The AM Forum
March 28, 2024, 08:56:33 PM *
Welcome, Guest. Please login or register.

Login with username, password and session length
 
   Home   Help Calendar Links Staff List Gallery Login Register  
Pages: [1] 2   Go Down
  Print  
Author Topic: Grid input network for K1JJ 813 rig  (Read 10746 times)
0 Members and 1 Guest are viewing this topic.
KK7UV
Member

Offline Offline

Posts: 76


« on: February 18, 2019, 03:37:08 PM »

I trimmed a piece of B&W miniductor to 40 turns as specified in the schematic.
It measures 12.5uH.  With the specified 150pF variable, it should resonate down to 3675.

If I wanted to take it as far down as 1870 khz, is it as simple as switching in another 430pF?  or should the inductance be increased to maintain the same Q?    How about operation up to the top of40m?

Should the grid coil be tapped and switched for each band (160 - 80 - 40) just as the output tank is?
Logged
W3RSW
Contributing
Member
*
Offline Offline

Posts: 3308


Rick & "Roosevelt"


« Reply #1 on: February 18, 2019, 04:47:58 PM »

Yes, sufficient  tapped miniductor for 160/80/40 and tapped 2, 4 and 6 turns or so on primary to match.
This is exactly what I did on my dual 813 transmitter.  


* 813s rig under deck.JPG (2022.58 KB, 2592x1944 - viewed 619 times.)

* link input circuit alt.JPG (1997.63 KB, 2592x1944 - viewed 644 times.)

* 813 rig in operation.JPG (1737.76 KB, 2592x1944 - viewed 563 times.)
Logged

RICK  *W3RSW*
K1JJ
Contributing
Member
*
Offline Offline

Posts: 8893


"Let's go kayaking, Tommy!" - Yaz


« Reply #2 on: February 18, 2019, 05:23:52 PM »

Nice 813 rig, Rick!  I never saw the bottom of it before....

I designed the 813s input grid tank and output plate tank without detailed specs  - thinking most builders will have different coils, caps and switches so that they would end up designing their own based on parts availability.  

I usually slap in a combination of L/C that looks close and then optimize the grid and plate tanks using the MFJ 259 analyzer using a load resistor (and clip leads) to simulate either the grid or plate impedance.

Another way is to just fire up the rig and set the taps for best performance... lowest swr on input and best power out and performance on output.

Rick, I like your idea about tapping the grid link for 2,4,6 turns. I have found that various rigs sometimes want to see 4, 5 or 6 turns with 5 turns being the most common. But a rotary switch makes it easy on the fly.  With no link tap switching, the grid tank Q has to match closer across the various bands to be able to use 5 turns only.

*** Another great method for grid tuning is a reverse pi-network.    (50 ohms IN) 1200 pF to gnd >   series tapped  coil   > 300pF to gnd (OUT to grid)   - works very FB.  I use this on all my linear amps and occasionally on AM plate modulated rigs. Lots of versatility there.

For the new project, for 160 grid tuning, I would use more L to get you down there instead of using 580 pF.  It will work either way, however, just a sharper tuning / higher Q when using a bigger C.

T
Logged

Use an "AM Courtesy Filter" to limit transmit audio bandwidth  +-4.5 KHz, +-6.0 KHz or +-8.0 KHz when needed.  Easily done in DSP.

Wise Words : "I'm as old as I've ever been... and I'm as young as I'll ever be."

There's nothing like an old dog.
KK7UV
Member

Offline Offline

Posts: 76


« Reply #3 on: February 18, 2019, 06:51:20 PM »

Thanks for those responses - very helpful.

What is the grid input Z of the 813 pair?   I may play around with some matching network software.
Logged
KL7OF
Member

Offline Offline

Posts: 2313



« Reply #4 on: February 18, 2019, 10:02:22 PM »

I built the grid tuner from the calif handbook...2 small coils and a little breadslicer....I got lucky and had it all in the junk box...40M 75M and 160...  I never got my tesla to go on 20M.....I think my leads are too long..I get oscillation as soon as I modulate...I'm sure if I put a couple bypass doorknobs on each side of a z-50 choke in the mod B+ line it would go away... 

