Using flyback/supression diodes on solenoid coils

[jashom1]

Fellow gurus,

We are combating issues with delays on a hydraulic proportional control system, and I have been informed (by a 3rd party) that the Plus+1 system does not like the output drivers connected to solenoids that have inbuilt suppression diodes.  I can't seem to find any information suggesting whether it is better to have them fitted with supression diodes or not.  Does anyone know an official answer to this?

Thanks!

John.

[Marbek_Elektronik]

You can use the PWM-Outputs, because they are "push/pull".
Thee are two MosFets with build in diodes.

[Marbek_Elektronik]

I have opened one old controller.
There is an Infineon PROFET® BTS 728 L2.
This are a few basic functions:
• Improved electromagnetic compatibility (EMC)
• Fast demagnetization of inductive loads

In datasheet, I read, that the maximum inductive load is 20mH @4Ampere or about 400mH at 1 Ampere).
We have driven a lot of relays at this pins and there was no damage.

I will ask helpdesk about this.


(In case of build in flyback/supressor diodes, I see no problem)

[oiltronic]

The official answer is in "Output Configurations for PLUS+1® Controllers Technical Information Manual", doc # 520L0922

(Dear Danfoss:  Please make these documents easier to find.  The new website is horrible for selecting products and finding technical info.)

Diode snubbers on a proportional coil?  If built-in, it doesn't sound like a proper proportional coil or one designed for your performance requirements.  Imagine the current speeding through a coil, then the circuit suddenly opening.  The current wants to keep going (inductance), so it diverts through the best path, which is the diode.  The current keeps merrily looping through the coil & diode and holding the solenoid in position until the energy is finally dissipated enough to let the solenoid go where it wants.  (In larger relays or contactors diodes can be a bad thing, because it means the contacts have a little gap between them for more time, which means a longer arc time.)    See also  Wikipedia's article on snubbers.

I have always relied on the snubber circuitry built into the Plus+1 microcontrollers & i/o modules, and I have yet to damage one.

[Marbek_Elektronik]

Thanks for this document.
They write, there is a diode inside. (DOUT)
But I have measured and there is no diode.
The only way, they minimize the negative inductive voltage, by switching of the output, is , they switch it slowly.
I think, if there is a big inductive load, this type of outputs may be damaged
and also if there is a big (EMV) coppling from one cable to another, may be the output will be damaged.

I prefer to use an optional diode to each relay. It is also good to minimize EMV. It is good to put diodes as near as possible to relays.

Tip: But be careful: They make the switch-off-time longer and this might shorten the life-time.

Mobile machines are sometimes as critical as industrial applications, even, if there are a lot of motors and relays.

[oiltronic]

They write, there is a diode inside. (DOUT)
But I have measured and there is no diode.

I think what you meant to say was, "I have measured and my particular measurement technique did not reveal the presence of a diode".  There are indeed diodes inside, and they actually can't avoid them.  They are probably talking about the intrinsic body diodes of the FETS.  These will have different characteristics than a discrete diode, and these characteristics will differ depending on the model of FET they are using.  But they are good enough for flyback protection.

One way to satisfy your curiousity could be to use an oscilloscope to observe the voltage at a relay contact driving a solenoid, then compare that with a measurement at the microcontroller output driving the same coil.

The only way, they minimize the negative inductive voltage, by switching of the output, is , they switch it slowly.

I don't think so, as that would eventually overheat the FET.  Circuit designers usually want their power FETs either 100% on or 100% off, and quickly.

[Marbek_Elektronik]

Sorry, I think you are wrong.
I am also a hardware designer and I used the same technique to switch off an inductive load without free-running diode.
The highside switch they use is the  Infineon PROFET® BTS 728 L2
simply illustrated: if output goes below ground, they give more voltage to the gate. If the output is to high, they minimize voltage at the gate.
The effect is, to drive slowly a MosFet down, so that output not goes to a big negative voltage.
Circuit designer don't switch the MosFet 100% on or off. The switching is regulated. If you have a resistive load, it is switched off very fast.
If it is an inductive load, it is  regulated very slowly. I will measure this and give you the time I measure.
The diode in MosFet does not helps here.
If you have a PWM-Output, you have in every MosFet a diode. If one MosFet was active and is shut off, the current is running through the diode of the other MosFet.
I prefer to use "push-pull" mode in this case, because if switching off one MosFet the current will at first run through the diode of the other MosFet and after a very short time
of e.g. 1µs this other MosFets is active and leads the free running current with near zero voltage drop and so they minmimize heat in MosFet (Synchronous rectification).

I have measured by my testing a DC-voltage ov -5V at the output. And don't want to test, by which voltage it will be damaged. But it shows, there is no diode from ground to output.

[jashom1]

Thanks a lot guys, you have given me enough evidence to try some coils with no flyback protection.  It's all a learning curve! 

[Marbek_Elektronik]

You can watch the voltage on oscilloscope to check your application.

