Diode Recommendation

rileyesi · Joined: 19 Dec 2006 Posts: 398

Hi.

Off topic post here.

I have a system with a sensor that is driven by a dc voltage supply. For several reasons, I need to raise the voltage potential level of the ground pin of my sensor above that of the power supply. So, I simply inserted a diode in between the two grounds thereby guarenteeing a relatively consistant voltage drop regardless of the current flow.

My problem is that I used diodes that were "just lying around" the shop and they failed in the field. Instead of diodes they are now just wires!

I am looking to buy new ones and I don't know much about the different diode types. These will be left outdoors so it will see temperature extremes. It will see no more than 5 VDC and 200 mA during normal operation. The actual voltage drop does not matter. Also, I am looking for the axial lead type (looks like a standard resistor).

As always, thanks.

Pete

AdrianJ · Joined: 16 Jan 2006 Posts: 2481 Location: Queensland

Your voltage drop will be only "relatively" constant ! Depending on the diode and current and temperature, you could expect anything from 200 - 700 mv drop. I hope you are not measuring the DC level from your sensor to that sort of accuracy.
But if you want to persist with this approach, any silicon diode capable of passing 200 ma should work. The good old 1N4001 series would be fine. 1N4004 is probably more common ( 400 v reverse, rather than 100 )

rileyesi · Joined: 19 Dec 2006 Posts: 398

Adrian,

Thanks for the reply.

The absolute value of the voltage drop is not important at all. I really only need a 0.03 V minimum.

The strange thing is the failures I already saw. I think the diodes I used are shottky types. I'm not sure why they failed, but I have a few guesses.

Physically, the best place to sneak these into the existing system was between a ground wire and the solder cup of the MIL type multi-pronged recepticle. For mechanical strength I used shrink tubing atound the wire-diode-solder cup junction. I think that maybe too much heat was used when soldering or possibly the heat from the heat gun used to shrink the tubing??

Anyway, thanks for the help.

Pete

Arera · Joined: 23 Sep 2007 Posts: 375 Location: Wuppertal, Germany

Hi Pete,
just some ideas related to your task:

-Is shifting the ground-level really that important (I´d avoid that if any possible!)?

-Any diode is temperature sensitive. Forward voltage lowers with temperature.

-Z-Diodes have a second effect: The reverse voltage rises with temperature.

-Both effect compensate at a Z-Voltage of 5V.

-That means: a 5V Z-diode used reverse gives a quite constant drop-voltage.

-There are circuits that meassure temperture with a 1N400x or 1N4148 as sensor!

-AFAIK it´s not possible to get 0,03V with a diode.

-There is an adjustable temperature-compensatet Z-diode: TL 431. It´s cheap and easy to use. 3-Pin TP92 case and to R´s to adjust. It works from
2.5V to 36V. Just google a little.

-Google for "precision bandgap voltage reference". You will find "Z-diodes" from 1.25V on, maybe even lower V. Bear in mind, they all need a certain minimum current to work, read the datasheets!

-Ask me and the other folks here around if you need further advice!

Good luck!
BTW: Give us one reason for shifting ground-level!

rileyesi · Joined: 19 Dec 2006 Posts: 398

Arera,

Thank you for the information. The reason for wanting the ground of my sensor to be different than the cround of my ADC is to sense when the sensor is missing, damaged, or if the signal cable is cut. By adding a diode between the two grounds, I am guaranteed a minimum voltage drop when the sensor is fine. If the sensor outputs drop below that diode drop, then I know there is a problem.

By the way, I decided not to use the diode design. I scratched my head and got smarter and am accomplishing what I need to do with my software. There is something I am not taking into consideration with the diode design. Every system I built in my shop worked perfectly, but every one of them failed when installed in the field! So, I am eliminating the diode and so, eliminating the possiblility that it will fail!

As always, thanks to all for the help.

Pete

Arera · Joined: 23 Sep 2007 Posts: 375 Location: Wuppertal, Germany

Hi Pete,

as you have seen by yourself, shifting groundlevel leads to unpredictable effects, that´s what I´ve said.
Now you found a better way to survey the sensor, fine!

Only to complete our thoughts about level-shifting:
I think it´s a good idea to add a few mV to your signal to detect wiring and funktion. To achive that, you can use a simple OpAmp. They are MADE to calculate for you! Your term is: V(signal)+V(constant)=V(line-signal).
That´s most easy with an OpAmp (goooooogle!).

Marc Arera

rileyesi · Joined: 19 Dec 2006 Posts: 398

Marc,

Thank you.

Pete