Vent

[W3SLK]

Steve said:

30 ma. grid drive for 2 class c 813's???   too much in my estimation

Funny, I've a pair running at 32 ma. and have been for sometime. I think they should see roughly 16ma a piece. Now I am using a brute force capacitance coupling from the Ranger but it seems to work FBOM.

[K1JJ]

Steve said:

30 ma. grid drive for 2 class c 813's???   too much in my estimation

Funny, I've a pair running at 32 ma. and have been for sometime. I think they should see roughly 16ma a piece. Now I am using a brute force capacitance coupling from the Ranger but it seems to work FBOM.

From my own tests, I'd agree to run them at spec or harder.  I run 20-22ma of grid current with a SINGLE 813. 

This is the test to make:  Load up your rig to it's sweet spot at max power and best efficiency using recommended grid current. (16ma/tube). Put an audio tone through and carefully study the modulated sine wave on the RF scope.  Increase the audio tone until you see flat topping or some other kind of limitation.  Be sure to have a negative peak limiter in place to prevent mod xfmr damage.

Now, while at this max audio point, slowly increase your RF drive to the final so that the grid current rises and watch the audio peaks as you increase audio further. You may find that the audio peaks get higher and cleaner when driving the final grid a little harder than specs.  At least my 813's showed this. I also run a little more screen current to eek out the highest audio peaks I can. I then load the final more heavily (LESS C2 giving less power output, less efficient, more plate current) and everything really comes to life.

I've found the same thing with 4X1's and especially using triodes as finals.

My 813's are original tubes and show no signs of pissbeat wear after years of driving them hard.  813's are amongst the most strappingly built tubes on the planet.

T

[W3SLK]

Tom said:

813's are amongst the most strappingly built tubes on the planet.

I'll drink to that Mr. Vu. The only ones I've ever seen go bad were those that 'fell' victim to gravity. 

[WA1GFZ]

I would feel good with some resistance across the input tank to put a stable load on the input. This will also lower the input Q a bit. I would think making it more stable.

[K1JJ]

I would feel good with some resistance across the input tank to put a stable load on the input. This will also lower the input Q a bit. I would think making it more stable.

Yep, that's a good idea, Frank.   Though, since I've been going to the trouble to neutralize my finals, I've never needed a swamping resistor across the grid. But w/o neut, yes, it is always required.   But on the bands above 20M, a swamp will help. You think a non-inductive 3K at 20W in series with a .001 to ground would do it?


Rick: Yes, love those 813's.  BTW, I'm glad you pointed out the grid meter polarity thing. For all these years it never dawned on me that the polarity  of the grid current meter was the same as the screen. I usually wire all my meters without bothering with polarity and then simply swap the leads once I fire them up to get them correct... :-) It's a 50-50 thang.

I guess the rule could be to start at the cathode (or gnd) and simply call that (-) and follow the path to each element and call that end the (+). The meter follows the same polarity direction. This works also for cathode current meters using gnd as a reference.  Another point is that the grid is drawing positive current in class C, just like the screen.

It's funny to me cuz for years, whenever I bothered to wire a rig up paying attention to the meter polarities, the grid meter was always wrong - I just figured I wired it backwards and never went back to check or think about it... duh  :-)

Tnx for the info, OM's.

Later -

T

[WA1GFZ]

Sounds good to me. Someday I want to build one of these 813 rigs.

[W1RKW]

Rig is on hold for a while until I get a few more parts to tinker with.
I'll do some measurements in the meantime.

Thanks for the help and suggestions everyone.

[W3RSW]

Tom JJ, et. al. - just got around to reading this again.

Tom per your request for last minute changes, I think a 25 to 50 ohm, say 20 watt resistor in the B+ line before the mod reactor would be in order. Since this is a bare bones, 1950's style rig with no auto cut-outs etc. this resistor helps to protect the tubes, power supply (allowing time for a breaker or fuse to blow) etc. in case of a plate short, etc.  The 20 watt rating sure won't sustain a short very long but it is beefy enough to keep up the good fight until a breaker trips.

