The AM Forum
March 28, 2024, 06:37:07 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 ... 5 [6]   Go Down
  Print  
Author Topic: 2x 813 amplifier?  (Read 129059 times)
0 Members and 2 Guests are viewing this topic.
N2DTS
Contributing
Member
*
Offline Offline

Posts: 2307


« Reply #125 on: November 23, 2009, 11:03:14 PM »

I dont have a lot of air on the tubes, just a muffin fan blowing up at the tube sockets, and the holes around the tubes are bigger than the tubes to allow the air to flow out.
The fan is not running at full speed, as it makes too much noise.

I think I will try the bias increase though.
I could js in about 20 more diodes for testing.

Brett
Logged
N2DTS
Contributing
Member
*
Offline Offline

Posts: 2307


« Reply #126 on: November 24, 2009, 10:39:06 PM »

Ok, I added as a test, 18 more diodes, another 12 volts of bias, and did some tests...
Results were mixed, while I could do 200 watts carrier out, 500 watts in, 300 watts dissipation in the tubes, peak power fell off quite a bit.

I tried every combination of input drive power and loading with no improvement.


Brett
Logged
K1JJ
Contributing
Member
*
Offline Offline

Posts: 8893


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


« Reply #127 on: November 25, 2009, 12:05:54 AM »

Sounds like the dead carrier power is 40% efficient - that's pretty good. What is it when you run it class B?

Do you have enuff drive peak power to hit the required amplifier peaks?  Just cuz the amp is in class C shouldn't reduce its ability to make peak power. Maybe your exciter is running out of headroom. It will take more drive, as I mentioned.

If you have another linear, you could try it as a test  IPA - but be very careful.

Also look at the input waveform on a scope. Is the input power still rising while the output power saturates?  Use a 1KC audio tone and look at the sine wave carefully for signs of saturation on both the input and output. That shud give you a good clue..

BTW, fired up the 4X1 X 4X1's tonight into the dummy load.  The 75M tank coil was too small - recalculated and gonna replace. Something in the screen circuit shorted to ground. Must be a cap - will check tmw.  The audio didn't work. I found I had wired the SS amp directly to the 4X1 grids...  Shocked  Should go thru the driver xfmr first.  Other than that and a few minor details like neutralization, it's coming along FB so far.

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.
WA1GFZ
Member

Offline Offline

Posts: 11152



« Reply #128 on: November 25, 2009, 07:44:09 AM »

My old 4-1000A linear had a string of zeners to bias it well into class c. This way I could get the best efficiency on CW. Every volt of bias shifts the drive signal down a volt so you need more drive in class C to hit your peak drive power, but your efficiency will go up and in class c and your tank will want more L. Tank Z for class c is Plate Volts/plate current / 1.2.  (1.6 for class AB)
Logged
W3FJJ
Contributing
Member
*
Offline Offline

Posts: 159



« Reply #129 on: November 25, 2009, 08:15:15 AM »

I've been reading this tread with some interest, as I have always have had trouble getting
good efficiency out of my homebrew amps.
You guys stated earlier the importance of using more C less L to get efficiency up.
I suppose that makes it a higher q circuit. I've always designed my HB amps using the ARRL
table for pi-nets, they use a Q of 12. Since my amps run a plate impedance between
3000-5000 ohms, the handbook  suggested L is 12.45-18.56 uh on 80m quite a far distance from the 7uh
Tom mentioned.  I'll be playing around with
this some, in the future...  Thanks.... and GL
Logged
K1JJ
Contributing
Member
*
Offline Offline

Posts: 8893


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


« Reply #130 on: November 25, 2009, 11:20:42 AM »

I've been reading this tread with some interest, as I have always have had trouble getting
good efficiency out of my homebrew amps.
You guys stated earlier the importance of using more C less L to get efficiency up.
I suppose that makes it a higher q circuit. I've always designed my HB amps using the ARRL
table for pi-nets, they use a Q of 12. Since my amps run a plate impedance between
3000-5000 ohms, the handbook  suggested L is 12.45-18.56 uh on 80m quite a far distance from the 7uh
Tom mentioned.  I'll be playing around with
this some, in the future...  Thanks.... and GL


Hi Chuck!

Long time, OM.

