Transistor Switch Voltage Drop

Thread Starter

mhinkamp

Joined Jan 30, 2016
8
Hi, I'm working on a smart sprinkler Arduino system, where I will use an Arduino to control electric 12V hose valves to control water flow. I'm using the 5V output of the Arduino through the base of a 2N2222 npn transistor to switch on an external 12V source to power the valves. I've attached an older schematic of how I'm trying to do this. The right side of the circuit is my concern. I've been able to get the switch working, however, the npn is consuming a considerable amount of voltage; when I tried one valve, that valve consumed only about 11V, and then trying two valves/transistors at the same time each valve only took about 10.3V. This is not a good pattern, and trying to connect all 6-7 valves would probably result in all not receiving enough voltage to open. I'm looking for suggestions for reducing the transistor voltage drop. I tried a similar circuit with a relay but the switching wasn't working at all. Thanks
 

hp1729

Joined Nov 23, 2015
2,304
Hi, I'm working on a smart sprinkler Arduino system, where I will use an Arduino to control electric 12V hose valves to control water flow. I'm using the 5V output of the Arduino through the base of a 2N2222 npn transistor to switch on an external 12V source to power the valves. I've attached an older schematic of how I'm trying to do this. The right side of the circuit is my concern. I've been able to get the switch working, however, the npn is consuming a considerable amount of voltage; when I tried one valve, that valve consumed only about 11V, and then trying two valves/transistors at the same time each valve only took about 10.3V. This is not a good pattern, and trying to connect all 6-7 valves would probably result in all not receiving enough voltage to open. I'm looking for suggestions for reducing the transistor voltage drop. I tried a similar circuit with a relay but the switching wasn't working at all. Thanks
How much current does each valve draw? Your circuit looks good if the transistors can handle the load current.
Are you using the DC latching Hardie-Richdel type valve?
 
Last edited:

#12

Joined Nov 30, 2010
18,224
You are supposed to give the base of a bipolar transistor 1/10th of its collector current to use it as a switch. If the controller can't deliver that much current, that's what mosfets are for.
 

hp1729

Joined Nov 23, 2015
2,304
You are supposed to give the base of a bipolar transistor 1/10th of its collector current to use it as a switch. If the controller can't deliver that much current, that's what mosfets are for.
Good words. I am trying to establish how much current the valves require. Manufacturers seem unfriendly with technical info.
 

Thread Starter

mhinkamp

Joined Jan 30, 2016
8
The valves each need 300mA; so that would make the base current need to be 30mA, which is high. So if I were to get MOSFETs for the job, is there a certain kind that would work best in this scenario? I would get NMOS type but I mean chip number
 

#12

Joined Nov 30, 2010
18,224
Computer took a dump. Forgive my weakness right now. Popular suggestion is 2N7000. If I remember correctly, that one will do 300 ma as a saturated switch. Please google the datasheet and read all about it.
 

Thread Starter

mhinkamp

Joined Jan 30, 2016
8
Computer took a dump. Forgive my weakness right now. Popular suggestion is 2N7000. If I remember correctly, that one will do 300 ma as a saturated switch. Please google the datasheet and read all about it.
Thanks, the 2N7000 says it can handle up to 400mA which is fine for me. As far as connecting this circuit, would I use a gate resistor analogous to the base resistor for the BJT? I'm not well-versed with MOSFETs
 

#12

Joined Nov 30, 2010
18,224
I don't think so, but sometimes people put in 100 ohms to avoid oscillation. I don't think that's a problem at your speed today.
Please know, I am guessing.
 

Thread Starter

mhinkamp

Joined Jan 30, 2016
8
Just for my understanding, this MOSFET circuit should allow more voltage to go to the valves than my BJT circuit did, correct? The original circuit worked fine with switching, just maybe not full switching. The voltage drop is my only issue
 

#12

Joined Nov 30, 2010
18,224
Any transistor used as a switch should try to disappear from the voltage equation. Mosfets are better at doing this than bipolar transistors.
Use the resistance from drain to source at the gate voltage you can supply, times the coil current, to find the voltage loss.
 

hp1729

Joined Nov 23, 2015
2,304
I don't think so, but sometimes people put in 100 ohms to avoid oscillation. I don't think that's a problem at your speed today.
Please know, I am guessing.
The resistor also protects the driving chip if the transistor should fail.
 

shteii01

Joined Feb 19, 2010
4,644
You seem to miss some information.
You tell us that when one valve is present, there is 11 V drop across the valve.
When two valves are present, there is 10.3 V drop across the valves.
What is the significance of this information? (Other then you don't like the numbers?)

I am looking at the 2N2222 datasheet.
When transistor is the switch, it either Off or On. When it is On, it is in saturation. So I am looking at the Vce voltage of 2N2222 when it is in saturation (Vce(sat)). The maximum Vce(sat) is 1.6 volts. In your circuit you have 12 volts at the top, then you observed 10.3 volts across the valve, 12-10.3=1.7 volts. These 1.7 volts is the voltage from Collector to the Emitter of the 2N2222, also known as Vce. Your Vce of 1.7 volts is very close to Vce(sat) of 1.6 volts, I would say that the transistor is operating as you have intended it to operate. You wanted it to be a switch and it is operating as a switch, the Vce observed by you proves it.

So basically from top to bottom you have 12 volts, then 10.3 volts across valve, then 1.7 volts from Collector to Emitter. Note that you SHOULD expect voltage from collector to emitter to stay at 1.6 volts, it will be a constant, you are not going to change this voltage. So. If you want to increase voltage across the valve, you will need bigger power supply. Lets say you get 15 volt power supply. Then voltage across the valve will be == 15-1.6=13.4 volts. Is that what you want? Do you want larger voltage across the valve?
 
