt push-pull stage. is it biased properly?

also, if i am going to use my own final amp stage (AB IRF510s) rather then your single ended irf510, do i need an L-Network between the bjt driver and the AB irf510 amp?

The 1 K resistor (R2) serves as the bias to the push pull stage. I came to that fixed value after some testing in my QRP transmitter. This value might not work in other setups. It's probably better to drop the 1 K and put a similar bias scheme in as you see for the VN-10
You have R3 as 5K fixed value, that should be set , so you have about 40 mA going through the VN-10

I don't know how your 510 amps input is designed. and the push pull stage might not be able to drive two FETS directly.
It is probably better to just design the drive with the single 510 and treat everything in separate units and proper impedance matching.

i got a copy of eagle CAD and some press-n-peel transfer film. i used frits' driver as a test board. it came out pretty nice: http://imgur.com/1ob6p.jpg

in the spirit of this thread, i am gonna try to put together a complete 5W QRP transmitter with pcb art and a parts list. i really like the kitsandparts.com style of modular circuit elements so maybe i will make a set of PCB images for each stage (VFO, amplifier, modulator, and filters). what do you guys think?

edit: i tried to redo the board layout to reduce the size down to 3"x2". i was able to do it but it is a lot more busy and i am not really able to draw a big ground plane on the board. how critical is it to draw a big ground plane?

I like it!  I especially like the modular approach.  This will allow a new builder to complete a section and troubleshoot it without having to worry about interactions.  Also makes circuit analysis much simpler, for those so inclined.

I'm with Ed  :)
That board looks nice Sol!

You can't AM modulate a Class AB RF stage with a transformer or series modulator - Must be a non linear stage -  class C, D, or E.

That's what I have always been told anyway?

hum.. I can't see why not.
The modulation basics are the same, double the voltage to the final = 100% modulation = 4X more output. This is true regardless the class of operation the amplifier stage is working in.
It might be very inefficient , because there's no use to bias the amp in Class AB.
..or am I missing something?

To those who are freely sharing ideas and schematics....

I will send FREE any of the parts below, if you are actually going to use them in a
low power AM rig.

just to keep it fair, and spread them around a bit, there are limits   :-)


I have a large quantity of IRL510 mosfets ( same as IRF510)  limit 10 pcs per person
A smaller quantity of IRFP260n ( limit 2 pcs)
and very limited number of 2N3055 ( limit 1 each)

Not that these are hard to find, or particularly expensive, but I can't use them all...

If you'd like some, let me know and send your address via pm or email....

( If you are outside US, may need a little poostage $)


Thanks for sharing!

Most any older ARRL Handbook will tell you  -  the amp power output " must vary as the square of the instantanious  plate voltage  in order for the modulation to be linear. This will be the case when the amplifier is operating Class C."

It indicates that the modulation will not be linear otherwise.

I have always been told that this is pretty important in properly modulating a transmitter.


BTW - does this "challenge" also cover receiving? I would suggest that that would make things much more interesting and also challenging.

SSbothwell

Your PC board layout will be disapointing I think. The lands are long and narrow. Thin lands will delaminate easily from the substrate when heated during soldering. If you try do use this same  type of layout using RF, you will be guaranteed to have stability problems. You need large copper ground planes and short lead lengths. Believe me, I have been laying out PC boards for 40 years and have discovered all sorts of layout problems. Most of the modern PC layout software assumes you are working on digital circuits . You have to learn the "tricks" that are available.

My favorite PC board software is Express PCB which is free. It will print to a standard printer. I use a laser printer to print the artwork on clear film and use positive resist board exposed in a simple UV frame. I develop it and etch with hydrogen peroxide/muriatic acid solution  - I used to use ammonium persulfate and it works a little cleaner - but the peroxide solution is fast and cheap. Long ago I used  ferric chloride - but  it is nasty and stains and is just messy. I hate the stuff.

If you are serious about making PC Boards, I suggest that you use the photo expose method. The iron on stuff works but the quality of the boards is not great. I can etch a .01" wide  strip using the photo process. From my experience, the iron on process, can't come close to that.

I have been working on a Vacker VFO today that I layed out out and etched this morning and may post some photos of the board later.

Have fun with your projects.