Logged
K1JJ
Contributing
Member
*
Offline Offline

Posts: 8893


"Let's go kayaking, Tommy!" - Yaz


« Reply #5 on: February 18, 2019, 10:24:10 PM »

I built the grid tuner from the calif handbook...2 small coils and a little breadslicer....I got lucky and had it all in the junk box...40M 75M and 160...  I never got my tesla to go on 20M.....I think my leads are too long..I get oscillation as soon as I modulate...I'm sure if I put a couple bypass doorknobs on each side of a z-50 choke in the mod B+ line it would go away...  


Hmmm... 20M parasitics...let's see, Steve. You probably already know of the following fixes.  l'm writing this more to help out some of the newer guys building rigs...  

Did it neutralize well with that grid circuit?  How far away from the grid pin is the grid circuit? Should be right next to it with stubby leads made of strap..

Try this trick: If not already, add a 2K non-inductive resistor 10W or whatever in series with a .001 cap from grid pin to ground. Reduce resistor to 1K and see if this cures problem.  The blower (if used underneath) may keep the resistor cool, so experiment with wattage.

Make sure the final's input cannot see the output...   The antenna relay - can the input see the output leads and contacts?   Everything shielded well?  Look over lead length in the final tank and shorten, use 1/2" strap, good quality bypass cap, try a different plate choke as a test, good groundplane for C1/ C2 and other topside parts....  Add another 2.5 mH choke and bypass cap stage to both the grid and screen DC leads.  Good quality, short leaded RF bypass caps, especially on screens.  Good quality and short RF fil bypasses.

** Make sure nothing is arcing when modulating on 20M. Darken room with big mirror behind the rig. Sometimes arcing can appear like parasitics.

My idea has always been to get it working well on one band higher than your intended use. IE, get it working well on 20M even if you don't use it there and it will be rock sold on 40M and lower. Same with power - gettting it putting out 1KW carrier full strap - and it will laugh at 500W, 140% modulation.

There's more if that doesn't help.   And read my slogan below:   Grin

T

Logged

Use an "AM Courtesy Filter" to limit transmit audio bandwidth  +-4.5 KHz, +-6.0 KHz or +-8.0 KHz when needed.  Easily done in DSP.

Wise Words : "I'm as old as I've ever been... and I'm as young as I'll ever be."

There's nothing like an old dog.
K1JJ
Contributing
Member
*
Offline Offline

Posts: 8893


"Let's go kayaking, Tommy!" - Yaz


« Reply #6 on: February 18, 2019, 11:29:00 PM »

Thanks for those responses - very helpful.

What is the grid input Z of the 813 pair?   I may play around with some matching network software.


Good question.

It would be R = E/I.     I= DC grid current of two tubes in class C from the tube data sheet.  But how is E determined?   Is it peak RF drive voltage above zero volts cutoff as taken from the tube data sheet in class C?  

Off hand I am not sure and didn't see anything quickly on Google.  Anyone know for sure?   I usually don't use the grid input impedance calculation when building a class C amp.

I suppose the power formula could be used...   R= P/ I squared.   For example:  Impedance =  10 watts drive - tank losses / grid mA squared  

T


Logged

Use an "AM Courtesy Filter" to limit transmit audio bandwidth  +-4.5 KHz, +-6.0 KHz or +-8.0 KHz when needed.  Easily done in DSP.

Wise Words : "I'm as old as I've ever been... and I'm as young as I'll ever be."

There's nothing like an old dog.
W3RSW
Contributing
Member
*
Offline Offline

Posts: 3308


Rick & "Roosevelt"


« Reply #7 on: February 19, 2019, 11:35:40 PM »

Z in to a class C amp. ; It's not that critical. A slight mismatch is easily overcome by drive voltage, at proper current, i.e., power.    Note the total 160 Meters Miniductor coil length in my amp yields approx. twice the 12 or 18 uH length of the your spec'd 80 meter coil length. This allows resonant capacity to be reasonably smaller.  Note also in my amp is a fixed (after setting) add'l cap. in parallel via switching mounted off to the side of the Plexiglas insulating/mounting strip for 160 meters. 

 More critical is the bias voltage to the 813's.  Properly setting the grid bias under load and according to the tube tables for a given tube will somewhat automatically request the proper drive voltage to overcome bias to cause proper plate characteristics.  Two 813's in parallel will of course require twice the grid current, hence half the value of a grid leak resistor called for one tube, etc.
The bias in my amp is a combination of fixed voltage and resistive grid 'leak.' Approximately Equal voltages across each when driven will yield minimum distortion.

