Need switching device that doesn't affect impedance measurement

[stojakapimp]

Hi all,

I have a number of underwater transducers that I want to measure the impedance of by using an impedance analyzer. I'd like to create an automated system that cycles the leads of the impedance analyzer to the various transducers, perhaps using something like a decoder or a number of computer controlled switches.

However, since I'm interested in measuring the impedance of the transducers, it seems like I need a switching device that doesn't affect the output signal from the impedance analyzer to the transducer. The output signal is fairly low-powered, but would I need to use something like a relay to do the switching? It seems like most IC parts have input and output impedances that would effect the signal and thus the impedance measurements.

If I use a relay, does it matter if it's an electromechanical relay or a solid-state relay?

Or, do you have any suggestions as to what a better option may be?

Thanks!

 

[berkeman]

Hi all,

I have a number of underwater transducers that I want to measure the impedance of by using an impedance analyzer. I'd like to create an automated system that cycles the leads of the impedance analyzer to the various transducers, perhaps using something like a decoder or a number of computer controlled switches.

However, since I'm interested in measuring the impedance of the transducers, it seems like I need a switching device that doesn't affect the output signal from the impedance analyzer to the transducer. The output signal is fairly low-powered, but would I need to use something like a relay to do the switching? It seems like most IC parts have input and output impedances that would effect the signal and thus the impedance measurements.

If I use a relay, does it matter if it's an electromechanical relay or a solid-state relay?

Or, do you have any suggestions as to what a better option may be?

Thanks!

Welcome back to the PF! It's been a while.

The way this is usually done is with electromechanical relays. You can get analog switches with pretty low ON resistance, but the signal needs to be biased between the rails of the analog switch devices. Your impedance analyzer is most likely ground-based in its measurement signal? How much series resistance could you tolerate adding into the measurement? Could you maybe calibrate out any series impedance that is added to the measurement?

 

[stojakapimp]

Welcome back to the PF! It's been a while.

The way this is usually done is with electromechanical relays. You can get analog switches with pretty low ON resistance, but the signal needs to be biased between the rails of the analog switch devices. Your impedance analyzer is most likely ground-based in its measurement signal? How much series resistance could you tolerate adding into the measurement? Could you maybe calibrate out any series impedance that is added to the measurement?

Thanks! Good to be back! Wasn't even sure if I still had an account but voila, I sure did.

The impedance analyzer has Low and High output ports, which I normally connect straight up to the two wires going into a transducer. So, I think to answer your question, the impedance analyzer isn't really ground-based but uses the voltage difference between the Low and High connections.

And I suppose I'll have to test how much series resistance it can handle. Our transducers usually have impedance around 20 Ohms when in parallel, but around 70 Ohms when not.

I'll try to get a hold of an electromechanical relay and see if that has much of an effect on the measurements.

Thanks!

 

[jim hardy]

i like to have that physical disconnect between separate devices.
Some electronic solidstate switches, when they lose power, lose the ability to block low level signals so their pins all become connected when power fails. That can be embarassing if it makes indications on a control panel go haywire because your test gizmo blew a fuse.. If you use solidstate switches check their datasheet for "power off input impedance".

But a relay that puts the circuit in its normal operating configuration when powered off fails gracefully.

Signals in the milliamp range are often switched with relays having gold plated contacts .

i've used these TQ's with good results

http://pewa.panasonic.com/assets/pcsd/catalog/tq-catalog.pdf

 

[Carl Pugh]

I would recommend reed switches.
Relays around water may have problems with the contacts making erratic connections.

 

[stojakapimp]

Thanks everyone for the input.

@Jim Hardy, I should have mentioned that aside from the impedance analyzer, the other output of the relay will be an echosounder, so something that emits and much more powerful signal, ranging from about 100 to 2000 W. So the switch would have to accommodate that as well.

@Carl Pugh, the switches won't be very close to the water. Our transducer cables are generally quite long, perhaps 50 to 100 feet. This will ultimately be used on a ship, where the transducers are attached on the hull and the impedance analyzer will be ran inside a dry lab.

 

[berkeman]

@Carl Pugh, the switches won't be very close to the water. Our transducer cables are generally quite long, perhaps 50 to 100 feet. This will ultimately be used on a ship, where the transducers are attached on the hull and the impedance analyzer will be ran inside a dry lab.

Have you considered how to protect the impedance analyzer from transients? Like surge hits from nearby lightning strikes, etc. And from RF pickup on the long cables from radio transmissions? The impedance analyzer input isn't generally protected from real-world transients on its own.

You could look at adding low-capacitance MOVs or other protection devices at the impedance analyzer inputs... Is your analyzer similar to the HP 4194?

 

[mdjensen22]

telecom or reed relays might work for you. With telecom relays you can typically get latching coils too which is nice because you don't have to constantly energize the relays (keeps your power requirements low if you decide to connect/disconnect all simultaneously).

Are you planning on completely removing the device when open (disconnect both wires)?
Can you handle open channels on your input or do you need to bias the signal somehow when the signal is disconnected?

 

[stojakapimp]

Have you considered how to protect the impedance analyzer from transients? Like surge hits from nearby lightning strikes, etc. And from RF pickup on the long cables from radio transmissions? The impedance analyzer input isn't generally protected from real-world transients on its own.

You could look at adding low-capacitance MOVs or other protection devices at the impedance analyzer inputs... Is your analyzer similar to the HP 4194?

The Impedance Analyzer we're using is the Agilent 4294A.

The analyzer is going to go on a ship and measure the impedance of underwater transducers as we travel. The transducer cables are grounded to the ship's ground, and thus I don't believe we need to worry about RF noise.

 

[stojakapimp]

telecom or reed relays might work for you. With telecom relays you can typically get latching coils too which is nice because you don't have to constantly energize the relays (keeps your power requirements low if you decide to connect/disconnect all simultaneously).

Are you planning on completely removing the device when open (disconnect both wires)?
Can you handle open channels on your input or do you need to bias the signal somehow when the signal is disconnected?

Thanks for the tip.

I think what I'm looking at now is getting a relay board that I can easily control via a computer. Will still have to decide on the type of relays to use, but I don't know if latching vs. non-latching will be a big issue.

As for completely removing the device, are you referring to the transducers or the impedance analyzer? Either way, I believe what I need is a DPDT relay, as each transducer has two inputs (high and low) and will either be switching between the two outputs from the impedance analyzer, or the two outputs from the echosounder (thing generating the transducer pings).

Thanks!