Lets do some math. A 1 Kohm resistor across a 50V voltage is Power = V*V/R, or 2.5 Watts. This is way too much heat for a component that is not designed for it, fire extinguishers optional. 10K is better, .25 Watts, but still a little warm. So you can use 100K ohm for .025 Watts (the 10 Kohm would work OK, I'm just conservative).

The more gain a transistor has, the less it loads the input, at least on a emitter follower design. Since we are trying to up the variable resistor that is being used as a voltage source feeding the transistor to reduce the resistors wattage dissipation it can source a lot less current to the transistor, because of the higher resistance. Upping the transistors gain compensates for this. Unlike most compensations, there is very little downside, many times to fix one problem you create several others when designing a circuit.

The general downside of a darlington is the increased base emitter drop, but in this case it doesn't matter. You can even use more than 2 transistors, as in a darlington triplet, for even higher gain values and it won't make much difference to this design, other than loading the adjustment resistor even less.

By the way, you can also use the emitter resistor to measure the current going to the load by measureing the voltage across it. They even make precision resistors with 4 leads in high wattage formats specifically for this function. So if you put this design in a box and spring for a meter you can have extra functionality.

 

unlv007,
You still owe me an explanation for the other thread, which I spent a fair amount of time on. What is the maximum expected NORMAL resistance for your load?

Or are you simply asking all of these questions without providing sufficient information to simply waste our time?

 

the maximum normal resistance is 1K. Thanks for your time. I highly appreciate your assistance.

 

thanks everybody for your posts.
I could obtain an LM317. I read from your posts that LM317 may not be able to provide 50mA. Can i connect a resistor of say 100 ohms to drop the excess voltage. MY Load is 1K and i need to pass const current through it from 1-50mA. THUS the compliance range of my CONST CURRENT supply=50V.I have a const voltage power supply to produce upto 60V.
With a const voltage supply=60V and LM 317 and resistance to drop the excess voltage that lm317 cannot withstand Can i make the const current circuit?
If this circuit does not work, please suggest a one that can be made with readily available components that i can buy right away at my place , Las VEGAS so that i dont have to wait.

 

No, the LM317 has limited input voltage.

If you desired up to 50mA through a 1.2K Ohm load (from the message you sent to me), then you would need to calculate for voltage.
E = IR
E = 50mA * 1200 Ohms
E = 60 Volts
All things being constant, that is the voltage that you will need at the input terminal to your experiment in order to assure a current of 50mA.

For a current of 1mA:
E = IR
E = 1mA * 1200 Ohms
E = 1.2 Volts

But for such a low current, if your source supply was putting out 60+ volts so that you could achieve the 50mA current, then the transistor would have to drop the remaining voltage across itself. We'll just say 60v for now.
P = E*I
P = 60*.001
P = 0.06 Watts
That's quite manageable. However, what happens if you have the output set to 50mA and your experiment decreases resistance to near-zero Ohms? The transistor must then dissipate that power:
P = E*I
P = 60*.050
P = 3 Watts
Now you need a transistor that can dissipate that much power, and has a collector voltage rating sufficient for the expected range.

For a Fairchild LM317, the maximum input to output voltage differential is 40v.

Besides, that regulator requires a minimum 10mA of current to ensure proper regulation. This would cause it to not be able to fulfill your requirement in the 1mA to <10mA range.

Now, if you returned to the other circuit that would use your original current supply, and clamp the output to 62v in case your experiment went awry and greatly increased it's resistance - that would work.

If you order parts from Mouser, they will arrive within 2 days. Texas is not far from you.

 

That's quite manageable. However, what happens if you have the output set to 50mA and your experiment decreases resistance to near-zero Ohms? The transistor must then dissipate that power:
P = E*I
P = 60*.050
P = 3 Watts
Now you need a transistor that can dissipate that much power, and has a collector voltage rating sufficient for the expected range.

For a Fairchild LM317, the maximum input to output voltage differential is 40v.

Besides, that regulator requires a minimum 10mA of current to ensure proper regulation. This would cause it to not be able to fulfill your requirement in the 1mA to <10mA range.

Now, if you returned to the other circuit that would use your original current supply, and clamp the output to 62v in case your experiment went awry and greatly increased it's resistance - that would work.

If you order parts from Mouser, they will arrive within 2 days. Texas is not far from you.

Thank you very much first of all for helping me to set up my lab asap.
Actually, due to the weekend, it will delay by four days, whereas i need to have my lab running asap.
If i use LM317, as per you problems will be -
1.Under zero load resistance- there is high power dissipated in LM317. However LM317 is only $2 cost, i can use another easily. That is why i want to build my cheap power supply.
2. THE MINIMUM CURRENT CAN BE 10mA, ACTUALLY having the lab running is more important.
3. I did not understand your comment
"For a Fairchild LM317, the maximum input to output voltage differential is 40v."
I can use my dual const volt power supply to produce 30+10=40v.
4. what is the maximum compliance voltage that LM317 const current power supply can achieve?
thanks once again

 

IN POINT NO. 2 i mean to say that i can relax the current range , 1-50mA is ideal but 10-50ma WILL ALSO WORK, my primary aim is to have lab running from tomorrow itself.
thanks for your very precious help in my times of need!!

