Call Search
     

New to Ham Radio?
My Profile

Community
Articles
Forums
News
Reviews
Friends Remembered
Strays
Survey Question

Operating
Contesting
DX Cluster Spots
Propagation

Resources
Calendar
Classifieds
Ham Exams
Ham Links
List Archives
News Articles
Product Reviews
QSL Managers

Site Info
eHam Help (FAQ)
Support the site
The eHam Team
Advertising Info
Vision Statement
About eHam.net

donate to eham
   Home   Help Search  
Pages: [1] 2 3 Next   Go Down
  Print  
Author Topic: Classic 6L6 Homebrew Transmitter Help  (Read 40097 times)
KG7RS
Member

Posts: 0




Ignore
« on: December 05, 2010, 10:48:43 AM »

Hi Folks,
Scratching my head here - hoping for some expert help. I just finished building a single 6L6 oscillator transmitter. The circuit is based on the 25-B circuit from the 1939 Stancor Hamanual and inspired by W0VLZ's "6L6 Classic" (google it online) which is the same circuit with changes to component values. My version uses the component values called out in the Stancor manual. My circuit uses a B&W 40 JEL plug in coil with end-link coupling to the antenna. It is tuned by a 140pf air variable cap. Using a 40-meter crystal.

The circuit is only producing roughly 2 watts output. I've used a 50-ohm dummy load and wattmeter as well as loaded a 25-watt light bulb, which glows a dull orange. Both methods would suggest the 2-watts on the meter are accurate. I would expect 8-10 watts from the transmitter. It appears to want to load to a higher output with careful adjustment of the variable tuning capacitor but then "craps out" and stops oscillating, output to zero. Minor readjustment of the tuning cap and the RF is back.

Plate voltage is 340vdc no-load and 305vdc under key-down load. Screen voltage is 185vdc under key-down load (25K series voltage dropping resistor from the B+ supply). I have a 150mA panel meter in series with the key to measure cathode current. With the output circuit at resonance and full output (roughly 2W), the cathode current is 30-40mA, rising to 80-90mA out-of-resonance.

I suspect the circuit needs some loading capacitance on the output link coil to obtain higher RF output. I'm going to buy a garden-variety 365pf air variable and connect it first in series, then in parallel with the link coil and see if that is the problem.

I'm used to pi-tank output networks rather than parallel-tuned/link-output early types such as this. Is this normal behavior for such an output network? Anybody agree with the idea of introducing more loading capacitance in the link? Am I missing something else? Any help from the tube gurus is greatly appreciated.

73, John, KG7RS, Mesa, AZ
Logged
W1BR
Member

Posts: 4188




Ignore
« Reply #1 on: December 05, 2010, 08:16:54 PM »

IIRC, that variable cap is used in series with the link, to tune out the reactance of the coil (series resonant.)

Have you tried swinging the loop in or out from the tank coil to see if changing the coupling does anything?

That screen voltage could be a bit higher, in my opinion. At 30mA cathode current your power input is about 10 watts. I'd expect to see at least five or six watts of RF coming out.

You might check to make sure you are tuned to 7 MHz, and not doubling to 14 instead!

Pete
Logged
N2EY
Member

Posts: 5079




Ignore
« Reply #2 on: December 06, 2010, 02:28:52 AM »

First, download and read this article:

http://www.mines.uidaho.edu/~glowbugs/tx/CrystalOscillators.pdf

It's a must for anybody dealing with crystal-controlled rigs, be they power oscillators or MOPAs.

Second, as K1ZJH says, you need the series capacitor for the link. I'd use an old 170/365 uuF unit from the junkbox.

73 de Jim, N2EY
Logged
KA5N
Member

Posts: 4380




Ignore
« Reply #3 on: December 06, 2010, 03:09:07 AM »

You would have an easier task and a much better operating rig if you would forget single tube (6L6) transmitters and add a 6AG7 (or 6V6 etc.) oscillator and use the 6L6 as an amplifier.  You won't get much more than 6-7 watts which is enough to make plenty of contacts.  Of course I would go a step farther and use an 807 amplifier(or even a 6146) with pi network output and with a suitable power supply be up in the 30 or more watt class.
It doesn't take much more effort and cost to build the bigger rig and you will have something really useful. 
Either fun it is a fun task and a learning experience.
Good Luck
Allen
Logged
KG7RS
Member

Posts: 0




Ignore
« Reply #4 on: December 06, 2010, 08:25:01 AM »

Thanks for the help everyone. Pete K1ZLH & Jim N2EY - I have a variable cap on order and will add it as series capacitance in the output link. I couldn't remember if the capacitance needed be in series or parallel and had planned to experiment both ways. Research I later did in old ARRL handbooks and Radio Handbooks does clearly define this issue how series capacitance acts to tune out the reactive component of the load.