[oiltronic]

Hey there jashom1, also ensure the PWM frequency is set correctly.  Forgot to mention that earlier.

And here's an old article illustrating a few things about coil inductance.

[pinias]

well,

this is just what I know after finished a master in power electronics some years ago.

few things about this:

1.- the PROFET BTS 728 L2 is not just a MOSFET is more like an IC, if you want to see a MOSFET datasheet I will recommend to look for the IRF1407, that is a datasheet for just a MOSFET.

2.- remind that a MOSFET is a solid state switch which means you move electrons for switching ON and OFF not mechanical parts therefore manufacturers try to minimize the avalanche for ON and OFF by different techniques, so the ON delay and the OFF delay can not  by controlled the IC.

the On state and the OFF state is a natural time for working. what you can do is switching ON and OFF with PWM as fast as posible which means you will create more heat which depends on application I have gone for up to 60 KHZ PWM frequency at 60 Amp as constant load.

3.- to set to ON a MOSFET you apply voltage to the gate, to set to OFF a MOSFET the voltage must be below the threshold voltage

4.-I use flywheel diodes in my applications and so far the response time is ok.


I hope this gives you another perspective

Regards

Antonio

[240glt]

Just a note regarding the release time of a large relay:

I had some problems with Gigavac contactors that was welded, but the make- and brake-currents was well below the limits. So I checked the release time with the oscilloscope. First with the coil driven directly from output 2.33 on a MC090-010 (0,5A high side driver, not a PWM output). The release time was about 25ms. Then I put a relay in between the output and the coil, to be sure that there is no suppression diode active. And the release time dropped to about 8ms. I have attached the two scope traces. Blue trace is coil voltage, and red is voltage over the relay contacts (to see when the contact was released).

Note that the gigavac has inbuilt suppression circuit, and should not be used with an external suppression. Something that I had not noticed before...

[Marbek_Elektronik]

@240glt: please, what relay have you used? datasheet?
Ok, do you think it is bad, if voltage is swinging from negative to positive in case of supression diode in relay?
I think, the voltage of supression diode is an important value.



Now, I have made some measurements and checked the datasheet again:
To your questions:
1. you can connect inductive loads, see datasheet page 9: 4A at 20mH or 1A at 400mH.
So in most cases of small relays, there will be no need to add an additional diode.
If the energy becommes bigger (1/2 L I²) the heat in IC will go to high. So: use an diode from ground to output.
Or if you often switch off the inductance, it is better to use a PWM-Output.

You see the voltage in Figure 2c by switching off: The negativ Voltage is about -25V at 12V Vdd or -13V at 24V Vdd.
The IC simulates an zerner-diode in this case. You see: the internal MosFet ist regulated between 0% and 100% !

EMV voltages and currents will also be limited by this effect, if they are not to big.

I hope, now are all questions answers!

One notice: I have used (1.) no free-running diode (2.) 4148     (3.) 1N4007
Between 4148 and 1N4007 there was no difference. 1N4007 is fast enough.
Choose a diode with min. the same current as load.

If you use no freerunning diode, switch off time is about 2,5ms,
with freerunning diode it is 10ms in my case.

[240glt]

Marbek_Elektronik: It was an Gigavac GX11 (or GX14, can't remember, but they are very similar). The datasheet is here: http://www.gigavac.com/pdf/ds/pp/gx11.pdf

Interesting output driver! It would be interesting to test that output with my gigavac relays. Your relay had four times faster release time with that output driver compared to just a diode? Nice!

What controller and output pins are using that output driver circuit? The output I tested was definitely just a free-wheel diode. But it was a 0,5A output that I used.

[Marbek_Elektronik]

Thanks,
CoilPower at GX11 is about 8W, so the current is also much bigger then my relais: Schrack RT424012. It is 12V and 360 Ohm, contact max. 2x8 Ampere = little relay, coil-power 0,4Watt.
So , I think GX11 have much more energy in the coil, which will heat the MosFet by switching off.

I have used a MC050 c1p31 DOut, max 3 Ampere.

I have an old MC050, which has an hardware error. I take it, to test outputs.
I watch voltage and current via oscilloscope and calculate the power loss in controller, c1p31.
It must be less then about 0,2 Joule, see my datasheet before!

For example:
the coil might have 24V and 300mA = 7,2Watt
if you measure:
37Volt between Vdd and negative Voltage of coil, and 150mA (average) and 30ms (simplified)
The MosFet-IC inside MC050 will get more warm by energy of 37V*0,15A*0,03s = 5,55 Watt * 0,03s = 0,166 Joule
That is a lot for the IC. A little bigger relay will damage the IC!

Another way to test might be to buy this ICs, Danfoss used, and test it alone and then conect your relay to controller.

At last: Danfoss controller are very good. But sometimes I don't found the maximum ratings of the outputs for inductive and capacitive loads.