Pat

Here is a Vacker VFO board that I built today. Works very well and stayed put better than I expected. Right now it is running at 5.0 Mhz. The unbuilt area on the right is a 5 pole lowpass filter and resistive attentuator.



 Here is the board after being developed

(http://i620.photobucket.com/albums/tt286/n4lta/IMG_1105.jpg)

Board after stripping resist with steel wool

(http://i620.photobucket.com/albums/tt286/n4lta/IMG_1107.jpg)


Board after plating with MG Chemicals Liquid Tin


(http://i620.photobucket.com/albums/tt286/n4lta/IMG_1108.jpg)


Completed and tested Vacker VFO. It currently is running at 5.0 Mhz and I am happy with the stability


(http://i620.photobucket.com/albums/tt286/n4lta/IMG_1109.jpg)

Below is a prototype 75 Meter high fidelity am tuner and audio amp. It was made to monitor the local net frequency. The front end uses a NE602 which is not known for high dynamic range so it has a very narrow bandpass filter. A different filter would be required to cover more than 50 Khz of the band. The IF amp uses two conventional IF transformers. The AGC is somewhat unconventional and works very well. The IF feeds a High Fi AM detector as described by G3YNH. It has a Sallen Key bandpass filter after the signal is detected Sounds very good. Bandwidth is about 10 Khz. The audio amp uses a couple of complimentary output transistors and is very clean at about 1.2 watts. It drives an 8" speaker easily without distortion.

Audio Amplifier

(http://i620.photobucket.com/albums/tt286/n4lta/IMG_1111.jpg)


High fidelity 75 meter tuner

(http://i620.photobucket.com/albums/tt286/n4lta/IMG_1110.jpg)



Pat

wow pat, those boards look really amazing!

how do you tin the board?

that first board i etched was using the default copper trace sizes. i am trying to put together a better one with larger traces. i'll post that soon.

i dont think have the resources to setup a dark room for film development, maybe someday.

Solomon,

You don't need a darkroom. The board is sensitive to UV light only. You can use a florescent bulb but the best way is to buy a low cost UV tube and build a wood frame with a glass top. Then you lay down the printed positive, lay the sensitive board on top and  put a cover on top (or build a hinged lid like I did)

Turn on the light for about one minute and thats it. You develop the board in normal room light. I can go from printer to finished, plated board in 15 minutes. I drill it after plating on a low cost drill press using resharped carbide drills - They are cheap  - use a #65 for IC and parts and a #60 or larger leads - like TO220 etc and a #55 and #50 is also nice for larger holes..

The exposure frame takes a few hours to build and then it's finished for a lifetime. It really is worth it if you plan to do IC type boards single sided. You can make traces thin enough to run between the IC pins easily.
This is useful when making microprocessor boards etc. Thin traces are not real useful for RF.

The plating is done with MG Chemicals Liquid Tin - It cost about $35 per quart but can do lots of boards. It is necessary when doing SM type work. I clean the board with steel wool (brillo) and that also does a good job of degreasing the copper and then immerse in the Liquid Tin for a couple of minutes. Then I take out the board and lightly clean (lightly or you will remove the tin) with brillo again. Then solder away.

The chemicals required are the positive developer (about $10 for enough for lots of boards) - Don't try to make your own with sodium hydroxide as described on the net (unless you get a formula for one with a buffer) - without the buffer - you will screw up many boards.  You need an etchant that can be made with two parts hydrogen peroxide (drugstore peroxide) and one part muriatic acid ( you can buy a gallon at Lowes for $6 - don't get the low fume version  - it has less concentration - you need about 30%. Pour the acid carefully into the peroxide - not the opposite. It is very cheap and works great - way better than that nasty ferric chloride - not quite as good as ammonium persulfate - but that has to be heated - (I used it in a pyrex dish on the range for 30 years.)

BTW - I layed out a board last week for the modulator - I want to give it a try. I have always heard that it is a good idea to modulate both the driver and final with a SS transmitter.  It is easier to get 100% modulation.


Pat

Again - Have fun

oh wow. i didnt realize the exposures are so simple. perhaps i'll put together a uv light setup in the future, but for now i have a bunch of untreated copper and press-n-peel paper to work through.

i'm attaching a pcb layout for frits' driver board. this one uses larger traces and a giant ground plane. is this a better layout style?

edit: i changed the design rules a bit and beefed up the traces. i've got both versions attached to this post now.
edit2: i just noticed the newer layout is missing a trace. i have to work on this board some more. pcb layouts are tricky!