Yesterday I experimented with the number of turns for the link, temporarily winding a link of insulated hook-up wire around the tank coil. Too many turns makes the resonance dip broad and output dropped. It turned out the best number of turns is 3 or 4, just as the original link on the B&W 40JEL plug in coil I'm using. Where the link is positioned on the tank coil (grounded end, center or top end) made little difference in output. Next I may experiment with a separate link coil in various proximity to the tank coil.

Once the loading issue is resolved, I will experiment with the value of the screen dropping resistor and note the effect. 185vdc seemed a bit low to me too.

Allan KA5N - Yes, I've built lots of MOPA's so I'm familiar with the benefits. The reason for this transmitter was to try something "new". I wanted to keep the project as '39-ish & beginner-ish as possible with the use of vintage parts wherever I could. The resulting transmitter is a thing of beauty - it looks very period-authentic and I spent lots of time carefully crafting it. I also layed it out for later changes, ie crystal socket is an octal socket which can accept a 6AG7, 6J5, etc separate oscillator tube. I left room on the front panel for a second output cap to convert to a pi-tank, etc.  Even with the loading issue yet unresolved, I can clearly see the shortcomings of using a crystal oscillator loaded directly to an antenna. Yes, it chirps like a sick bird. The CW note can be greatly improved by careful adjustment of the tank capacitor and although chirp is still present, not so much that I wouldn't put it on the air.  I also noted that output frequency shifts very slightly with plate tuning, which was noted in old technical write-ups. Once plate tuning is set, I did not notice DRIFT or instability though. I suspect these shortcomings will become exacerbated once the transmitter is able to load up to higher output with the addition of link capacitance , but we'll see how it goes.

A more important issue that concerns me is high crystal current. I have some crystals purchased from AF4K, many of which are retrofitted with small modern crystals inside vintage FT-243 cases. This is a fantastic service to hams wanting to run older equipment but it's well known that such single tube transmitters are hard on crystals. I in fact blew two of my "modern" FT-243's while experimenting with this project. It seems fine with genuine old-stock FT-243's though. Anybody know an acceptable level of crystal current for various type fundamental crystals? For this reason alone there's a good argument to use a MOPA rather than a single oscillator transmitter like this.

At any rate, it's been an enjoyable project. Thanks to all for the input. 73, John, KG7RS

Logged
KA5N
Member

Posts: 4380




Ignore
« Reply #5 on: December 06, 2010, 10:21:26 AM »

OK John

Sounds like you did a good job of planning ahead.  I have aproximately 50 FT-243, only problem is that they are nowhere near a ham band.  I have several 7300 kHz units.  We used to use them on 10 meter fox hunts at 29200 kHz back in the good old days.  Most everyone used homebrew crystal controlled AM transmitters with about 10 watt output and either comercial mobile receivers or convertors.  Alas transmitter hunting went to 2 meters and is  now pretty much a thing of the past in my area (Austin, Tx).
Have fun
Allen
Logged
W1BR
Member

Posts: 4188




Ignore
« Reply #6 on: December 06, 2010, 10:29:36 PM »

What kind of oscillator circuit is it?

A Pierce oscillator can develop high crystal currents. I'd avoid using a Pierce oscillator with a 6L6. Regarding FT-243 with miniature crystals hidden inside, I know the seller claims they are lower loss and better at handling RF current that the original crystals. I tend to disagree.  One trick used by old timers was to add a small lamp in series with the crystal to limit the current. Probably something like a #47 would work, but I'd consult an older handbook for more precise info since my memory ain't what it used to be..

Pete
Logged
N2EY
Member

Posts: 5079




Ignore
« Reply #7 on: December 07, 2010, 02:48:53 AM »

What kind of oscillator circuit is it?

A Pierce oscillator can develop high crystal currents. I'd avoid using a Pierce oscillator with a 6L6.

The article I referenced earlier gives a lot of data about crystal current with various tubes and crystals.

It must be remembered that before WW2 hams mostly used the big round crystals that fit a 5 pin socket. Crystals like the FT-243 were "miniature" and developed to get the most from limited supplies of quartz.

One trick used by old timers was to add a small lamp in series with the crystal to limit the current. Pete

The pilot light was used to indicate crystal current, not limit it. And it was a cheap fuse. A #47 is too high current (150 mA), you want a #48 IIRC (2 volt, 60 mA).

73 de Jim, N2EY
Logged
W1BR
Member

Posts: 4188




Ignore
« Reply #8 on: December 07, 2010, 10:52:08 AM »

Hi Jim

You inspired me to dig out the older handbooks. I did see one reference in an old RSGB handbook that confirms 60mA. They also claim that the crystal should be a X cut, and that neither BT or AT cuts are not suitable if subjected to high RF current.

