Author Topic: Square wave across solenoid without anti-parallel diode overshoot  (Read 1403 times)

0 Members and 1 Guest are viewing this topic.

Offline ricko_uk

  • Frequent Contributor
  • **
  • Posts: 536
  • Country: gb
Hi,
I have a solenoid connected to Vdd on one side and to ground via a transistor and across it the classic antiparallel diode. When I switch off the transistor the voltage "overshoots" the rail by the diode voltage drop.

Two separate questions:

1) how can I make it so that there is no overshoot (i.e. the equivalent of an ideal diode)? Even putting a resistor still causes the overshoot

2) the current in the solenoid tends to maintain its direction, does it mean that the magnetic field is always in the same direction (when the transistor is on) and just no magnetic field at all after the transistor has been turned off and the energy recirculated and dissipated? Or does it briefly reverse polarity?

Many thanks :)
 

Offline langwadt

  • Super Contributor
  • ***
  • Posts: 2294
  • Country: dk
Re: Square wave across solenoid without anti-parallel diode overshoot
« Reply #1 on: June 01, 2020, 08:44:25 pm »
The current and thus the magnetic field is in the same direction while the energy is energy recirculated and dissipated.
The higher you allow the flyback voltage to be the faster the solenoid will release, that's why for things that need to be
a fast like fuel injectors you cannot use an antiparallel diode
 

Offline duak

  • Super Contributor
  • ***
  • Posts: 1009
  • Country: ca
Re: Square wave across solenoid without anti-parallel diode overshoot
« Reply #2 on: June 02, 2020, 05:20:28 am »
Here's a couple of ways to limit the overshoot voltage:

1.1:  instead of the anti-parallel diode connecting back to the power supply voltage, connect it to a voltage about a diode drop lower than the supply vlotage.  Note that this may not be very efficient as this second voltage will have to be able to absorb the recirculating current from the solenoid.

1.2:  have the transistor switch clamp the overshoot when releasing the solenoid.  ie., when the voltage on the transistor's collector or drain reaches some predetermined level, have the transistor turn back on again but just enough to hold the voltage down to where you want it.  A zener diode between the collector and base (or drain and gate) is one way.  Note that the transistor will have to dissipate a good portion of the energy stored in the solenoid.

 

Online Gyro

  • Super Contributor
  • ***
  • Posts: 6284
  • Country: gb
Re: Square wave across solenoid without anti-parallel diode overshoot
« Reply #3 on: June 02, 2020, 09:49:43 am »
Is the slight overshoot a problem for you? As long as you're not driving the relay directly from a GPIO on a micro (where it might stress a protection diode), a driver transistor should easily accommodate it in its Vce rating.

If you are looking for faster solenoid release, then using a series combination of a diode and zener (or bidirectional TVS) across the coil will result in faster collapse of the magnetic field. This will also overshoot more (Zener+diode drop) which needs to be accommodated in the transistor rating. A properly sized RC snubber can also be very effective, but again you will see some overshoot.

The problem with putting a clamp across the switching transistor instead of the coil is that the current path for the back EMF then has to pass round a larger loop (through the supply rails). This can cause more of a noise spike, compared with putting it directly across the coil.

Again, the question is, does this well clamped overshoot really pose any problem to you?
« Last Edit: June 02, 2020, 10:02:47 am by Gyro »
Chris

"Victor Meldrew, the Crimson Avenger!"
 

Online Zero999

  • Super Contributor
  • ***
  • Posts: 15121
  • Country: gb
  • 0999
Re: Square wave across solenoid without anti-parallel diode overshoot
« Reply #4 on: June 02, 2020, 12:43:25 pm »
Just a couple one question: why?
Normally it's better to interrupt the current through the solenoid, as quickly as possible, to switch it off, as fast as possible, which will result in a high voltage overshoot, but isn't a problem, as long as the transistor can withstand it.

Using another transistor for the high side, forming a half bridge, will virtually eliminate the overshoot.
 

Offline Le_Bassiste

  • Regular Contributor
  • *
  • Posts: 198
  • Country: de
Re: Square wave across solenoid without anti-parallel diode overshoot
« Reply #5 on: June 02, 2020, 05:05:58 pm »

1) how can I make it so that there is no overshoot (i.e. the equivalent of an ideal diode)? Even putting a resistor still causes the overshoot


well you basically answered your own question: use an ideal diode  :-DD
this may help you:
1) look for synchronous rectifier drivers.
2) use a synchronous buck controller to drive the high-side and low-side MOSFETs.
3) use a (for example , MOSFET-) halfbridge and drive it with complementary signals.
all three methods use the same principle: "short-circuit" the freewheeling diode during all the times when the transistor that is used to switch the solenoid ON is in its OFF state.

as others already have pointed out, this will vastly slow down the energy decay in the solenoid.
 
 


Share me

Digg  Facebook  SlashDot  Delicious  Technorati  Twitter  Google  Yahoo
Smf