Schematic Check Print 813 Deck

[KB7HTV]

Decided to post this, as a sacrificial cow
This is REV 0
This is my first junk box AM transmitter.
Poured over the designs on this Board
The handbooks of old, and pieced this togeather
if something is flawed, let me know

HTV

[KA8WTK]

It looks like the blocking cap (1000, 5Kv) is in the wrong place. At the moment it will block B+ from reaching the plate of the 813. It should block DC from the Pi network.
There is something unusual about the grid connections, but I don't remember the pinout of the tube well enough to put a finger on it. Go here and take a look: http://www.amwindow.org/tech/htm/813/813.htm

Bill

[KA8WTK]

BTW, I like K-1 and R-2 to bias off the tube. I use the same setup on my 813 and 4-400 transmitter. I think you forgot to draww the ground at the end of R-2/K-1.

Bill

[Ralph W3GL]

Yes, Bill is correct, move the top connection of RFC 1 over to the left side of
the .001/5kv cap.

Also the ground is missing from the relay shorting area of the cut-off (stand
 by bias) resistor there in the center tap return...

I think you are slightly overdoing ( making it too complicated, why?) the bias
with the cathode resistor as well as the VR controlled grid bias circuity.
Remember KISS... There are better, simpler ways...

Pin 4 is the control grid, pin 3 is the screen and pin 5 to ground...

[W2PFY]

it looks like your using the beam forming plates as grid one? They are usually grounded as I remember. (Pin 5) I never seen this grid used in this way.

Pin four is the control grid.

[WA2IXP]

The control grid connection pin # needs to be corrected as well as the plate blocking cap placement  as mentioned. also I believe the bypass on the grid rf choke needs to go to the bottom of the coil and maybe reduce the 1000uuf cap some if you don't get enough feedback for nuetalization. The neutralizing cap and the bypass cap form a voltage divider with the rfc isolating it from the bias supply. After you get her up and running you might experiment with some grid leak as well as your regulated bias. Has worked better for me on modulated amps. Good luck with your project seems, like you've got it down pretty well,
                                    Jay-

[WA2IXP]

center tap of filament transformer shown on wrong winding.

[WB6NVH]

The neutralizing capacitor can be connected on the cold side of the blocking capacitor, at least that's the way I have done it.

VR tube in bias line seems needless.

[w3jn]

If you want adjustable bias, R4 should be downstream of the VR-90.  An adjustable resistor at the point where you have it is worthless because all it's doing is adjusting the input voltage to the VR-90.  When the VR-90 fires, that's what your bias voltage will be.

+1 on the other comments.

[K1JJ]

Always good to see guys design their own stuff and give it a go.

I noticed the grid L1 / 250pf parallel circuit appears to be floating (thru the RF choke) with no RF reference to ground to drive the grid.

Add a 500pf or 1000pf cap from the cold end to ground to create a voltage divider for the neutralizing cap to work into - and also give the RF grid circuit a ground return. This may then require more than 10pf of neutralizing capacitance - try and see.

As Bill suggested, here's a sample neutralizng circuit to work from:
http://www.amwindow.org/tech/htm/813/813.htm


Update: (whoops - Jay mentioned this already in a few posts above)


T

[K5IIA]

how about getting screen voltage from the modulated b+ from a dropping resistor?

[WA2IXP]

very good point brandon, or maybe using a small choke in series with the screen supply.

[KA2DZT]

You have many mistakes in your layout and design.  First you need to improve your drawing of schematics.  The grid is always just above the filament and to the left, the screen grid is next and to the right, then the surpressor grid also to the right (which is grounded), then the plate at the top.

The dc blocking cap goes between the plate choke and the pie network.  You have it between the choke and the plate.  You're missing a 250-300pfd cap from the junction of you're grid RFC and input coil, the cap goes to ground.

You need to supply screen grid voltage directly from the modulated 1500V B+ through a 40K 150watt resistor.  You can use a combination of power resistors to make up the 40K 150 watter.  You should add a clamp tube circuit to the screen grid to protect the 813 if there is no drive.

You're missing the center tap on the filament xfmr.  I see you're placing the plate current meter in the B+ lead.  I do not like that design although commonly used.  I would place the current meter in the filament center tap lead to ground.  You will be reading the total of the plate current, screen current and grid current.  This design keeps the HV voltage off the meters.

You do not need the bias supply.  Just put your grid resistor from the grid RFC to ground, through your grid current meter.  The grid resistor should be 13K about 10-15watt.  Also can be made up from a combination of resistors.

I'll have to go back and take another look at your drawing to add more comments.

I just finished building a single 813 rig, completely my own design including winding the coils, the 813 worked perfect on the first try.  With proper design and construction there is no reason why your 813 rig shouldn't work on its first try.  Just to add a note; if you see smoke or hear sparks, you made a mistake

I'm not sure how you're placing the rig into standby.  I shut off the HV power supply (with relays) on standby.  It is fired up, on transmit, through a step-start circuit.

My comments are based on the fact that you're building a plate modulated AM rig.

Fred

[KA2DZT]

Took another look at the drawing.

