QuadZilla sees first light - Quad 4X1's - pictures

[KL7OF]

Tom... Is it considered to be class A operation with the low (20 watt) drive and 30% efficiency and class B with the higher drive and 65% efficiency...without the bias changing....?  I see you have that thing idling at 250ma with no bias...What will change if you apply bias?  (Other than the idle current) If you put a bias supply in and apply the bias all the time vs bias during recieve (standby)only... What about grid leak?     I'm considering building a 2 holer.

[K1JJ]

Tom... Is it considered to be class A operation with the low (20 watt) drive and 30% efficiency and class B with the higher drive and 65% efficiency...without the bias changing....?  I see you have that thing idling at 250ma with no bias...What will change if you apply bias?  (Other than the idle current) If you put a bias supply in and apply the bias all the time vs bias during recieve (standby)only... What about grid leak?     I'm considering building a 2 holer.

Hi Steve,

Glad you axed, cuz I just got done with some extensive tests and settled on some operating parameters for now...

Well, actually, true class A is idling so that the entire 360 degrees of the drive signal stays above cut-off and conducts 100%.  I was probably more in AB cuz the plate current was swinging upwards, but the drive was so low that it acted like class A efficiency.   I think when the amp is driven harder, it draws grid current, so the efficiency goes up.  But you're correct, class is mostly predicated by idle current.

BTW, to unkey, just open the fil CT with a relay. Look at the Telsa 813 circuit. For this linear, the diodes are in the circuit all the time, RX and TX.

Anyway, here's what I came up with...  I'm talking about class AB2 now. Running the amp at 6kv was way too much voltage for AM operation. Maybe it wud be OK for ssb, but on AM the plate dissipation was exceeded too easily and it made just too much power... :-)  I found that running it at about 3500V was perfect. For four tubes, this gives a dull orange plate color, with an idle of about 300 ma. (about 75ma X 3500V = 262W per tube) Considering they are 1KW dissipation tubes, this is FB.  It makes plenty of power output at 3500V - plenty.. :-)

HUZman sent me some old IMD specs for the 3-500Z showing the IMD was much better at lower plate voltage, as low as 1500V. I know some guys think the opposite, but maybe a lower voltage is better. I will just make it up in current.  Real high voltage bugs me anyway.

Another thing I've been noticing for years and have talked to a few people about:  On AM, when adding a lot of bias to a linear, until the tubes go below cutoff (just below class B - barely into class C) the linear operates very well on AM. The AM carrier seems to bias the tube into conduction and operates linearly when modulated. But on ssb, there is severe cross-over distortion cuz the amp stays below cut-off until a loud voice peak comes thru.

To confirm this, I experimented with adding diodes in the CT of the fil xfmr as bias to class C, and see that the tube plates glow duller on AM (less power dissipated) with the higher bias. This is probably simulating a class C grid/screen modulated stage.  Anyway, I loaded it heavily, brought the amp down into B/C operation and it operates very, very nicely on AM. I think the efficiency is up around 65% on voice peaks, from what I can tell.

I tried driving it with the MOSFET MRF-150's and the power output was tremendous - and clean. So, I have a lot of choices here on AM - to run it class C, B or class A - with or without the MOSFet driver.

I don't think grid leak is necessary with this config - it is so much easier to use a string of diodes in the CT and short some out to change the idle level for different purposes. For example, to go on ssb, I wud just short them out and it wud idle at 300ma - perfect for ssb. In CW, let them run in class C.  The tones thru the amp are very clean as is, so I don't think grid leak wud help to linearize it any more on AM.

It needs a lot of air, but running it at a lower voltage and splitting the dissipation amongst many tubes helps quite a bit to quiet things down. I found when using a moderate amount of compreesion of AM, the amplifier (8' away) creates just a small amount on on-air blower noise. Quite acceptable. With no compression, there is no noticable on-air noise.

So, I'm a happy camper. This is one of the few projects that fired up right away and had no problems or crap outs. That's a rare event, trust me.. :-)

Tell me more about the two holer plans ??

T

[K1JJ]

Steve -

After thinking it over, I'm going to try biasing the tubes far into Class C for AM (using more dodes in the CT) and see if I can raise the carrier efficiency.  It will act much like a standard grid modulated class C "efficiency" amplifier.

Also, the more bias it has, the more drive is needed. So good thing I have the MOSFET driver amp. I figure the efficiency will be the equivalent to a standard grid driven class C amplifier with standard grid bias and screen voltage that gets modulated via an audio transformer connected in series with the grid and/or screen.  But this will be a GG class C amplifier that gets its grid and screen modulation thru the RF signal. Since both grids are grounded, it is a composite grid and screen modulation system.

