NEW Project - 6146B Plate modulated by Hammond 1628SEA and 85W SS Amp

[KD1SH]

I love the idea of a tube window, but it's way more fun with tubes whose plates glow.

[W3SLK]

I don't think I can recall when a 6146 has been quite 'show-cased' like that!

[KD1SH]

For "showcasing" tubes that don't glow, I prefer the 807, 'cause it's got that classic shape and buzzardly look, like my wife (the classic shape part, not the buzzardly part).

[K1JJ]

Ya know, I was thinking the same thing.  It's still not too late to switch to another tube type. No chassis drilling yet.  The 6146B will certainly do the job, but it is so common and plain. No color in the plate unless something is wrong.

In contrast, I once built a single 3C24 /  24G  old buzzard triode rig from the 1930's.  It's a tiny tube only 4" tall and 1 3/8" diameter. About the size of a 2E26.  My rig was one modulated by a pair. It had 2KV on it or 4KV+ peak under modulation. Insane, but within extreme pissbeat ratings. The tube plate is rated at only 25W of dissipation but a single one will put out 47W, 60W or 90W depending on the voltage.

I could easily do 1250V with the supply I have.  It would need to be neutralized, of course.

The best part is how that plate glows under normal power... just like a 304TL.  The plate is so tiny but the efficiency is marvelous.

I could even put in a pair -  for a conservative 60 watts out, easily exceeding a single 6146B but more fun to watch.  Plus a better match for the 5K Hammond xfmr.   I'm not kidding... the filaments  light up like a floodlight.  The modulators and final plates of those 3C24s / 24Gs glowed like a hot electric stove burner, and that was normal color. It was doing near 88% final plate efficiency.

They are very fragile tubes and out of 15 I have, 10 have bad gas (purple glow) and 5 are still FB as far as I know from using them 10 years ago.

It IS a possibility and gets me more excited than a 6146B..  [yawn]



Check out these specs and typical operating parameters:  (check out page 3, plate modulated service)

https://frank.pocnet.net/sheets/201/3/3C28.pdf


(The pics below are taken off the web -  not my rig)

[KD1SH]

Now that is a nice glow. Looks like a couple of Coleman mantles in there.

[w9jsw]

4-65?

[K1JJ]

After doing the calculations on the 3C24/24G,  6LF6 sweep tube and  4-65...   other than the lack of plate glow, I'm disappointed to find the 6146B is still the best fit for the job.  I won't get into details, but that old 6146B is hard to beat.   

Also, the 6LF6 sweep tube (I have four) has huge emission and peak current capability for big audio peaks.

Still thinking about it.

T

[N4LTA]

I have a couple of RS 1003. Specs for class c telephony are as follows ( as well as I can interpret German)

800 volts at 130 mA plate
300 volts at 25 mA screen
Grid bias  at -90 volts 14 mA
Plate dissipation 34 Watts

power out at 100 % modulation 70 watts

Tube max anode dissipation 60 Watts.

Is the 1628SEA up to modulating this tube or should I use the larger 1642SEA. I have a 1642 and a 1628 coming.


Pat
N4LTA

[Opcom]

a small disgression on your sigline:

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

Every time I see this I want to dive into LTspice..

[K1JJ]

Hi Pat,

I would say the 1628SEA should have no problem fully modulating the RS 1003, being a 30 W dissipation tube.  Your simple RF final plate impedance is about 6K. That is pretty close to the xfmr 5K.

From my own experience with Summer Breeze, the 4D32 rig; the 1642SE modulated a 150W carrier quite nicely. So interpolating down to the smaller xfmr with your smaller tube makes sense.


I've pretty well decided on sticking with my original plans of using a 6146B with a variable supply that will go up to 850V easily.  Frank / GFZ runs his four 6146Bs at 800VDC under load and is getting out 250W carrier. This is about 62W each. He has big positive peaks to boot.  He uses four sweep tubes in P-P parallel as modulators, so has plenty of mod power.