Looks good. Build one and see how it works


PC Layout is somewhat an art. You learn a lot of tricks as you go. I enjoy doing it almost like working a puzzle. I like to see how compact I can squeeze thing in.

You can check and double check and then an error pops up right in the middle of building it. I'd be embarrased to show you my box of boards that one day I hope to salvage the parts out of. The monitor receiver that I show in the photos is an expansion of a compact board I made last week. It had a high frequency oscillation that I could not track down. I ended up doing a new layout with a little more wasted space.

Pat

Nice boards!

What is the local net freq?

Did you ever get your log periodic up?

here is a type3 mixer i put together. i plan to use this in a superhet receiver to go along with my 35w transmitter. :)

i am etching a driver circuit now for my irf510 amp.

edit: and now the driver board is attached. i dont have the proper dril bits to finish these boards but i should have them in a day or two.

The local AM net (OF Net) is on 3715 Khz on Sunday afternoon at 3:00 pm EST and at 9:00 am on Tuesday mornings. Most of the guys run 100 watts or less.

Negative on the log antenna - It's still standing in the garage corner in the box. I have had trouble getting someone to put it up for me. The local club used to be a good source but the members as myself are suffering from gray hair itis - After falling off a ladder up 17 feet 2 years ago while putting up a dipole - I no longer climb. I will probably have to pay a professional to put it up. I  back burnered it when the upper bands were slow. Now I need it. Someone said there were 2  guys in Columbia, SC who would work for $100/hour together and I may have to go that route.

Sol,

I like those boards a lot. Plenty of ground plane. I think they will really work well. My software won't do that. I may have to look at yours.

Pat

OK. I've checked into the OF Net once or twice. IIRC, Terry, N4RQ is a semi-regular. He used live about 40 miles from here.

I might be convinced to give you a hand with the log antenna. I won't volunteer a certain AMer in NC, but he might help too. You can drop me a PM to speak further.

frits, i put together the driver. it seemed to be working great. using a 12V/1.5A power supply i was getting ~4w out from 100mW input, but then i decided to try connecting it to a 13.8V/3A power supply and i think i burnt out the irf510. oops.

i am using a completely inadequate heatsink on the irf510 and i positioned the irf510 poorly so the heatsink is bumping against the l-network capacitor.

how big of a power supply should i be using?

edit: i replaced the irf510 and hooked it into a 12V/750mA power supply. now i am getting perfect  7.5VRMS output from a 1.1VRMS input. the waveform looks perfect.

this will be perfect to drive my push/pull irf510 amp.  :)

Nice work Sol!
The FETS Drain is connected to the casing, better to isolate the heatsink, you don't want DC and RF flowing over it  ;)
The FETS current depends on the load from the L-Network and the level of bias,
in your case 750 mA should be OK, but better to not max out your power supply.

Great progress!
Frits

Steve,

Did you get my pm? I sent it (I thought) but it isn't shown in my sent messages.

Sol.  Good work on the amp driver. When you modulate it  - I'd run it in class C as most of the literature states that you need a nonlinear final (square law) to get linear am modulation. That should make it a little harder to drive, but it will run much cooler and be easier on the output devices. You may want to consider modulating the driver also. Running class C will be more stable also, and will do away will the bias circuit and thermal problems there.

You need a power supply that can deliver the current that is needed and not "squat down" when the amp is driven for lond periods - the heat sink needs to be big enough to keep the device from getting smoking hot when the key is down for a minute or so since it is goind to be AM. This is with the amp fully modulated and it may be a challenge to keep the device cool at full rf power and fully modulated.

Have you blistered you index finger tip yet? You probably will. I have branded myself with a TO39 circle many times testing how hot a device is.



Pat

thanks for the kind words. i've got the modulator circuit built but i used the wrong package size for the pots. i need to go find some 16mm pots from the electronics shop. i'm heading over there in a minute.

how do i determine when the circuit is running in class C?

in class C, the driver shouldn't draw any current when their is no input signal, right?

No current on the final with no signal. Usually a resistor or choke to ground from the base works well with a single transistor.  You can probably adjust the bias until the bias current is just barely cutoff with no drive signal. Should find some examples in the literature for CW transmitters.

