LEAK NOISE CORRELATOR
Q I volunteer for my local water department. The pipes are all cast iron — four and six inch — within the water district. We have several and expensive breaks each year, mainly during the winter and spring thaw. Our water operator currently locates these leaks with a sounding device he acquired years ago. As you can imagine, it is labor-intensive and often we have to dig up whole sections of road just to find a leak or broken pipe.
Lately, a new device has come on the market which can accurately locate the break. The only thing about it is the cost is out of sight. The principle on how these leak noise correlators work has been described as a microprocessor-based system that uses cross-correlation to determine the time difference between signals.
The ping sensor with an amplifier maybe ideal for this, although I do not know. Any help you could give with the design and assembly of one of these devices would be greatly appreciated.
— Howard Epstein
A The system that I envision includes a digital storage oscilloscope because the human brain is the best correlator you can find. I found a USB powered unit on eBay that you could use with your laptop computer for about $30 (shipping included). The item number is 220651026871.
From the microphone attached to the pipe, I would put a band pass filter to eliminate low frequency traffic noise and limit the upper frequencies so the correlation will be easier. Follow this with a rectifier circuit (envelope detector), then to the scope. The scope is two channels so you can measure the time difference between signals, and then apply the formulas to determine the distance to the break.
There will have to be some experimentation, and I am willing to work with you to optimize the system. To start, assume that a 300 Hz to 1 kHz band pass will attenuate traffic noise and provide a usable signal envelope. The acoustic sensor could be a microphone epoxied to a ring magnet. The circuit for each sensor would be something like Figure 1.
I expect the sensors will be attached to a fire hydrant. You will need to measure the velocity of sound in the pipe. To do that, make the measurement in a section with no break. Hit one hydrant with a small hammer and measure the time for the sound to reach the other hydrant. The velocity (V) is: V = T/D, where D is the distance between hydrants. I have derived the equations that determine the distance to the break; refer to my sketch in Figure 2.
If you want to use an RF link, Qkits has a transmitter (TX433B) for $9.95 and a receiver (RX433) for $7.95 ([url=http://store.qkits.com]http://store.qkits.com[/url]). You can’t use two transmitters on the same frequency, so I recommend that you hard-wire one of the sensors to the correlator. You will need antennas and an amplifier for the transmitter modulator at least.
MODIFIED BEEPER/LED CIRCUIT
Q I saw a schematic in N&V for a beeper/LED and would like to add an SCR to it, but don’t know if it’s placed in the right position.
Please indicate if the SCR is correctly inserted into the schematic (refer to Figure 3 in the answer); if not, please modify it.
A The SCR will not work because the worst case holding current is 3 mA and there will not be that much current when the 555 output is high. Even the best case (0.3 mA) probably won’t work. The only way I see to do this is with a latching relay, as in Figure 3.
Q Back a year or two, I designed a Theremin tube circuit. The PITCH and VOLUME coils are a kind of transformer or modified RF coil. The pitch frequency is about 350 to 450 kHz. The volume operates at a 250 kHz frequency. My request is: Can you give me data on these coils for the pitch and volume RF coils? It must be sensitive to hand capacitance. The BFO is also needed for data to wind coils.
The two IF transformers are 455 kHz radio types which operate as a filter for the first and second harmonics. Can these be made out of the RadioShack cores #276-104? Also, the volume detector transformer needs data on this to wind coils too. Can RadioShack core #276-104 be made from this core too? I’m trying to get a prototype built, but I need coil windings on all of these. If you can help me out on the coils/transformers for this design, it will be greatly appreciated.
The inductor in series with a capacitor helps make the antennas more sensitive by resonating at the coil’s frequencies. If I made any mistakes in this design, just let me know. It started out as something to work with.
— Craig Kendrick Sellen
A I didn’t print your tube circuit because I do not fully understand it. The series L-C at the antenna will not work because it is not a complete circuit. I don’t understand the use of a 455 kHz IF after the mixer; I would use a low-pass filter (see Figure 4). The RadioShack core will not operate at the high frequency; a ferrite AM antenna core would work but is difficult to cut to length.
The Theremin operates by beating two oscillators such that the difference in frequency is in the audible range. One oscillator (the VFO) is varied by hand capacitance near the antenna. A high L low C design is necessary for this to work. The other oscillator is fixed (the BFO); a higher C–L ratio is best for this oscillator to improve stability.
The volume can be varied by running the signal up and down the triode transfer curve or by directly varying the VFO amplitude. Figure 4 is arranged to vary the VFO amplitude by controlling the current. A frequency discriminator or slope detector is needed for this to work.
I will raise the frequency to about 20 MHz in order to use easily wound coils; refer to Figure 5. The handbook formula only works for single-layer solenoid coils; that is the limitation. I simulated one of the oscillator circuits; the frequency was much lower than I calculated, but the frequency is not important as long as the VFO and BFO are the same. The volume oscillator is at 15 MHz so it does not interfere with the VFO.
Two oscillators at nearly the same frequency will pull together unless there is good isolation. I am hoping that Q3 and Q7 will provide sufficient isolation, and coils L1 and L4 must be mounted at right angles to each other and as far apart as practical.
In Figure 5, Q1 is the VFO oscillator and Q2 controls the current, and thereby the volume. I was surprised to learn that more current gives lower volume; I had expected the reverse. L3 in the volume circuit is tuned lower than L2 so that hand capacitance increases the DC output and reduces the volume. It might be better to tune L3 to be the same frequency as L2, and then hand capacitance will increase the volume.
I had initially used a diode mixer but the output was too low for the envelope detector; an amplifier is needed (now I see why you had an IF transformer). I changed the diode to Q9 which provides gain and also doubles as an envelope detector. The Q8 circuit is a three pole, low-pass filter at 20 kHz.
When winding the coils, the wire size is not critical, but if the size is larger you won’t get 77 turns in one inch for L1. In that case, use 77 turns close wound. I recommend drilling a hole in the coil form at the length specified and wind from hole to hole. Coat the coil with shellac or varnish for stability. NV
Figure 3 correction is in the downloads for June 2011 Q & A.
June 2011 Q&A Correction
Figure 3 correction