Power required to drive an 813 in grounded grid linear?

[K1JJ]

Does someone know how much power it takes to drive an 813 in GG linear operation?  (cathode driven, both grid and screen grounded) What is the gain in dB of the stage with about 3KV on it?  This info is not available in the Eimac data sheet.

I'm looking to build up a compact, convection cooled linear amplifier (no fans or blowers) and need to adjust the number of tubes by the drive power I have available.

Thanks.

T

[w4bfs]

hi Tom .... about 30 to 35 W in will give 300 to 350 W out ... tube condition will determine onset of gain compression .... will need a bit of neg fb above 20 mtrs to keep stable ... if you need to operate 15 mtrs or higher would recommend neutrailzed grid driven (worse imd figures) or different toob

[kb3ouk]

I wonder if running them with voltage on the screen might increase the gain some?

[Steve - K4HX]

I seem to recall Bill, KD0HG published some triode specs and maybe even some curves for the 813 in Electric Radio back in the 90s.

You can glean some info from previous triode 813 gg amp designs.

http://bama.edebris.com/manuals/b&w/l1000a/

http://n4trb.com/AmateurRadio/GE_HamNews/issues/GE%20Ham%20News%20Vol%2014%20No%206.pdf

[K1JJ]

Hmmmmm.. so an 813 in GG is good for only about 10dB of gain.   I was thinking of four in parallel. That's at the hairy edge for a 100 watt pep ricebox to drive.    But I have a 200w pep FT-1000D, so maybe a 250 watt carrier is possible with 25 watts carrier drive and nothing is stressed.

I'm looking to build a simple "summertime rig"  with no blowers, to fill in when it's too hot to run the 4X1 plate modulated rig.

BTW, from what I've experienced and read, adding screen voltage to GG does not make it easier to drive cuz we must also increase the grid bias to idle it back down, thus making more drive required to overcome this higher bias.  I suppose it depends on the tube mu too.

Thanks for the info!

T

[W1RKW]

Tom,
My old 813 GG had about a gain of 10 using 2 tubes and about 2200V on the plates.  No fan on the buggers either.

[W2VW]

At 3000 volts there should be a little bias on the control grids anyway. You're gonna get way more than 10 dB gain at 3 kv  until the grids succumb to the punishment.

[KK4YY]

Tom,

I've been looking into the same configuration as a project here, so I hope you get some good answers!

From what I've read, and my limited understanding, the gain can be from 10 to 14 dB. Higher gain with an input matching network and with the higher plate voltage. With 3KV you'd be near the higher end but you may need to bias the CG.

So my guestimate would be about 12 or so dB gain...  15 watts in 200 watts carrier output for 4x 813s with no forced air. Ricebox ready.

Adding a screen voltage apparently does not increase gain as the SG is still at RF ground, but makes the amp cleaner IF the voltage is extremely well regulated (as with the G3SEK tetrode boards).

The B&W LA-1000  and the GE "GL-813" amps have no input matching networks.

Hope this helps.

-Don

[K1JJ]

At 3000 volts there should be a little bias on the control grids anyway. You're gonna get way more than 10 dB gain at 3 kv  until the grids succumb to the punishment.

Interesting, Dave.  Yeah, I was thinking 3KV might make the difference. I know a guy who ran four 813s in GG with 4KV. He claimed it did 2KW pep out but the plates were stressed when he talked too long...

Actually, some years ago I bought twelve  813 sockets (China) to build a silent GG amp, no fans.   Never did anything with it. But I thought six in a circle would be a cool project too.  Maybe with 3500V it wud do a VERY conservative, cool and silent  375 w carrier with ~ 20 watts of carrier drive, I dunno.

This has morphed from the 833A plate modulated project I was considering last week.  The 4X1 (Fabio II) has become so complex with the elaborate sequencing and new grid/screen/plate shutdown circuitry, I just didn't want to go thru it again with another plate modulated rig.  Thus, the desire to keep it simple and silent with convection cooled GG 813s.

T

[K1JJ]

Don,

12-13 dB sounds reasonable.  

I always use an input tuning for better IMD and easier drive requirements.

Interesting on your thought that screen voltage will make it cleaner. Is this idea from actual measurements or common sense? Considering it already gets NFB from the GG config, that wud be a bonus.  I could easily share an electronically regulated screen voltage from my 4CX-350FJ linear, say, 300 volts reg.   What screen voltage wud you suggest at 3500V HV?   Any idea what the bias wud need to be?  I wud use diodes in the cathode to keep it simple and stable as I usually do.

