Sonar application 1000W

Hi Ed

I bought an amplifier EVM from TI and tried a similar application to drive a transducer via a transformer and it failed. The reason for failure turns out to be due to the charging current of the transformer during the first cycle causing the short circuit protection on the device to trip and shut down the device. See my earlier post regarding this.

BR

Manjula

Hi, Ed,

No, we have not looked at sonar applications in any great depth (booooo!). I have heard of many customers trying to use our audio amps for these kinds of applications, but I don't know how successful they have been.

-d2

I am looking at a similar application to drive a pinger transducer with up to 1KW at frequencies up to 15KHz. I would be interested in an off the shelf solution or perhaps if anyone has schematics and/or PCB artwork for a tested prototype it may help shortcut development time. The TI evaluation kit for the LME49810 does not seem to provide for boosted output drive. I would be interested to hear about the outcome of this requirement for a 1000W device which would be suitable for my application.

As your transducer bandwidth is small, there is little problem with using a square wave drive into the matching network.

An approach I have taken in to use a tuned transformer to tune out most of the fixed capacitance of the piezo ceramic transducer, and to precede that with a L match to extend the bandwidth somewhat, you usually end up with a series resistor to lower the Q, but that is reasonable at these duty cycles.

Last time I did one of these it was a 10 - 20KHz design running ~1KVA (500W real, possibly a little more at resonance), with a simple H bridge driving the transformer, the art is in designing the matching network to be either resistive or slightly inductive across the operating range, this is easier to deal with then a capacitive reactance.

A reasonably high DC bus voltage really helps matters, as it makes the tuned transformer primary far more reasonable, 12V in for a KV out with the secondary magnetising inductance I needed resulted in a 1.5 turn primary and a gap measured in mm!

Note that matching network voltages get very scary, very quickly, careful modelling is advised as is some care in construction and test.

HTH.

Regards, Dan.

Thank you for the suggestions.

This may be a suitable design once one has identified the most suitable frequency for the application.

(1) If it is necessary to dynamically change this frequency during experimentation, would it not make sense to use a power amplifier to drive the transducer?

(2) If so, I would value your thoughts on using the 125W AN-1850 LME49830TB Ultra-High Fidelity High Power Amplifier Reference Designs in bridge to get more power directly into a 200 ohm transducer transformer.

(3) If I could get a few of the LME49830 Integrated Amplifiers, it looks like I could get the components locally to produce a few prototypes for test purposes. I could not find this reference design in kit form which would save some time. Are you aware of such availability or even a third party source for the Printed circuit board blanks or even populated  boards for the reference design. I will have another look on line to see if I missed it on the TI site.

Regards,

Jock

.

You can use an amplifier, but care is required lest the load capacitance below resonance causes the feedback loop to become unstable, for audio band work in the lab we used to use an old Crown MA5000VZ in bridge mode which would push ~100V into a transformer primary all day long and was fairly well behaved.

You usually need to redesign the matching if trying to move a high power transducer over more then about an octave anyway (Look at the admittence/susceptence curves), and sometimes the transducer itself needs a bit of a redesign (Thickness mode, and acoustic matching layer are frequency dependent in simple plate or sphere designs, Tonpills elements you usually need to tweak tail mass, and preload values.

Regards, Dan.

Thanks for your advice, Which I read a second time after a few months of part time experimentation. Your advice worked perfectly. A stereo Dixon MA400 gave me 2KW RMS bridged into 4 ohm fine at the less than 1% duty cycle required for my target frequency range and the octaves could be bridged with binary wound primaries and secondaries on matching transformers to a low frequency piston type transducer with a frequency maximun around 7KHz. It struck me that if we can reduce the equations to that of a simple pressure wave interface considering the built in filtering inherent in the piston transducer,the use of a square wave similar to the way it is achieved in variable speed drives would as you have suggested surely work. In which case the MA400 route was a good way to go anyway because It is very helpful to design and test the transformers. I wonder if the guys who have looked as magnetic hi-fi amps using superconductive magnetic materials have not considered the applicability of the technology to the dolphin loudspeakers too? I am sure there is a fun project for someone there. While I would appreciate any further useful inspiration perhaps along the lines alluded to above, thanks for your helpful ideas.

Regards, Jock