Help:Digitally controlled current source

[abicash]

Thanks a lot Mike

I think I can use your scheme.I will go ahead with this.
Please be there if there are some problems ahead

Thanks and regards

 

[abicash]

Hello

Digging this out after a long time

I tested the circuit and it is fine i can linearly vary current till an ampere with a sense resistor of 1 ohm.

Although the 358 has a typical offset of 3 mV which creates an error of 3 mA
(which is okay but not accurate).I have to have a control step of 10 mA (i.e. i have to set currents from 50 mA,then 60 mA,70 mA...up to 1000 mA) for a set period in steps of 50 millisecond up to 500 millisecond.
I see that accuracy depends on a lot of factors and prominent is the offset voltage of the op-amp.How can i nullify this w/o using a dual power supply on the op-amp?
I am also thinking on using a low offset op-amp like the OP07 or TLC4502 (TI)

But before that i need to confirm that other offset balancing techniques wont work on LM358.

Awaiting precious replies as always

Thanks and regards

 

[Roff]

I think your best bet is something like OP07. TLC4502 is basically a 5V op amp, and might not drive your MOSFET.
There are ways to compensate for the offset voltage of the LM358, but the only ideas I came up with required some kludgy differential current sources, which would require a lot more parts, and probably cost more, than an OP07.

 

[abicash]

I think your best bet is something like OP07. TLC4502 is basically a 5V op amp, and might not drive your MOSFET.
There are ways to compensate for the offset voltage of the LM358, but the only ideas I came up with required some kludgy differential current sources, which would require a lot more parts, and probably cost more, than an OP07.

Thank you Roff for feedback.
Yes i figured that TLC4502 will require another few parts to cater to 12v (maybe a totem pole at its o/p?)
Anyways i am sure NOT to use a LM358.But OP07 will require a dual power supply (or a single ended supply will do?)
I am thinking about AD620,although a bit expensive.

Am i on the right track?

Thanks and regards

EDIT: Can you suggest some good opamps for the purpose?

 

[Roff]

Thank you Roff for feedback.
Yes i figured that TLC4502 will require another few parts to cater to 12v (maybe a totem pole at its o/p?)
Anyways i am sure NOT to use a LM358.But OP07 will require a dual power supply (or a single ended supply will do?)
I am thinking about AD620,although a bit expensive.

Am i on the right track?

Thanks and regards

EDIT: Can you suggest some good opamps for the purpose?

You're correct about the dual supply.
If you can get LT1677, I think will work well for you. Linear Technology says they have five in DIP8 packages for US $2.80 each.
See attachment for schematic.

 

[abicash]

You're correct about the dual supply.
If you can get LT1677, I think will work well for you. Linear Technology says they have five in DIP8 packages for US $2.80 each.
See attachment for schematic.

Hi and thank you so much for the information Roff

I am checking on availability for this part here.I hope i get it.
What all precautions i need to take while designing the board for this?

Thanks and regards

EDIT: LT1677 is not available the supplier asked me to check with LT10xx .I am checking this.Maybe you can throw some light.

 

[Ubergeek63]

Thanks a lot Mike

I think I can use your scheme.I will go ahead with this.
Please be there if there are some problems ahead

Thanks and regards

just use an LED driver. they are meant to supply current sources and many are switching regulated controlled by either analog or digital means.

 

[abicash]

Hello again

I have added a schematic of my design here.

There are several blocks,of which i have shown two blocks.

One is the CC source as mentioned by Roff (although i am yet to decide on the op-amp part in place of LT1677) and the subsequent is the differential instrumentation amplifier using TI TLC4502 (the parts which i have).

My intention: The RL load is a resistor having a temp.coeff (50 to 180ppm/C)
Initially i have to calibrate the circuit.For this I will pass a set current for a very short duration of 10ms.I will get the differential voltage by passing it through the diff. amp.The samples will be averaged in an ADC and stored.The stored value is then converted to an analog value with a DAC (not shown in scheme)

Now during actual test, the current will be the same as used in calibration but with larger time (say 500ms).The stored analog value of calibration is set as Vref (not shown in scheme) through the DAC to another stage of op-amp (not shown).Due to larger time for which the current flows ,the temperature of the Rload and thus Resistance changes so voltage drop also changes (above the calibrated value).
This change in drop is in microvolts.Which needs to be further amplified and measured with a 12 bit ADC for 2k samples.

I have used the application circuit mentioned in the TLC4502 datasheet Page23,Fig.37

The diff.amp. in schematic has to have a gain(A) of 1 since i only need to get the differential voltage across the Rload (RL) and further amplify the difference between the set calibrated value and the time modified value (as mentioned above) to maybe A=1000.
If you check the formula for Vo then for A=1 i get nonsense values for the resistors (i only need Vo=VRL)

What is wrong? Am i doing something wrong here?

Sorry for the loooooong post

Please get back

Thanks and regards

 

[Roff]

You are trying to run the inputs to the TLC4502 at 12V, but it has a 5V supply. The input range of the TLC4502 only goes up to Vcc-2.3V=2.7V!
I'll look at your problem again today as time allows.

 

[Roff]

Why can't you buy LT1677 directly from Linear Technology?

Also, can you use surface mount parts?

 

[abicash]

Actually the Linear parts will get a lot of time to arrive.Though i can get AD620 tomorrow.As i said the particular part is out of stock for now.

 

[Roff]

How are you controlling this circuit? Are you using a computer, or a microcontroller?
Is there dead time between current pulses?
If so, I have an idea that would allow you to use a less exotic op amp.

Did you read post #29?

 

[abicash]

And i can use surface mount parts..Preference is for through hole dips though

 

[abicash]

(Am sending this through my mobile device hence the brevity)
TL4502 is just an idea since have these parts lying around.The current pulse is just on demand one time affair.But can be regular depending on user.I m using a silabs MCU to control all this and send to the computer on UART

 

[Roff]

Here's the idea for using a less exotic op amp. Since op amp input offset voltage causes a bigger percentage current error when the current is low, we use a larger value sense resistor on the low range and scale the DAC output up by a factor of 10. This might also allow you to use a lower resolution DAC.
If the 3mV offset of the LM358 is still too high, we can probably find something that will work that is actually available.

EDIT: corrected schematic

 

[abicash]

Sorry had missed the schematic..Hard to see on a mobile phone Now your post makes sense.Forget my comment about DAC scaling..
But anyways i would prefer a normal approach using a low offset opamp.

 

[Roff]

Thanks Ron for replying.I dont think thats a very good idea. Since the maximum scaling i can get is 3.3v from the onchip DAC.I can use AD620 as mentioned earlier.Also this is half the problem addressed.Can u read my long post where i explained whats the bigger picture?

As I said, you scale the DAC to the range. You never need more than a volt at the op amp input.

I read your long post. I'm trying to address one problem at a time. As I read it, your calibration scheme does not eliminate the input offset voltage problem.

 

[abicash]

You r right.The calibration scheme does not eliminate offset problem.The only other thing is that i need currents till 1amp

 

[Roff]

You r right.The calibration scheme does not eliminate offset problem.The only other thing is that i need currents till 1amp

Sorry, I was thinking you said 500mA max. The circuit with the switched sense resistors will go 0 to 1A with 0 to 1V in from the DAC. I changed the note on the schematic to reflect this.
Is 300uA max error at 50mA acceptable? If so, the LM358 will work. If not, there are other op amps available with lower offset voltage.

 

[abicash]

Thankyou Ron for taking interest in my problem
i think 300uA is acceptable (i dont have the asking specs here nor do i remember ) but what other opamps can be used?