50 Ohm 4000W dummy load

[steve_rb]

I need a 4000 W dummy load for RF test and calibration but they are very expensive. I was wondering how difficult is building one myself buy cutting a normal household radiator to 50 ohm? Will this work or need other considerations?

[Frank Sanns]

Dummy loads are rated at continuous dissipation. A 300 watt dummy load can take a KW for 30% duty cycle.

If you are just matching impedance to tune your amplifier this is one thing but what about your final load? How are you measuring its impedance?

What about keeping the RF where it needs to be. You can get RF burns well below 100 watts and as you approach a KW or more, this is frightening if you do not know what you are doing. Items that you touch or are close to you can become antennae and get you. You also have the worries of the radiation to yourself. Radio waves are not harmless and with an isotropic radiator at the frequency that you propose, the distance that you should be from the source is probably a hundred feet or more.

You also have the responsibility of not putting RF into the air. In most of the world, you need a license to put more than a couple hundred milliwatts into the air. If you think think the Federal Communications Comission will not know you are operating your machine outside of anechoic RF chamber, you are mistaken. You will be hunted down and fined if you survive your initial experiments. In many ways this is a more egregious infraction than running your own dental x-ray machine since the RF carries and can cause harmful interference potentiallly worldwide.

Just some words of caution.

Frank Sanns

[zhgreader]

of cause you can by parallel and series to get the power.
but rf 4000w is too large and easy radiation. and the load will has distribute parameters. is it 4w[4000mW] or 40w? not 4000w?

[Chris Bradley]

Do you need 50ohm for calibration purposes, or are you just after a load?

Also, how long are you proposing to operate at 4000W? The other responses apply, but there again if it is for a ms then 4J would likely be absorbed by a small 50W resistor. However, again as said, if you use a wound resistor then you will likely have uncontrolled mismatching going on.

So - is the test still appropriate? What are you aiming to prove?

[steve_rb]

It is 13.56 Mhz RF with 0-1270 W adjustable and it's manual says for system check and initial start up I need 4000W dummy load. I will not use more than 500w for my fusor but since I have purchased this unit second hand I want to test it's full functionality first before using it. Because I have purchased it in the condition of working status and if something is wrong I will take it back. This was the only cheap unit I found through a friend of mine purchased for me.
Purchased unit with a 150W dummy load. Measured load ohm and it shows 50 ohm. This means I can increase power up to 1270 and I will have no reflected power but beyond 150 watt load will start getting hot and I can't run at this power for a long time but. Load is Bird 8135 model and is oil cooled.

What will happen if I increase power to 1270 W into this 150W load for only about 10 second just see RF is fully functionl?

I know about regulations and that is why I am asking here to get all info before doing anything wrong.

[John Futter]

Steve
put a 100 meter roll of RG58 coax in your bath filled with water put your bird and dummy load on the far end.
make sure that the coax is submerged especially at the generator end of the run.

i've tested 5kW 88-108 transmitters this way until I built a load.

I've also used a load that doesn't get hot by using a quarter wave vertical and radiating the energy, side effects included all the fluoros in the garage coming on including the dead ones and a friend of mine hearing it 50kM away at a similar strength to the normal FM broadcast stations --best done at nite when the reg boys are in bed.

[Carl Willis]

This is expensive high-power gear and is not well-suited as introductory equipment for the RF neophyte. Prepare to encounter the smell of burnt human flesh--your own. I don't mention that to scare you off, just to point out that it is commonplace for unshielded RF power to be quite mobile and surprising in its habits, in a way that will make your fingertips a wee bit shorter and very much darker after your first encounters. People who start learning with little 50-watt radio transmitters only get a smattering of peppercorn-sized divots in their flesh on Day One. You have "kicked it up a notch," as Chef Emeril would say.

Do you have a manual for the ENI (and are you planning to follow it to the letter)? If not, you should get one from them. Human smoke is no biggie for me, but I do worry about that poor piece of equipment.

At my job, we sometimes use dummy loads to dissipate up to 50 kW of RF power in the VHF band. The loads for this purpose have a cheap ~100W-rated carborundum film resistor that is axially mounted in a tube through which water (~10 GPM) passes in a turbulent-flow regime. These loads are very expensive to buy, even though the resistor is cheap. They are expensive because of the need to keep the impedance of the transmission line constant all the way down to the resistor element, and because of the complicated hydraulics around the resistor. I think at 4 kW / 13.56 MHz the flow requirements and line impedance requirements would be much less stringent, and if you just fashioned something to grip a carborundum film resistor in a piece of pipe and force water over it with a Reynolds number above 10,000 or so, you would probably have no difficulties.

-Carl

[steve_rb]

Good advice John. Have you made load your self? Can you post or email instruction for calculating and making let say a 5kw 50 ohm load?

[Chris Bradley]

I think it would be irresponsible not to state, on a public forum, that this isn't the way to do things.

