Thats a great idea, to use hot water heating elements. I was troubleshooting a burnt element this weekend, that was measuring open. The specs are typically 4000 watts for 240 VAC, so doing the math that comes out to 14.4 ohms DCR. You would need to series 3 to get 43 ohms. There are some caveats. The 60 Hz impedance is probably not the same at 4 Mhz or higher. You would most likely need to do some compensation (tuning) to get the thing to have low reactance and keep your tx from tripping off on high reflected. Otherwise audio and RF performance of your rig will suffer trying to drive a load not in the center of the Smith Chart. The other thing is that you need to measure the hot resistance with them in water. Maybe that can be done with a heater, flip off the breaker and stick an ohmmeter on the terminals right away to see what the hot R is. This would change the value of your load too.
I remember my college roomate loading up an electric skillet at 1 MHz. I gave a AM station $50 for a semi-burnt 7.5 kW Ohmspun load they used to use for 5 kW, that was made in a cast aluminum frame. It had the resistive wire 'toaster elements'. After reconfiguring it and removing the burnt elements, i got a nice 1 kW load that has plenty of headroom. I stuck a fan on the end, and added a series mica cap block in series with a type UHF connector. It is a perfect match at 1260 KHz right now, but I can move it up to 3.88 with a change of C. In the winter I stick it in my shack and in summer I put it outside in the adjacent carport, and hope the dogs don't come lift their leg on it while I am on air. Having a vector impedance meter, network analyzer, OIB, or a low frequency antenna analyzer is helpful when checking homemade or modified loads. My first load using some solid carbon rods and a BC rig would trip off the TX VSWR protection after about 10 minutes of carrier, due to the tempco of the load as it got up to boiling temperature. 73 John K5PRO