T

[K1JJ]

a small disgression on your sigline:

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

Every time I see this I want to dive into LTspice..

Patrick,

I solved the audio filter problem by picking up a DSP  DBX  DriveRack unit.  It does brick wall audio filtering in any bandwidth desired as well as EQ, limiting, etc.  About $100 used. Such a sweet box and what a deal.

T

[K1JJ]

This is the fun part - visualizing where everything goes and what parts are missing.

The big beer can caps will be mounted underneath, as well as the input L/C  matching network lying in front of the front panel.  The switches, fuses and lights will be on the front lower panel also mounted underneath.  Except for the meter wires, there will be very little wiring on top, except for the RF tank.  Notice the RF tank is extreme overkill. Who knows... someday I may replace the 6146B with a pair of 4D32s using two Hammond 1648SE transformers.  

I like the 6146B choice more and more. It will be like a Ranger 3 but built to last and a little on steroids. I heard someone say an old, unrestored Ranger went for $500 on ebay today, so I feel better about the whole thing... :-)

The 85W SS audio module will be mounted underneath with a heatsink on top with some air from the muffin fan. I'm hoping to get a slight breeze on the 6146B somehow. The muffin fan will use one of the small Variacs, so will be inaudible. The bigger Variac will control the HV supply, up to at least 850VDC under a heavy load. The three beer can caps will work out to be 833 uF at 1050 volts in series.  The HV transformer will be in full-wave CT config and is rated at 300mA. The 6146B will be pulling about 140 mA, so all is conservatively rated.

Notice the transformers in a line. There is a LOT of weight there. I will add a reinforcing angle rail underneath to firm up the chassis under the transformers. The transformers from right to left: (back view)  Hammond 1628SEA, 6.3VAC 6146B filament, SS Audio module transformer and the +- 800V CT HV xfmr for the 6146B.  There is also a fixed bias transformer to be mounted underneath.

T

[K8DI]

DDS rf source, and 6146 grid…. What is the plan for in between?  The grid, per specs, needs a half watt, at a decent voltage. Of course the DDS is well below on both counts, so how will you get the rf power and drive voltage?

Ed

[K1JJ]

DDS rf source, and 6146 grid…. What is the plan for in between?  The grid, per specs, needs a half watt, at a decent voltage. Of course the DDS is well below on both counts, so how will you get the rf power and drive voltage?
Ed

Good question  Ed....  shows you understand the lashup.

Background:  I have a DDS chain that presently drives the 4X1 plate modulated rig and the 813s X 813s.  The DDS drives a simple 11N90 MOSFET using an input 50 ohm unun toroid and an output toroid at about 24V VCC. (I had the 11N90s lying around)  That puts out a few watts, maybe 5W, IIRC.  I then go into  a quad MRF-150 kit amplifier that gives me an easy few hundred watts drive available for any big rig in the shack.

So for this 6146B project, I will simply duplicate the 11N90 circuit (and DDS) and include a drive level pot to drive the 6146B gird thru its 50 ohm L/C  input circuit.  I will also have a switch to select and drive the 6146B with any external driver in the shack directly into its L/C network without the DDS or IPA involved.  (to use the FT-1000D driver, SDR driver, old DDS, etc.)

T

[N4LTA]

Tom

I think I will build a rig with the RS 1003 similar to this one. I spent a little time gutting the old amp from the fire and ended up with some good salvage. Will build a separate
power supply with the salvaged parts. I have a Hammond 800V center tapped power transformer - 278CX and a 300 mA 10 H choke for the plate supply, a 15H 125 mA choke for the regulated screen circuit and I'll use a small 125 volt transformer for the bias supply. I'll regulate the screen with a couple of VR150/0D3s. I salvaged three 270uF 600 volt snap in caps that I'll use as filters for the plate supply(2) and the other for the screen supply. Also salvaged a plate choke that should work. I'll use an "economy" power circuit for the 800 volt plate supply and 400 volts for the supply to the VR tubes for the screen.