 

If you are still using your previous current supply for the source, and your experiment should unexpectedly rise in resistance as you have mentioned that it may, the current source will raise the voltage across even this new current regulator until the programmed current is satisfied - however high that voltage is, up to it's 600V limit.

If you are using a different power supply, then you will have to reduce it's maximum voltage output to whatever the maximum specifications are of the particular voltage/current regulator IC is that you can get ahold of. This will naturally limit the maximum current that you can supply to your experiment.

 

I am not using the old power supply any more and it is less likely that the experiment i will do tomorrow and recently will become awry. Therefore right now i am seeking a short term help.
bY the way i can purchase 2N 4403 and also 2n 3906 from kiesub electronics here at las vegas. These are for the circuit that bill posted, which i attached below.

Will these two work? you have made references to these two in this thread but i want to make sure that they work. kindly mention a few more easily available transistors that i can buy

 

The absolute maximum differential between the input and the output of the Fairchild LM317 voltage regulator IC is 40V. National Semiconductor's LM317 is also 40V maximum. If you know that your experimental load will have a range of between, say, 500 Ohms and 2k Ohms, then you could use it.

To use an LM317 as a constant current source, you connect a resistor between the Vout and adj terminals, and use the adj terminal as the current source. The Vin terminal is connected to your voltage source.
Vref = 1.2
Iout = Vref/R1
0.8 Ohms < R1 < 120 Ohms
If R1 = 24 Ohms, Iout = 50mA
If R1 = 120 Ohms, Iout = 10mA

 

The transistors that you mention will not work due to the power dissipation requirements.

Use an LM317, and a heat sink. The LM317 can handle up to 20W if it is heatsinked properly. You could use a computer processor heatsink if you drill a hole and use a sheet metal screw to attach the IC to it. You should use heat sink compound. If there is a fan on the heatsink, either connect it to 10v-12v power or remove it. Note that the tab of the LM317 is electrically connected to the Vout terminal.

 

While you're at it, see if you can pick up a TIP122 (NPN) and a TIP127 (PNP)

These are complementary Darlington ICs that can handle up to 65W and 100V if they are properly heatsinked.
They should be commonly available.

TIP120/TIP125 are complementary but rated for only 60v.
TIP121/TIP126 are complementary but rated for 80v.

 

It's too late; after 2AM here. Tomorrow.

 

these are three more constant current circuits, that might work.
Basically i want to build 5 circuits at least so that atleast one will work and my lab can keep working.
If you think these circuits are okay, please suggest alternatives to these devices so that i may buy them

1) circuit1- TIP41C and 2N5401
2) circuit2-TL071 , LM 385-2.5
3)CIRCUIT3-TIP42
I dont know if these are rare or common devices, but backups will help.

 

The first circuit involving LM317 is called LM 317 bootrapped circuit . The output voltage of LM317 range from 1.2V to 37V whereas i need ideally 50v but i can bring it down to 45V at the most. Will this bootstrapped circuit work?

 

I highly appreciate if you could please give your response to my question above today morning so that i may buy the components you recommend before the store closes

 

You know, I'm really a bit put off by this.

Your constant current circuit has been discussed repeatedly over the past couple of weeks, and at great lengths - yet you kept giving information that wasn't correct, or that for some reason you needed to change in the middle of the discussion.

For instance, suddenly the resistance of your load went from 10K, to 4K, to 1.2K, now to 1K, you at first needed VERY close regulation to 50mA, with a range of 1mA to 50mA adjustment, then 10mA to 50mA, then everything changed again, and at the very last minute you want advice on three MORE circuits that would have to be simulated in detail so that you can buy the components for them this afternoon?

Does this seem like a reasonable request, were it made of you?

 

Your constant current circuit has been discussed repeatedly over the past couple of weeks, and at great lengths - yet you kept giving information that wasn't correct, or that for some reason you needed to change in the middle of the discussion.

I think he may be getting confused by simultaneous input from multiple electronic forums. John

 

I think he may be getting confused by simultaneous input from multiple electronic forums. John

I don't doubt that one bit. However, this thrashing about is really getting ridiculous.

I simulated the first circuit. I did not have a TIP41C in my library, so I substituted a 2N3055. Severe oscillations were encountered 3.5mS into the simulation. A 10nF capacitor from the LM317 adjust terminal to ground stabilized the circuit, and it seems to function in a range of 10mA to 49mA output.

I really don't have time for the other two circuits.

 

I am sorry but i was nervous as my instruments were malfunctioning and we lose research funding in such cases. however i thank all of you for time and would close this thread