[KK4YY]

Tom,

I got that IMD reference from here...

http://www.w8ji.com/grounded_grid_amplifiers.htm

"In an attempt to improve linearity and reduce drive power requirements, I built a screen and bias supply. These tubes required about 130 volts of control grid bias when screen voltage was applied. After designing and constructing a very good bias and screen supply, vacuum tube regulated with 6146 regulator tubes, drive requirements barely changed. Of course IMD products were reduced greatly because the tubes had much better ratios of screen to control grid current (control grid current went to zero mA)."


Also the original G3SEK article in QEX is a must read.
http://www.muenster.de/~dl5qe/qexartic.pdf

-Don

[KK4YY]

Also, here's the schematic of what I was planning as I was thinking through this. A work in process.

-Don

[KK4YY]

Tom,

As for screen and bias voltages, I have those same questions unfortunately. I'm hoping to find the answers in this tread. I have done a lot of searching (which is way easier than learning) but even though the 813 has been in many designs, running them cathode driven with screen voltage (tetrode mode) doesn't seem to be very common. The issue in the past was getting the screen voltage regulation very very tight. With modern electronics this is do-able so it's a mix of old and new tech for the venerable 813.

-Don

[K1JJ]

OK Don -

That's good info about tetrodes.  I've been thru it before when I built my super clean 4CX-350FJ grid driven amp.   I have all the screen snubbers and arc precautions as well as shut downs.  Considering that the tube is VERY clean (-55dB 3rd) when no grid current - but degrades when grid current begins, I will run the 813s the same way, except GG.

I was surprised to see in modulator AB1 service, the Eimac 813 datasheet suggests 750 volts! for the screen at only 1500 plate HV. It stays at 750 volts even when at 2500V.  That seems quite excessive and I wud prefer to run it at 400 - 450V.... and run the cathode drive JUST before grid current starts. (GG)

BTW, at 750V they are suggesting 54 ma of screen current at full drive. That's 40 watts of power. Times four tubes is 160 watts.  Guess there's no free lunches.

Considering what I wish to do, that is, to drive it conservatively in AB1, I have laid out six sockets and tubes and like the looks of it...

T

[W2VW]

Guy I know had a "contest" amp which ran over 4 kv and somewhat less under load on a quad of 813s. I calculated dissipation for each tube was 500 watts during a steady carrier tuneup!

Damn things worked every year for the Sikh Eew Whilwide.

[KK4YY]

Tom,

Wow, six 813s? Got 'em lined up like soldiers? I was looking at using just three (that's how many I have).

I was hoping to find a smooth way of switching from AB1 with 150 watts out and little (quiet) air movement for a clean output and then "flip a switch" for 300 watts with lots of air and poorer IMD when the need arises (as is sometimes the case). If I can change HV to keep the same load impedance I won't have to re-tune the output... that's the thought anyway. Should give more gain at the same time by doing that. It would be too cool to have a switch on the front panel labeled "Quiet Mode / Riot Mode". This may be a bridge too far for me, but I can dream 'eh?

At the rate you build, I'm sure you'll have yours on air before I finish my final schematic so I look forward to peeking in at your progress on the build.

-Don

[w7fox]

Hello fellow 813 users.  I have had a linear running two 813's for a long time...28 years.  I run 2800 volts on the plates, and use a pi net on the input and output.  My power meter stinks but it says something between 600 and 700 watts output.  It was very easy to overdrive with my Heathkit SB-101, and very abusive on the grids, so I tapped down the grid toward the cathode until it can take all the power the Heathkit can give.  No more overdriving.  The circuit is in various places, but simply consists of a tapped homemade choke similar to filament chokes.  The screen is driven directly, that is, its grounded, but the grid gets about half drive.  Good luck, and try not to abuse those 813's.  Oh yes, I forgot to mention I use a muffin fan and usually show no color on the plates.

73, Chris

[K1JJ]

Well, I finally decided to build six 813s in a grounded grid configuration with 400V regulated on the screens.  The screen voltage will make it run cleaner with no grid current. This is better than both grids at DC ground since grid current tends to degrade IMD. I figure an easy 13dB of gain which means 20 watts of carrier in and 400 watts of carrier out.  Using 3200V will give an easy X6 pep of headroom.