Regarding the grid loading with a resistor my single 813 shoebox had swamping but took the same amount of power 8- 10 watts to drive it as my twin 813 rig w/o a swamping resistor.  Yes, the twin 813 is neutralized and I find no instability.  But a light swamp shouldn't hurt anything,  I'd go for more like 5k to 8k rather than 8k if people feel they must have swamping as long as they use neutralization.  With 3k one probably wouldn't need the complexity of a neutralizing network on 160, 80 and 40 anyway.

BTW, my twin 813 uses the classic, low capacity, HV neut. capacitor of a few pf from plate to bottom of grid circuit with a fixed 800 pf from there to ground.  I assume you used the higher capacity variable grounding cap because HV neut. caps are hard to find.

Your circuit's beauty lies in the cathode circuit relay switching, receiver noise muting, etc. borrowed from later 'linear' technology I'd guess.  And yes men, you have to test your power supply transformers for approximate voltages under load. None of our transformers out here match perfection, even JJ"s were probably resistored up, etc. to get him what he wanted.

For example my VR-75 dropping resistor turned out to be 1.5k instead of JJ's 1.0k.  Don't be afraid to deviate. Use what you have but measure, measure, measure, particularly the screen voltage under load.  350 to 400's fine.

Too many 'etc.s' but you understand 

[K1JJ]

OK Rick -

Thanks for the suggestions.

I like adding the power resistor in the B+ as an additional safety factor - will do.  I actually use them in my own rigs too.

The 5K RF grid swamp is a reasonable value.  I've noticed that about 2K-3K works well with un-neutralized rigs, whereas, the higher value is better when neutralized to minimize drive requirements.



John, I've updated the list of schematic changes for you... The rig worked FB as is, but this should bring it into the 21st century... :-)

Thanks again.

1) At the far right, where the plate current says "= ?"  -    put in, "Load final to 250ma - 500ma, depending on power output desired."
2) See the "300pf neut cap" adj ?   - add a 500pf  1KV fixed capacitor across it in parallel.
3) Increase the two ".001 1KV" screen caps voltage rating to .001  2KV.
4) On the power supply schematic, reduce the screen choke rating from 500ma to 200ma.
5) On the power supply schematic, reduce the 30K bleeders in the screen supply from 30K to 7.5K  (Rick's suggestion)  That's 11 watts dissipation total using two 20 watt resistors. A good safety factor.
6) Transmitter schematic: See the "B+ 2KV" label?  At that point add  a 25 ohm, 20 watt wire wound resistor in series with this B+ line so that one end of the resistor connects to the old B+ point and the other end gets the, "B+ 2KV" label.
7) Add a 5K, 20 watt non-inductive resistor in series with a .001 1KV disc capacitor from one tube CONTROL grid pin to ground. These are additional components. (add note: "Swamping resistor usually not required unless final becomes unstable, especially on 20M and above.")

T

[W1RKW]

I've had some time today to sit and go over this thread and previous emails from those who have been helping me with this project. I just want to say your insight and experience is not going unnoticed despite me not saying so or responding to those who have offered their knowledge and suggestions.   Thanks everyone for the advice and encouragement.

I'm still plugging along, not progressing as fast as I like but still progressing nonetheless. Slow is a good thing anyway.  I know this isn't kids stuff and can be lethal if a mistake is made and that 's my problem, sometimes I lose sight of that fact but I know I have to walk away from it and do to avoid short cutting and having an accident. Many years ago I got popped and suffered a pretty severe burn on my hand from not paying attention. Suffice it to say I got lucky.

Additionally, I've taken a little time today also to look at the rig and rehash some of the assembly. 

Here's some info based on the thread discussion and questions posed.

