• WARNING: Tube/Valve amplifiers use potentially LETHAL HIGH VOLTAGES.
    Building, troubleshooting and testing of these amplifiers should only be
    performed by someone who is thoroughly familiar with
    the safety precautions around high voltages.


I am new at designing amplifiers so this high school kid is building a high power audio amp for me. He is planning on using a 3cx3000 tube for the final stage. He is trying to get a used one from the local radio station. He is going to use the LM3886 amplifier as a driver. He figures on getting 500 - 1000 watts if he uses a 2000 volt transformer.

I was wondering where is the best place to buy the tube if the station won't pan out? He said many of the tubes out of stations have lots of life left because they are supposed to put out more than 3000 watts.

Thanks for helping a newbie.

Hi Frank,

Your designer seems to be out of his depth on this one - There is no way that an LM3886 could swing enough voltage to make any sort of attempt at driving a 3cx3000. I guess he has looked at audio power and current output rather than voltage swing.

I don't know of anyone who has successfully built an audio amp with one of these tubes. The 3cx3000 series are designed to produce kilowatts of rf power for transmitter or rf heating use. Trying to get one stable enough for audio would be no easy task.

Other considerations would be cooling (designed for forced air cooling) and power supplies (huge transformers required) also, what would you use for an output transformer?

Biggest problem is safety - 2000v is lethal stuff and requires expert knowledge of materials and layout which from his other design ideas, I suspect your designer does not have.

If I were you, I would call a halt to all this before he fries himself :hot:

Start with something simple - there are loads of ideas here on diy audio and elsewhere on the web - make sure you understand how a circuit works before trying to build it.

Good luck
"The RF amp runs class C but it is class A for audio amp. I really don't think the tube can run in class A operation at all."

They can run in class A. The Continental 317-C transmitter uses them to modulate the screens of a pair of 4CX35000C tubes.
Continental runs them in class A.

Cooling them is another matter. They require lots of air flow.
3CX3000 amp

I am all for building monster audio amps, BUT this is too far over the edge, even for me. Assuming that you had an expert, who could deal with the required cooling and serious voltage, you could never get an output transformer made for this amp, that could cover the entire audio range. If you are still considering this amp, start searching for the output transformer first. Consult with a few transformer vendors to see if anyone could build you a transformer. Assuming a single ended amplifier, producing 500 to 1000 watts would require a tube dissipation of 1500 to 3000 watts. Using your figure of 2000 volts, gives an idle current of about 1 amp. The insulation must stand about 4000 volts. This is easilly a 50 pound 1000 dollar piece of iron! If you can get the output transformer made, then you need power and filament transformers. Hey those should be easy.

You are definitely not going to drive this tube directly with a chip amp. You might drive it with a chip and a driver transformer, but now you need another custom transformer. I would look at my PowerDrive circuit and then design a bigger one, this tube is going to suck some serious grid current.

As far as the comment that this tube can only be run class C, don't believe it. I have never met a tube that couldn't be run in class A. Many of my bigger amps were made with RF tubes. Most of them were never speced for class A operation. The 833A and 811A imediately come to mind.

Keep in mind that this amp (with the required transformers and cooling) would occupy a small room and require its own air conditioner.

If this is where you want to start out, whats next? 1.21 Gigawatts? Seriously, build a smaller amp first!
I thank you for your help. I might have given bad information, he was planning on using several LM3886 because he said you can get lots of power by using several and he was planning on using an torroid transformer to get the drive voltage up high.

The cooling of the tube was something his dad was going to do for him since his dad has a machine shop at home and he works for a company that provides cooling fans for RF equipment.

The output transformer is the biggest concern for him but his uncle was the head design engineer for a transformer company called Triad (or something like that name) and he talked about a power core double E liminate winding the primary on double bobbin design. The primary winding wire is 18 guage and the secondary was a multii-cable type design. I think he said something about square wire or something like that for the output. The whole transformer was going to be vacuum varnish impregnated and the winding method took into account the capacitance of the winding.

The weight of the transformer is of some concern but the kid feels the results will be something worth the hassle.

I want to thank you all for your advice. Does anybody know where I can get the tubes?



Paid Member
2003-06-12 7:04 pm
Maine USA
I think the plan is insane.

Why build a 500 watt SET? out of a tube not noted for audio linearity? With apparently a super-high Mu, suggesting poor damping and real trouble in the OPT winding?

The only thing I like is getting the former Triad guy on board.

> where is the best place to buy the tube

www.pentalabs.com about $900 each.

