开云体育

Flow calorimeter for RF/Microwave power measurement

Neil Smith G4DBN All Messages By This Member #141 开云体育 Jeff's reference to the Oct '63 "Electronics World" article reminded me that I'm working on a project for my Youtube channel for a balanced flow calorimeter that can measure to 750 W at 2.5 GHz and 250 W to 3.5 GHz and potentially higher with waveguide loads. I'll be using two identical thick film RF power resistors mounted to copper blocks with embedded fluid channels.? The plan is to make it self-calibrating as much as possible with a facility to swap DC and RF inputs. I may just use my 30 gallon transmitter watercooling tank as the source and sink. Even at a total of 1.5 kilowatts DC plus RF, it takes almost 7 minutes to raise the tank temperature by 1 C.? Keeping the total temperature rise to 5 C requires 5 litres per minute flow rate. I'll use a 200 volt DC 4 amp variable power supply to provide the comparison power. I considered using a positive-displacement tandem pump to give a consistent volume flow rate with both loads fed with identical temperature coolant. As I'm not really looking for more than a few percent accuracy, it might even not require dual loads, but where's the fun in that. Having seen this: I wonder if it's possible to build an instrument with good precision in the 1-100 watt range using very small thick-film loads? I may use a non-aqueous coolant to avoid corrosion, and have a heat transfer coil immersed in the water tank. Ideally I want the measurement to be independent of flow rate, fluid characteristics, power level and ambient temperature and just act as a self-balancing bridge with DC in one side and RF in the other, with volts/amps DC as the output and a reasonable settling time of maybe 20 seconds at 750 W if I can get the thermal mass low enough. Has anyone built such a device?? How about advice for making it as independent of physical variables as possible?? What I'd like is as pure an RF power to DC volts/amps converter as possible. -- Neil G4DBN On 12/08/2022 17:05, Jeff Green wrote: toggle quoted message Show quoted text While for RF, the information on this site might be relevant. ?
Dave Wise All Messages By This Member #142 开云体育 Hewlett-Packard 434A. ? toggle quoted message Show quoted text From: [email protected] <[email protected]> On Behalf Of Neil Smith G4DBN via groups.io Sent: Friday, August 12, 2022 11:22 AM To: [email protected] Subject: [Test Equipment Design & Construction] Flow calorimeter for RF/Microwave power measurement ? Jeff's reference to the Oct '63 "Electronics World" article reminded me that I'm working on a project for my Youtube channel for a balanced flow calorimeter that can measure to 750 W at 2.5 GHz and 250 W to 3.5 GHz and potentially higher with waveguide loads. I'll be using two identical thick film RF power resistors mounted to copper blocks with embedded fluid channels.? The plan is to make it self-calibrating as much as possible with a facility to swap DC and RF inputs. I may just use my 30 gallon transmitter watercooling tank as the source and sink. Even at a total of 1.5 kilowatts DC plus RF, it takes almost 7 minutes to raise the tank temperature by 1 C.? Keeping the total temperature rise to 5 C requires 5 litres per minute flow rate. I'll use a 200 volt DC 4 amp variable power supply to provide the comparison power. ? I considered using a positive-displacement tandem pump to give a consistent volume flow rate with both loads fed with identical temperature coolant. ? As I'm not really looking for more than a few percent accuracy, it might even not require dual loads, but where's the fun in that. Having seen this: I wonder if it's possible to build an instrument with good precision in the 1-100 watt range using very small thick-film loads? ? I may use a non-aqueous coolant to avoid corrosion, and have a heat transfer coil immersed in the water tank. ? Ideally I want the measurement to be independent of flow rate, fluid characteristics, power level and ambient temperature and just act as a self-balancing bridge with DC in one side and RF in the other, with volts/amps DC as the output and a reasonable settling time of maybe 20 seconds at 750 W if I can get the thermal mass low enough. ? Has anyone built such a device?? How about advice for making it as independent of physical variables as possible?? What I'd like is as pure an RF power to DC volts/amps converter as possible. ? -- Neil G4DBN ? On 12/08/2022 17:05, Jeff Green wrote: While for RF, the information on this site might be relevant. ?
Neil Smith G4DBN All Messages By This Member #144 开云体育 Thanks Jeff, that's interesting about the temperature-variable specific heat capacity of water. I'm assuming that for a dual-load device, the flow rate and SHC are automatically compensated so long as you wait for a steady-state endpoint.? Mass-flow calorimeters are certainly a different matter, there I'm sure the expansion and any temperature-related effects on the SHC needs to be allowed for. The heat exchanger in the HP424 design coolant loop is a source of error that probably needs a DC-RF changeover to calibrate it out, also as you suggest, the pump might generate a little heat, although if the pump is gentle and isolated from the motor, that wouldn't be a problem. The idea of using two cylinder pumps driven together seems a reasonable solution, and very easy to calibrate for identical volume flows over timescales of a few seconds.? I don't think spring/ball or flap valves would suffice though, I rather imagine driven valves would be needed for ultimate precision.? If the input flow is from a large reservoir and the output goes to a separate tank, that would eliminate the variance of temperature between the sensors and eliminate the heat exchanger.? If the input reservoir was held at a constant temperature and was insulated and well-stirred, that would be another step towards making the system independent of external influences.? You could even adjust the flow rate to keep the output temperature within a narrow range so reducing the errors from conductive heat loss through the insulation around the loads and blocks. Taking it to the extreme, you could enclose the blocks in a copper box and heat it to same temperature as the blocks, all but eliminating conduction losses I don't actually need a precision power meter. This would be a project for the sheer joy of making a precision instrument in my home lab and machine shop, and might even get me a few views on the 'choob. -- Neil On 12/08/2022 19:37, Jeff Green wrote: toggle quoted message Show quoted text An EE friend reminded me water expands with heat and the watt meters used in TV stations have a chart for calibration. If I'm understanding it right, because the water is less dense with temperature rise, a correction factor must be applied that takes in to account this factor. ? The dummy load has a resistor with multiple water channels and water is circulated by a pump. ? This is a grossly wrong example, the real world will be a lot less. ? At 40F say it takes 100W to raise the water to 45 degrees, but it will only take 95W to raise it to 50F, 90 to raise it to 55, 85W to raise it 60F. ? The above example overstates the effect by insane amounts to demonstrate. I was told the rate of change is linear and about 2% for ever 10 degrees Fahrenheit. ? Somewhere there should be charts that give the expansion per degree of temperature rise. ? My EE friend suggest we check to see if there is a member who worked in a TV broadcast facility for clarification. ? I remember physics practical labs where we placed a 4 sided paddle in an insulated water bath and measured the temperature rise caused by the rotating paddles. ? I wonder if the water bath in a TV transmitter power meter would show a temperature rise from the mechanical energy added by the pump.

Previous Topic Next Topic