Hope to get started on the power supply tomorrow.

I have attached a few photos of the old amp chassis and the salvaged parts


Pat


.

[K1JJ]

Pat,  

Cool on your plans to build one of these rigs.  Keep the updates coming!

**  I was talking with Rick / W8KHK.  He always has good advice regarding my rigs. He suggested that I use TWO 6146Bs to have more emission and lower impedance to get better linearity... and not have to beat on one tube.

In addition I could make a rectangular hole for the tubes and install an aluminum plate that could hold the sockets  -  and the plate could be switched out for some other tubes in the future, if desired.

I then thought, hey, I could always unplug one tube and run a single 6146B as planned if I wanted to experiment and try this.  So here we are. I have TWO 6146Bs standing side by side. All of the metering and supplies will be basically the same and the rig can now do from 20W (lightly loaded one tube) to as much as 100W carrier out with two. Or 40 watts (two tubes)  just loafing as a clean driver for the 4X1 linear amp. I like it! This is a good compromise.

T

I still need to clean up the rough edges with a file and get a window bezel mounted:

 

[K1JJ]

BTW, when she is done,  I want the new Commander (Ranger 3)  to basically look like this.  I still plan on a two-tone panel paint job and lettering similar to a Ranger. (shown is 813s series modulated rig)

T

[K8DI]

In addition I could make a rectangular hole for the tubes and install an aluminum plate that could hold the sockets  -  and the plate could be switched out for some other tubes in the future, if desired.

T

This idea is my plan for the next RF deck I build. I had been amassing parts, but put it aside when I happened on the broadcast rig I’ve been working on.  The one additional thing I planned is to make the socket mounting plate mount from underneath, so it can be below the chassis on spacers. Some tube types need to poke thru a hole for shielding, and with a fan mounted underneath, I could also ensure tube base cooling. To control airflow and improve shielding a second plate with holes fitting the tube envelopes could be top mounted to close off most of the big rectangular hole.

Ed

[K1JJ]

Hi Ed,

Yep, the tube mounting plate and adding another one on top to control the air flow and RF shielding is interesting. I will give that some thought.

This may be common knowledge to some, but Rick brought up another interesting point that I've never thought of -  the height of the chassis (in my case 6" high) has a big effect on how strong the chassis is to flexing/ collapse.  I was concerned that the four heavy transformers in the rear needed rail reinforcement underneath.  But as it turns out, a 6" high chassis is VERY strong!  Compare this with a chassis that is only 1/4" high. It is almost a sheet of flat metal and would bend easily. But a chassis that is 10" high would be like a rock and could handle big weight due to more metal cross-sectional area.  Like bending a 4" pipe compared to a 1/8" tube.  The chassis is still strongest at the ends and weakest in the middle of the span, of course, so put the biggest transformers closest to the ends and outer edges.

I noticed this effect by picking up my 6146B rig with the transformers sitting on the back. With the  existing gusset side rails, it could be supported by the front panel if needed, though most of my rigs sit on a table or rack supporting plate.  This is something I never thought about before -  regarding overall strength based on chassis height.


BTW, my choosing this 6" chassis was not intentional. It was just the best deal I could find on eBay for a Bud chassis. I was looking for a 3-4" but lucked out here.

T


Clear example:  Imagine pushing down on the rear of this 6" high  chassis trying to collapse it.  (even with these shorty side gussets)   I tried standing on the chassis and it was solid.

[N4LTA]

Tom

Are you planning to use a similar input circuit as you used in the 4D32.  I am looking to do something similar with the R 1003.

Anything that I should keep in mind?

Also - is that 41H of  total reactance in the screen lead?

My screen impedance is about 12K ohms  (300/.025).  Hammond makes a
157G rated 30 H at 40 mA. and I have a 9H 50mA rated choke. In series I should get 39 H which will take me down to 50 Hz. This should be enough, I think?