I've seen anedoctal reference that the lamp, besides being an indicator, also provides some current limiting-- but given the nature of the references they could be fact or old ham lore.

But if the lamp blows, I'd say it did it's job as the ultimate current limiting device. Grin

73

Pete
Logged
N2EY
Member

Posts: 5079




Ignore
« Reply #9 on: December 07, 2010, 06:57:57 PM »

Take a look at this article:

http://www.mines.uidaho.edu/~glowbugs/tx/CrystalOscillators.pdf

and you'll see a lot of good info.

Including the following:

Maximum crystal current for grid-plate oscillator:

6AG7: Less than 30 mA
6F6: 50 mA
6V6: 60 mA
6L6: 90 mA

Maximum crystal current for Tri-Tet oscillator:

6AG7: 35 mA
6F6: 40 mA
6V6: 45 mA
6L6: 80 mA

Maximum crystal current for modfied-Pierce oscillator:

6AG7: Less than 30 mA
6F6: Less than 30 mA
6V6: Less than 30 mA
6L6: Less than 30 mA

There's a lot more in the article. What's important is that they actually measured the crystal current and other factors, and optimized the circuit values for the various tubes.

The 6AG7 is clearly the best tube choice and the modified-Pierce is clearly the best circuit choice.

73 de Jim, N2EY
Logged
W1BR
Member

Posts: 4188




Ignore
« Reply #10 on: December 07, 2010, 07:12:27 PM »

Thanks Jim

I'll have to remember that resource, and I have the old QSTs on CD as well. Good stuff.

Pete
Logged
KG7RS
Member

Posts: 0




Ignore
« Reply #11 on: December 07, 2010, 08:20:26 PM »

Hi all,
I received the 365pF variable cap I ordered today and setup to test the transmitter with some series loading capacitance in the link coil. Success! With a 50-ohm dummy load, the transmitter is producing just a tad under 6-watts output. Plate voltage approx 300V key down. Cathode current at resonance of 40mA. 12-watts input, near 6-watts output, just as expected. I experimented with adding additional fixed loading capacitance using silver-mica caps but as the old song says - "that's all there is and they ain't no more"! In fact, the variable loading cap required about 2/3 mesh and any additional loading resulted in declining cathode current draw at resonance and obviously lower RF output. It's surprisingly stable and the CW note has minimal chirp. I could convert the companion power supply from choke-input to capacitor input and realize higher plate voltage, but 6-watts is plenty. Less concern over other nasty characteristics developing. It's been a great project and a valuable learning experience.

Photos of my completed transmitter and power supply are at:
http://www.getnet.net/~corsair510/25b1.jpg
http://www.getnet.net/~corsair510/25b2.jpg
http://www.getnet.net/~corsair510/25b3.jpg

Yeah, the tuning knob is a little on the large side, but I'm looking for a similar knob with a smaller 2-3/4" diameter dial skirt. I have a smaller one that's brass (doesn't match the nickel hardware) and has a 270-degree scale (180-degree for the cap needed). The tube shown is a modern Sovtek from the local guitar store and is temporary. I previously bought an ST-envelope 6L6G tube from a well-known antique parts dealer in Tempe, AZ at nearly $47 and represented as "new". The first one had the getter silvering flaking off inside the envelope leaving shiny (and presumably conductive) glitter inside the tube. The replacement they provided had nearly no emission and was obviously well used. Hopefully the third time is the charm as I await replacement number 3.

Jim, N2EY - That article you reference is an EXCELLENT primer on crystal oscillators. Thanks for the link. I've been studying it the last couple nights.

With regard to the sensitivity of the miniature crystals in FT-243 cases supplied by Brian, AF4K:
Brian and I have had lengthy email discussions about the current withstanding of these miniature crystals. He maintains that it's poor practice to allow crystal current in excess of 40mA but has seen evidence of his crystals withstanding nearly twice that level. While experimenting with my transmitter, I installed the time-honored 60mA bulb in series with the crystal as a rough check of crystal current. If there was any glow to the bulb filament, it was barely discernible, whether the tank is in or out of resonance. I don't know exactly how I ruined 2 of Brian's crystals, but I most certainly did. I feel these miniature crystals are not suitable for a "power" oscillator transmitter that is run hard. Mine runs at a conservative plate voltage and no sign of excessive crystal current so I will carefully try another one and see how it goes. Brian actually offered to replace the two blown crystals at no-charge! Of course I wouldn't take him up on the offer, but it speaks volumes about his customer service....Good guy. I have been successfully using his crystals with an HT-40, DX-40 and Viking II without issues.

73, John KG7RS

Logged
W1BR
Member

Posts: 4188




Ignore
« Reply #12 on: December 08, 2010, 06:20:47 AM »

Very nice workmanship, John! A First class "old buzzard" CW rig!  Glad it is working as it should.