It will be a little harder to drive but the efficiency should be quite a bit better.

Pat

i finished assembling the modulator.

frits, what would be the proper procedure for testing this beast?

does the 100Kom pot on the power input control the peak output of the modulator?

edit: i am pretty clueless on how to test the modulator. i hooked it up to a 12V power supply and checked the Vout on a DMM. it read 10V regardless of how i adjusted the pots. i also tried inputing a 200Hz audio signal and that didnt seem to change the output.

i'm nervous about hooking it up to an amplifier without first verifying that it is working properly and is set for the right peak output so that it does not damaged the amplifier.

any suggestions?

Hi Sol, No that sets the carrier level.


The basics of AM modulation, at carrier level with no audio applied your voltage should be half of your maximum voltage. If you double the voltage (under the same load) you have 4 times the output power. And that is what you want to achieve. If you apply 12V to the modulator , set the output to 6V (carrier level) and feed it some audio, you'll see that the audio peaks reach almost 12 V (there's some loss in the 2N3055)
Hook up your scope to the emitter of the 2N3055, you don't have to put a load on it , it will show on the scope (if not put the probe on the base). Now with the carrier potmeter all the way down, there should be little DC out, slowly turn the the potmeter to 6 Volt and put some audio into the circuit. Watch the audio being carried on the DC.



Perhaps it is better to modulate the drive circuit and leave the amplifier as it is?
In that case you have to make sure the amp is really linear.
If you decide to modulate the amp keep in mind that that a higher current needs to be modulated and a single 2N3055 might not be able to handle it.

If you do try it on your amp, monitor the current to it and the temp of the 3055, if something fails, it's likely the 3055. It's a good idea to stock up on a few of them  ;D

Frits

frits, thanks for the explanations. i wont have a chance to test anything until tomorrow but i will report back with my results.

so i'm having some problems with the modulator circuit. when i first power on the circuit it jumps to 10V regardless of the carrier pot setting. it then drops down to around 2V.

at this point, if turn the carrier pot to full resistance between the wiper and ground then i see a 750vrms dc signal on the 2n3055 emitter.

if i turn the carrier pot to zero resistance between wiper and ground then i see a ~2V DC signal that turns into a transforms into a ~9mhz sine wave every couple seconds.

do you think the circuit layout is causing instability? i didnt use a ground plane on this circuit layout.

Sounds like, you have to put a load on the circuit. a 100 ohm resistor should work nicely or a 12 V light bulb.
http://youtu.be/-vqEPzr8HCY (http://youtu.be/-vqEPzr8HCY)
If you measure at the base of the 2N3055, over the 467 Ohm resistor, you should be able to test it without a load on the 2N3055
as of the voltage jump, I've noticed that to , when the circuit gets powered it jumps the voltage for a split second. That might be the TL071 going in "high" state for a moment.

oh. i was measuring off of the 2n3055 emitter.

i just put a 100ohm load across emitter->ground and now the carrier is a solid dc signal but i am only able to adjust it between 800mV and 2V when measuring across the emitter.

if measure across the 467ohm resistor on the 2n3055's base then i see a 500mV  10-20MHz waveform that doesnt change when i adjust the carrier pot.


i tried inserting an audio frequency sign wave from a function generator and it is definitely modulating the power supply. this is good news. :) i just need a higher voltage carrier.

Hi Sol, This circuit should be able to (almost) match the DC input on the 2N3055's output, by adjusting the carrier potmeter to full open.
I have three of these circuit up and running, with no problems. If you're sure
about the circuit design, try lowering the 100K resistor a little bit and see if
you can reach the 12 Volt out.
I have to ask, are you doing a realtime test or an LT-Spice simulation?

i was using a real circuit board. i think the problems might have been with my board layout. i designed the board to include a large groundplane and thicker traces and the new one seems to be performing better.

i'm attaching side by side comparisons of the two boards.

with the new board i am able to set the carrier voltage to whatever percentage of the total power supply input. tried inserting an audio waveform and was able to modulate the signal successfully.

i set the audio input to a 440Hz sinewave at +4dBu and was able to produce a nice clean 6V modulated carrier. :)

i want to try modulating the driver board next but i only have one signal generator. i need one to act as vfo and one for the audio test tone.