So drive power to the class C will have the impedance portions almost automatically rolled in the approx. six requirements outlined in Eimac's "Care and Feeding of Power Grid Tubes"   All six are simplified by following any good design in ARRL's handbooks, etc.  Just use the circuit components specified for a given tube; then double some values, halve some values, etc. as necessary for parallel tubes such as JJ's.  That's basically what he's done.

The six, all of which have equations, some incorporating impedance derivation if your interested:
1.)  The power consumed by the bias circuit.
2.) Power dissipated in the grid due to rectified grid current.
3.) Power consumed in the resonant grid circuit.
4.) Power loss due to transit time,  (not much in usual HF circuits)
5.) Power consumed in that part of resistive component of input impedance due to cath. lead inductance.
6.) Power dissapated in the tube envelope due to dielectric loss.

So if you have sufficient driving power to overcome all these in the published circuits, your on your way. Wink

So that's the input part of it.

Now wait 'till you look at properly setting the screen voltages, currents under modulation. What fun you'll have.  Again, properly following ARRL handbook or "The Radio Handbook" construction will save you a
lot of grief.  Neat tricks like audio inductance in the screen or tapped screen windings on modulation transformers all are nice to be understood for various modulation techniques of fully plate modulated class C amplifers. 

Everything's related; should be fun to get the Eimac power grid tube book or similar , familiarize yourself with the constant current plots/characterics, load lines, and the analog tube performance computer equations.  The actual electrons travel in analog paths in real life. Wink 
Numerical simulation is just that, be it in Fortran IV or all the later variations.  Be interesting to see if
modern software incorporates all the transfer functions of real tubes and their related real life RF cicuit functions beyond simple differential transfer equations.
Logged

RICK  *W3RSW*
KK7UV
Member

Offline Offline

Posts: 76


« Reply #8 on: February 24, 2019, 12:00:30 PM »

Has anyone tried this input network with a wound toroid?  Which mix?  Ferrite or iron powder?   I'm stunned by the price of B&W coil stock.
Logged
K1JJ
Contributing
Member
*
Offline Offline

Posts: 8893


"Let's go kayaking, Tommy!" - Yaz


« Reply #9 on: February 24, 2019, 12:31:14 PM »

Has anyone tried this input network with a wound toroid?  Which mix?  Ferrite or iron powder?   I'm stunned by the price of B&W coil stock.


It's not scientific, but first I'll tell you how I do it caveman style....  Wink

To wind and optimize a donut unbal to unbal toroidal transformer, I first find a #43 or #61   1" to 2" diameter donut. Not critical for low power use.  There are many combinations that will work 160-10M.  

Then experiment  using # turns on the input and turns on the output just like any transformer.   Put an MFJ-250B (or equiv) into the input with a resistor (grid simulation) load on the output and keep trying turns combination ratios until it covers 160M - 10M with a good swr.   It may take 3 turns pri and 15 turns sec or whatever.  Ground both cold leads together. The mix may not cover the whole freq range.
 
It will not have any selectivity like a tuned circuit does, but it will work as a broadband transfomer once you get the right turns ratio. You will have to estimate the input impedance of the grid you are driving until you fine tune it in the actual circuit. A warning: Lack of selectivity may be a problem if you are using a dirty DDS VFO.(<-60DB down)  My sharply tuned 4-1000A  grid circuit cleaned my Chinese DDS up quite well.


There are other more sophisticated and stable designs published using fixed 9:1 ratio, 16:1, etc., on the web that will also work FB.

Look at page 8 for a 9:1 UnUn sample that will work in a transmitter 50 ohm > grid application:  (50 ohms to 450 ohms UnUn)

http://www.philipstorr.id.au/radio/technical/Balun%20and%20UnUn.pdf

T
Logged

Use an "AM Courtesy Filter" to limit transmit audio bandwidth  +-4.5 KHz, +-6.0 KHz or +-8.0 KHz when needed.  Easily done in DSP.

Wise Words : "I'm as old as I've ever been... and I'm as young as I'll ever be."

There's nothing like an old dog.
KD6VXI
Contributing
Member
*
Offline Offline

Posts: 2648


Making AM GREAT Again!