The thing is in GG, usually the drive signal does not get modulated - and is added to the output. The result is it will not hit negative -100% modulation.  BUT in this case it WILL be modulated via the RF signal and added to the output power, so 100% negative is achievable.

The GG stage is also very stable compared to a grid driven amp.

BTW, you axed about grid leak bias...  in GG, you want to strap all grids/screens directly to ground with copper strap. Grid leak wud require a floating/bypassed ground, so is not as good for stability. Use the diodes instead - but you might try a grid bias test anyway just to see what it does.

I plan to add a rotary switch to be able to select (short out) any combination of diodes in the string to go from class C to AB.

This could be an interesting experiment. From initial tests, this could be a great way to have a band agile amplifier with reasonable efficency on AM while maintaining tremendous fidelity and positive peaks from the low level balanced modulator driver.

I'll let ya know how tests work out.

T

[KL7OF]

Tom...vERY INTERESTING....I have been experimenting with a single GG 4x1 using a relay to switch in grid leak on idle and ground the CT on xmit...I just acquired a fil trans that will lite two tubes so I want to try the duece...So far I have only used 4 KV  I am want to try some lower Voltages...  I really want the low distortion clean signal  LOUD AM amplifier that I can drive with the 102 for push to talk..
     Are you monitoring the grid current?  Have you meausured the voltage drop(s) on your bias diode string?  I like the input tuning T...  Heavy duty....

[K1JJ]

Tom...vERY INTERESTING....I have been experimenting with a single GG 4x1 using a relay to switch in grid leak on idle and ground the CT on xmit...I just acquired a fil trans that will lite two tubes so I want to try the duece...So far I have only used 4 KV  I am want to try some lower Voltages...  I really want the low distortion clean signal  LOUD AM amplifier that I can drive with the 102 for push to talk..
     Are you monitoring the grid current?  Have you meausured the voltage drop(s) on your bias diode string?  I like the input tuning T...  Heavy duty....

Steve,

Just to clarify, when I say, "idle" I mean the quiescent, no signal current, when the amp is keyed. To put the amp in standby on receive, just lift the CT off ground, as shown in the Tesla 813 amp schematic. No need for grid leak bias unless you want to experiment with it on transmit only. But, in GG, the screen/grid shud be strapped to hard chassis.  Use diodes only - for bias.

Good on the FT-102 driver - FB... also the fil xfmr to handle 42 amps... :-)

My grids are strapped directly to gnd, so I'm not monitoring grid currents. It doesn't really matter cuz there's nothing that can be done about them anyway in GG, as long as their ratings are not exceeded. The Eimac specs seem to say GG is OK on the grids when the tube is driven to max.

I will measure the diodes' voltage tmw. I have about 40 diodes in series now that just about cuts off the amp when keyed, no signal.(idle) This is about 19 volts positive on the cathode = -19V on the grids ref to cathode.   I'm gonna add another 50 diodes and see what a total of 45 volts does for class C.

Talk wid ya tmw -

T

[KL7OF]

Tom ..OK on the difference between  quiescent and standby idle...It makes sense to me now...I await your results...

[K1JJ]

Tom ..OK on the difference between  quiescent and standby idle...It makes sense to me now...I await your results...

OK Steve -

I plan to get it on the air over the next few days for some shake down runs. We'll let ya know.

BTW, another benefit of the grounded grid configuration is the RF negative feedback generated by the cathode drive power - cleaner. This power is not wasted and even shows up in the output.

Also, I just realized that a GG amp running in this triode configuration for RF AM is no different than running them triode connected as modulators for AM audio. In triode connected modulator service they sound wonderful.... just like the triode connected 813's.

I'm cornvinced that a GG 4X1 with class C bias is a sleeper config for RF AM use.   Very simple circuit and supporting infrastructure.

T

[K1JJ]

Here's some shots of the bias diode board that goes in the filament transformer center tap. By shorting out diodes, it changes the tube's operation class from class AB to B to C.  The rotary switch gives about 11 bias positions.  The bias is rock solid at all currents since it is the drop across the diode junctions. (About 0.5v per diode drop)  I used fast recovery 3A 1KV diodes cuz that's what I had available.  Any diodes will do with at least 1A rating. The voltage rating can be as low as 50V per diode.

I measured the bias voltage when the amp is keyed and it is selectable form 0 to 23 volts.