maybe i should build a two tone audio test oscillator? is that the standard way of testing radios?

i put together a basic Twin T oscillator as an audio source for testing the modulation. i'm attaching images of the unmodulated and modulated signals when powering the driver board off of the modulator board.

how can i measure the modulation index and how can i check make sure there is no distortion in the modulation?

hi frits. are you sure that when the carrier is modulated it should reach 12V?

i have the carrier set to 6V and if i turn the audio up the audio then it clips at 7Vrms (8V Pk-Pk).

edit: hmm. i think something weird is going on with the driver. i just noticed that my variable power supply wont go above 8V went hooked up to the driver.  ???

i tried using a 13.8V/3A fixed supply and the VN-10 and IRF510 started smoking.

Hi Sol,
In the design stage, I recommend to monitor the current at all times. I've learned the hard way by blowing up 2N3055's with my 30W RF Deck.
As of the modulator , yeah it should be no problem. The best thing is to set the carrier potmeter full open, take note of the voltage output (and current) and then set it back with carrier potmeter to half voltage, while doing that keep an eye out on the current. Repeat that step a few times and make sure the max output stays the same. If not, see if it is the power supply or the driver circuit.

As of the driver circuit;
The VN-10 and the 2N push pull transistors should be connected to a fixed voltage, you only want to hook up the output of the modulator (through a RFC) to the IRF510.

Mind you.. I've never been really satisfied with the AM produced with
this driver circuit. It most be something in the linearity and classes
of operation. But it is a great circuit to boost VFO power up a to a
few Watts. (I turned you on to this circuit because my original
understanding was that you wanted to modulate your amp)

If you still like to modulate the driver,
I recommend using a dedicated driver IC , the IDXF604PI, so much more easier, just drive the IRF510 with that. This chip is good up to and a little beyond the 80 Meter band. take a look at the schematic of my 80M RF deck, you can do the same with the 510 and probably directly drive the 604 chip with a VFO

Frits

i think the modulator might be fine and the problem is with the rf driver.

i suspect the voltage limiting has to do with it triggering the crowbar protection on my power supply.

if i connect the driver to a 13.8V/3A fixed power supply then the VN-10 and the IRF510 immediately start smoking.

i also noticed that when the VN-10 gets burnt out, the power supply voltage (when using my variable power supply) is no longer limited.

hey Frits, its been a while since i posted in here but i wanted to give you a little update. i havent had as much time to work on electronics projects lately but this week i did get a chance to sit down at the work bench.

i tried using your modulator circuit with a 100mW AB amp design i found in "experimental methods in rf design."

here is the AB schematic: http://i.imgur.com/0Ucf4.png
and here is my 'ugly style' board: http://i.imgur.com/hasZW.jpg

i'm using a homemade hartley VFO (dont have a photo or schematic on hand currently) to drive the amplifier.

here is my modulator board (i think i posted a photo of this board in this thread earlier): http://i.imgur.com/dLT5Y.jpg

here is the output from the AB amp WITHOUT using the modulator circuit: http://i.imgur.com/lVa3K.jpg

when i set the modulator board to 6V carrier and power both the 2n3904 the 2n3966 from AB amp with the modulator then i get this really distorted waveform: http://i.imgur.com/YKD30.jpg

if i input a 500hz sinewave into the modulator and turn up the audio level then i end up with this waveform: http://i.imgur.com/I5txm.jpg

any idea why the waveform is getting so clipped when i try to use your modulator circuit?

Hi Sol,
I think, lowering towards 6 Volt (carrier) changes the class of operation. Looking at the schematic of the 100 mW amplifier. Try modulating Q2 only (V3), keep Q1 at 12 Volt (V2).
Now with Q2 at 12V, you want to see 100 mW out (if that is the max) and at 6 Volt (carrier voltage) about 25 mW.
You may have to experiment with the feedback on Q2 (R5 and C6) or eliminated it completely. (this might however lower the maximum output of the circuit)

Frits W1FVB

ok, i've been kinda following this thread, and i got an idea of something i want to do with some of these designs. i have a grounded grid amp that uses a quad of 572Bs and it needs about 15 to 20 watts to drive it. however, it doesn't have a tuned input, so my icom 718 will just barely put out enough power to drive it, and my yaesu ft-901 has to be retuned everytime i adjust the amp tuning or the antenna tuner. what i want to know is can i make a broadbanded output network on a low power solid state transmitter that instead of having an impedence of 50 ohms over its range, the impedence matches the input impedence of the grounded grid amp.