Thanks

Pat
N4LTA

[K1JJ]

Hi Pat,

I have been using that circuit, an L - C input to the high impedance grid for class C AM rigs.  For linear amps, I use a C-L-C reverse pi-network to go into the 50 ohm cathode of the usual GG amp.

The L-C is very simple and works well once the input coil is optimized for each band via a small rotary switch.

BTW, I did up a new schematic for the (2) 6146Bs that includes the entire rig and supplies, etc.  I want to check it a few more times and then post it so I can get some more critical eyes on it before I build.  You can use it as a guide for your  R 1003 rig.  

I have about 50% of the metal work finished. It's all coming together FB and will now be in the class of a DX-100 or Apache with two tubes.  I still plan to run it "Ranger 3" class -  around 50-60 watts for cleanliness.  

I'd like to make a viewing window brass bezel, similar to what Chuck did on his 4-400A rig.  That is a cool looking add-on.

Oh, I see you added another question....  good you asked.  I do NOT plan to use a screen choke on this build. A screen dropping resistor from modulated B+ to screen is more linear and not prone to audio series resonances like a iron core choke.  On BIG rigs the power loss is too much, but even an 813 rig can do well with a screen dropping resistor.   You will eliminate the screen supply and choke.  You can also use some fixed grid bias to limit final current when drive is gone. Plus the screen will never be in danger because it will be drawing current only when the plate voltage is on. The resistor will be flat from 1 Hz to  15 KHz, no problem.

I have gotten MANY rigs to perform flawlessly using a screen supply and choke, but recently a friend pointed out that the dropping resistor may be a better way to go with these smaller rigs.

T

[N4LTA]

Tom,


I look forward to building this rig along with your build.  Getting rid of the screen supply and the regulator tubes will save a lot of space. I am trying to put all of this in a 12 x 12 chassis.

I have a power supply PC board that I have used in the past. It uses snap in caps as will save quite a bit of room. I plan to use two 270 uF 600 volt caps and a 10H choke to keep the voltage under control. I'll use a 800 Volt  transformer (Hammond 278X) for the plate supply.

Microchip makes a simple low cost high voltage regulator in a T092 package. With a high voltage transistor, it make a nice variable regulator for up to 400 volts or so. I have some PC boards for it that I used for the  power supplies for my transmitter in the AWA Bruce Kelly. Should work fine for a grid bias supply. I may try it. The tube data shows -90 volts for AM operation of the R1003  with 800 volts on the anode and 300 volts on the screen.

Looking forward to see your preliminary schematic.


Pat

[K1JJ]

Tom,


I look forward to building this rig along with your build.  Getting rid of the screen supply and the regulator tubes will save a lot of space. I am trying to put all of this in a 12 x 12 chassis.

I have a power supply PC board that I have used in the past. It uses snap in caps as will save quite a bit of room. I plan to use two 270 uF 600 volt caps and a 10H choke to keep the voltage under control. I'll use a 800 Volt  transformer (Hammond 278X) for the plate supply.

Microchip makes a simple low cost high voltage regulator in a T092 package. With a high voltage transistor, it make a nice variable regulator for up to 400 volts or so. I have some PC boards for it that I used for the  power supplies for my transmitter in the AWA Bruce Kelly. Should work fine for a grid bias supply. I may try it. The tube data shows -90 volts for AM operation of the R1003  with 800 volts on the anode and 300 volts on the screen.

Looking forward to see your preliminary schematic.

Pat

Hi Pat -

Yep, if you build the rig it should optimize out well. Frank/GFZ has been sharing his parameters. He is getting 85% plate efficiency and 250 watts out with 2KV peak voltage on the plates with four tubes.  He is using a very large relative value for the screen dropping resistor, which I want to experiment with.  Very FB running rig.

12" X 12" for your rig may be a little small, depending on if that includes the power supplies and all iron.  

