A 200 Watt Linear Amplifier -- design thoughts
The Pic-a-Star transceiver I'm building will probably start out with a QRP final running no more than 5-10 watts from some cheap transistors. There are actually several final designs I'm thinking of based on parts I already have or are cheap to buy. I have some MRF476 and MRF475 transistors, a pair of the former is good for about 5-6 watts output, the latter ones about 20 Watts. Many hams building the Pic-a-Star have built a popular final that uses a pair of RD16HHF1 RF MOSTFETS driven by a pair of RD06HHF1's (though the final could be driven by a cheap pair of bipolar transistors as well). This amp is good for between 15-20 watts output. I also have a pair of MRF477's that are good for about 50 watts output, there is an old handbook design using these. Many hams have built amplifiers using IRF510 switching transistors. A pair of these can deliver over 50 watts output at 28 volts, but the IMD isn't low and the transistors are quite fragile at this power level. I'm could also use some junk box 2N2219's and get about 5 watts out of them. About the same as the MRF476 CB transistors.
The "standard" amateur transceiver usually runs between 100-200 watts output. The old pair of 6146's found in almost EVERY rig from the late 1950's through the early 1980's was in this category. By the 1970's color TV sweep tubes had become the norm, a real dumb move as the price on these bottles climbed and they were never as rugged as the 6146's. Power tetrode - Beam power tubes in the 20 - 30 watt class such as the 807 and 6146 families can run at about 200 watts input for a pair in the linear modes, producing about 100-120 watts output (same as their audio ratings). In grid driven mode only a watt or two of drive power is needed. (Class AB1 actually requires NO driving power, just voltage. Still the driver must consume some power.) Untuned, passive grid driven mode puts a low value resistor across the input which DOES eat up some power, but stabilizes the amplifier. Depending on the value of this resistor drive power can be just a watt or two, or ten watts or more. Grounded grid amplifiers using tubes such as the 811A or 572B require between 15-50 watts of drive (depending on how many of what kind of tubes are used). A single 811A will need about 15 watts of drive to deliver up to 200 watts output. A pair of 811A's or a single 572B will require about 30 watts of drive to produce up to 400-500 watts output.
Vacuum tube (valve) amplifiers are rarely considered for amateur transmitters these days, except at the highest power levels. 150-200 watts of power can be obtained from a pair of bipolar or power mosfets these days for a cost of under $200 for all the parts (less power supply). It can be done at 13.8 volts power, so the rig is good for mobile as well. Some amplifier designs run at 28 or 50 volts, these are a bit easier to design as far as matching circuits do to the higher output impedances. However all solid state rf power amplifier designs are broad band with fixed input and output impedances. They MUST be run into a 50 ohm antenna with a low swr present on the feed line. You will need an antenna tuner / matchbox to correct any deviation from this condition on your antenna. Vacuum tube amplifiers usually have tunable tank circuits that run at a limited bandwidth. They suppress unwanted harmonics to some degree and don't require as much external filtering as solid state circuits. Their tank circuits can usually handle some degree of miss match, usually as much as 2:1 or even 3:1 SWR ratios, and still deliver near their rated power output to the load. Tube amps are not for the faint of heart though. Power supply voltages of several hundred to several thousand volts are common, there is no such thing as a 'slight electrical shock' from these power supplies! While hams have safely built such equipment for over a half a century accidental electrocutions were not unknown. Common sense and awareness of the dangers keep one safe.
I was toying with the idea of building an amplifier using one or two 811A's. There is a nice toroid power transformer from an E-Bay seller with a 120/240 volt primary and a dual 550 volt secondaries. It also has two 6.3v 4A filament windings. This transformer was designed for audiophool tube amps, but would work nicely in the power supply of a small 811A linear. It's a 400VA (CCS 50-60hz) transformer, but could handle about 600VA in ICAS at 60hz so it's barely able to handle TWO 811A's, and will run cool with one. Perfect for the 811A linear design from the '67 handbook.
OTOH, I have a nice NOS TV transformer in the junk box. It's a 660VCT @ 330MA unit with three 6.3v and one 5v filament windings. In a bridge configuration with a capacitor input filter it should deliver about 850-900 volts under load. I also have a baker's dozen type 1625 tubes. Some NOS, some used pulls from old ARC-5 transmitters. They are a good match for the available power transformer. The 1625 is of course, an 807 tube with a 12.6 volt heater and a 7 pin medium base. The 807, is a 6L6G with a plate cap, internal RF shielding and a low loss 5 pin base. The 6L6 family are low gain screen types, IE: they can't be driven to more than 70 ma or so at their max rated screen voltage without driving grid #1 positive. This limits their class AB1 power output to about 55 watts or so in CCS (600 volts plate), about 73 watts in ICAS (750 volts plate) for a pair of tubes. In class AB2 with the grids driven positive they will deliver up to 120 watts ICAS (about the same as a pair of 6146's in AB1). The 6146 has a higher gain screen, it can be driven to at least 120 ma plate current in AB1 at lower screen voltages than the 6L6 family.
