Voltage spikes and use of protection diodes in aircraft systems

[RFguy]

Here you go folks, on a typical master relay or starter solenoid. This is not my fastest scope, but fast enough.

First pic, what happens without any spike protection :

 

Note in the bottom left corner, the voltage per vertical division (vertical scale) 
This is 50V per division.  The negative spike is off the scale, > 250V  negative, and ! 100V positive.

Nasty ! This is what kills electronics. And the length of the transient is long (a few microseconds)  = lots of energy. 
The time divisions are 5uS per division.
These long voltage pulses can rip through the whole wiring harness...

 

 

***************************************************************************************************

Second pic - with a 1N5408 rectifier diode across the coil,

Vertical scale is 10V /division.

negative spike is at -1V. positive spike is at about 33V. Much better ! 
And the spike length is very short- less than 200nS. Most electronics designed for 12V would survive this most of the time.

Not bad. The time divisions are 200nS  (0.2uS)  per division.

 

 

**********************************************************************

Third pic - This is with my recommended 1N5819 Schottky rectifier diode across the coil. 
Maximum excursion is about +22V. Negative excursion ~ -0.2V
and its fast - it is all over within 200nS !

 

 

**************************************************************

 

Fourth Pic- using a  18V UNIdirectional TVS voltage transient diode across the coil. Type 

It is designed to clamp -1V in one direction and +18V in the other direction.

It's designed to be placed on a voltage bus and deal with all nasties.

These diodes are designed for when there is LOTS of energy to be dumped. Alternators can  dump lots of energy if disconnected.

Notice the behaviour- clamp is around 0Volts on negative, and peak about 20V positive. But has substantial ringing. 
Length of the ring is about 1uS. Use unidirectional, not bi-directional. (they come in both types) 

 

*********************************************************************

 

So when to use the rectifier diode, and when to use the TVS diode ?

The TVS diode can be placed across the 12V bus and will clip the positive and negative excursions.

A rectifier diode placed across the bus will only clip negative excursions- the rectifier diode must be placed across the relay or contactor- that is it must be placed across the inductive component that is generating the high voltage spikes.

 

So the summarize :

Place across relay or contactor or solenoid : 1N5819 , or 18V TVS diode (band to positive side)...or  1N4007 (2nd choice - good) or 1N5408 3rd choice.

Place across 12V bus : 18V TVS diode

Place across alternator battery connnection regulator output  : 18V TVS diode.

When using TVS diode, buy the unidirectional, not the bi-directional.  Band to Positive.

 

-glen english

 

 

Edited August 13, 2022 by RFguy

[RFguy]

I'll add that for TVS diodes, it is best that go at the source of the trouble, or on the gear itself to protect the gear.
 

There place where they are  well used is on the gear itself- The voltage spike HAS to pass the diode before getting into the gear. The diode will strip off the spike .

I put them into Microair radios when I come across them.

 

Compare to say the TVS diode just being in the wiring somewhere- the voltage spike might not have to go past the diode, and it wont be effective.

 

 

[nomadpete]

Very nice CRO pics.

Have you looked at the current associated with those spikes?

Personally, although I love protection diodes or TVS, I would be tempted to wire a fusible link or fuse in series with it - although failures are rare, the consequence of a short suddenly occurring across your solenoid would be unpleasant.

(Such addition would have to be monitored)

[nomadpete]

PS

Do they still fit a dkhead diode across the dc supply inside radios? In the old days of CB radio, they were common, and would at least suppress the negative going spikes.

[RFguy]

I have never seen a TVS diode go show except when it wasd grossly over dissipated- IE where I connected a 24V battery across a 18V  TVS.

The TVS is designed for TRANSIENTS (spikes) - that's in the name.. it's not designed to clip off continuous 24V on a 12V rail, for example. 

 

So, I beleive the concern of a TVS going short circuit is unfounded and astronomically unlikely in the service.

 

There is plenty of energy, it's in the coil and you can work this out  = 1/2 LI^2.  some of it ends up in the diode, some of it ends up in high frequency risetime losses.

 

Whatever wiring is connected to it must be capable of blowing the fuse of circuit breaker.  That wont generally be a problem for well engineered aircraft

Yes, the idiot diode is still in most equipment, fortunately. In many cases the idiot reverse protection diode has been replaced with a unidirectional TVS diode because in one direction, it functions as a reverse protection diode and in the other direction, it functions as a good spike remover.

 

 

Edited August 14, 2022 by RFguy

[pluessy]

Eliminate the source of the spike where possible and replace coil relays/solenoids with solid-state versions.

[Geoff_H]

40 Years ago when I started working in power stations we used relays for control.  We recorded spikes of nearly 5kV.  We imposed a 5kV spike tolerance on the new electronic control gear.  We didn't even consider that with the relatively slow switching of electronics a maximum tolerance to 1kV spikes. That was until Siemens forced us to accept the 1kV tolerance.  The electronic control systems worked great.  The moral of the story is avoid relay switching spikes and go for electronic switching as much as possible. 