Screen power supply choke, DC resistance: 102 ohms   Not sure if this is low enough.  Will determine voltage drop across it and provide info on next update.

Screen secondary DC resistance: ~70 ohms

Screen bypass caps: .001@6kV  I think I'm OK there Rick.

Installed a fixed padding cap at grid input: 500pF recommended. I found double that value was necessary when I was having difficulty with the grid tuning circuit last year.  The question on grid peaking.  The peak is sharp.  Will add recommended resistor input tuning tank to broaden peak.

Don't have the resistors needed for the screen bleeders yet so rig is on hold for now.
bw

[W1RKW]

Tom,
In step 7 you say to connect a 5K in series with one of the .001uF caps.  Those caps are the screen grid caps.  Is this were you want the swamping resistor? Or on the control grid?

[K1JJ]

Tom,
In step 7 you say to connect a 5K in series with one of the .001uF caps.  Those caps are the screen grid caps.  Is this were you want the swamping resistor? Or on the control grid?

Hi Bob,

The 5K in series with a .001 cap is a NEW set of components. They go from one control grid tube pin to ground. (The tubes have their grids in parallel with short leads, so one set is enuff. I will add "control" grid to step 7.


Almost anything series/paralleled can be clip leaded into the screen supply to provide heavier bleeding. No need to stop on that account for exact parts.

BTW, consider trying the screen resistor from plate I mentioned above. It will completely eliminate the screen supply and mod choke to tell you quickly if that really is the problem.

T

[W1RKW]

OK.  Understood on the 5K and .001's. 

Unfortunately, my resistor and clip lead supply is limited in QTY and size. 

[WB2YGF]

OK Tom

Check 'em out

John

[WA1GFZ]

John,
You have the swamping on the wrong grid. it should be on the control grid.

[K1JJ]

OK John -

The power supply looks fine.

The transmitter looks good except for the 5K and .001.   Those go from the CONTROL grid to ground, instead of the screen to gnd. The control grid is the first element above the filament as shown on the schematic. You can probably connect the parts to the control grid connection going between the two tubes, and then to gnd.

If anyone else sees something to correct, speak now or forever hold your log.


Other than that, looks good, John - TNX, OM!

T

[ka3zlr]

How many Turns is the tank Coil and Diameter, data could be placed right under 1/4" copper...or above and spacing...

[K1JJ]

How many Turns is the tank Coil and Diameter, data could be placed right under 1/4" copper...

Hi Jack,

There's so many ways to make or buy a coil that I left that up to the builder. Some want to optimize for one or two bands, while others use plug in coils or some want 160-10M in a compact package. I recommend someone check out the handbooks for suggested tank designs.

However, I'll tell ya what I do here.  I usually wind a 4" diameter coil  about  9" long. I make sure it has enuff inductance to cover 160M.  Dimensions are not really critical at all. Then I estimate where the taps shud be for each band and connect them to the band switch. Once the rig is running, I fine tune the taps by trying all the different parameters looking for maximum power out, maximum effciency and best audio peaks. Tap, tap and tap. There usually is a particular tap for each band  where you find the sweet spot for any set of tube voltages.

I find this to be the best way to get maximum performance in the real world. So, as you can see, what coil I have to work with I can usually optimize the parameters.  However, try to make the coil length between 1 to 2 times its diameter for best Q.  This ratio will vary for each band, of course.  For the higher bands, (10-20M) making a separate coil will bring that L-D ratio back in and also permit you to use a larger diameter coil stock.

Then again, many simply use the handbook charts and calculate the coil sizes before building. That works FB, of course. You must know your expected plate impedance.   It doesn't allow for real world problems that crop up, like stray capacitance or other random things, but is close enuff.

T

[WB2YGF]

OK John -

The power supply looks fine.

The transmitter looks good except for the 5K and .001.   Those go from the CONTROL grid to ground, instead of the screen to gnd. The control grid is the first element above the filament as shown on the schematic. You can probably connect the parts to the control grid connection going between the two tubes, and then to gnd.

