Constant Current Sink

Hello Harish,

Does this require bidirectional current?  250 mA at 70 ohms is 17.5V.  The power dissipation in the load is either I^2*R or V^2/R.  250 mA at 70 ohms is 4.4 watts in your flow meter.  If you use a linear driver the dissipation in the driver will be (Vin-Vload)*250 mA.  A switching solution will not do this, it will be more efficient.

Best Regards,

Hi Irwin,

Irwin Nederbragt said:

Does this require bidirectional current? 

Yes it requires bidirectional current, done through MOSFET H-bridge as shown below.

I am currently using the linear driver LM317. As you said the linear driver dissipates significant power which results heating of the driver. To reduce the heat I have added a buck regulator MCP16301 to lower the power supply. When the excitation coil resistance is 75 ohm, I adjust the voltage output (Vexc) from buck regulator as;

Excitation coil resistance = 75 ohm

Rds of MOSFETs = ~ 1 ohm

Rsense = 10 ohm (to set 250mA constant current via LM317)

Vexc = 75*0.25 + 1*0.25 + 10*0.25 + Head Room of 0.5V = 22 V

By doing this I significantly lower the dissipation of linear driver.

But, my problem arises when the excitation coil resistance lowers for some other flow meter, lets say 65 ohm. Then the same circuit values dissipates heat in linear driver. I do not want to check every flow meters' coil resistance to change component values in each case. I want the circuit and its component value to be universal for a range of excitation coil resistance ( 65 ohm to 75 ohm ).

For this, I need to change linear driver with buck regulator and then I came across the constant current LED driver IC solutions as LM3404.

As per my understanding, since the internal mosfet Rds (max) as 0.75 ohm + internal regulator dropout of 0.3V should dissipate power for 250mA constant current = (0.25)^2*0.75 + 0.3*0.25 = 122 mW.

As I understand, I will also need fast turn-on/switching of output. Do you think that LM3404 will work for my application efficiently?

Regards,

Harish KS

Hello Harish,

Yes, you could do it that way.  You could look at the TPS92515, LM3409 as well.

It looks like you plan on having the H-bridge determine current direction and the switcher providing the current to the H-bridge, that would probably be the easiest though.

What do you mean by fast turn-on time wise and when (when powering up, etc?)

Best Regards,