BNC switch - expand the IO number of your lab instrument

[martinLink]

Hi guys,

Link to the BNC switch project website www.linkbone.com

I want to share with you the project that I have been recently working on. The BNC switch with remote access via 10/100 Ethernet or RS-232 interfaces for measurement and test automation. This is an analog signal multiplexer allowing to do connect up to 15 signal points to one common signal line. This is a passive BNC switch for analog or digital signal routing. It can be used for making measurements with an oscilloscope at multiple test points. With signal generator connected, it can be used for routing system clock or any other signal.  Please have a look at the attached picture with BNC switch example application.

Here are some preliminary specs/features:

    - Bidirectional signal routing
    - DC to 50Mhz bandwidth
    - Accept voltages up to +/-24V with 1.5A load per channel
    - 50Ohm channel impedance
    - Http server with user configuration Web page
    - Telnet server and RS-232 command line interface
    - RC5 infrared remote control
    - Touch screen on front panel for manual control
    - 19-inch rack muting system

There is also a dual BNC switch version with two 1-to-15 multiplexers/demuliplexers signal networks. The device can be also used for video/CCTV switching. I have also made an XLR switch version for audio.

Currently first lot of BNC switch boxes is in PCB assembly. I post more information with signal measurements when I will have the devices ready.

In the blog section I have prepared few application notes with code examples:
http://linkbone.com/coaxial-switch-with-remote-ethernet-rs-232/
http://linkbone.com/bnc-switch-remote-box/

Please let me what do you think about it? Do you find such a device useful? What kind of additional functionality or information you would like to get? I appreciate your feedback.

PS I hope I will not be taken as a spammer. I was reading this forum for a very long time. But finally I have decided to register to tell you about my project.

[tld]

How much capacitance would you add by routing a signal through this?

And what about channel-crosstalk?

tld

[rdl]

You have the power switch installed upside down.

[T3sl4co1l]

I also wonder about reflections along the backbone.  But I suppose being rated to only 50MHz, that would be part of the point.

What's transmission loss and channel-to-channel and output isolation?  Is there a way to guarantee response time over Ethernet or RS232?

Tim

[martinLink]

Thanks for your comments.

The BNC switch has a 4-layer layer PCB for signal routing. This limits the crosstalk between the channels is around -45dB or less at 20Mhz. I will measure the capacitance and post it later. Here you can see the 10Mhz square wave signal connected directly to the scope and through the BNC switch. As there was no termination at the line ends the BNC switch signal shows bit higher reflections.

T3sl4co1l: You are right, the 50Mhz BW is because of reflections on the internal switch network. However the 50Mhz limit is for the worst case input 'O'. For example 'A' input can have around 80-100Mhz BW.

The BNC switch acknowledges the state change during execution the text commands over RS-232 or Ethernet interface. The complete network switch time is round 100ms. There is also an web page for controlling the device:

Im also working on the 8x8 BNC matrix allowing to connect any 8 inputs to any 8 signal outputs. The Dual BNC version is shown here:

rdl: thanks for the hint with the power switch 

[tld]

Partly put of curiousity, partly as a suggestion... Any reason it isn't broken into segments?

You have Port P as an always attached, but it would be a lot more flexible if you could divide the whole thing into 2-8 segments.

As an example, if you cut it down the middle, used port H to (internally) optionally bridge the two halves, and added a port for the other half, similar to your current Port P.

tld

[w2aew]

One caution...  You show the unit being used with scope probes (presumably 10x probes). This won't work, because it's unlikely the probe could be compensated for the extra capacitance of the line from the switch matrix to the scope.

[martinLink]

tld: Inside of the BNC switch there is quite extensive switch matrix. It allows to plug in up to 92 signal relays. One configuration which Im also considering to build is 2-to-14 BNC switch. It allows independent connection from any of A to N inputs/outputs to O or P ports. I attached the picture of PCB connection network:

w2aew: you are right, the BNC switch box should be used for direct signal connection e.g. via coax. I attached another application example of using the device as a coaxial switch for clock distribution. Of course care should be taken to avoid reflections from different line endpoints.

[TiN]

Nice project, but practical application depends on price. There are various switch units already on market, like Keithley 7001/7002 series, which possible to get for few hundred USD used.

[Mechatrommer]

You have the power switch installed upside down.

here, that is correct. btw, this project is a luxury.

[martinLink]

TiN: what price range would be appropriate for a new BNC switch unit? This might be a good market indicator