« Reply #10 on: February 25, 2019, 08:26:35 AM »

Years ago I built a grid driven 4cx5k (following the lead of Rich Measures, though he didn't realize what I was building it for).

Problem was, I wanted to include ten meters, but I wanted a broadband input. Nobody seemed to care about anything higher than 20 metros then, and people running qro in 15 seemed like....  Well, not many around I found.

Everyone I spoke with then (circa 1999) said use two different tuned inputs.

Someone (don't remember who) came up with using the flat style torroids instead of a binocular core.  And said stack two different mixes, tape them together, then wind the transformer.

FINE BUSINESS!    Ended up with a single cap that tuned 80 to 10.

So if you want high and low bands, 43 and 61 mix stacked would be what I'd be looking at. I don't remember the exact size of the torroids I stacked, but they where not the 2.4 inch monsters we see now.

My xformer was only hit with appx 250 watts peak, but I doubt yours will need that much drive!

My bigger issue was cooling the grid R.  Mine was severely undersized (thinking cantenna, I used a 450 ohm 50 Watt non inductive.  OK for ssb.  For heavy am use I had to go up to (iirc) a 150 Watt, air cooling it)

About ten years later, I upgraded to be able to use the amp everywhere it was designed.  Should have done it long before.....

--Shane
KD6VXI
Logged
KK7UV
Member

Offline Offline

Posts: 76


« Reply #11 on: March 03, 2019, 02:20:22 PM »

So let's say I had three different input networks, one each for 160, 75, and 40m, mounted on a board with relays.

Rather than a broadbanded toroid transformer with taps, could each network have a toroid transformer with optimized windings and have it be narrow-tuned just like the air-wound coil input network?

see attachment





* gridtuneidea.gif (15.29 KB, 966x941 - viewed 376 times.)
Logged
K1JJ
Contributing
Member
*
Offline Offline

Posts: 8893


"Let's go kayaking, Tommy!" - Yaz


« Reply #12 on: March 03, 2019, 02:59:26 PM »

Yes, that should work FB.  You're basically replacing the mini-ductor coils with toroids - and using the same basic circuit. (with coil selection switching)

I have used toroids in tuned circuits like that -  as well as seen some big amplifiers like Alpha use toroids in their plate tank circuits.

Experiment with a grid resistor load in place of the real grid circuit and you should be able to find a good 50 ohm input. Then optimize/ fine tune it when in the real circuit, if needed.

I didn't study your 160-40M relay switching logic closely, but it appears to be a workable design.  (To make it easier to follow, put the toroid relay contacts into the 40M position, as the control switch indicates)

As long as there's no wire arcing on the core, you should be OK. Probably insulated wire rather than enameled wire would be best.

BTW, that RF choke in the grid circuit should be a regular 2.5 mH air-core type. The line above it indicates a magnetic core.

T

Logged

Use an "AM Courtesy Filter" to limit transmit audio bandwidth  +-4.5 KHz, +-6.0 KHz or +-8.0 KHz when needed.  Easily done in DSP.

Wise Words : "I'm as old as I've ever been... and I'm as young as I'll ever be."

There's nothing like an old dog.
KK7UV
Member

Offline Offline

Posts: 76


« Reply #13 on: March 03, 2019, 04:16:40 PM »

Thanks Tom.   I fixed the switch/relay position.

I'll try some bench experiments. 
Logged
w9jsw
Two shots of Whisky
Contributing
Member
*
Offline Offline

Posts: 592



« Reply #14 on: March 04, 2019, 07:29:54 AM »

Tom,
 
Is this approach any more stable if you try to achieve 20M as well? I have preferred toroids as well for my other projects.

I wonder how tight I could make it on a PCB...thinking... 

John
Logged
K1JJ
Contributing
Member
*
Offline Offline

Posts: 8893


"Let's go kayaking, Tommy!" - Yaz


« Reply #15 on: March 04, 2019, 11:05:33 AM »

Tom,
 
Is this approach any more stable if you try to achieve 20M as well? I have preferred toroids as well for my other projects.

I wonder how tight I could make it on a PCB...thinking...

John


Hi John,

As far as stability, it's really the same basic circuit except for toroids being used as the cores rather than air, so should work the same.  All will depend on how tight you can make the layout and relay contact connections -  and proximity to the tube grids.  Maybe you can do a better job of it with toroid layout.  Short leads is what it's all about.  Remote relays are usually tighter than bandswitching, so give it a go.

