H Bridge Noise Protection

Lightfoot

My Plan Of Action:

I have a tortoise switch machine (a gear drive electric motor for moving a model railroad switch or turnout). I want to construct a reliable H drive circuit that toggles a relay for half the bridge and toggles off for the other. I attempted this before and experianced noise (some crosstalk between the relay and tortoise coils or something) issues. I made an H drive out of 2N3904 and 2N3906 transistors (The current drawn from both the relay and Tortoise is around 75ma (dont know exact off hand)). An SX chip will supply a signal for the base's of the transistors. The relay is for controlling an electric current for a model railroad siding (When the switch is open (pointing towards the siding), the train will run into the siding, when it is closed (pointing towards the main line), the current will shut off isolating the train). I have a schematic of what I did before below (I made it on the fly so don't be suprised if it contains errors):



Thank You All [noparse]:)[/noparse]

▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
Well well, I'm seeing things, three of them.

-Stanley Blystone

Post Edited (Three of Them) : 6/8/2006 10:03:12 AM GMT

PJAllen

· So you're not going to vary the speed of the motor?· Just full-blast in one direction or the other?

· Or are you figuring on PWMing A, B, C, D?· If you do that, then you'll be PWMing the motor AND that relay (coil) you have in parallel with it.· That could be interesting.

Lightfoot

There is no reason to PWM a tortoise . It is geared so the turnout opens and closes at a realistic speed (HO scale speed). Instead of a 12 volt relay I will select a 5 volt and wire it to the SX pin that activates the appropiate H bridge half. Still, with this design, I am concerned about noise. Would a capacator along with the flyback diode (across the coil) make a good supresser?

Schematic:



▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
Well well, I'm seeing things, three of them.

-Stanley Blystone

PJAllen

Tortoise -- What do I know from a Tortoise?·

As shown, you ought not attempt to drive a relay coil directly from an SX-Pin.· The only noise, I believe, will come from the motor, not the relay coil itself.· Maybe the relay contacts will make a glitch if they switch some heavy current, which you could dampen with a Cap.· You might consider using opto-isolators (4N32) between the SX and the H-Bridge (using separate, isolated power supplies.)·

Lightfoot

I don't see why driving a relay from an SX pin is harmful as long as the coil sinks less then 20ma, or however much an sx pin can safely supply. If I run into trouble here I can just use a transistor switch. I can understand a relay glitching from high currents. Where should I put the cap and what value should I use?

Thanks [noparse]:)[/noparse]

▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
Well well, I'm seeing things, three of them.

-Stanley Blystone

Post Edited (Three of Them) : 6/9/2006 5:16:32 AM GMT

pjv

Hi Moe;

You are partly right and partly wrong.

Sure, the SX pin can drive a small relay directly, as you said, up to 20 ma or so at 5 Volts, but it is when the relay (inductor) is turned off that the trouble begins. An inductor with even 1 mA of current flowing through it, will, when switched off rapidly, generate a voltage spike proportional to the current flowing and the inductance of the coil. The "glitching" will occur regardless of the amount of current; it need not be "high current", its intensity will vary with current. One way to mitigate that spike is to provide an alternate current path for the spike to dump into, and that you have handled correctly in your circuit with a "catch" diode.

Regarding the capacitor, I have no idea why you have included that. I believe it's presence to be detrimental as it will increase loading on the SX pin, and slow down transition time. Unless you have some particular reason, I'd get rid of it.

Cheers,

Peter (pjv)

Lightfoot

I always put flyback diodes on inducticve loads. Without them, the spike will kick a transistor's butt. I put the diode in because I assumed the transistor in the SX is not internally protected. As for the cap, I should have no noise problems if I just have a relay and flyback diode?

▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
Well well, I'm seeing things, three of them.

-Stanley Blystone

PJAllen

Jameco (and others, I'm sure) has DIP-relays.·

Their part number 138430CK are pretty neat: 5V relays with a·coil resistance·of 500ohms, that means only 10mA drive current -- and they have an internal free-wheeling/catch (protection) diode.· The output is rated 1/2A, 100VDC (it's DC, because there isn't "creepage & clearance" enough for AC with this package.)

pjv

Hi Moe;

For running the relay, the diode is all you should need. In fact, I believe the capacitor to be undesirable unless it were very small. For the purist it would have the beneficial effect of lowering the frequency of any self-resonance the coil might have....... but that's a bit of a stretch as then we should also consider the relay's self capacitance, inductance of leads etc. All way out of the ballpark for practical situations.

Cheers,

Peter (pjv)

Bo

Yo,

I am working with an EV200 Czonka by Kilovac. It's a power electromechanical relay which operates at 9-36VDC and operates on 1.7W at steady state. Would anyone happen to know if a 1N4002 (or 4001, etc.) diode would be an appropriate fit to dissipate back EMF voltages when the relay is opened? I've attached spec sheets of both components.

thanks,
Bo

pjv

Yo, Bo;

I'm a bit confused by your question. The relay specs state that there is no back EMF generated to the outside world, so if that's the case you would need no diode....... we ARE talking about the coil right ??????

For greater certainty however....... if you did want to use an external coil quenching diode, that diode would need to handle the current that was flowing at the instant the coil was switched open. Assuming 12 volts at it's rated 17 watts is about 1.5 Amps. According to your sheets, a 1N400X is rated for a 8 mSec surge at 30 Amps.

So it should handle it.

Cheers,

Peter (pjv)

Bo

Hi Peter,

I'm designing for the czonka w/o the internal coil economizer, so it would require an external flyback diode. I'm not sure where you saw a spec for 17W@12VDC when the relay opens, but I guess the way to find this out would be to take some oscilloscope measurements. Thanks though.

regards,
Bo