# Power supply circuit

Thread Starter

#### StephenMG

Joined Feb 7, 2008
10
Hi Everyone,

I talked to my friend about his project last night and got some more details of what he needs.

He would be operating this circuit on 3 to 4 9volt Batteries. But here is the thing. He needs the circuit to provide a constant 1.5 milliamps of current.

Let me explain a little more about this project. He works in a Micro-Lab doing testing on ultra pure water. For this project he has a beaker filled with pure water. Then inserts two rods into the water and applies DC voltage to the rods. As the conductivity of the water rises so does the current. After this process is done, he then tests the water in a TOC analyzer. What ever that is.

The problem is that the current after so many hours will go way over 1.5 milliamps. He needs it to stay put at 1.5 milliamps.

So, he wanted to know if a circuit could be made that would provide a constant 1.5 milliamps. Meaning before the rods are inserted or after the rods are inserted into the water it would provide 1.5 milliamps current.

Thanks for all of your help guys,
StephenMG

#### SgtWookie

Joined Jul 17, 2007
22,210
Well, in order to get a decent design on it, what is the maximum voltage he wants to see?

In order to have a 1.5mA current through the rods, if the resistance between the two rods is very high, it's going to take a heck of a lot of voltage to cause 1.5mA current to flow.

I = E / R
or,
Current = Voltage / Resistance
In order to keep current constant when the resistance changes, the voltage has to be increased or decreased correspondingly.

A quick glance seems to indicate that a Linear Technology's LT3012 might be a good choice of a low dropout regulator, there's a constant brightness/constant current LED source application in the notes, but it's late and I'm too tired to figure it out at the moment. An LM317 would not be appropriate, because it requires a minimum of 10mA current to regulate the load; and in this case that would have to be wasted. I'll give it a shot in the morning.

Thread Starter

#### StephenMG

Joined Feb 7, 2008
10
Hi Everyone,

The maximum voltage would be 27 - 36 volts DC and the minimum would be 9 volts DC. Again, it would be running from 9 volt battries.

Thanks again guys,
StephenMG

#### studiot

Joined Nov 9, 2007
4,998
Meaning before the rods are inserted or after the rods are inserted into the water it would provide 1.5 milliamps current.
Hang on there just a cotton picking minute!

Are you asking for there to be a constant current of 1.5 mA between the rods or where?

What are these rods made of?
what is their separation? Is it fixed?

I am concerned about the idea that 1.5 mA will flow when the rods are not inserted, what do you mean? Sure no current will flow when the rods are air separated?
The statement only makes sense if the rods are actually the probes of the TOC meter and the current is being passed by other electrodes?

As the conductivity of the water rises so does the current.
This is correct (Ohms law) but as the purity of water increases the conductivity falls.

I would guess TOC is 'total oxygen content'.

Thread Starter

#### StephenMG

Joined Feb 7, 2008
10
Hi,

Yes, he is looking for a constant current of 1.5 milliamps between the rods. If you touch the two rods, you have completed the circuit.

But in this case, when he places the rods in the water, the water has almost a 0 conductivity. When he applies voltage to the rods in the water a recation takes place and the conductivity goes up. As the conductivity goes up so does the current.

Problem is it goes up past 1.5 milliamps and he needs it to stay at a constant 1.5 milliamps.

It's almost like the circuit would have to short itself. That way when you hook the 9 volts to it the current of 1.5 milliamps is there.

I'm not good a drawing guys but I drew what he is doing and attached it to this message.

#### Attachments

• 19.5 KB Views: 22

#### KMoffett

Joined Dec 19, 2007
2,886
StephenMG,

Attached is a 1.5mA constant current source that should work. This is the output section of a circuit I designed for use in a clinical environment. The 24v-36V supply would work in place of my 60V battery. To adjust, just place a 2mADC current meter across the output terminals and adjust the pot for 1.5mA. Or leave the meter in series to assure that resistance in the conducitivity loop doesn't overcome the headroom of the current source.