At 3500 volts on the amplifier:  (this is my recommended voltage for AM operation by experimentation - 3,000 volts will do OK too)

At 0 volts the amplifier idles at 500 ma.  The next step is 250 ma and then 125 ma. This is good for ssb operation.


For AM, I found a bias of about 18.5 volts works well. This cuts the amplifier off at about 5ma of idle resting current. (no signal)
The higher biases are for CW operation in hard class C. It gets increasing more difficult to drive the amp above 18 volts of bias, so 18v is a good compromise for AM operation - if you have plenty of drive available.  

Steve: You will probably want to run slightly less bias with two tubes when using the FT-102 as a driver. Maybe 12 to 15 volts will do for you. The less bias, the easier to drive, but then the plate dissipation is higher.

Pics below of the diode board:

[K1JJ]

Latest pic:

Notice the HV meter mounted on the top. I wanted to make it REAL clear... :-)

It's so easy to work on with that side Plexiglas panel removed...

T

[WA1GFZ]

Tom,
You should find a way to monitor air pressure inside the chassis to get an idea where you stand when in QRO to make sure you don't trash a tube. Also would be good to know the air temp above the tube. A thermal couple on a small ua meter would be easy to calibrate. I have a spool of thermal couple wire around here somewhere. fc

[N3DRB The Derb]

that is bad brotha. That there gonna wipe out the moonwalk. 

[Fred k2dx]

Tom ..OK on the difference between  quiescent and standby idle...It makes sense to me now...I await your results...

OK Steve -

I plan to get it on the air over the next few days for some shake down runs. We'll let ya know.

BTW, another benefit of the grounded grid configuration is the RF negative feedback generated by the cathode drive power - cleaner. This power is not wasted and even shows up in the output.

Also, I just realized that a GG amp running in this triode configuration for RF AM is no different than running them triode connected as modulators for AM audio. In triode connected modulator service they sound wonderful.... just like the triode connected 813's.

I'm cornvinced that a GG 4X1 with class C bias is a sleeper config for RF AM use.   Very simple circuit and supporting infrastructure.

T

Tom, I have buried somewhere some interesting articles on variations of typical  tetrode/triode amplifiers. You may be interested in:

"The Class 'C' Linear Amplifier" by D.O. Mann w6hly/w3mby
in CQ, March & April 1964

"The Cathode Driven Linear Amplifier" by Bill Orr w6sai and William Sayer wa6ban
in QST, June & July 1967

"The G2DAF Linear Amplifier" by C.W.O. Joseph Dempsey wa6jcm/4
in CQ, March 1966

Some  very interesting reading. Bill Orr discusses a simple way to control RF negative feedback, power gain, linearity etc. In "The Class 'C' Linear Amplifier" D.O. Mann talks of peak efficiency of 85%, 20 db gain, -30db IMD w/o negative feedback...

If these CQ articles aren't available (QST is online I think for American Radio Ripoff League members) I could scan and try to post them, or mail them to you on cd.

As Don KYV mentioned elsewhere, a reduced carrier/controlled carrier am rig driving something with huge PEP capability could be an interesting exercise. The Drake T4 was very effective on am driving a linear with a carrier of about 8 watts that would talk up to 50 Bird Watts.  Then run the amplifier essentially at ssb levels. Many corntester SSB rigs transmit more than that much power from uncontrolled blower noise!!! The Drakes were capable of decent frequency response without a lot of effort.



Hoooooooooooola!

[K1JJ]

Tom,
You should find a way to monitor air pressure inside the chassis to get an idea where you stand when in QRO to make sure you don't trash a tube. Also would be good to know the air temp above the tube. A thermal couple on a small ua meter would be easy to calibrate. I have a spool of thermal couple wire around here somewhere. fc

Frank -

Yes, safety features are the next step. I like your idea about thermal alarm/shutdown. I have a sensor diode hanging over my other linear's final which turns on an alarm when it gets too hot.   The reason I installed it was cuz I once ran it w/o air - the blower hose pulled out. It got so hot that solder weeped out of the fil connnections.  I will do something for the 4X1 rig too.

Also need to build up that input overdrive snubber for the MS-150 SS amp.


Derb: "The Moonwalk" wud be a good name for a rig too... :-)   Right now I still have "Mr. Ugly,"  "Dr. Love" and "Fabio" running in the shack.


Fred:  Those sound like very interesting articles. Sounds like someone "discovered" class C AM linears way before I did... :-)  I remember reading that article by the British guy somewhere before.  Tell ya what... if you cud scan those articles on your printer/scanner and email them to me, that might be easier than CD/mailing. Let me know - thanks much.  (email OK on QRZ.com or profile here)  About the super-modulation: The controlled carrier thing sounds interesting cuz it wudn't overload a diode detector like a 200% balanced modulator would, but then the hedrodynes get funky. We'll have to experiment with the best way to use the headroom.