If anyone else see something to correct, speak now or forever hold your log.


Other than that, looks good, John - TNX, OM!

T

Sorry, I figured out which was the control grid (the one with the signal and not the screen voltage). Then as I got focused on the trees, I forgot about the forest.

I was going to do this in MSPaint like I used to fix the meter polarity.  Turns out, they disabled the ability to add text in the Vista version of MSPaint. Bast***ds.

I found another program loaded on the machine called "Microsoft Digital Image Starter Edition 2006 Editor". Kind of Photoshop-like, but it seems faster and easier.  Still, it was a learning curve figuring out how the tools work and the idiosyncrasies of the program.  NEXT time, it will take me a LOT less time.

One thing I hate about Photoshop is its native format (to preserve the layers) makes the file like 100 times bigger than the original.  MS Editor uses something called PNG Plus.  Files are bigger, but not by a ridiculous amount. (850K vs 250K, for example)

If I wanted to do this thing RIGHT, I would re-do the whole thing in PCAD (schematic capture).  Only thing is, I don't have any tube symbols in my library and would have to create them.  Project for another time perhaps.

John

[K1JJ]

Hi John,

Yes, it looks good now.

Interesting how you were able to get into that schematic file using generic software. 

Well, let's sleep on these changes for a few days to make sure nothing else pops up before sending it to Gary to repost.

Thanks again for the efforts, John!  I'll bet there will be a few more of these rigs built as a result of the dialog here. It's a strapper. That pair of 813's will run 1KW out carrier at 120% using 3KV. I've done it - into a dummy load, of course... :-)


T

[Steve - WB3HUZ]

Once all the changes are made, I'll redraw the schematic (correctlY) and put it with the original article on The AM Window Web Site.

Thanks for all the input and corrections.

[W7XXX]

30 ma. grid drive for 2 class c 813's???   too much in my estimation......Steve

Hi Steve ... are you thinking a single 813 at 30ma? I run my single 813 with 22ma grid drive, necessary with the -180v fixed bias. The screen supply is from a dropping resistor off modulated HV, so the grid drive of 22ma is also necessary to keep the screen at 400v and 40ma, when loaded up at 400 watts input and 300 output.

[K1JJ]

Once all the changes are made, I'll redraw the schematic (correctlY) and put it with the original article on The AM Window Web Site.
Thanks for all the input and corrections.

OK, Steve -

Oh yeah, that's right - this project is on the original AM Window Website. I had forgotten.. :-)

I'd say the schematics are all set to go now.

Why do you want to redraw them?  They look OK to me in their present form below.... 

T

http://amfone.net/Amforum/index.php?action=dlattach;topic=18117.0;attach=9575
http://amfone.net/Amforum/index.php?action=dlattach;topic=18117.0;attach=9571

[Steve - WB3HUZ]

With all the editing and saving, they are getting a little fuzzy. I'm picky. I like nice clean schematic, if I can help it. And I have so much free time! I will put up the current versions until I get the new ones done. Many thanks John.

[WB2YGF]

With all the editing and saving, they are getting a little fuzzy. I'm picky. I like nice clean schematic, if I can help it. And I have so much free time! I will put up the current versions until I get the new ones done. Many thanks John.

Glad to help.

The original image was only 96 dpi.  It's amazing its not even more fuzzy.  I had to create new text, which when printed, stands out as less fuzzy.  With the components, I was able to create object blocks by copying the existing fuzzy parts so they blended in better,  Unfortunately, I had to double the size of the pic to add the text, because the original scaled down text is about 2 or 3 points and the program could not make text that small, then I reduced the image back to the original size.  I was actually expecting more distortion than I ended up with.  It would be nicer redrawn, but of course, it's more important to convey the correct information than be pretty.

What do you use to draw schematics Steve?