Beware the many suffixes: they may be cooling method, they may be Mu, or who-knows-what?
Extremely large tube amplifier

If you are still serious about building the biggest tube amplifier on the planet, I would seriously think about what you are considering. Please read the electrical SAFETY information this forum, and on my web site before proceeding. If you do not understand EVERYTHING, DO NOT PROCEED, you really could be KILLED. Understand that any amp this big WILL cost SERIOUS money, and generate a LOT of heat.

You seem to be fixated on using the 3CX3000 tube in a single ended configuration, with a chip amp driver. None of this has been done berore. Your chances of success are virtually zero and the risk (and cost) is extremely high. I have been building tube amplifiers for about 40 years. I don't believe that the output transformer, or the chip amp driver could be successfully built.

I have built a single ended amplifier that puts out about 200 watts. The picture and info is on my web site. It uses an 833A tube, and the output transformer for that amp is a challenge. Getting flat frequency response at the 200 watt level from a single ended transformer is very hard, even for an expert transformer designer. A single ended amp of this, or higher power is definitely NOT a beginner project.

The chip amp driver is another problem. The low impedance (4 to 8 ohms) output of the chip, must be matched up to the high impedance input of the tube's grid. You will need a transformer to do this. The problem is that the grid impedance is not constant with drive power or frequency. This will DRIVE A CHIP AMP NUTS. The best case will be poor frequency response, the likely case will be blown chips.

You are looking for a source of tubes, which implies that you don't already have them. Why then do you want to use the 3CX3000? At $900 each there are plenty of cheaper choices that are much better suited for audio. The ceramic triodes are easy to damage, and at this price, I would be seriously upset if I fried one. If you could beg a used one from a radio station, sell it and use the money for the parts that you are giong to need. Tubes like this often have a core value, so getting a free, or cheap one is unlikely.

If you are determined to build a big single ended amplifier, consider using 2 845 tubes in parallel for each channel. The power output would be about 80 watts per channel with an 1100 volt power supply and 180 to 200 mA of current. The output transformer (5000 ohms) is off the shelf. New Chinese tubes can be found for $30 to $40 each.
For maximum power an 833A tube could give you about 200 watts if pushed hard. Power supply is 1500 to 2000 volts at 250 to 300 mA. The output transformer is a challenge due to the high voltage and current. New Chinese 833A's are about $100 each, and good used ones made by American or European vendors are about the same price on Ebay. I have built both prototypes of these amps using the 845SE schematic that is on my web site. They are both buildable designs, although I would NOT consider either for a first amp due to the LETHAL voltages involved.

If you are convinced that you just must build a kilowatt tube amp, then the only path to success that I can see is a push pull amp. A pair of 833A tubes can provide 1500 to 3000 watts in class B. You will get less power in class AB. Since the DC is cancelled out, this transformer can be built, although it would be a custom design. A power supply of 2500 to 4000 volts would be required. Push pull 833A's were commonly used for modulators in radio stations in the 1940's.

If you do not understand everything that that I have said above, find someone who does before even thinking about building anything like this. At the least you could toast a lot of expensive parts. Worst case you could toast yourself, and you can't buy a new one of those. This is serious, 2000 to 4000 volts at these current levels is INSTANTLY LETHAL on contact. You don't get a second chance.


2004-06-21 8:04 am
big smoke
As crazy as it sounds, the 1953 Eimac catalogue has these operating recommendations for the similar 3X3000A1. For a B+ of 3000 volts, -600 on the grid, biased at 665 ma, an effective load plate-to-plate of 1170 ohms, 555 volts peak grid input per tube with zero (!!!!!) input power, two tubes push-pull will suck 10 kwatts into the plates and return 4000 watts at 2.7% THD at the speaker terminals. As Eimac cautions, "before feedback". If you're comfortable 4000 volts of B+ will put you in front of Your Personal Maker just as surely as 3000 these specs can be greatly improved: 6000 watts out @ 1.8% THD. They really were the Greatest Generation.

The plate curves look like hell's 6c45. Incidentally, 6000 volts B+ nets 10,000 watts out, but that's getting just silly.
If you are still serious about building the biggest tube amplifier on the planet, I would seriously think about what you are considering.

He isn't even close:

The huge, high-level modulator was also water-cooled. It could easily make 400 kW audio with both of its largely redundant modules simultaneously cranked to full rock and roll. Fortunately for the survival of civilization, this case was rare. Each module used four more of the biggest tubes made, bringing all of RCA 1's RF and AF output tube complement to twenty, with a total cost of $34,000 in 1930s' depression dollars. Even more mindboggling were the two modulation transformers, one per module, each 37,000 pounds, oil filled, and 10 feet high. It's possible that these two Westinghouse reactors were in fact the world's largest transformers for a brief time - "heavy iron" indeed!