i was wondering if it was a problem that i was modulating both transistors. i isolated Q2 powered it off the modulator. When i turn up the carrier i get a nice clean sinewave: http://i.imgur.com/DCoDD.jpg

i am still having trouble with modulation. the signal gets really distorted as i turn up the audio input.

here is a photo of the modulated carrier: http://i.imgur.com/MuxAr.jpg
and here is a photo of it in XY mode: http://i.imgur.com/xzzvZ.jpg

i noticed that if i reduce the carrier to the point where it looks distorted (such as in my previous post) the trapezoid pattern looks more like a trapezoid (but everything still looks pretty distorted).

I'm curious to see, what the circuit does without R5 and C6 installed. (Keep the modulator on Q2 only) Output might be way down, but hopefully no distortion.

You can certainly design an L-Network impedance matcher, if you know the amps input impedance , but it wouldn't be very broad banded. Using a transformer probably gives you a better range. I'm certainly no expert on this, maybe you should post this question as a new technical Q&A post. I'm sure there are other people on this board with more experience then me  ;)

hi frits. my previous tests had been done without much attention to details. i decided that before i try rebuilding or changing the modulator i should go through one more round of tests where i record everything i do and i use an amplifier that is not entirely home made. i used an hfprojects PA-100 which you can see here: http://www.hfprojectsyahoo.com/sla5wamp.html

here is the unmodulated output from the PA-100: http://i.imgur.com/XT78K.jpg

i'll just walk you through my test process of testing the modulator. maybe i am making some critical mistakes that i dont realize.

1. connect the modulator output to my scope input with a dummy load (50ohm). adjust the carrier level until it is half the power supply voltage (6V). http://i.imgur.com/9MCj8.jpg

2. adjust audio input level to carrier for largest possible Pk-Pk signal without it clipping at the highest desired audio frequency. http://i.imgur.com/6ykiO.jpg

3. connect modulate output to power input of PA-100 amplifier and observe signal with varies audio frequencies inserted into modulator.

5kHz: http://i.imgur.com/bEt2i.jpg
500Hz: http://i.imgur.com/ktkZS.jpg
50Hz: http://i.imgur.com/Zc0DL.jpg

-does this behavior seem normal? why is there way less modulation at higher frequencies?

-i can physically hear the audio signal coming from the modulator circuit. i have no idea what is doing this but is it anything to be concerned with?

-where should i check the audio if i want to  look at the trapezoid pattern in XY mode? should i check it at the modulator output or the output from my audio signal generator?

Hello Sol,
It's easier to tell what the modulation is doing if you slow down your scope to 1 ms, then you can actually see the modulated audio frequency.

If you have a dual scope you can put one channel on the audio signal and the other on the RF output and compare the signals. I've attached an example image, the brighter trace is the audio input (speech in this case). If you compare the peaks you'll notice that they are not perfectly the same.

It's OK to hear some audio coming out of the circuit, most be the 3055 buzzing  ;D

Anyway, I'm curious to see what it looks like if you slow your scope down. (The last image you've posted of the 50 Hz signal looks very strange)
Personally I've never played around much with a XY setup. And I know there's some good information to be gathered out of that. Maybe somebody else can jump in on that subject.

oh cool. i never realized that was how people got those cool views of the modulation.  i also just found the peak detect display mode on my scope. it should help with the crappy cell phone photos.


heres images of the modulation at 50hz, 500hz, and 5khz. i used the same calibration method i used in the previous post (set signal generator to 5khz and set audio source voltage to maximum undistorted signal).

50hz: http://i.imgur.com/p8g7B.jpg
500hz: http://i.imgur.com/FDEz9.jpg
5khz: http://i.imgur.com/xzJcl.jpg

looking at the signal at 1ms makes it clear that i am achieving full modulation at 50hz but i am not getting anywhere close to it at higher frequencies.

Do you keep the audio input level the same across all three tests? Thinking, maybe the circuits low frequency response is better then
at higher frequencies?
If so, increase volume on the 500 and 5K test and see if it modulates better.