I suspect John is simply reflecting amusing anecdotes over his own "minor-spectrum-infractions" and given his skill and experience understands the limits and negligible degree of interference he may cause.

If you only want this for a system check and initial start-up, then just miss those stages and presume it works. Do not, please, intentionally stick 4kW into a 1/4 wavelenth!

You'll know soon enough if it is working if your *low power* experiment doesn't light up.

[steve_rb]

Chris

I have to check it and I want to return it if it is not working. It may work at low power but at high power it may fail. Beside I have to spend a lot for the fusor chamber and vacuum system and if it fails I will loose what I have paid for the chamber too. I have limited budget and have to be careful with how to spend it.

I have though something myself for making a 50 ohm 3 Kw dummy load. I have purchased a normal resistive 3KW element (the one is used for electrical heaters). Its resistivity is about 50 ohm. It is spring shaped and about 50 cm length. I have pulled it all the way and now it turned into a 6m wire with 50 ohm resistivity. I want to use ceramic bids to cover the whole 6 m and insert it into a 6m strait copper tubing ( about 5 mm diameter). This way I have made a coaxial 50 ohm resistive resistor which will not radiate anything to the air. If it gets hot no worries because ceramic bids will stand heat. This will be a cheap simple 3 KW dummy load I need. I don't see any problem here. Any comment or advice about this?

[Dustinit]

It is not pure resistance that you are talking about.
The 6m length represents a large inductance effectively in series with the resistance.
This adds to the impedance and will increase the impedance to a few 100ohms for the RF.
Inserting it into the copper pipe would help if the conductor size/ pipe size matches 50 ohm as this would make it a lossy transmission line which would be ideal. Putting ferrite beads on the wire is not a good idea as this would also increase the impedance and send reflections back into your rf gen (increase vswr).
As the element wire is probably thin, the copper pipe inside diameter would also have to be small making insulation to the walls difficult when the element is dessipating so much power. Ceramic beads would be required.
This sounds like a very difficult solution.
I have used a heap of light globes as a load but the impedance does change a bit with output power. This may be the best / easiest solution.
Dustin

[steve_rb]

Dustin wrote:
> It is not pure resistance that you are talking about.
> The 6m length represents a large inductance effectively in series with the resistance.
> This adds to the impedance and will increase the impedance to a few 100ohms for the RF.

Where is the inductance? I don't understand. If I would use the element as it is there would be an inductance because of its toroidal shape but I have made a 6 m long wire (0.1 mm diameter) out of a 50 cm long torodial element by pulling the ends and straitening it all the way and I don't see any inductance in there.

> I have used a heap of light globes as a load but the impedance does change a bit with output power. This may be the best / easiest solution.
> Dustin

What kind of globe are you talking about (w=? , v=?) and how did you connect them (series or parallel) . Could you give me a schematic?

[Chris Bradley]

steve robinson wrote:
> Where is the inductance? I don't understand. If I would use the element as it is there would be an inductance because of its toroidal shape but I have made a 6 m long wire (0.1 mm diameter) out of a 50 cm long torodial element by pulling the ends and straitening it all the way and I don't see any inductance in there.

Any length of wire has an inductance. It is a function of how long a propagating wave (a wire is a waveguide, just a very bad one) gets to the end, sees a discontinuity of impedance and so reflects some of the energy back. It then bounces around in that wire.


> What kind of globe are you talking about (w=? , v=?) and how did you connect them (series or parallel) . Could you give me a schematic?
What kind have you got? V=IR, P=V^2/R: So if you have 100W bulbs you'll need 40 of them. If they are 240V then R[when hot]=V^2/P=576ohm (disregarding subtelties of AC power). 10 in parallel will therefore be around 50ohm, so sounds like a 50 ohm 4000W arrangement would be 2 serial sets of 20 in parallel.

Initial resistance is much lower, but light bulbs are the amateur RF termination of choice for new experimental circuits as they are PTC and so will adaptively soak up transient power.


But I remain concerned over your safety on this. I recommend you find a radio ham amateur to assist/advise the work.

[steve_rb]

Bulbs are also tungesten wires and the same thing you mentioned above should also happen here. Only lenght is shorter because resistivity is higher. In that case how about using some high watt and high resistivity resistors (old large high watt resistive variable resistors perhaps are the best) and then build a metal shield for them.

Another question: How would you prevent RF not to radiate to air with the bulbs?

[Doug Coulter]

Steve,
anything that makes a magnetic field when current is flowing has some inductance. Long things, or things that are wound into circles are especially bad, worse the bigger the circle -- most old EE handbooks give the inductance formulas for making your own inductors, and that can be really informative.
You can use a bunch of resistors in parallel, which is what a lot of the pro loads do. This puts the inductance of each in parallel, which cuts the net inductance down, just like it cuts the resistance down.