BTW, I found another used 4X1 (4PR-1000) from a guy that may work out and replace the soft one in the lineup. I'll know in a week. Actually I bought two used ones just in case.  Hope to have the new amp shaking on the air this weekend.


73,

T

[Fred k2dx]

OK Tom. Do you need the Qstreet Bill Orr article too?

Have you considered balancing the bias/idling current between the tubes? Granted the tubes are rugged, especially the 4pr1000's... did it with six 6lq6's once, it just seemed the thing to do.

Another advantage to your amp is the multipile tubes lower the plate Z. 4x1's get pretty hi @ 6kv but you're dividing it by 4. The high total plate capacitance would become a factor on the higher bands... but on 160 and 80 it's gotta be one impressive amp. How stiff is your ac supply? hehe

[w4bfs]

Hi Tom and others ... this has been an intresting thread, esp about op an am tx with 'leenyar' amp that is only partially linear ... you mentioned experimenting with a mix of controlled carrier to bal mod to high level plate mod rigs ... I have (to me) an new additional wrinkle .. I have come up with a way to vary the amount of controlled carrier from none to "some" in a plate modulated tx .. some tbd .. I am incorporating it in my current construction project ... I hope to have it ready to test soon ...73 ...John

[K1JJ]

OK Tom. Do you need the Qstreet Bill Orr article too?

Have you considered balancing the bias/idling current between the tubes? Granted the tubes are rugged, especially the 4pr1000's... did it with six 6lq6's once, it just seemed the thing to do.

Another advantage to your amp is the multipile tubes lower the plate Z. 4x1's get pretty hi @ 6kv but you're dividing it by 4. The high total plate capacitance would become a factor on the higher bands... but on 160 and 80 it's gotta be one impressive amp. How stiff is your ac supply? hehe

Hi Fred,

Yes, the Bill Orr article about ways to control negative feedback sounds like something I may want to do -  If it's not too much trouble... TNX.

On balancing the tubes - I wud have done that but I'm using ONE fil transformer to feed all four tubes. The bias goes in the fil CT, thus no individual biasing. I don't want to lift the nice thick ground straps to the screen/grids and play with external supplies either.  BUT, three of the 4X1's show nearly the same plate color during idle and full power, so hopefully the forth tube I get to replace the soft one will match and balance with the others reasonably.

I found with 6KV it was too "animal" for me.  3,500v under load gives a beautiful dull orange plate when running full carrier with moderate air. I think I have the plate impedance matched right too at that voltage cuz of the good efficiency. I think I found the sweet spot of the lashup, including the MRF-150 driver amp.

John:  I'd be curious how you will be swinging the controlled carrier. Is it a grid driven amplifier with control of the screen and/or grid voltage you're accomplishing this with?  Might be a little harder with straight GG unless it is done in the low level stage, like at the FT-102 driver. Now that wud be fun to experiment with I think.

The only thing with controlled carrier is the annoying up and down sound of hedrodynes when using it.... :-)  Everything is a compromise or I guess we'd all be using the same thang eventually.

T

[WA1GFZ]

Tom,
I had a few emails with Rick Measures on NFB in GG amps and he told me there was problems with the Orr method.
At 6 KV the tank circuit would need more L running at higher z.

[WB2YGF]

I hear you in there using Quadzilla on 3885.  Nice signal

[kb3ouk]

Heard you, Steve-HUZ, and Don at 20 over here around nine thirty. Steve-QIX was 40 over,. Quadzilla sounds really good.
Shelby KB3OUK

[K1JJ]

I here you in there using Quadzilla on 3885.  Nice signal

John and Shelby:

Thanks much for the nice reports, OM!  These were the first contacts made with the new amp.
I was beaming west with the high Yagi all night, so glad you were hearing me well locally.

The new amp ran flawlessly all night. Real pleased.  The only problem so far is one of the vacuum variable shafts froze up and needed a new shaft. (Don't be talking 'bout Shaft)  I found that at 3500V, Quadzilla is about equivalent to my 4X1 X 833A plate modulated rig, (Fabio)  but easier to use...and it's prettier to look at... :-) 


Big group in there tonight on 3885 and still going on. Maybe 20-25 stations including Calif and Wash state.   Just love to hear that AM DX come in.