Ken

#### Attachments

• 8.7 KB Views: 46

#### studiot

Joined Nov 9, 2007
4,998
Yes I was getting around to something similar to Ken's. His has the advantage of a modicum of adjustment.

Points to remember.

When not immersed may transistor will drop the full battery voltage so must be rated accordingly.
For this reason do not try the FET circuit referred to in post #2.

The polarity may be reversed by using an NPN transistor and reversing the LM385.

Either way the pass transistor needs to be about 500 mWatts or greater rating.

#### Attachments

• 74.5 KB Views: 29

#### SgtWookie

Joined Jul 17, 2007
22,210
KMoffet's design should work just fine for your application

You'd have to order the parts, though.

Attached is another variation on the same theme, but using commonly available parts that you could pick up at your local Radio Shack.

In this case, the Zener diode is replaced with an ordinary red LED, and the transistor is a commonly available 2N3906. Using the 2N3906 limits your maximum input voltage to 40; any higher and you'll burn up the transistor. Fresh 9V batteries CAN exceed 10V, so just use three in series.

There is a 10K pot in parallel with a 680 Ohm resistor; this gives much more precise control over a narrow output range, but beware that getting the wiper arm of the pot towards the low end of it's resistance range will rapidly increase the output current. Start with the pot in the center of it's travel, and make small adjustments with the probes shorted.
[ETA: this schematic has been superceded by an improved version, further down in the thread]

#### Attachments

• 12.5 KB Views: 31

#### nanovate

Joined May 7, 2007
666
TOC is total organic carbon. It is used to detect organic contamination in the water.

When not immersed may transistor will drop the full battery voltage so must be rated accordingly.
For this reason do not try the FET circuit referred to in post #2.
Of course you need to choose parts that are rated for the operating conditions. The FET in the circuit is rated to 25V so if you are going to use 36V then you need a FET that is rated higher. The circuit itself is fine though. As far as dropping the full supply voltage goes... I think it is more of a concern when the probes are immersed since the transistor needs to "drop" voltage to regulate the current down.

The LM385 in Ken's post will exhibit better stability and accuracy over temperature and is a better solution.

#### KMoffett

Joined Dec 19, 2007
2,886
SgtWookie,

To be on the safe side, using standard Radio Shack components (good choice), I'd repace the 680Ω resistor/10KΩ pot, with a 470Ω and 1KΩ resistor in series between the emitter and +supply, with a 1KΩ pot paralleling the 1KΩ resistor. Range then is 1.0mA to 2.0mA (though very nonlinear), with no chance of exceeding the max collector current if the user accidentally (what's the likelyhood of that ) turned the 10KΩ pot so it shorts emitter to +supply.

Ken

#### studiot

Joined Nov 9, 2007
4,998
more of a concern when the probes are immersed
The water will always present a reasonable resistance across the probes so dropping a large part of the voltage.

I was referring to the poster saying he touches the probes together. Whether deliberate or inadvertent this is the worst case as all the battery supply, less the base emitter circuit will be across the transistor.

Oh and thanks for the def of TOC

#### SgtWookie

Joined Jul 17, 2007
22,210
SgtWookie,

To be on the safe side, using standard Radio Shack components (good choice), I'd repace the 680Ω resistor/10KΩ pot, with a 470Ω and 1KΩ resistor in series between the emitter and +supply, with a 1KΩ pot paralleling the 1KΩ resistor. Range then is 1.0mA to 2.0mA (though very nonlinear), with no chance of exceeding the max collector current if the user accidentally (what's the likelyhood of that ) turned the 10KΩ pot so it shorts emitter to +supply.

Ken
Hey Ken,
I actually already did a mod somewhat akin to your suggestion, as the possibility of overcurrent wasn't sitting well with me.

I added a 220 Ohm resistor in series with a 1k pot, which will be for coarse current adjust. In parallel with the 1k pot is a 10k pot for fine adjust.