I kinda like the novel relays/toroidal approach simply because it's different and space age…  almost like using the GFZ MOSFET driver rather than tubes and transformers.  :-)

T
Logged

Use an "AM Courtesy Filter" to limit transmit audio bandwidth  +-4.5 KHz, +-6.0 KHz or +-8.0 KHz when needed.  Easily done in DSP.

Wise Words : "I'm as old as I've ever been... and I'm as young as I'll ever be."

There's nothing like an old dog.
w9jsw
Two shots of Whisky
Contributing
Member
*
Offline Offline

Posts: 592



« Reply #16 on: March 04, 2019, 12:07:41 PM »

I was thinking the same thing...

I laid out a schematic. Looking for a good relay that can handle 100vdc on the contacts.

Anyway, here is the schematic with a wrong choice relay.

This would be super-simple to make a PCB for it.

John

* Schematic_813-Grid-Input.pdf (50.16 KB - downloaded 188 times.)
Logged
K1JJ
Contributing
Member
*
Offline Offline

Posts: 8893


"Let's go kayaking, Tommy!" - Yaz


« Reply #17 on: March 04, 2019, 12:48:17 PM »

Looks good.

Though there is a lot to it.  And, the match swr will not always be 1:1 like an adjustable network would be.

Thinking further...an alternative might be using a reverse pi-network, 50 ohms in, > ~~  anything out. (3 adjustable components)    1200 pF cap to ground > small roller inductor in series  >  300pF cap to gnd.    Adjustable  1:1 swr 160 - 10M.   Install it and you are done, no test tuning, etc.

I use this on all of my linear inputs and for some class C rigs. It would work just as well for your 813 class C finals.  I believe Chuck uses the same circuit in his 4-400A X 3-500Z  rig we recently featured in your rig building thread.

The difference is one knob (and 1 switch) to tune vs: 3 knobs (and no switches) -    The 3 knobs will tune you 1:1 exact swr whereas the one knob will be close but not perfect most of the time...


T
Logged

Use an "AM Courtesy Filter" to limit transmit audio bandwidth  +-4.5 KHz, +-6.0 KHz or +-8.0 KHz when needed.  Easily done in DSP.

Wise Words : "I'm as old as I've ever been... and I'm as young as I'll ever be."

There's nothing like an old dog.
w9jsw
Two shots of Whisky
Contributing
Member
*
Offline Offline

Posts: 592



« Reply #18 on: March 04, 2019, 01:12:04 PM »

I like this idea. What is a good example roller inductor. I already have a couple of air variables for the 1200pf and 300pf caps.
Logged
K1JJ
Contributing
Member
*
Offline Offline

Posts: 8893


"Let's go kayaking, Tommy!" - Yaz


« Reply #19 on: March 04, 2019, 02:48:01 PM »

Here is the general type you need:

https://www.ebay.com/itm/B-W-Barker-Williamson-Rotary-Roller-Inductor-Assembly-0-15-uh-3-x-2-Dia-2/392251173134?hash=item5b53fde90e:g:HDYAAOSwNSdca18S:rk:16:pf:0

Lapointe Industries is another good brand. (out of biz)

Look on eBay. "roller inductor"   There are often some there.   Expect it to be about the size of  a soup can with as many turns as possible to cover 160M.  It will be almost like the pi-network coil you find in any 100 watt rig.    30 uH or so should cover 160M. As many turns as possible.

T
Logged

Use an "AM Courtesy Filter" to limit transmit audio bandwidth  +-4.5 KHz, +-6.0 KHz or +-8.0 KHz when needed.  Easily done in DSP.

Wise Words : "I'm as old as I've ever been... and I'm as young as I'll ever be."

There's nothing like an old dog.
w9jsw
Two shots of Whisky
Contributing
Member
*
Offline Offline

Posts: 592



« Reply #20 on: March 04, 2019, 08:11:58 PM »

I may do the Pi net if I can find a good roller.

However, while this relay/inductor circuit seems complicated, this is easy stuff on a PCB.

I find it interesting that we use large components to achieve inductance when a toroid can do it much more compactly. And then we try to go through extra effort to stuff it under the tubes and attach it as close to the grids as possible. The PCB approach could be much more compact. Maybe even a breakthrough approach?