73

Pete
Logged
AC5UP
Member

Posts: 4546




Ignore
« Reply #13 on: December 08, 2010, 12:40:02 PM »

Very nice workmanship, John!

I second that emotion... The black crinkle finish is not only boatankerly accurate, but it looks good.

Now, as to flavors of the 6L6: If you want to be true to the original RCA design you're looking at a metal tube. Good luck in finding one. Probably worse than your experience in trying to score an ST (Coke bottle) glass version of the 6L6. I've been rebuilding some Seeburg juke box amps and can tell you the metal and ST versions are about 3/4" taller than the modern style and might look odd peeking out above the front panel. On my jukebox amps the top cover will not clear the ST or metal tubes. I've tried both.

BTW: In terms of vacuuminus pulchritude the 837 is hard to beat... No kidding, it's as visually impressive as an 807 but with a style of its own.

http://www.tubecollector.org/main/837.jpg

Output is 20+ watts in class C telegraphy, much taller than a 6L6, needs a 7 pin ceramic socket, plate cap, 12.6 volts on the heater. I have a small stash of these in the garage and sommmmmmmmmmmm daaaaaaaaaaaaaaaaay...........
Logged

N2EY
Member

Posts: 5079




Ignore
« Reply #14 on: December 10, 2010, 02:46:33 AM »

Hi all,
I received the 365pF variable cap I ordered today and setup to test the transmitter with some series loading capacitance in the link coil. Success! With a 50-ohm dummy load, the transmitter is producing just a tad under 6-watts output. Plate voltage approx 300V key down. Cathode current at resonance of 40mA. 12-watts input, near 6-watts output, just as expected.

Excellent!

The actual plate input is somewhat less than 12 watts (40 mA times 300 volts) because the screen and grid currents are included. So if the grid and screen currents add up to, say, 10 mA, you're actually running 9 watts in and getting 6 watts out.


I could convert the companion power supply from choke-input to capacitor input and realize higher plate voltage, but 6-watts is plenty.

You probably realize this, but for a filter to be choke-input requires that the first choke be of at least critical inductance. For a 60 cycle full wave single phase application, critical inductance is equal to the number of kilohms of the load. For the varying load of a CW or SSB rig, not just any choke will do.

What this means in practical terms is that you have to consider the load both key-up and key-down. It also means that, unlike capacitor input, the bleeder resistor or unkeyed load has to be part of the filter design. If that isn't done, the voltage will soar in the key-up condition. Special "swinging" chokes were made which have much higher L values at low current.

I don't know what the bleeder load is in your rig; I'll assume 25 K. Such a resistor will draw 12 mA at 300 volts

With key down, the load is 300 volts and 52 mA, which works out to about 5.75 K. So a choke of at least 5.75 henries is needed for the critical value with key down.

With key up, the load is 25 K because that's the bleeder resistor. So a choke of at least 25 henries is needed for the critical value with key up.

If the above rules are not followed, the rig will still work. You just lose some of the advantages of choke input, such as good regulation. Where the rules become important is with mercury-vapor rectifiers, but that's another discussion.

What choke and bleeder are you using?

Less concern over other nasty characteristics developing.

The rig is a grid-plate oscillator. It uses the tube and wiring capacitance for the grid-to-cathode capacitor, saving 1 part. Very classic design.

I think the reason the crystal current, cathode current and power input are so low (for a 6L6) is that the screen voltage is only 185 volts. With screen-grid tubes, plate current is much more dependent on screen voltage than on plate voltage, and 185 is quite low for a 6L6 with 300 volts on the plate. But if all you want is 6 watts out, that's not a problem. Besides being easy on the xtal and keeping the input down, the 6L6 will last almost forever at such low power.

Note that with a simple dropping resistor for screen supply the screen voltage rises to the full B+ when the key is up. This can be a cause of chirp. One solution is to use VR tubes to regulate the screen voltage; a pair of 0C3s and the appropriate dropping resistor will give 210 volts. Other VR combinations can be used as well.

If VR tubes are too much complexity, a simple resistive voltage divider will work too.

Either scheme will act as a bleeder resistor if properly designed, solving that problem too.


It's been a great project and a valuable learning experience.

Photos of my completed transmitter and power supply are at:
http://www.getnet.net/~corsair510/25b1.jpg
http://www.getnet.net/~corsair510/25b2.jpg
http://www.getnet.net/~corsair510/25b3.jpg

WOW! Looks GREAT! Love the black crackle paint!

Good luck with it.

73 de Jim, N2EY
Logged
Pages: [1] 2 3 Next   Go Up
  Print  
 
Jump to:  

Powered by MySQL Powered by PHP Powered by SMF 1.1.11 | SMF © 2006-2009, Simple Machines LLC Valid XHTML 1.0! Valid CSS!