Maybe I'm reading your plans wrong, but you should not need any regulated supplies... there is no fixed screen voltage, the plate is stiff and if you bleed the fixed grid bias heavily with a bleeder resistor, it will not move around on you.     The only possible problem is the grid fixed supply and I wud tackle it only if it becomes unstable by charging up from RF rectification or whatever.

You may not need the 10H choke either, considering that's a good amount of capacitance you have in the HV supply.

I'll get that schematic posted probably tmw once I  stop making changes... :-)


** I ordered some brass bar stock and brass thumb-wheel screws from McMaster-Carr today. Gonna make up some brass bezel viewing windows  ala K1KW's 813 rig.  That should spruce up the new girl... :-)

T

Chuck's 813 /572B rig showing brass bezel:

[N4LTA]

Tom

Thanks for the info. I am slimming things down.

I was thinking originally using two VR150s to regulate the screen. I can take that out now.

I was thinking about and using a regulated grid bias supply. It is probably not worth the problems.

The choke was to keep the plate supply from rising at light load. I do not plan to make it variable with a variac. I had a bunch of small ones good for 3  amps or so, but lost them in the fire.

It is overkill, but I like to make the plate supply stiff using a choke and loading it to the critical value. I can drop the choke I think.

With the choke, I'll need to keep 80 mA load with a bleeder resistor to keep the voltage at 800 volts DC with a 10H choke. That is a lot of wasted power and space. Probably not a good idea to dump 64 watts of heat into the chassis.

I have two 600 volt filter  caps in series for the plate filter, so I have some room to find a suitable bleeder with a capacitor input filter that keeps the maximum voltage below 1200 volts

With no bleeder, the 800 volt AC will charge the cap to about 1131 volts which is safely below the rating of the two filter caps.  The voltage will drop to near 800 volts when keyed, When in idle or unkeyed, the higher voltage should not be a problem I think?

So it looks like I can reduce two VR tubes, a large heavy 10H choke, the screen power supply filters and the screen power supply choke also the screen audio choke.. I may get it in a 12 x 12 x 4 chassis yet?

Pat

[K1JJ]

OK Pat -

Normally when using plate modulation and a screen supply with a choke, the DC screen current during transmit is constant, so no need for a regulator. But since you will be using a dropping resistor from the modulated B+, the regulator is still not needed.   Screen regulation is mostly used in linear operation because the screen current varies widely and may cause distortion if not regulated.

A solution for your HV soaring when the rig is unkeyed can be fixed with a step start in the 120VAC line. Refer to the schematic when I post it. Use a ~20 ohm resistor in the 120VAC line that gets shorted out after a fraction of a second. Also, a Variac for the HV is very important for testing and later having control of the overall rig power, etc..  I would highly recommend a 3A or larger Variac.  You can always turn it all the way up to "bypass" its slight loss later on.

A 1200V capacitor voltage rating is also too narrow when soaring to  1131 V DC in the HV supply.  The step start and Variac both can help here.  850V max is OK if the caps are healthy with equalizing resistors that balance the load well.

Maybe you can stuff a lot underneath the 4" X 12" X 12" chassis,  but it will be a battle servicing and living with later.  Don't forget the LV transformer for the 85 watt audio module, filament xfmr, heatsink, etc.  More room is always appreciated.  As is, my own chassis of 17" X 15" X 6" is starting to fill up.  But we always end up getting it all in no matter what the size, huh?   :-)

Keep the HV bleeder current as low as possible. Wasted heat. Used mainly for safety. The AM carrier helps a lot to pull steady current anyway. Acts like a big bleeder when in transmit. But the step start is the ultimate in safety, power economy and preventing the HV from soaring.  IE, the 20 ohm step start resistor limits the secondary voltage long enuff until the carrier draw takes over.  

These pre-build discussions are very valuable to make less mistakes later on. I've already corrected a few things in my schematic that I spaced out...

T