0
\$\begingroup\$

This is a follow up to this issue: Damaged TPL7407LA Transistor Array (*Updated With Scope Testing Images*)

enter image description here

We are experiencing both reverse polarity and transient voltage spikes in our automotive circuit application pictured above where we've added protection for both situations but are still experiencing aforementioned behavior when an adjacent input to our device is being switched off (oscilloscope input 3 - PARK).

enter image description here

enter image description here

The only feasible possibility that we can consider at this point is the fact that the diodes may not be reacting quick enough these conditions.

Here are the reverse polarity and TVS diodes we are using, respectively.

Reverse Polarity Protection Diode

TVS Overvoltage Protection Diode

Curious to hear what you folks have to say on this topic.

\$\endgroup\$
  • \$\begingroup\$ Turn, park & brake signals are slooooow. Throw an RC filter on each line. \$\endgroup\$ – brhans Sep 29 '18 at 2:02
  • \$\begingroup\$ @brhans Can you please explain? I guess I don't understand why our existing protection circuitry wouldn't resolve these issues let alone reasoning to add an RC Filter? If you have any good resources that would help me understand, that would be greatly appreciated. \$\endgroup\$ – Jacob Anderson Sep 29 '18 at 2:12
  • \$\begingroup\$ Add a series RC snubber shunted to device and to 0V and shorten all grounds to < 5cm incl probe you got resonance at 1.5MHz e.g.100 Ohms and 0.1uF \$\endgroup\$ – Tony Stewart Sunnyskyguy EE75 Sep 29 '18 at 2:31
  • \$\begingroup\$ If turn signal is 1m long or 1uH then capacitance at 1.5MHz is on the order of 10nF. at 10 Ohms so series R from diode in transition running on might be 1 Ohm for a Q of 10. Meaning unless you define your reactive loads and reactive cables, ringing is to be expected with a mismatched switched load \$\endgroup\$ – Tony Stewart Sunnyskyguy EE75 Sep 29 '18 at 3:04
  • \$\begingroup\$ The TVS has a maximum clamp voltage at 22V. Are you sure your input can tolerate that? Your diode is not Schottcky and will have a significant junction capacitance, you may find this is problematic, and it may even be the cause of some of the oscillation you see. \$\endgroup\$ – Jack Creasey Sep 29 '18 at 4:49
1
\$\begingroup\$

In the first waveform I see a single spike related to 'Park' being switched OFF.

In the second waveform I see an intense ring wave, cause unknown.

I spent 15 years working at a surge suppressor plant and the hard lesson was that no matter how expensive or sophisticated the suppressor was the best suppression was right at the device itself. There is a nasty equation that haunts us all.

For every foot/300mm of distance from the suppressor the voltage rise could be 100 volts.

That meant some surge suppressors had to be at the point of use, including their ground connections, and in some cases 100uH to 1mH inductors in series.

The point I am making is that for any suppression to work at the scale of switch's and relays the suppression devices had to be at the point the surge was generated.

For every inch/25.4mm the suppressor was from the surge creating device (including ground return) the voltage would rise by 8.5 volts. If high enough some ringing effects would occur as there was nothing close enough to dampen them.

Whatever you use to clamp surges need to be right at the point the surge is created, including diodes, TVS diodes, capacitors, inductors. Else your wasting your time and money and the frustration builds up.

\$\endgroup\$
  • \$\begingroup\$ I guess for lightning and it's effects coupled into cable you need a big surge suppressor in the sky …..would this be a Cloud based TVS? \$\endgroup\$ – Jack Creasey Sep 29 '18 at 4:41
  • \$\begingroup\$ @JackCreasey. Lightning rods approximate that. \$\endgroup\$ – Sparky256 Sep 30 '18 at 22:01

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.