I plan to wind my own fil xfmr on an old Variac form. The primary is 120V and I'll just add 10 turns or so to make the secondary 10V at 30A.

Notice the cathode zeners which will allow class AB1, B and class C.  I especially want the class C position for the high-efficiency AM linear use. (Refer back to a previous thread about class C AM linear operation)

I plan on a single, quiet, large muffin fan to suck air up and out of the cabinet. Should be a silent amplifier, which is my main goal.


Here's a first draft, hand-drawn schematic below.  Anyone have suggestions or comments?

Do I need a separate RF choke to each 813 screen -  or is one screen choke, community-shared, as drawn, OK?

T

* Update:  I forgot to add .01  filament bypass caps to ground. The ant relay normally-closed contacts need to be joined together for the receiver to work. Also, an electrolytic cap to bypass the zener string.

[N0WEK]

Nice!

Now all you need is a large glass tube to surround the circle of six 813s and provide good viewing and chimney action.

Maybe cut the top and bottom off of one of those 5 gallon water cooler bottles?

I assume you're going to put the plate choke in the center of the circle?

[W2PFY]

Hi Tom, what will you use for the filament choke? Roll you own,etc?

[K1JJ]

Hi Terry,

As I do for all my linears, I run two parallel filament leads through a stack of large toroid donuts. Usually about 12" of core stack length does it.  

The current requirement is 30A at 10V, so there is less voltage drop when using less wire vs: the coiled choke method.  Possibly less chance for a filament choke resonance too.

Greg, interesting idea of mounting tubes in a circle surrounded by a circular glass form.  Haven't decided on layout yet.  I'd like a design that will pass some air thru the bottom tube seals too.


T

[N0WEK]

Hi Terry,

As I do for all my linears, I run two parallel filament leads through a stack of large toroid donuts. Usually about 12" of core stack length does it.  

The current requirement is 30A at 10V, so there is less voltage drop when using less wire vs: the coiled choke method.  Possibly less chance for a filament choke resonance too.

Greg, interesting idea of mounting tubes in a circle surrounded by a circular glass form.  Haven't decided on layout yet.  I'd like a design that will pass some air thru the bottom tube seals too.


T

You could put all the filament and input circuits in a lower chassis, the tubes and plate choke in the glass tube and the output circuits on an inverted top chassis with a mirrored rear chassis in the back connecting the two. I would be really pretty and would isolate the input from the output. Maybe a small, quiet muffin fan in the lower chassis to push air through the sockets and tower into the upper chassis.

[w4bfs]

sextet of 813 - sounds like modern art exhibit .... look at the BW 1000 amp for the negative fb stub / fil choke arrangement .... without it even 40 mtrs may not be stable .... gl 73

[K1JJ]

sextet of 813 - sounds like modern art exhibit .... look at the BW 1000 amp for the negative fb stub / fil choke arrangement .... without it even 40 mtrs may not be stable .... gl 73

You make a good point about stability.  If built well, I think six would be OK up to 40M in GG without additional RF NFB, but higher might be a problem.

HOWEVER, I am still considering other choices.  The main objective is to build a silent amplifier that relies mainly on convection cooling. (And looks good.)   I have another single 4-1000A with socket, chimney and fil xfmr that I think could be configured to run with a slight breeze thru the chimney to cool the seals. At about 4KV in GG linear, it wud do everything the six 813s could do and be more stable.  I wud also use about 500 volts of regulated screen voltage and run zero grid current for maximum cleanliness.  Driven by the FT-1000D, with preset roller inductors in the input and output, vac variables and turns counters, it wud be able to row around the bands quite quickly.

A single 4X1 wud be about 1/2 the fil power of the 813s too.

Another idea is four 4-400s in GG with screen voltage, run basically the same way and quiet.  They wud really sing at 4KV, whereas a 4X1 is just loafing.  4-400s wud need about the same fil pwr as the 813s.   So the single 4X1 looks attractive.  In GG it already has NFB and the screen voltage will make it even cleaner.

The good thang about glass tubes is they radiate a lot of heat thru the glass and are OK as long as the seals are reasonably cooled. In contrast, an external anode ceramic tube MUST have forced air - and are noisy.

Still thinking..

T