The problem with using LEDs is that the voltage drop across them can vary considerably from batch to batch, unlike the Zener in your circuit. However, I believe that this new configuration sufficiently addresses the overcurrent problem of the last configuration, yet leaves enough adjustment room with a "fine tweak" capability due to the parallel 10k pot, which will have roughly 10x the sensitivity of the 1k pot.

#### Attachments

• 20.5 KB Views: 33

#### KMoffett

Joined Dec 19, 2007
2,886
The water will always present a reasonable resistance across the probes so dropping a large part of the voltage.

I was referring to the poster saying he touches the probes together. Whether deliberate or inadvertent this is the worst case as all the battery supply, less the base emitter circuit will be across the transistor.

Oh and thanks for the def of TOC
Studiot,

A short circuited output in a constant-current regulator is not exactly a "worst case" condition...well maybe maximum power dissipation...36V x 0.0015mA = 69mW. The way to "set" the current in this circuit is to "short the output", with an ammeter (ideally zero resistance) in series. As long as the transistor's maximum Vbe, Ic, and Pw are not exceeded, it's fine.

Ken

#### studiot

Joined Nov 9, 2007
4,998
I have always understood that connecting 30 volts across a 25 volt transistor does not yield good news.

#### KMoffett

Joined Dec 19, 2007
2,886
Studiot,

In SgtWookie's circuit, the 2N3906 has a Vce max of 40V, not 25V. This is a little close for a 36v supply, but still within it's ratings. I would look for a little higher Vce, but that's just me. You must realize that this an active circuit that limits the Ic to 1.5ma, despite the load resistance. Yes, the full supply voltage, minus 1 volt across the emitter resistor, will be impressed across the C-E junction with a shorted output. But, the transistor actively limits the Ic to 1.5mA.

Ken

#### SgtWookie

Joined Jul 17, 2007
22,210
Also, you'll notice that in the schematic I posted, there are only three 9V batteries, for a maximum potential of roughly 30V, or about 3/4 of the maximum Vce of 40.

Certain rechargeable 9V batteries can read over 11V after charging; that's why I limited the circuit to three batteries.

Even with the 220 resistor limiting max current to 3.9mA, that's still around 1.17 Watts with 30v in the circuit with either of the pots turned to 0 Ohms and would result in a very short life for the transistor. However, if the rather explicit current calibration procedure were followed, this should not happen. Leeway had to be left for the variability of the LED voltage, which could be anywhere from 1.6v to 2.2v.

It was simply a circuit to get him up and running until he could order parts from afar to build Ken's circuit.

#### KMoffett

Joined Dec 19, 2007
2,886
30v x 0.0039A = 0.117W ... you slipped a decimal ... so still in very safe territory, and you have a very workable circuit.

Though I didn't indicate it, the 200Ω pot in my circuit was a 10-turn trimmer.

Ken

#### studiot

Joined Nov 9, 2007
4,998
When not immersed may transistor will drop the full battery voltage so must be rated accordingly.
For this reason do not try the FET circuit referred to in post #2.
Comments are welcome but I do wish people would read posts properly before commenting.

Ken, You had already offered a superior version of the standard CC configuration that I was sketching. I was simply trying to offer some additional help to a poster who may not be familiar with 'sizing' transistors for a job. this is particularly pertinent to the later suggestion of a 60 volt supply, which would also destroy a 2N3906 ( although funnily enough I have recently received a chinese batch with a higher rating). The TIPP32 I suggested in my sketch also has a Vce of 40.

I clearly warned that the FET circuit posted by another used a FET that has a voltage rating that would lead to its destruction if connected to 3 off 9volt batteries in series. The voltage rating warning applies to any transistor, FET or Bipolar used.

#### KMoffett

Joined Dec 19, 2007
2,886
Studiot,

Sorry...didn't see a reference to the previous FET post, so wasn't sure where the 25v came from.

Ken