LT3092 Programmable current source

[grimmjaw]

Hallo eevbloger,

I currently playing LT3092 Programmable current source.
I'm using it as a current source by driving it with PSoC5LP.

Please refer to pic for the circuit .

My problem is when I set the Vdac to 0V, the R1 read ~50mV which mean  that
5mA is flowing throug the load.

When I set the Vdac to 40mV , R1 read 90mV, and it seem like the is an offset of 50mV at Vdac.
Does anyone knows  what happening here?
The set pin has a 10uA flowing out of it.Maybe this is the culprit?

[c4757p]

The only way they ever show it being used in the datasheet is with a programming resistor. It looks like it should work when driven by a DAC, but have you tried doing it the way the datasheet recommends?

Though I think it would be massive overkill for a non-floating, DAC-driven current source, that can just be an op amp...

[grimmjaw]

I followed the schematic from  the datasheet ,page 16.

I know it is an overkill, but it should be working, and LT seem to be very proud of this chip 

[PA0PBZ]

My problem is when I set the Vdac to 0V, the R1 read ~50mV which mean  that
5mA is flowing throug the load.

* Ground the set pin, what do you measure across R1 now?

* Use a resistor to set the current, is it correct now?

* Using the DAC, what voltage do you measure on the set pin?

[tszaboo]

Do you have at least 1.2V on the "in" pin?

[c4757p]

I followed the schematic from  the datasheet ,page 16.

Damn, how did I miss that one?

I know it is an overkill, but it should be working, and LT seem to be very proud of this chip 

The "Typical Applications" section of a somewhat expensive chip almost always lists a bunch of implausible uses in the hope that you will use it for everything. Sure, use it as a DAC-controlled current source! Hell, try to use it as a decoupling capacitor, we don't care, as long as you buy a lot!

It looks like a nice, handy chip for a resistor-controlled current source, because the reference and all is built right in, and the specs are decent. But when controlling it with a DAC, most of those nice specs become completely irrelevant!

[grimmjaw]

okay, to answer few question

I got 0mA when set pin is grounded, and 50mV when Vdac is attached (Vdac set to 0V)

Vcc is 5v should be okay.


@c4757p: It sound like LT is some evil corporation lol.
 

[PA0PBZ]

...and 50mV when Vdac is attached (Vdac set to 0V)

What voltage do you measure on the set pin when the DAC is set to 0?
My guess is that it is not 0V.

[grimmjaw]

Vdac is set to 0V, but you are right, after connected to the set pin I got 50mV, which put the LT3092 to 5mA output

[PA0PBZ]

Vdac is set to 0V, but you are right, after connected to the set pin I got 50mV, which put the LT3092 to 5mA output

Thought so, I think the dac is not capable of sinking the set pin. An opamp should be able to solve that.

[grimmjaw]

@PA0PBZ: sorry for the noob question, could you elaborate

[PA0PBZ]

A voltage follower, like this:



Use the opamp with positive and negative supply, otherwise it will not go to 0V.

What dac are you using?

[grimmjaw]

duh  ...I though some op amp+transistor combo..anyway will give it try. thanks for the tip

i'm using build in dac in PsoC5lp

[minime72706]

Just thought I'd chime in with a small piece: The LT3092 has a near-identical topology to the LT3080, which is a voltage regulator that is often discussed here.
I think they made the set current so low so that it can be safely overridden by applying a voltage source (DAC output, preferably buffered?) to the set pin.
I make this assumption because such a low set current often requires larger resistor values than I might use in a precision circuit. Johnson-Nyquist noise and all that.

[PA0PBZ]

i'm using build in dac in PsoC5lp

There you go:

8.10.2 Voltage DAC
For the voltage DAC (VDAC), the current DAC output is routed
through resistors. The two ranges available for the VDAC are 0
to 1.02 V and 0 to 4.08 V. In voltage mode any load connected
to the output of a DAC should be purely capacitive (the output of
the VDAC is not buffered).

[grimmjaw]

yup, that was the culprit.
I'm using an internal volatge follower and set was reduce to 5mV,
it seem that I need an external opamp with negative supply to get 0V.

Thanks for the help