You can wrap a grounded shield around things to keep the RF inside -- just don't make big holes in it anywhere, and make sure it's really well connected to itself -- just touching in one place isn't a good connection even if an ohmeter says it is -- again because of inductance -- it's going to look like a single turn of wire, so it'd better be well shorted together at the ends.

You can you know, also run low power -- almost nothing needs that kind of power except fast RF sputtering of really big things, or an inductive furnace heater. My ion source uses 20 watts or so....and it's plenty.

[tligon]

When I was a boy (Ligon leans on his cane), we used incandescent light bulbs to load our CW transmitters, and we LOVED it! But that was maybe 80 or 40 meter ham stuff at a hundred watts or less.

We even tried this at EMC2, with microwave power. However, once we finally got the thing to work a little, it turned out we were radiating several hundred watts at 2.45 GHz, which tended to make the people in the front office scream about their computers screwing up, and it was a bad idea.

Steve, paging thru this I know you are looking at RF, but I don't know what frequency you are using. It makes a big difference.

Those who could afford a few bucks for a proper amateur radio load in the 10-80 meter band range bought a Heathkit Can-tenna. As shipped, this was basically an empty paint can with an RF connector on the lid and a large resistor inside, designed to be as non-inductive as practical to allow the frequencies to be pushed. I don't recall the upper limit. As-shipped it had relatively limited power capability, possibly a hundred watts or so. If you filled it with mineral oil you could apply a kilowatt for a few minutes. Richard Hull could probalby spot ten of these in his first minute in the gate of the average hamfest.

The strategy can probably be extrapolated to 4 kW. Larger can, more capable resistor (more likely many resistors in parallel for a net resistance of 50 ohms), and shorter on time. A more thermally conductive coolant, including possibly distilled water, might work ... watch the temperature carefully. But if this is for UHF or microwaves, man, that's not amateur stuff at those power levels.

I will say, without recommending it, that after the computer interference incident at EMC2, we ran the microwave cable into one of the microwave ovens we had gutted for its magnetron and power supply, and put a 100W incandescent bulb in there. Who knows the actual impedance, but we managed to get it to light up brightly to a strange blue (sounds a bit like some Tesla coil effects), without causing interference. The source was another microwave over supposed to be capable of 1 kW. After we found a cable that could take it (LMR-400) we were able to run this test without producing obvious interference. So one might consider an old microwave oven as a possible RF shield for the dummy load. Most models these days are intended for continuous use at 1.5 kW or so.

[tligon]

Ah, cool, of course the manual for is free on-line.

My browser does not allow me to give the link to this particular .pdf file conveniently, but it should pop up on any search engine with no problem. The copy I found was on a long path off www.cnyara.com/_manuals/index.php.

Heathkit HN-31 Cantenna Manual

[tligon]

Just for fun, I was noodling in my head over the prospects for building a 50 W power resistor for an RF load.

Typical space heater wire has to be coiled to make a radiant heater. I have made load resistors with ribbons of nichrome which we happened to have on hand where I worked, but to handle this power level would take a number of strips in parallel, with a length that would just not be practical without winding it around something. That produces more far inductance than a straight length.

So I thought of something I once saw in the Electric Spacecraft Journal, a little specialized science journal Richard Hull inherited and I subscribed to for a while. It is a mix of some saucer kook stuff with some really good papers, and some interesting stuff in between. One of the devices that draws a bit of saucer kook interest is the Caduceus Coil. This is essentially a pair of coaxially wound coils turning opposite directions. It has peculiar impedance considerations. It might make an interesting wire wound dummy load, but beware its reported side effects of emitting tacheons and producing time travel.

Anyway, looking it up brings up the related topic of non-inductive wire wound resistors, which might suit the dummy load nicely, providing you don't accidently miswire it into a time travel device.

http://www.fluxcap.com/HDR_Caduceus_Coil.htm

[lutzhoffman]

What about using a tank of water with an appropriate electrolyte?

[Doug Coulter]

Ionic conduction in water is a bit different from that in metals, and has time issues. It's not fast so the impedance will go all over with frequency. Since water has a dielectric constant of 80 or so it's not going to be a resistive load...(but that changes with frequency too).

My Terman book shows all kinds of ways to use wire to make non inductive resistors -- don't have to wind it into a coil at all. Zig-zag etc on a mica holder and this way/that way windings do well, and can be immersed in oil fine. This scan doesn't even cover all the ways to do this.

Man, all you guys gotta get some good EE handbooks and actually read them -- don't just put salt on them and eat them.

This Terman book is about the best ever and is still around on used book sites. Only goes to WWII but really, many things haven't changed since. Add some good stuff on semiconductors and you've got it all.
Same concepts apply anyway, just different active device characteristic details.

[Dustinit]

Or you could just put some capacitance across it to make it parallel resonant effectively making it appear purely resistive via extremely low Q.
This would be even easier I guess.
Dustin.