I see how you're putting the rig into standby.  You're biasing off the 813 with high value cathode resistor and the relay.  You're missing the ground connection at this relay.  Also place a arrow head on the relay's moving contact arm.

Using this method, how are you shutting off the modulators??  The commonly overlook problem with this design is what happens to HV voltage power supply when the rig is on standby.  The voltage on the lightly loaded HV supply will increase, maybe to an unsafe level.  Something you have to think about.  Are the HV filter caps of a high enough voltage rating to handle the higher standby voltage??  This is the reason why I like to shut off the HV supply on standby.

Fred

[KC2ZFA]

Fred, we wanna see a real schematic or it didn't happen 

[The Slab Bacon]

Fred, we wanna see a real schematic or it didn't happen 

Fred actually showed up on the WFDN about a week or 2 ago running the "imaginary 813 rig". It's no longer imaginary!! (I didnt think I would live long enough to get to hear it! ! )      

[KB7HTV]

Guys, thanks for the comments, I will try to mark the drawing up with your comments I will reply to this with a new drawing showing integration of your design ideas into the sketch, I will be wiring soon I hope, with proper caution if I zorch something nobody will get hurt.

HTV

[Ralph W3GL]

  Very good Fred...

  Very comprehensive and I agree with using a clamp tube on the
  screen to pull its voltage down on loss of drive.
  
  A 6Y6 is what I use, plate connected to the screen side of the
  series resistor and grid connected at about 10K above the ground
  side of the  grid load of the 813.  Cut the clamp tube off with drive;
  no drive on clamp grid, tube conducts pulling final screen down to
  safe value on final tube...

  You want to select the series feed resistor of the 813 for around
  400 volts at about 15/16 ma when the tube is fully loaded. And yes,
  a high wattage dropping resistor is required there.  

   In one 813 final I built 60+ years ago I used a fixed 40K/150W in  
   series with a 15K/150W with an adjustable tab to set the screen voltage.

   Be very carefully in there, remember that's lethal voltage on that tap!

   Fred, scratch out your schematic for us and lets see what your final
   looks like at this stage.

   I missed that filament transformer CT on the primary side on my scan
   of the original drawing, sorry...

[The Slab Bacon]

I have read over everyone's suggestions and agree with all of them except for the biass / cutoff scheem.

Keeping with the KISS theme, why bother with a clamp tube?? Simply biass the 813 well down below the cutoff point and if it loses drive it will simply cut itself off and go to sleep. That is what I did with my 4X1 transmitter and it has been happy as a clam for more than 10 years now. It just requires a little more RF drive to the grid.

Also the heavier grid drive and increased biass will really help make the munky swing!! (and improve plate efficiency a bit as well)

Also, if you're building the thing why not use a pair of 813s instead of a single. Even
if you dont run any more power than a aingle, it helps with the "loaf along" factor, or will allow you to make more power out with a lower plate voltage.

Just my $.02 worth.

[KB7HTV]

Why not a pair of 813's, Hey I only scrounged up an art-13 mod transformer, I sold a 350Watt unit 2 years ago, thought I would never get around to it, but here I am back at it, maybe it will burn up and need to be replaced.

Have a hole in the chassis for the second 813, put a piece of aluminum over it as a patch. I will try to draw this animal up after work and post a second version of the thing I need to carefully read all these comments and absorb them. some are obvious flaws or misdrawn stuff, other stuff I need to study further as I attempt to correct the schematic. Thanks again all, for taking the time to consider and correct my humble efforts

HTV

[W1RKW]

Is it me or are the screen voltage connection and control grid bias connection reversed?

[W8ACR]

I agree with K1JJ regarding the grid circuit. I think there needs to be a ground point in there somewhere. I would use a split stator capacitor with the rotor grounded. Of course to get 250pf one would have to have 500pf per side. Alternatively, one could use a smaller variable cap with a padder. This is what I do with my single band 160 meter rig. I believe that grounding the rotor also helps suppress VHF parasitics. 100pf should be enough grid tank capacitance. I believe my B&W 160JEL grid coil is resonated by about 80pf at 1.9 mHz.

One can also ground the center point of the grid tank coil. There was a very interesting thread a while back started by K9ACT regarding the drive requirements being less if the coil is grounded vs the capacitor. Do a search on the topic "Mystery Grid Drive" to see this thread.

Ron W8ACR

[KA2DZT]

Fred, we wanna see a real schematic or it didn't happen 

Top Secret

[KC2ZFA]

Top Secret

I'm having trouble putting up an antenna (for the time being)...so
I think I'll hang my dummy load from the balcony and try to raise you...
after all we're only 12 miles apart. I've got two homebrewed xmtrs
and a dx-40 I repaired and no frigging antenna

[KA3EKH]

Don’t get me wrong, I am not trying to be a jerk or insult you but the amount of errors like the wrong placement of the plate blocking capacitor, driving the suppressor grid instead of the control grid and no DC return in the cathode circuit all remind me of the very old test you had to take for the second class commercial license and have to identify the problems with the circuit, once again I am not poking fun at you but I found it fun to identify the errors. Would almost like to see someone draw up some more circuits with errors and see who catches them. Thanks for posting this and wish you success with your project!
Ray F.