Last edited:

Thread Starter

mhinkamp

Joined Jan 30, 2016
8
You seem to miss some information.
You tell us that when one valve is present, there is 11 V drop across the valve.
When two valves are present, there is 10.3 V drop across the valves.
What is the significance of this information? (Other then you don't like the numbers?)

I am looking at the 2N2222 datasheet.
When transistor is the switch, it either Off or On. When it is On, it is in saturation. So I am looking at the Vce voltage of 2N2222 when it is in saturation (Vce(sat)). The maximum Vce(sat) is 1.6 volts. In your circuit you have 12 volts at the top, then you observed 10.3 volts across the valve, 12-10.3=1.7 volts. These 1.7 volts is the voltage from Collector to the Emitter of the 2N2222, also known as Vce. Your Vce of 1.7 volts is very close to Vce(sat) of 1.6 volts, I would say that the transistor is operating as you have intended it to operate. You wanted it to be a switch and it is operating as a switch, the Vce observed by you proves it.
Maybe I wasn't quite clear. I now understand the BJT does consume a relatively large portion of the voltage in the manner I'm using it, my question was more directed towards a route of not losing so much voltage. I'm not married to using the 2N2222 in this circuit, it's just what I had lying around and gave me a starting point to pose my question here
 

#12

Joined Nov 30, 2010
18,224
I'm not married to using the 2N2222
Some bipolar transistors are much better at saturation voltage. I've seen them rated at 0.3 volts, but only if you can give their base 1/10th of the collector current. You have no chance of doing this with a microprocessor. You can either use two transistors in a "double invert" circuit or go straight to a mosfet. The mosfet is simpler, cheaper, and more effective.

(2N4403 is a good saturation transistor, but still not better than a mosfet.)
 
Last edited:

shteii01

Joined Feb 19, 2010
4,644
Maybe I wasn't quite clear. I now understand the BJT does consume a relatively large portion of the voltage in the manner I'm using it, my question was more directed towards a route of not losing so much voltage. I'm not married to using the 2N2222 in this circuit, it's just what I had lying around and gave me a starting point to pose my question here
Yes, I realize that 2N2222 is not a requirement. You will have 1.6 volt drop from Collector to Emitter. Your valve then will have 12-1.6=10.4 volts, so there will be 10.4 volts across the valve, you actually saw 10.3 volts, close enough. Is that a problem? Do you have the data on the valve? How much voltage should be across the valve?
 

Thread Starter

mhinkamp

Joined Jan 30, 2016
8
Yes, I realize that 2N2222 is not a requirement. You will have 1.6 volt drop from Collector to Emitter. Your valve then will have 12-1.6=10.4 volts, so there will be 10.4 volts across the valve, you actually saw 10.3 volts, close enough. Is that a problem? Do you have the data on the valve? How much voltage should be across the valve?
The valves each need 12V to turn on; they can be heard slightly clicking on with the 10-11V, but it is not as loud of a click as with a straight 12V. This worries me when I apply water pressure to them, in terms of them being able to effectively push water without constricting the flow too much. The valves have 39 ohms coil resistance which is where the 300mA number comes from
 

shteii01

Joined Feb 19, 2010
4,644
Some bipolar transistors are much better at saturation voltage. I've seen them rated at 0.3 volts, but only if you can give their base 1/10th of the collector current. You have no chance of doing this with a microprocessor. You can either use two transistors in a "double invert" circuit or go straight to a mosfet. The mosfet is simpler, cheaper, and more effective.
I generally agree with #12.
I just reread Uno spec, Arduino says that pins on Uno rated to supply 20 mA, and not to exceed 40 mA. Meaning there is some gray area from 20 to 40 mA where you would be OK.
You say that valve needs 300 mA. That means Uno must supply 30 mA. According to Arduino this is doable, but not recommended.
 

shteii01

Joined Feb 19, 2010
4,644
The valves each need 12V to turn on; they can be heard slightly clicking on with the 10-11V, but it is not as loud of a click as with a straight 12V. This worries me when I apply water pressure to them, in terms of them being able to effectively push water without constricting the flow too much. The valves have 39 ohms coil resistance which is where the 300mA number comes from
Ok. Now things are clear.
You have two choices:

Choice 1.
Replace the 2N2222 with transistor that has lower voltage drop. #12 mentioned BJT with lower Vce like .3 volts. You can try this, this would mean that valve will get 12-0.3=11.7 volts. 11.7 volts might be enough.
No matter what you do, there will be some voltage drop across the transistor so your initial choice of 12 volt supply is what got you in trouble.

Choice 2.
Replace the voltage supply. Get something larger than 12 volts you have now. Something just a few volts higher. 14, 15 or a few volts higher, whatever is easy to find where you are.
 

ian field

Joined Oct 27, 2012
6,536
You are supposed to give the base of a bipolar transistor 1/10th of its collector current to use it as a switch. If the controller can't deliver that much current, that's what mosfets are for.
How would that work with a BU208? - ISTR: the gain spread is something like 2 to 8.
 

#12

Joined Nov 30, 2010
18,224
The 2N4401 is the npn version that you aren't going to use.
(Spooky voice.)
You are getting sleepy.
You are ordering some mosfets.

[Sleepy/off]

There is a butt load of great mosfets that make this job look like playing in a sand box. You can order some TO-220 mosfets so big a fuse will be necessary, if you have some money for your hobby.
 
Top