I try to use a Q of 12 on all my amps too.   In fact I just went thru it last night for my new 4X1 class C amp and found I need about 18uh or so for 75M. (I had about 8 uh in there and found it worked FB on 40M... Wink  I use the following coil calculator and pi-network calc on the web.

http://www.qsl.net/wa2whv/radiocalcs.shtml

http://www.captain.at/electronics/coils/


By going between the two, you can find just what you need to tap off using the particular coil you have on hand or need to wind.

I did my 4X1 coil by seat of the pants and found it was way too small. I was used to the big linears with low plate impedance and forgot I was dealing with a high impedance again... Grin

My comments about using a lot of C meant that I had always underestimated how much C was required to give a Q of 12. In the past I always used too much L cuz I was using breadslicers and never seemed to have enuff C to do the job.  Stick with these calulators above and the tank will work out FB. (Or your ARRL charts)

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 #131 on: November 25, 2009, 11:30:53 AM »

My old 4-1000A linear had a string of zeners to bias it well into class c. This way I could get the best efficiency on CW. Every volt of bias shifts the drive signal down a volt so you need more drive in class C to hit your peak drive power, but your efficiency will go up and in class c and your tank will want more L. Tank Z for class c is Plate Volts/plate current / 1.2.  (1.6 for class AB)

HMMM.. maybe that's the problem Brett is having trying to get PEAK power outa of the newly biased class C amp on AM. IE, the tank impedance needs more L for the peak power impedance match.

BTW, Frank, Brett is experimenting with putting his linear biased into class C for AM use. Read back on my posts here about using the carrier as the "linear bias." It worked for me in the past to achieve a higher dead carrier efficiency (like 40% carrier, 65% peaks)  - and we're hoping he can verify these findings too.

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.
N2DTS
Contributing
Member
*
Offline Offline

Posts: 2307


« Reply #132 on: November 25, 2009, 03:40:15 PM »

I was talking with Stu today and he said I have no dip because the Q is too low, so I did some experiments.
I increased the inductance till the caps (tune and load) were almost nothing, I got a dip in plate current, but noplace near maximum power output.
Otherwise, no change in efficiency, peak power, etc.


I think my peak power is down somewhat for 2 reasons, 2000 volts on the plates, and a lack of tuned input.

In the experiments i did with the bias increase, I tried increasing the plate voltage, various settings of drive and loading, without getting the peaks back.
Normal operation gives 150 watts carrier and 800 watts pep.

Brett

Logged
N2DTS
Contributing
Member
*
Offline Offline

Posts: 2307


« Reply #133 on: November 27, 2009, 10:28:47 PM »

More amplifier questions...
Never having had one before, how do they usualy handle the grids?
I have each grid, screen, supressor grid going to ground at the tube socket, the grid and screen through a 10 ohm resistor and a .01 cap (in paralell) to ground.

Would it be better to connect all the grids of each tube directly together and then to ground, all the tubes to a single point ground, or does it not matter?

In qso with Stu and Dave today, dave thought it was a neutralization problem, which I never heard of in grounded grid amplifier, so I am thinking maybe its how I have the grids grounded...

Brett

 
Logged
K1JJ
Contributing
Member
*
Offline Offline

Posts: 8893


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


« Reply #134 on: November 27, 2009, 11:50:09 PM »

Brett,

On all my linear amplifiers, including two commercial 3-500Z amps, I always tie the grids together with wide copper strap and go directly to a single point on the chassis.  None of my GG amps have needed neutralization  and were always stable. Bias is accomplished through the fil CT, of course.

I never liked those funky inductors or caps in series with the grids for supposed neutralization. Seems like a good place for instabilities.  

As I mentioned before, I usually don't pay attention to the plate current meter in a linear except for idling current. For tuning, I just keep an eye on the power output and watch for a sharp tuning peak when the Q is designed correctly.

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 #135 on: November 28, 2009, 01:37:57 AM »

Brett,

On all my linear amplifiers, including two commercial 3-500Z amps, I always tie the grids together with wide copper strap and go directly to a single point on the chassis.  None of my GG amps have needed neutralization  and were always stable. Bias is accomplished through the fil CT, of course.

I never liked those funky inductors or caps in series with the grids for suppoused neutralization. Seems like a good place for instabilities. 
T


Thank you!  I do the same, and any amplifier darn near that crosses my bench, I throw the BS away.

