Hey all,

I'm actuating a bunch of dc latching solenoids for irrigation that were once part of battery timer units. I've hooked up arduino -> shift registers -> l293d and have one solenoid per channel.

Solenoids are 9-12VDC , 5-10ohm (can't find spec. sheet but similiar solenoids are around this resistance). I'm running them at 12V because there will be large voltage drop on some of the irrigation lines.

I was getting all worked up about the l293d overheating (some really hot and I burnt one) and was planning on piggybacking 2 extra l293d on each to get up to 1.8 A per channel , then I realized that these solenoids are normally powered by battery, and that the click that they are making when I have them hooked to a 1.2A power supply is way louder then the tiny click I hear with the 9v alkaline battery turning them on/off .

A. My local electronics store only has up to 1/2 watt resistors, so I was wondering if I put a bunch of 1/2 watt resistors in parallel (20 maybe?) if it would be sufficient to limit the current to the solenoid so I don't need the extra l293d's.

B. If so, what ohm value should I start with. Does 20x 1k ohm 1/2 watt sound right? I can either take off / add them and make sure the l293d isn't getting warm.

C. All the solenoids have a common ground (terminal) so will putting current limiting resistors on this one line fix all of them (I'm only ever powering one at a time) , or am I completely wrong about doing things this way?

Any help is much appreciated,
-tekgen

 

More than likely, the problem is you're running 9V solenoids with 12V.

 

Thanks for replying, sorry I read that and forgot to include they are rated 9v-12v

 

I really don't like the idea of using resistors for this, but if you do, you only need a couple ohms at around 3W per solinoid. Any more than that, and the solinoids would probably not engage.

 

I really don't like the idea of using resistors for this, but if you do, you only need a couple ohms at around 3W per solinoid. Any more than that, and the solinoids would probably not engage.

No harm in trying it? Is there?

I'm open to other suggestions though? The power supply isn't an issue as I can always find one with higher current , its just the l293d's. I could make a proper H-Bridge with 10A mosfets , but with the diodes and everything it will be bulky and complex and probably more expensive than just 3 l293d's stacked together. The resistors are the simplest/cheapest route I can see right now, is all.

Remember that the solenoids are normally driven from a 9v alkaline battery. How much current could they actually need then ? 400mA tops? the battery only gets kinda warm when using the timer circuit.

 

I'm open to other suggestions though?

You could consider testing a lower power supply voltage. Bring the 12v down to 9 with a linear regulator and see if your relays still click, they should, and if now the IC stays cool. This route could be easier than putting a resistor on each line. You'll probably need a heat sink on the regulator, but again there'll only be one.

And BTW, something doesn't add up. A 9v battery won't supply much current at all for very long, and its voltage will sag to 7v or so quickly if loaded. So the solenoid shouldn't really draw more than 100mA, IMHO. You should check this to make sure you don't have something else going on.

 

We had a similar situation with the electric tape brakes on a punched tape reader (yes, very old tech). The brakes were not stopping the tape fast enough resulting in read errors. The voltage to them was increased. Then they started getting too hot. The answer was to put a resistor in series with a cap parallel to the resistor. When activated, the cap gave the brake coil a good hard jolt, which stopped the tape, then charged up as the resistor started to limit the current. Worked like a charm. You will need to determine what current is needed to hold the solenoid on and using ohms law, determine a resistor value that will give a little more than that with 12 volts applied.

 

Thanks for all the responses!

You could consider testing a lower power supply voltage. Bring the 12v down to 9 with a linear regulator and see if your relays still click, they should, and if now the IC stays cool. This route could be easier than putting a resistor on each line. You'll probably need a heat sink on the regulator, but again there'll only be one.

And BTW, something doesn't add up. A 9v battery won't supply much current at all for very long, and its voltage will sag to 7v or so quickly if loaded. So the solenoid shouldn't really draw more than 100mA, IMHO. You should check this to make sure you don't have something else going on.