[Geoff_H]

Sorry just realised it was 50 years ago I started work in Power Stations.  I am getting old and forgetful. 

[nomadpete]

Geoff, switchyard back EMF involves spikes several orders of magnitude greater than any tiddly little 12volt solenoids or starter motors.

 

When a switchyard circuit breaker trips there may be 330,000 volts and thousands of amps at play.

 

Nevertheless we should be protecting our fragile and personally expensive electronics on our aircraft. Especially when it is so simple to do.

[RFguy]

But it is all relative, You don't plumb the switchyard in 14 AWG wire and Thrystors that are the size of a pin head. You plumb in aluminium bar and have thyristors that are the size of hockey pucks or a Discus.

 

The degree of difficulty and potential for damage are the same.

 

 

Edited September 2, 2022 by RFguy

[Geoff_H]

54 minutes ago, nomadpete said:

Geoff, switchyard back EMF involves spikes several orders of magnitude greater than any tiddly little 12volt solenoids or starter motors.

 

When a switchyard circuit breaker trips there may be 330,000 volts and thousands of amps at play

I am talking about the control systems inside the power systems not the 330kV or 500kV systems, they are very specialised systems.  I was the commissioning engineer for the first Eraring 500kV switchgear.  Another item that makes large spikes is the size of the wiring.  The original power stations used 24/48V systems with 7/0.29 wire, modern systems use 1mm or less.  Larger wiring gives larger spikes.  

[Jabiru7252]

On 13/08/2022 at 11:41 AM, RFguy said:

Here you go folks, on a typical master relay or starter solenoid. This is not my fastest scope, but fast enough.

First pic, what happens without any spike protection :

 

Note in the bottom left corner, the voltage per vertical division (vertical scale) 
This is 50V per division.  The negative spike is off the scale, > 250V  negative, and ! 100V positive.

Nasty ! This is what kills electronics. And the length of the transient is long (a few microseconds)  = lots of energy. 
The time divisions are 5uS per division.
These long voltage pulses can rip through the whole wiring harness...

 

 

***************************************************************************************************

Second pic - with a 1N5408 rectifier diode across the coil,

Vertical scale is 10V /division.

negative spike is at -1V. positive spike is at about 33V. Much better ! 
And the spike length is very short- less than 200nS. Most electronics designed for 12V would survive this most of the time.

Not bad. The time divisions are 200nS  (0.2uS)  per division.

 

 

**********************************************************************

Third pic - This is with my recommended 1N5819 Schottky rectifier diode across the coil. 
Maximum excursion is about +22V. Negative excursion ~ -0.2V
and its fast - it is all over within 200nS !

 

 

**************************************************************

 

Fourth Pic- using a  18V UNIdirectional TVS voltage transient diode across the coil. Type 

It is designed to clamp -1V in one direction and +18V in the other direction.

It's designed to be placed on a voltage bus and deal with all nasties.

These diodes are designed for when there is LOTS of energy to be dumped. Alternators can  dump lots of energy if disconnected.

Notice the behaviour- clamp is around 0Volts on negative, and peak about 20V positive. But has substantial ringing. 
Length of the ring is about 1uS. Use unidirectional, not bi-directional. (they come in both types) 

 

*********************************************************************

 

So when to use the rectifier diode, and when to use the TVS diode ?

The TVS diode can be placed across the 12V bus and will clip the positive and negative excursions.

A rectifier diode placed across the bus will only clip negative excursions- the rectifier diode must be placed across the relay or contactor- that is it must be placed across the inductive component that is generating the high voltage spikes.

 

So the summarize :

Place across relay or contactor or solenoid : 1N5819 , or 18V TVS diode (band to positive side)...or  1N4007 (2nd choice - good) or 1N5408 3rd choice.

Place across 12V bus : 18V TVS diode

Place across alternator battery connnection regulator output  : 18V TVS diode.

When using TVS diode, buy the unidirectional, not the bi-directional.  Band to Positive.

 

-glen english

 

 

Howdy. My cheap digital CRO lets me save the screen to a USB stick so I don't have to photograph it. Also, if you put a diode across a solenoid or relay to stop the spike, should not the spike be limited to 0.6 volts?

SDS00006.BMP

[RFguy]

Kevin,  500mV/div or 5V/div (with x10 probe) ? solenoid stand alone with just a cliplead put on the coil ?

5mS/div, you are probably seeing the L/R product of the relay solenoid.  

A rectifier diode across the solenoid should clamp , once it eventually turns on, at ~ -0.7V yeah. There will be some overshoot due to turn on delays.  The unidirectional TVS will try and clamp ~ 1V one way and whatever the clamp voltage is the other way and bounce between .  Bi directional type will bounced both ways.