Mechanical Engineer needs relay or diode

[willcambridge]

I am a mechanical engineer, designing a small aerospace device with a simple electronic circuit. I was working with an EE but he left with work uncomplete, and now I have to finish in a hurry. The device is meant to function in a chain: think of christmas tree lights. I need to select a component to act as a switch: in the event that a key component in the device fails, causing an open circuit, I need to bypass that component and keep the circuit intact.

I looked, and Christmas tree lights do this with a resistor in parallel to the component (in their case an LED). That causes power burn, which I want to minimize if not avoid.

I have a tight form factor: no more than 0.130" in thickness for the component as well as the board. Laterally, there is more space.

Electronic requirements are up to 3 A current, and < 1.5 volt drop across a single component. However, in the moment of failure, voltage could be higher, i.e. the stacked voltage drop of all units in the chain. Maximum in that case would be 28 V.

I understand my options are 1) a relay, which draws no power; and 2) a diode, which would draw power. We ruled out a switching transistor based on complexity required to drive it. I want a very dumb passive component. This is an analog circuit.

My question specifically relates to relays - I am hoping someone can tell me if there is a component in the size range required. My searches show large components that look like they were cutting-edge in the 50's. I feel like there must be something smaller I am just not finding.

Your help is much appreciated.

 

[MacGyverS2000]

You may have a tough time finding a 3A relay that will fit into a sub-100mil space (I'm assuming a 30mil board thickness).

A diode will pretty much bypass most circuits, but it'll do that permanently, so there's not much point.. the LED and parallel resistor work because the LED is practically a short compared to the resistor. If your "backup" is a diode already, the only thing that would work is a circuit that has a lower resistance than the diode.

Doing this with a single passive given your requirements is going to be reaaaaally tough. Give us more information on what you're trying to accomplish and what the circuit does and maybe we can give you a more useful solution.


Dan - Owner

http://www.Hi-TecDesigns.com

 

[ScottyUK]

With only 1.5V drop across each 'component' (is it top secret?) you could bypass each with three forward-biased silicon diodes in series, giving a forward volt-drop of about 2V over the string of diodes, without causing any effect on the 'component'. The diodes would not conduct unless the mysterious 'component' burns up open circuit, in which case they would drop an additional 0.5V or so compared to whatever burned up.


----------------------------------

If we learn from our mistakes I'm getting a great education!

 

[melone]

Based on your size and power requirement, it sounds like a transistor is what you need (PNP or PMOS). Bias the transistor off of some output voltage from your circuit. In other words, if the circuit is outputting voltage, which means it is working, disable the transistor bypass.

Ensure you don't have any reverse leakage paths created by your transistor (insert a forward biased diode to prevent it).

3A is still a lot of power, which would translate into heat. Does your enclosure allow for adequate heat disapation?

 

[willcambridge]

Thanks for the responses. The mystery component in each unit on the 'chain' is a thermo-electric device with internal resistance that results in the above specs.

I have seen reference to much smaller 'numerical relays' used on microprocessors. I am guessing that is a digital technique and is not applicable, but I'm not sure. I am just surprised that macro-sized relays do not seem to have micro- or surface-mount counterparts in the way that components such as resistors and transistors do. (Yes, I have taken one electronics class and have enough knowledge to be dangerous)

If a relay is a no-go, can someone supply product recommendations on a small diode?

Thanks.

 

[ScottyUK]

Assuming surface mount, the SMBYT03-400 from ST Microelectronics is a typical small 3A 400V diode. 'Small' is relative - you won't get much smaller because the current rating is quite high for a SMT diode without additional heatsinking. There are other similar types - check Digi-Key or RS Components or your local supplier of choice. The package style is a TO214AB which is about as small as you'll get for a 3A diode.

If you can cope with through hole diodes the smallest you will get are the glass bodied types, although these run pretty hot.


----------------------------------

If we learn from our mistakes I'm getting a great education!