It is possible to run the 807/1625 tubes at higher plate and screen voltages. With up to 400 volts on the screens the tubes can be driven to 110 ma plate current in AB1 to deliver the same output that they deliver with lower screen voltage in class AB2. Plate and screen dissipation remains about the same. In the February 1963 issue of QST such an amplifier was described. It used a pair of type 807W's with 1100 volts on the plates and 400 volts on the screens to run 230 watts input. Claimed output power was 150 watts PEP. With the transformer I have the tubes wouldn't be pushed quite that far. I'm looking at about 850-900 volts plate under load with 366 volts on the screens (A pair of 0B2's and an 0A2 regulating the screen voltage.) I should be able to get 120 watts output from a pair of tubes, but I'm thinking of going for broke and put THREE of them in parallel to get 180 watts output. Three 807 class tubes in parallel is about the limit for a sane design. It should work fine on 20 meters and below. Power will start to drop off on 15, and 10 will be less than optimal. I'm thinking of using a pair of toroids in the tank circuit, a T200-2 for 80 and 40 with a T200-6 on 20-10 meters. The grid circuit will use a 1:9 balun on a type 43 ferrite core (1:3 turns ration, autotransformer). The grids will be loaded with TEN 4700 ohm 1/2 watt resistors in parallel. This will present the driver with a 52 ohm input impedance. (Try finding new 1 or 2 watt carbon resistors these days. Building one out of 1/2 watt units is the only affordable way).
Once upon a time 1625's cost about 25 CENTS each. Literally MILLIONS of them were made during WWII and the resulting surplus at the end of the war still hasn't been used up. Prices for these bottles at hamfests are usually about $5 - $10 NOS and $1-$5 used. Tested used tubes go for more, pigs in pokes for less. Ebay sellers often ask for the moon and stars and will have their offerings go unsold .... unless some audiophool snatches them up. If you are willing to buy in quantity you might be able to find a bunch of 1625's for under $5 each. The lot I have came from hamfests and a few junked ARC-5's and ART-13's.
The same "trick" can be pulled with 6146's. I wouldn't run the 6146 at more than 300 volts on the screens, 250 volts is more usual. These bottles will run stable in the passive grid mode and will run the same power as the 1625's. They won't be as cheap, but MIGHT be easier to find. 6883's or 6159's (12.6 volt or 26.5 volt heaters) can be found a bit cheaper, just wire up the required filament transformer windings in series. The 1625 / 807's should be sub-mounted so the bottom of the tube plate sits just a hair above the chassis. This adds a bit of shielding that keeps these tubes stable without neutralization. Maybe not needed in the passive grid mode, but with three bottles in parallel it's a good idea. The 6146's have a metal shield base that does the same thing when pin 8 of the tube is grounded. I've also got a bunch of 6146's so I could build a swappable sub-chassis to allow for changing out tubes.
The "standard" amateur transceiver usually runs between 100-200 watts output. The old pair of 6146's found in almost EVERY rig from the late 1950's through the early 1980's was in this category. By the 1970's color TV sweep tubes had become the norm, a real dumb move as the price on these bottles climbed and they were never as rugged as the 6146's. Power tetrode - Beam power tubes in the 20 - 30 watt class such as the 807 and 6146 families can run at about 200 watts input for a pair in the linear modes, producing about 100-120 watts output (same as their audio ratings). In grid driven mode only a watt or two of drive power is needed. (Class AB1 actually requires NO driving power, just voltage. Still the driver must consume some power.) Untuned, passive grid driven mode puts a low value resistor across the input which DOES eat up some power, but stabilizes the amplifier. Depending on the value of this resistor drive power can be just a watt or two, or ten watts or more. Grounded grid amplifiers using tubes such as the 811A or 572B require between 15-50 watts of drive (depending on how many of what kind of tubes are used). A single 811A will need about 15 watts of drive to deliver up to 200 watts output. A pair of 811A's or a single 572B will require about 30 watts of drive to produce up to 400-500 watts output.
Vacuum tube (valve) amplifiers are rarely considered for amateur transmitters these days, except at the highest power levels. 150-200 watts of power can be obtained from a pair of bipolar or power mosfets these days for a cost of under $200 for all the parts (less power supply). It can be done at 13.8 volts power, so the rig is good for mobile as well. Some amplifier designs run at 28 or 50 volts, these are a bit easier to design as far as matching circuits do to the higher output impedances. However all solid state rf power amplifier designs are broad band with fixed input and output impedances. They MUST be run into a 50 ohm antenna with a low swr present on the feed line. You will need an antenna tuner / matchbox to correct any deviation from this condition on your antenna. Vacuum tube amplifiers usually have tunable tank circuits that run at a limited bandwidth. They suppress unwanted harmonics to some degree and don't require as much external filtering as solid state circuits. Their tank circuits can usually handle some degree of miss match, usually as much as 2:1 or even 3:1 SWR ratios, and still deliver near their rated power output to the load. Tube amps are not for the faint of heart though. Power supply voltages of several hundred to several thousand volts are common, there is no such thing as a 'slight electrical shock' from these power supplies! While hams have safely built such equipment for over a half a century accidental electrocutions were not unknown. Common sense and awareness of the dangers keep one safe.