Yes, this causes MANY an argument on amplifier reflectors, but I can say this:  I've NEVER had a parasitic in any of my amps with the grids directly grounded.  Not to say that it can't or won't happen, but (knock on wood) nothing thus far.  The YC165 is great like that, no suppressor needed, bolt the grid to the chassis.

People that put the grid "LC" circuits don't take this into account...  You now have an additional c. 26 ohms of resistance (DC) in line with the actual reactance.  Since this WILL effect the operating parameters of the tube, WHY WOULD YOU PUT A 26 OHM RESISTOR IN LINE WITH YOUR BIAS SUPPLY?  The grid is the return for the CT bias in the 3-500 style amplifiers.  Yeah, I like having a variable bias supply on my amplifiers.

NO, ground the grids and be done with it.

--Shane
Logged
KD6VXI
Contributing
Member
*
Offline Offline

Posts: 2648


Making AM GREAT Again!


« Reply #136 on: November 28, 2009, 01:40:10 AM »

More amplifier questions...
Never having had one before, how do they usualy handle the grids?
I have each grid, screen, supressor grid going to ground at the tube socket, the grid and screen through a 10 ohm resistor and a .01 cap (in paralell) to ground.

Would it be better to connect all the grids of each tube directly together and then to ground, all the tubes to a single point ground, or does it not matter?

In qso with Stu and Dave today, dave thought it was a neutralization problem, which I never heard of in grounded grid amplifier, so I am thinking maybe its how I have the grids grounded...

Brett

 

Brett,

As Tom says, do a DIRECT bond to the chassis.  I do this on all my grounded grid amplifiers, as well as many tetrodes (operated at DC ground), and don't have a lick of problems.  From 4CX250s to the 4x15, I've not had a parasitic go off in one of my amps.  I'm not a magician, nor do I know for a FACT why, but they are all stable.

Having that 10 ohm resistor in series with the grid just meant that you introduced 10 ohms of DC resistance in line with the bias of the tube.  Makes for nice linearity, having the bias jump around a few volts in both directions, ya know?


--Shane

Logged
N3DRB The Derb
Guest
« Reply #137 on: November 28, 2009, 07:14:54 AM »

FWIW, I received a gates 2X 4-400 FM rf deck from 7FTO and I thought the way they wired the sockets was pretty nice. 1/2" wide silver plated copper strap for every connection tied together with 2/56 nuts and bolts ( I presume copper) and then soldered. even if the solder would melt,
you'd still have a decent mechanical connection holding it together.

I'll take a pic of the sockets in a bit. I say ground the grids with strap to the chassis. No funny stuff, just ground em.

remember,soldering to a cad steel chassis is only as good as your plating is thick, couple of mills.  For a really good ground connection, drill a small hole in the chassis and put a press to fit wire through the hole and lay both ends of the wire flat against the chassis. treat it like you would treat a plated through hole on a pc board. solder down both ends on both sides after trimming. Now you have a ground to the steel underneath the plating. You could use small nutz and bolts as well. the important thing is bonding to the steel and not just the plating lying on top.
Logged
N2DTS
Contributing
Member
*
Offline Offline

Posts: 2307


« Reply #138 on: November 28, 2009, 12:39:14 PM »

Thanks guys.
I fixed it, but might redo the grids direct to ground with strapping.

What I did was add another plate choke and cap under the chassis, as that is shown on many grounded grid amps, and I also added some spring clips to ground the metal 813 shells.

Now I DO get a dip at peak power output, otherwise, the amp runs the same, same power out, same current and efficiency (or very close).

So one or both of the above eliminated some feedback between the plate and grid circuits.

Brett

 
Logged
K1JJ
Contributing
Member
*
Offline Offline

Posts: 8893


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


« Reply #139 on: November 28, 2009, 01:03:38 PM »

Aaaah.....  you just triggered a memory here -

I once built a linear that had VERY lazy tuning like that on 160M. It worked "OK" but did not have a plate dip at all and the power out was so-so.  I simply added a few more inches of plate choke windings and the amp came to life on 160.

Evidently in my case,  the inductive reactance was too low for 160M and some power was coupling into and the tank was being influenced by the power supply path to ground, killing the tank Q. 

With your four 813's in parallel, it would seem the requirement for a large value choke would be reduced, but you can't argue with success.

