Hi. Will it be ok to use this diode 1n5711as a flyback diode on a 12V relay ?
If you had to choose the best diode for a flyback use, which one woul it be ? no matter the cost.
Thank you.
If you had to choose the best diode for a flyback use, which one woul it be ? no matter the cost.
Thank you.
15 milliamp Schottky? If the relay draws more that that it may be undersized. It’s job is to to discharge the inductive energy and prevent over stressing the driver transistor/IC, and a generic 4001 is just as good. Don’t overthink it. High speed or low Vf doesnt really buy anything here, and a 1 amp diode can take more of a whallop than a 15 mA one.
Sometimes I connect the flyback protection across the switch transistor instead of across the relay coil. Since the transistor is, after all, the thing which can be damaged by high voltage during the flyback event.
In this example I'm using a transistor rated Vce_max = 300V, and clamping its Vce to 75 volts. Very very safe.
_
In this example I'm using a transistor rated Vce_max = 300V, and clamping its Vce to 75 volts. Very very safe.
_
Attachments
Thank you for this Mark I will probably copy your technique. Diode wise I think I will go for the standard schottky 1nxxxx, nothing to gain it seems ?
But diode speed isn’t important. The coil L-R time constant is HUGE compared to Trr. Don’t sweat it.
Thank you all of you guys. I like to have only one part stocked (it allow me to have cheaper prices if I order 10 or more).
If you had to choose a single schottky for low voltage winding, 12v relay, I was going to stock some SB540 but it might be too low voltage for the spike on relay coils. I am thinking about one low vf 75 or 100v schottky to do it all.
So if you were to choose only 1 schottky which one would it be ? is it even smart to be this cheap and stock only one schottky ?
If you had to choose a single schottky for low voltage winding, 12v relay, I was going to stock some SB540 but it might be too low voltage for the spike on relay coils. I am thinking about one low vf 75 or 100v schottky to do it all.
So if you were to choose only 1 schottky which one would it be ? is it even smart to be this cheap and stock only one schottky ?
Last edited:
1n4002. It has to be good for the relay current (~200mA+) and the relay voltage (12V+), and a fast diode is a bad idea. There is nothing fast about relays. This is an application where you use the most common thing so that you do not create a sourcing issue. Now if you want a PWM flyback power supply that switches at 100KHz+, then yes, fast is good but not a tiny signal diode.
No need to go overboard: a relay rarely draws more than a few 100's of mA, and even a humble 1N4148 can take 4 amp of Ifsm, which is the relevant rating for such an application.
In principle, there is no need for speed: the only important parameter would be the forward recovery, but any diode made in the last fifty years would be more than suitable, HOWEVER: even if the reverse recovery is irrelevant under normal conditions, it can become important when things go astray.
When the relay is chattering, it is possible that the control transistor turns on before the coil is completely discharged, and it then sees a dead short for a few µs if the diode is a slow one.
This is no problem for a single event, but if it repeats 50 or 100 times each second, the transistor and diode will overheat and eventually die. That is something I have seen in some communication equipments.
With a 1N4148, there is no such risk.
Note that any diode will slow the release time of the relay. If that is important for you, you should opt for Mark's solution, or something equivalent: a varistor, a diode connected to a higher voltage rail or a suitably chosen RC circuit
The diode prevents the spike, it doesn't see a voltage spike, it just steers current. I'd choose something rated at 1 or 2A or so, maybe 40 to 60V,
you'll find the pulse current ratings for ones like this are 10's of amps and thus pretty robust for various applications.
I have a voltage controller, 12V relays, 1N4148 diodes, has held up more than 20+ years.
Relays are 30A DC or so, pretty big coils compared to audio relays mounted on PCBs..
Easily available, and reliable.
Relays are 30A DC or so, pretty big coils compared to audio relays mounted on PCBs..
Easily available, and reliable.
Attachments
Hi . My few cents to add. Diode in parallel to relay winding must withstand two things : relay operating voltage , and relay current . So for 48v 100ma relay as example you need to choose diode with more than 48v reverse voltage and more than 100ma current .
Regarding that circuit with D9: the diode needs to be across the coil, not the transistor. Placing it across the transistor protects nothing and puts the diode itself at risk of being destroyed.
The purpose of a freewheel diode is to provide an alternative route for the coil current - if the current in the coil is forced to drop rapidly to zero it will fight back generating whatever voltage is necessary to keep that current flowing, which can be 100's or 1000's of volts.
Thus the diode is placed backwards across the coil.
The purpose of a freewheel diode is to provide an alternative route for the coil current - if the current in the coil is forced to drop rapidly to zero it will fight back generating whatever voltage is necessary to keep that current flowing, which can be 100's or 1000's of volts.
Thus the diode is placed backwards across the coil.
First thank you all I really understand now. May I ask what is the technique that provide the lowest EMI noise ?
Naah, the only thing that flyback can hurt is the switch (transistor). A 75 volt zener across a transistor rated for 300 volts, keeps everything within spec. Self's power amp book covers this, see Figure 24.32. His accompanying text above that Figure says
BTW, the schematic snippet shown in post #14, has been built and tested in real life. It's a small piece of the Azul headphone amp, which I demonstrated at the Burning Amp Festival in 2019. It worked nicely and attendees who listened to it, liked what they heard.
_
This is a good point to stop and consider exactly what we need to do: the aim is not 'suppress all back EMF' but rather 'protect the transistor'.
BTW, the schematic snippet shown in post #14, has been built and tested in real life. It's a small piece of the Azul headphone amp, which I demonstrated at the Burning Amp Festival in 2019. It worked nicely and attendees who listened to it, liked what they heard.
_
Last edited:
A very interesting discussion but I did some measurements some time ago with a fast oscilloscope and I saw that if you want very little noise propagated on the power supply, the circuit with a single diode is not enough. The final circuit has become quite complex with many components so that no noise is visible on the power supply.
The zener diode solution provides for a faster release of the relay contactor as it allows the relay coil to breath a bit more freely when it commutates than with a freewheeling diode directly in parallel with it.
The schematic shown above is sub-optimal in the way the zener is connected as it now depends on how stiff the supply rails are, that is because the rails are the ones doing the clamping action of the overshoot and I would be worried with noise shot into to supply rails.
A zener + diode (+ optional R) series circuit directly in parallel with the relay is a better solution as in this case the supply rails will now be excluded from the clamping action as it is now kept local around the relay only.
If the turn off release time is not critical then there is no need for a zener solution, am suspecting another potential problem using zener might be the EMI emission as it allows for a larger voltage spike, therefor a freewheeling diode is in most case a good enough solution.
The schematic shown above is sub-optimal in the way the zener is connected as it now depends on how stiff the supply rails are, that is because the rails are the ones doing the clamping action of the overshoot and I would be worried with noise shot into to supply rails.
A zener + diode (+ optional R) series circuit directly in parallel with the relay is a better solution as in this case the supply rails will now be excluded from the clamping action as it is now kept local around the relay only.
If the turn off release time is not critical then there is no need for a zener solution, am suspecting another potential problem using zener might be the EMI emission as it allows for a larger voltage spike, therefor a freewheeling diode is in most case a good enough solution.