I was toying with the idea of building an amplifier using one or two 811A's. There is a nice toroid power transformer from an E-Bay seller with a 120/240 volt primary and a dual 550 volt secondaries. It also has two 6.3v 4A filament windings. This transformer was designed for audiophool tube amps, but would work nicely in the power supply of a small 811A linear. It's a 400VA (CCS 50-60hz) transformer, but could handle about 600VA in ICAS at 60hz so it's barely able to handle TWO 811A's, and will run cool with one. Perfect for the 811A linear design from the '67 handbook.
OTOH, I have a nice NOS TV transformer in the junk box. It's a 660VCT @ 330MA unit with three 6.3v and one 5v filament windings. In a bridge configuration with a capacitor input filter it should deliver about 850-900 volts under load. I also have a baker's dozen type 1625 tubes. Some NOS, some used pulls from old ARC-5 transmitters. They are a good match for the available power transformer. The 1625 is of course, an 807 tube with a 12.6 volt heater and a 7 pin medium base. The 807, is a 6L6G with a plate cap, internal RF shielding and a low loss 5 pin base. The 6L6 family are low gain screen types, IE: they can't be driven to more than 70 ma or so at their max rated screen voltage without driving grid #1 positive. This limits their class AB1 power output to about 55 watts or so in CCS (600 volts plate), about 73 watts in ICAS (750 volts plate) for a pair of tubes. In class AB2 with the grids driven positive they will deliver up to 120 watts ICAS (about the same as a pair of 6146's in AB1). The 6146 has a higher gain screen, it can be driven to at least 120 ma plate current in AB1 at lower screen voltages than the 6L6 family.
It is possible to run the 807/1625 tubes at higher plate and screen voltages. With up to 400 volts on the screens the tubes can be driven to 110 ma plate current in AB1 to deliver the same output that they deliver with lower screen voltage in class AB2. Plate and screen dissipation remains about the same. In the February 1963 issue of QST such an amplifier was described. It used a pair of type 807W's with 1100 volts on the plates and 400 volts on the screens to run 230 watts input. Claimed output power was 150 watts PEP. With the transformer I have the tubes wouldn't be pushed quite that far. I'm looking at about 850-900 volts plate under load with 366 volts on the screens (A pair of 0B2's and an 0A2 regulating the screen voltage.) I should be able to get 120 watts output from a pair of tubes, but I'm thinking of going for broke and put THREE of them in parallel to get 180 watts output. Three 807 class tubes in parallel is about the limit for a sane design. It should work fine on 20 meters and below. Power will start to drop off on 15, and 10 will be less than optimal. I'm thinking of using a pair of toroids in the tank circuit, a T200-2 for 80 and 40 with a T200-6 on 20-10 meters. The grid circuit will use a 1:9 balun on a type 43 ferrite core (1:3 turns ration, autotransformer). The grids will be loaded with TEN 4700 ohm 1/2 watt resistors in parallel. This will present the driver with a 52 ohm input impedance. (Try finding new 1 or 2 watt carbon resistors these days. Building one out of 1/2 watt units is the only affordable way).
Once upon a time 1625's cost about 25 CENTS each. Literally MILLIONS of them were made during WWII and the resulting surplus at the end of the war still hasn't been used up. Prices for these bottles at hamfests are usually about $5 - $10 NOS and $1-$5 used. Tested used tubes go for more, pigs in pokes for less. Ebay sellers often ask for the moon and stars and will have their offerings go unsold .... unless some audiophool snatches them up. If you are willing to buy in quantity you might be able to find a bunch of 1625's for under $5 each. The lot I have came from hamfests and a few junked ARC-5's and ART-13's.
The same "trick" can be pulled with 6146's. I wouldn't run the 6146 at more than 300 volts on the screens, 250 volts is more usual. These bottles will run stable in the passive grid mode and will run the same power as the 1625's. They won't be as cheap, but MIGHT be easier to find. 6883's or 6159's (12.6 volt or 26.5 volt heaters) can be found a bit cheaper, just wire up the required filament transformer windings in series. The 1625 / 807's should be sub-mounted so the bottom of the tube plate sits just a hair above the chassis. This adds a bit of shielding that keeps these tubes stable without neutralization. Maybe not needed in the passive grid mode, but with three bottles in parallel it's a good idea. The 6146's have a metal shield base that does the same thing when pin 8 of the tube is grounded. I've also got a bunch of 6146's so I could build a swappable sub-chassis to allow for changing out tubes.