How to correct relay noise

sorasit46 · Joined: 07 May 2007 Posts: 69

16F876A
I'm use output_pin to drive relay for supply dc motor. Some time (not any time) relay work on-off-on-off it has noise to cause program hang/error. I try correct by
1. use cer103-104 connect motor pole, not work
2. use cer103-104 connect relay contact, not work
How I can solve this noise?

regard

PCM programmer · Joined: 06 Sep 2003 Posts: 21708

See these posts about using relays with PICs:
http://www.ccsinfo.com/forum/viewtopic.php?t=30336
http://www.ccsinfo.com/forum/viewtopic.php?t=30001
http://www.ccsinfo.com/forum/viewtopic.php?t=23634

sorasit46 · Joined: 07 May 2007 Posts: 69

Matro · Guest

Just use MOSFET transistors instead of relays.

Matro.

sorasit46 · Joined: 07 May 2007 Posts: 69

Ttelmah · Guest

While MOSFETs are a good way to go, You need similar protection from induced currents for these, to a relay. The relay adds the problems of possible contact bounce when switching on, and arcing when switching off, but in both cases, when driving an inductive load like a motor (and worse, a motor, presumably with mechanical inertia, so there will be a generator effect when power is removed), you _must_ design the power circuit to handle these properly. The MOSFET, will make certain things worse (the sheer switching speed generates more RF problems, unless the gate current is very carefully controlled), and also thre is a tendency for people to 'rely' on the internal diode created by the structure, which in general has much _lower_ power ratings than the device itself, resulting in a common failure mode (some special devices like Hitachi power MOSFETs, are explicitly manufactured with internal diodes that match the switching elements, but unless this is known and verified on the data sheet, external diodes must be used).
Remember that there is energy contained in an inductor, when you switch 'off', this must go _somewhere_, and it is down to you to ensure that the route it takes can handle this without creating problems.
Look at Microchip's AN898.

Best Wishes

Guest · Posted: Sat Mar 22, 2008 4:36 am

Got PIC · Joined: 21 Mar 2008 Posts: 10

Is the relay connected directly to the OUTPUT Pin on the CPU?
You may be drawing way too much current from the CPU pin. You should alway isolate the CPU Pin from what ever you are driving, especially if you are driving an inductive load like a relay. The inductive kick back alone could cause you lots of damage/problems…

baltazar_aquino · Joined: 16 Mar 2008 Posts: 27

sorasit46 · Joined: 07 May 2007 Posts: 69

Matro · Guest

I often control MOSFET directly with a PIC. But never for DC motor driving.
But here if your MOSFETs just control the spin direction and not the speed, I don't really see a problem. But the motor control will be not easy during inertia phases.
Matro.

Ttelmah · Guest

There are 'logic level' MOSFETs, which have lower gate drive voltages than nrmal types, which can be directly driven by a PIC. However, even with these, when driving at high speed, the actual output drive ability of the PIC, will result in very slow switching, and may give problems. The typical 'power' MOSFET, has a very significant capacitance present (often over 1000pF). To drive the gate through the transition region quickly, requires a suprisingly high current capability in the driver involved.
On the existing circuit, do you have a diode to trap the energy in the relay coil, when the transistor switches off?. If you look at the 'equivalent circuit', of generic 'logic driver' IC's (like the standard ULN200x family), you will see that in every case, they have a set of diodes from the outputs, to one pin of the IC, which is meant to connect to the +ve rail, to trap the switch off energy in coils. This is _essential_. Without it, if you drive a transistor, with a logic pin, the 'odds' are that when you switch off, the voltage on the junction between the transistor and the coil, will rise to a massive level, and often breakdown through the transistor to the driving logic pin, and destroy the chip driving the pin...

Best Wishes

baltazar_aquino · Joined: 16 Mar 2008 Posts: 27

Power MOSFETS are inherently more expensive than bipolar transistors. However, you get less quiescent power consumption with MOSFETS , if cost is a concern, you can also use power transistors in double push-pull fashion to create a full bridge. This will give you direction control also of the dc motor(inject complementary signals at the bases and you can have current direction change) - and you get rid of the relay problem. Do you have inductive line filters in your supply rail? In some cases, capacitor alone cannot damp the low frequency noise -in such a case you have to use a filter coil. But still you have to use clipping diodes at the motor terminals the opposite leads are tied to the rail to protect your transistors. With proper filtering , maybe relay is just sufficient for your application.

sorasit46 · Joined: 07 May 2007 Posts: 69

baltazar_aquino · Joined: 16 Mar 2008 Posts: 27

you can use the complementary pair NA51W/NA52W . Or the equivalent 2N6121/2N6124 pair. But I dont' know exactly your load current. You just mentioned the relay rating (8A/24V). With these transistors, you can safely switch 3A (4Amax) up to 40V(max) VCE. If the rating is not enough for you, you can use the whooping 2N3055/MJ2955 pair (TO-3 package). In both cases you have to buffer first your PIC pins with smaller signal general purpose transistors (complementary pair also).