I know something doesn't add up, I tried running them directly off of a 9v battery but to no avail, and I tried taking a timer circuit apart, but its full of cement or something inside, which I tried to dissolve in some acetone but it doesn't seem to want to give up the goods. I'm thinking theres some caps/resistors in there that allow the solenoid to go on/off because it takes a few seconds after you turn the dial before the timer actually does anything.

Like I said though , all the solenoids have a common terminal (one line connects to all of them, so I'd only have to put the resistors on the common terminal and it (should work for all of them).

I'll grab a 9v wall wart though and try with that instead. These will eventually be run off of SLA 12v batteries though , so current limiting will have to be built in.

We had a similar situation with the electric tape brakes on a punched tape reader (yes, very old tech). The brakes were not stopping the tape fast enough resulting in read errors. The voltage to them was increased. Then they started getting too hot. The answer was to put a resistor in series with a cap parallel to the resistor. When activated, the cap gave the brake coil a good hard jolt, which stopped the tape, then charged up as the resistor started to limit the current. Worked like a charm. You will need to determine what current is needed to hold the solenoid on and using ohms law, determine a resistor value that will give a little more than that with 12 volts applied.

I think you are a bit confused, the solenoids are nice and cold and these are latching solenoids, meaning they need a few milliseconds in either polarity to switch on and off , with no current in between (no holding current)

And to be clear, the solenoids are switching great right now, its just that the l293d's are getting really hot (still touchable hot, barely) for just that split sec while they are on, then cooling down in between. I just wanted to temper the load a bit so I don't burn em out or have to use 2.

I'm also considering turning their 'on' time down from 500ms to 200ms or lower.

 

So after you guys pointed out the whole 9v battery thing sounded odd, I was thinking it was a long time ago I tried the 9v battery, so I just tried it again, and it switched the solenoid perfectly (no heat on l293d OR the battery). I'm thinking I didn't have the motor supply caps last time (i had tried just doin battery directly to solenoid like an idiot) and that now it has ample time/current whatever to get the job done.

So now knowing that my circuit only needs 9v, and from an alkaline battery no less (although the power led did dim a bit when the motors switched), how can I design it to work properly with a 12v sla with a large amount of available current?

remember that these are irrigation solenoids for water valves , so I guess I should have not been stupid and realized they should draw almost no power (http://blog.makezine.com/2009/01/20/how-sprinkler-solenoid-valves-work/ ) because the plunger is just a tiny little thing.

 

I admit ignorance of your driver IC but I think something isn't right. If your load doesn't draw enough to make it hot, it may be a problem with the inputs. Maybe you should post your schematic to get another set of eyes on it.

 

the l293d is a 2 channel h bridge driver circuit capable of driving a motor or solenoid both forward and reverse at 600ma continous or 1a peak for 100us. basically each side has 2 logic inputs, a logic motor enable, 2 terminals which can accept a motor. the chip takes supply power up to 24v i think. if i send hi-low or low-hi to the inputs the current flows from the motor supply power either forwards or reverse polarity to the motor/solenoid.

the 600ma output is far higher than a 9v battery produces, so the solenoids must not need all that current i reckon. their low resistance must be drawing far more current than required is the only thing i can think.

 

the 600ma output is far higher than a 9v battery produces.

I just shorted out a 9V battery with my ammeter and read 6A.

 

Solenoids are 9-12VDC , 5-10ohm ... I'm running them at 12V

9-12VDC across 5-10Ω gives a range of 900mA to 2.4A.

Why are you trying to limit the current? They are 5-10Ω for a reason. If it needs 2.4A to switch reliably, then it needs 2.4A. I would understand if they were maintained, but they aren't, they're switched on impulse, then shut off.

how can I design it to work properly with a 12v sla with a large amount of available current?