We can use different cores per band to fine tune the Q. If I was a smarter engineer I could maybe do it with pin diodes to remove the relays entirely.

I have been happy with the D3002 relay for it's compact size and good contact ratings. They make one in that footprint for 24V that can use my shack 24V supply.

I am going to proceed with building a test PCB so I can see how it runs on the bench.

John
Logged
K1JJ
Contributing
Member
*
Offline Offline

Posts: 8893


"Let's go kayaking, Tommy!" - Yaz


« Reply #21 on: March 04, 2019, 08:49:38 PM »

OK John -

I just realized that the reason I originally went with that mini-ductor published grid circuit is to accommodate the neutralization feedback.  With the new reverse pi-network I suggested, it would have to be floated like the other circuits to handle the neutralization.

HOWEVER, tetrodes have so little grid to plate feedthru capacitance (just 0.25 pF per 813 tube)  neutralization is usually not needed.  Make sure there is no external g-p coupling.  I always neutralize my tetrodes though many times when testing there is very little signal to balance out. The screen is grounded for RF and shields the grid from the plate, thus only 0.25 pF G-P.    Triodes have no screen shield and will almost always take off if not neutralized.  I think Chuck told me his 4-400A rig is not neutralized -  and he operates it up on 10M... which is a compliment to his layout.

The  ~2K-3K  carbon loading  resistor (in series with .001 cap) from grid to ground will also have a big effect on stability. Reduce the value if need be.

Parasitics are a different animal and neut will usually have little or no effect.  Always keep layout tight and use good components and parasitics will be tamed.

So my suggestion is to either go with the reverse pi-network without neutralization - just make it tight with strap leads and shielding.  Ground the grid input and output variable caps and float the inductor as desribed earlier.  No neut cap connection to the plate.

Or try your toroid circuit and use neutralization as originally intended.   It will not always be a perfect 1:1 swr.  If your driver has a pi-network or tuner, then the small 50 ohm mismatch will not be a problem. If solid state and fixed, then you will have to see.  The fixed input link # turns for each band will be your main adjustment.

T
Logged

Use an "AM Courtesy Filter" to limit transmit audio bandwidth  +-4.5 KHz, +-6.0 KHz or +-8.0 KHz when needed.  Easily done in DSP.

Wise Words : "I'm as old as I've ever been... and I'm as young as I'll ever be."

There's nothing like an old dog.
WA4WAX
Member

Offline Offline

Posts: 415


« Reply #22 on: March 04, 2019, 09:30:25 PM »

Here is an 813 rig article.  Check out the input circuitry.

One could parallel two I suppose.

https://www.americanradiohistory.com/Archive-Electronics-World/60s/1962/Electronics-World-1962-02.pdf
Logged
KK7UV
Member

Offline Offline

Posts: 76


« Reply #23 on: October 13, 2022, 08:58:10 AM »

Finally getting around to building the grid tuning network for the 813 rig.

What voltage rating/plate spacing is needed for the grid tuning variable cap?  The old handbooks suggest "grid bias voltage plus 20% of the plate voltage" which would be 400-475v for my setup.   An old Cardwell catalog says 350v per .01".   I have a couple 150pF variables I could choose from with plate spacings of .026 and .030. which appear more than adequate.

However, I noticed in an old post by Rick W3RSW and his similar build that his grid tuning variable appears to much closer plate spacing (see photo) and I presume you have had no problem with yours Rick?

Steve



* W3RSW input.JPG (47.11 KB, 477x429 - viewed 149 times.)
Logged
W1RKW
Contributing
Member
*
Offline Offline

Posts: 4410



« Reply #24 on: October 13, 2022, 10:14:57 AM »

Steve,
If I recall correctly, I use the standard 365pf variable in the input section without issue.
Logged

Bob
W1RKW
Home of GORT. A buddy of mine named the 813 rig GORT.
His fear was when I turned it on for the first time life on earth would come to a stand still.
Pages: [1] 2   Go Up
  Print  
 
Jump to:  

AMfone - Dedicated to Amplitude Modulation on the Amateur Radio Bands
 AMfone © 2001-2015
Powered by SMF 1.1.21 | SMF © 2015, Simple Machines
Page created in 0.085 seconds with 19 queries.