 

i tried that but if i increase the input level at all past this point then it starts to distort really badly at higher frequencies.

if it is not to inconvenient for you, would you be able to post photos of the modulation when using various frequency sinewaves? it would be nice to compare results and see if there is a problem with my board.

The best way to test the modulator board is to test at DC level. Forget about any RF parts and hookup up a simple load. Get a 12V 5W or 10W automotive light bulb and measure across the bulb at DC level.
This way you know for sure the board works OK.
http://youtu.be/-vqEPzr8HCY

Try that, and if it still shows up, not being capable of reaching 100% , I'll be happy to compare some results.

Frits

i just tried that test using an 8.3W bulb for a car brakelight. i hooked up the lineout from my computer to the audio in of the modulator and played some music. the bulb glows but it is pretty dull (especially compared to if i sort a 12V/3A power supply with the bulb) and it doesnt change brightness much at all. i guess i should try in the other room using my stereo receiver and see what happens.

same situation with my home stereo. on my scope i can see the audio modulating the carrier signal and it is 6VRMS but it is only 600mV Pk-Pk.

(in the attached image 0V is at the bottom of the window.)

Sol,
It looks like there's a problem with the circuit. It shouldn'T have any problems "swinging" the voltage to a double.
I hooked up my scope and made a little movie this morning:
http://youtu.be/uttBWvyDK1w (http://youtu.be/uttBWvyDK1w)

(You can see that my DC at carrier level is not nice and flat under load of the RF deck! However this isn't noticeable
on the audio from the receiver)

I expect that something is not working right with your circuit. I'll attach the latest schematic of the circuit and that is basically
the one I'm using in the video. (I just noticed that the 10uF cap symbol after the 5K6 should be an electrolytic)
Maybe check pin 2 & 3 to the op-amp?

hi frits. sorry i dissipeared for a couple days. i got busy with work and school.

i am gonna go over the my modulation circuit tomorrow and try to figure out what is going wrong.

just to be clear, if the modulator is working correctly and i am inserting an audio sinewave into the carrier i should see a 12V Peak to Peak signal?

No problem,
The audio signal "rides" on the carrier. If it peaks reaches 2 times the carrier voltage you've achieved 100% positive modulation
And with the circuit , yeah set the carrier voltage half of the power supply voltage and the audio peak should be able to reach the 12 volt again. However there is a little loss in the 2N3055 ,to compensate , just turn the carrier level a little bit lower.
Keep testing it on the light bulb

Hope you find the problem,

Frits

SOL,

Let's say you expect a 5 Watt output from the circuit. (Assuming Class C or E final)

This is about 450 mA of current at 12 V peak.

Rload = V^2/(2XP) = 144/10 = 14.4  or 15 ohms.

Recommend you place a 15 ohm resistor between emitter of 2N3055 and ground rated at 5 Watts or more, and measure emitter voltage.

Also check your voltages at pin 7 and 4 of the IC. 7 = 12V (or whatever supply voltage) and  4 = zero.

As the 100k potentiometer is moved up toward 12v, you should see the emitter voltage of the final pass transistor 2N3055 rise as the potentiometer wiper goes toward 12V, since the wiper supplies voltage to the non-inverting input of the opamp. Stop at say 6 V.

Apply audio. Voltage at emitter should vary between I would think about 2V and 10 Volts in step with audio.

Phil

It came to my attention that the feedback resistor in the modulator schematic was incorrectly listed as 50K instead of 150K.
The 50K value is most likely to much feedback to the TL071 opamp.

Here's the updated schematic:
http://files.myopera.com/Frister/projects/modulator.jpg (http://files.myopera.com/Frister/projects/modulator.jpg)

Hope you didn't toss it in the garbage yet Sol  ;)

very interesting. my board has been sitting on my desk forever.i assumed i design it incorrectly and havent gotten around to checking it against the schematic. i'll try switching out the feedback resistor and see what happens.

hi frits. i tried assembling the modulator in LTSpice  but it doesnt seem to be operating correctly.

the audio signal gets biased to ~30V at the op-amp input and the op-amp output is a flat ~29V dc signal.

does this circuit look right?

simulation needs a load resistor between L1 and ground
Bias the op amp positive input so the output is just under 50% VCC of your final
Add a resistor between the positive input of the op amp and ground to pull the voltage down. this can be a trim pot to make it adjustable.