Now add the copper grid straps to groumd, "paint it blue, and ship it"... Grin   (an old VZR saying)

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.
W2VW
Contributing
Member
*
Offline Offline

Posts: 3489


WWW
« Reply #140 on: November 28, 2009, 01:17:54 PM »

More amplifier questions...
Never having had one before, how do they usualy handle the grids?
I have each grid, screen, supressor grid going to ground at the tube socket, the grid and screen through a 10 ohm resistor and a .01 cap (in paralell) to ground.

Would it be better to connect all the grids of each tube directly together and then to ground, all the tubes to a single point ground, or does it not matter?

In qso with Stu and Dave today, dave thought it was a neutralization problem, which I never heard of in grounded grid amplifier, so I am thinking maybe its how I have the grids grounded...

Brett

 

Brett, I did not think is was a neutralization problem. All I said was to look at the dip verses max power out to get a feel for how well neutralized the thing is.

Lose the 10 ohm resistors and bond all grids together to ground as mentioned.

Did the existing plate choke show any signs of distress?
Logged
stevef
Guest
« Reply #141 on: November 28, 2009, 02:10:54 PM »

Here's how I did my 813 grids.


* grids1.JPG (36.73 KB, 519x389 - viewed 1222 times.)
Logged
N2DTS
Contributing
Member
*
Offline Offline

Posts: 2307


« Reply #142 on: November 28, 2009, 09:55:30 PM »

Very nicely done on the grids!

No distress on the plate choke, its one of the old B+W ceramic jobs, but many amps I look at use 2 chokes because they say its hard to uncouple the low impiedance plate from the power supply.
I don't think the B+W choke was designed for such low impiedances.

I will try the grid change next.

Brett
 
Logged
WA1GFZ
Member

Offline Offline

Posts: 11152



« Reply #143 on: November 29, 2009, 12:21:23 AM »

B&W 800 is only 90 uh so would be a bit low for 160. 2 in series should be fine. B&W made an 801 that was 200 uh and the same size. I used one in a 4-1000A rig to convert it to 160M. I bet the rig would have worked at full power running a low plate Z. You are lucky it didn't fry. They are easy to rewind though.  I still think your tank Q is a bit low.
Logged
W3RSW
Contributing
Member
*
Offline Offline

Posts: 3308


Rick & "Roosevelt"


« Reply #144 on: November 29, 2009, 11:39:37 AM »

...and I too.

If he hasn't already, Frank usually mentions  turning the plate tune cap. over for a shorter connecting strap.   Cool

Logged

RICK  *W3RSW*
N2DTS
Contributing
Member
*
Offline Offline

Posts: 2307


« Reply #145 on: November 29, 2009, 07:10:56 PM »

I did not try turning the cap over, I may later.
The Q should be ok, I used more inductance and have about 75 pf used in the tuning cap.

I bonded the grids all together and then to ground, no change in operation of the amp, which seems to be working fine.
Not an AM power house with 500 watts of plate dissipation, but it gets me from 22 watts carrier and 100 watts pep to 175 watts carrier and 1000 watts pep at well below the maximum tube ratings.

That seems to make quite a difference, but it was sort of amazing how well I was heard at 25 watts.

Brett


Logged
WA1GFZ
Member

Offline Offline

Posts: 11152



« Reply #146 on: November 30, 2009, 09:02:54 AM »

Adding inductance will reduce the Q.
Flipping the cap over will help with stability and ability to resonate on the higher bands.
I've seen some nice amps that I know won't work on 10.
Logged
N2DTS
Contributing
Member
*
Offline Offline

Posts: 2307


« Reply #147 on: November 30, 2009, 09:20:57 AM »

I dont plan on using the amp above 20 meters, and don't think I would get on 20, but I might want to check into the flex net up there on ssb...

I have tried lots more and lots less inductance in the pie net, does not seem to make much difference in power or operation.

I have to get some more thick wire, for 40 meters I use copper tubing, but on 80 meters, I cant seem to get enough inductance in the space I have using tubing, without making it very large around.

I did make one tubing coil as long as I could, it might work, but at a high Q, I forget how it tested.

In no case do the coils seem to get warm...

Brett
Logged
Pages: 1 ... 5 [6]   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 0.048 seconds with 19 queries.