Frank:
Yes, I did add more tank inductance when I tried it at 6KV+ to compensate for the higher impedance requirement.  Still, the plate dissipation was too high on AM and I didn't like being on the hairy edge of destruction... :-)   3500V under load is a nice, controllable level with dull orange plates for OB xmissions. I increased the air flow with the Variac and see very little blower noise on the scope when using it on the air, so all set. Quadzilla is basically finished for now. The next challenge is making an effective input power limit snubber for the MosFets. I will be trying your design.

Later -

T

[WA1GFZ]

Tom,
I was listening between 10:00 and 10:30. Conditions were great. Now you know why I like running afterburner mode. The 4CX3000A is almose 4, 4-1000As. at 4000V I can do the same thing. KISS. now that i have a bit more floor space showing it might be time to put the linear on the bench and do a little work on it when I not playing with the 4 other projects I have going on.   

[W1ATR]

Sweet Jebus, that first pic of the closeup with fils lit is beautiful. Hope you don't mind if I shag it for a wallpaper on the pc.

Regarding the post where you melted the slobber out of the pins on another amp, or TX, whichever, how about a pressure sw setup to babysit the blower. I never had any faith in sail switches, thermal devices in the outlet air are way to slow to prevent nuking the tube base, but a pressure sw cutting the control voltage will kill the juice if the blower starts to slow down, or air pressure gets lost in any way. On my old 2x4-400 leenyar, I had a pressure switch series wired into the coil holding voltage for the mains contactor. It was setup with a bleed off tee to keep it accurate and it knocks out the entire unit if the box pressure drops below 12"WC.

Just a thought.

If your interested in going past just the thought, all I would need to know is what the minimum box pressure needs to be for your super rig, measured with a manometer, and I could look up the right p-switch to do the job.

73

[w4bfs]

... I have (to me) an new additional wrinkle .. I have come up with a way to vary the amount of controlled carrier from none to "some" in a plate modulated tx .. some tbd .. I am incorporating it in my current construction project ... I hope to have it ready to test soon ...73 ...John

John:  I'd be curious how you will be swinging the controlled carrier. Is it a grid driven amplifier with control of the screen and/or grid voltage you're accomplishing this with?  Might be a little harder with straight GG unless it is done in the low level stage, like at the FT-102 driver. Now that wud be fun to experiment with I think.

The only thing with controlled carrier is the annoying up and down sound of hedrodynes when using it.... :-)  Everything is a compromise or I guess we'd all be using the same thang eventually.

T

Tom, as far as I know this may be new ground ... I don't have access to the old magazines so it is possible that this is a "revisit" of an old idea and is relatively simple

I decided a while back that having control of the amount of controlled carrier would be a good thing .. I think 3 db would be unnoticed, 6 db would be noticed by some, etc.  The reason for going after this is when following the plate mod low power tx with an amplifier, controlling the carrier  level when unmodulated reduces heat production (3db = 1/2, 6db = 3/4 )

this may wind up being a compound scheme for increased + direction modulation index

let me see how this works and if promising I will release to the public domain

idea first disclosed to Jack, WB8BFS January 2009

73 John WB4BFS

[K1JJ]

Sweet Jebus, that first pic of the closeup with fils lit is beautiful. Hope you don't mind if I shag it for a wallpaper on the pc.

heheheh - Glad you like that pic, Frank!  It's my favorite of the whole batch too. I reposted it below.  It would make a good screen saver for sure.

OK on the air sensor switch. That may be a good way to go.  I'll give it all some thought.


John:  Yes, for controlled carrier operation, 3db would be a transparent amount of carrier swing and also take a huge load off the steady state carrier by using 1/2 power. I like it. Think I'll talk with the HUZman and Bill/DUQ  about modifying the NE-602 circuit so that it can be goosed with a DC offset controlled by the audio. I suppose a certain audio threshold cud trigger a 3db carrier increase. It sounds simple enuff in theory to me....   This wud help boost integrated carrier efficiency and tube color reduction, markedly.


Frank:  Yep, your GG 4CX-3000 is probably the simpest way to duplicate the 4X1 amp. For viewing, you cud always mount a single 833A and use a few miirors looking at each other to make it look like a dozen.    I tried a large, long mirror behind the 4X1's the other day and it was stunning. You cud not tell the difference - there were eight in there... :-)

T

[Steve - WB3HUZ]

Sounded great last night Tom. That inline four is the way to go. It works well and looks good doing it. You were at least 10dB louder than anyone else and the audio sounded great. Good conditions too. The west coast guys were rolling in here at 10-20 over 9.

I checked out at about 1:30 AM and there were several still in there.