They will draw 2.4A max, you don't have to worry about them trying to gobble up all available 600amps of a car battery.

remember that these are irrigation solenoids for water valves , so I guess I should have not been stupid and realized they should draw almost no power (http://blog.makezine.com/2009/01/20/how-sprinkler-solenoid-valves-work/ ) because the plunger is just a tiny little thing.

yeah, are designed to draw almost no power in the long run, but for brief impulses, they are designed to draw gobs of power, relatively.

the l293d is a 2 channel h bridge driver circuit capable of driving a motor or solenoid both forward and reverse at 600ma continous or 1a peak for 100us. basically each side has 2 logic inputs, a logic motor enable, 2 terminals which can accept a motor. the chip takes supply power up to 24v i think. if i send hi-low or low-hi to the inputs the current flows from the motor supply power either forwards or reverse polarity to the motor/solenoid.

My opinion is that your L293D is ill equipped for the task.

 

A bit of an aside, but related to the topic nevertheless:

The flipper solenoids on a pinball machine have two windings, a big one and a small one. When the flipper button is pressed, both coils are powered to give a good kick to the ball. However, as the solenoid reaches the end of travel, an integrated switch breaks the connection to the large coil, leaving power only to the small one. This way the flipper can be held in the up position without overheating the solenoid.

I'm wondering why the solenoids in this application are not set up in a similar manner?

 

A bit of an aside, but related to the topic nevertheless:

The flipper solenoids on a pinball machine have two windings, a big one and a small one. When the flipper button is pressed, both coils are powered to give a good kick to the ball. However, as the solenoid reaches the end of travel, an integrated switch breaks the connection to the large coil, leaving power only to the small one. This way the flipper can be held in the up position without overheating the solenoid.

I'm wondering why the solenoids in this application are not set up in a similar manner?

automobile starter solenoids are the same way. The valves OP is using are not the same, they are actually better. They are latching valves, so instead of a big coil and a small coil they have just a big coil, and once the big coil does it's job, it doesn't need a smaller coil to hold it in place. It simply remains in place via mechanical means, and zero power is needed to keep it there. This is why I don't understand trying put resistance in series with it. in some scenarios where no second coil is available and where a large switching current is needed, but only a small holding current, resistance will be placed in series with the coil, with a capacitor. This allows for high switching current to flow from the capacitor, followed by low holding current through the resistor. That's totally not the case here, and adding resistance in series with the coil is going to effect the switching. limiting the current while switching in a bench test might prove to be a different thing than in real life. Let that valve sit out in the sun, with water in it for one summer and get full of rust - see if it still switches when you limit it to 25% of it's required power (600ma vs. 2.4A).

 

sorry about the delay.

strantor, this is why im asking the question about the l293d. i realize they probably arent i deal so im asking what is. would piggybacking not work? they can handle bursts up to 1a so 3 l293ds piggybacked would work right? they only cost me 1.30 ea whereas the power mosfets for a proper h-bridge would cost 0.65$ and i need 4 plus some diodes.

im willing to fork over the 50 cent diff if thats all it takes.

or am i missing some really simple way to do this?

i appreciate the input.

 

sorry about the delay.

strantor, this is why im asking the question about the l293d. i realize they probably arent i deal so im asking what is. would piggybacking not work? they can handle bursts up to 1a so 3 l293ds piggybacked would work right? they only cost me 1.30 ea whereas the power mosfets for a proper h-bridge would cost 0.65$ and i need 4 plus some diodes.

im willing to fork over the 50 cent diff if thats all it takes.

or am i missing some really simple way to do this?

i appreciate the input.

I don't really know about the l293d in parallel. You could try it, or maybe someone else here has tried it and can speak up. If it were me, I would do the MOSFETs because it's a sure bet. Buying the MOSFETs right off the bat is cheaper than buying a bunch of L293Ds, finding out that doesn't work, and then buying the MOSFETs.

 

ok i already have about 10 l293d but i need to get this thing switching 20-30 solenoids. i know mosfets woill be the proper option now.

brainstorm though; can i use a single h bridge to give polarity then have the control lines go to a spdt relay for each solenoid? i can get 4-5a relays for 1.50$ . i'm trying to cut costs here but i still want to make sure it works properly. only one solenoid will be powered at a time. would this work?