Need stable 0.3V drop at 100uA

[Leo Bodnar]

I need a low current (100uA) two terminal 0.3-0.5V zener-like voltage drop device.  Modern, tiny, available in SMT.
Any ideas apart from Schottkies?

Background: I have bought a 1000 TLV1117LV30 3.0V LDOs by mistake - instead of TLV1117LV33 3.3V ones.  Luckily, I was able to get TLV1117LV33 too.

My plan is to use TLV1117LV30 3.0V LDOs elsewhere on a non-critical 3.3V LDO duties having inserted a Schottky (or whatever vox populi says) in the ground pin.  This is a low power LDO so its ground current is probably in tens to hundreds microamps.

Cheers
Leo

[bdunham7]

How stable is the ground pin current?  I don't know the architecture of the particular LDO, but if it is like older fixed linear regulators it just has an internal voltage divider and a resistor might do.

[Benta]

The adjustable TLV1117s just seems to be a fixed one with the resistor divider lasered away.
An adjustable resistor in ballpark range in series with the "Ground" pin should do the job.

[srb1954]

I need a low current (100uA) two terminal 0.3-0.5V zener-like voltage drop device.  Modern, tiny, available in SMT.
Any ideas apart from Schottkies?

Background: I have bought a 1000 TLV1117LV30 3.0V LDOs by mistake - instead of TLV1117LV33 3.3V ones.  Luckily, I was able to get TLV1117LV33 too.

My plan is to use TLV1117LV30 3.0V LDOs elsewhere on a non-critical 3.3V LDO duties having inserted a Schottky (or whatever vox populi syas) in the ground pin.  This is a low power LDO so its ground current is probably in tens to hundreds microamps.

Cheers
Leo

The fixed voltage versions of the TLV1117 have a significant ground pin current which probably comes from a resistive divider on the output to set the voltage. It appears that this resistive divider is proportioned to sink the minimum load current required this type of regulator, removing the normal requirement for a minimum external load.

Since the ground current is not that well controlled (typ. 5mA max 10mA) it may require some select-on-test components in the ground leg to adjust the output voltage to an acceptable tolerance range.

[Leo Bodnar]

Thanks, guys.  I can run the LDO through a selection of output current scenarios and measure the ground pin current.
Indeed, a resistor might be the easiest solution!
Thanks again
Leo

[magic]

TLV1117LV has an insignificant quiescent current of 70µA which increases significantly to 600µA at full load
(adaptive bias it is or whatnot, I remember homework threads about it on the forum)

At room temperature BAT82 seems about right.

[SiliconWizard]

Of course you may also just consider running your circuit @3.0V instead of 3.3V. We know nothing about it, but I bet you do. Unless you use the power rail as a voltage reference (for instance for ADCs or DACs), running 3.3V CMOS ICs @3.0V is usually no problem whatsoever (check your datasheets). The simplest solutions consist in doing nothing at all. =)

[EPAIII]

A forward biased, Schottky diode seems to be the obvious answer, but you may need to add a resistor to provide more current. Here are just a couple I found with 300 mV forward Voltages:

https://www.comchiptech.com/admin/files/product/CDBQC0130L-HF%20RevB681116.pdf

https://my.centralsemi.com/datasheets/CFSH05-20L.PDF

The forward Voltage will vary somewhat with temperature, but should be stable enough for your use.

[Terry Bites]

I'd try returning them to the supplier and putting the right part in. I've always had sucess sending back parts to reputable suppliers in the UK RS, Farnell /CPC, Digikey, Mouser etc.
You'll be digging a techical and finacial hole for yourself by adding resistors or diodes.
There are substitutes for the TI part. ON make the NCP1117LPST18T3G Mr Mouser has 3k in stock as of today. Same spec but cheaper.

https://www.mouser.co.uk/ProductDetail/onsemi/NCP1117LPST18T3G?qs=kbSN2QblFfJvIdaUG0xRow%3D%3D

[Scottie]

You could also experiment with running the running the regulator LM317 style. In this case vref would be 3V. Run the feedback divider at current high enough such that the regulators gnd current is negligible. But then you have to worry about stability.

[Leo Bodnar]

Thanks for the advice but you'd need to be mad to return any parts these days!
I got them directly from TI for £0.20 each.  I'll use them alright.
The LDO you suggested has 1.4V dropout voltage.  They won't work on USB bus.
Cheers
Leo

I'd try returning them to the supplier and putting the right part in. I've always had sucess sending back parts to reputable suppliers in the UK RS, Farnell /CPC, Digikey, Mouser etc.
You'll be digging a techical and finacial hole for yourself by adding resistors or diodes.
There are substitutes for the TI part. ON make the NCP1117LPST18T3G Mr Mouser has 3k in stock as of today. Same spec but cheaper.

https://www.mouser.co.uk/ProductDetail/onsemi/NCP1117LPST18T3G?qs=kbSN2QblFfJvIdaUG0xRow%3D%3D

[Terry Bites]

My mistake.
I looked at the Iq vs Io chart. Anything you put in series with the GND pin is going to make Vo wander with Io.

[David Hess]

How much complexity are you willing to put up with?

A schottky diode is the easiest solution and the output will not be that far off.  Variation of the ground pin current can be reduced by adding more current from the output with a single resistor.

Alternatively a two resistor divider and single supply operational amplifier can drive the ground pin from the regulated output to 0.3 volts restoring the functionality of the adjustable version.

[Leo Bodnar]

Schottky should be fine. This is for non-critical digital circuitry use.

If not - schottky with extra current injected from the input side (resistor, BJT, JFET.)

If not - 0.3V low impedance external voltage lift using BJT or tiny opamp.

Cheers
Leo

[David Hess]

If not - schottky with extra current injected from the input side (resistor, BJT, JFET.)

Since the output voltage is fixed compared to the common pin, a resistor from the output to the schottky diode will provide a fixed and stable current.

[SiliconWizard]

Schottky should be fine. This is for non-critical digital circuitry use.

In this case, you still didn't say why operating it at 3.0V, as I suggested, would be a problem.

[Terry Bites]

The Schottky diode Vf tempco gets added to the otherwise great tempco of the TLV1117 which may be undesirable.

Still it costs pennys to try it out.

[David Hess]

The Schottky diode Vf tempco gets added to the otherwise great tempco of the TLV1117 which may be undesirable.

It does, but how much accuracy is required?  It will still be closer to 3.3 volts than a 3.0 volt regulator.

[Leo Bodnar]

I use parts that won't be happy at 3.0V level like TCAN330, etc.
Leo

In this case, you still didn't say why operating it at 3.0V, as I suggested, would be a problem.

[SiliconWizard]

I use parts that won't be happy at 3.0V level like TCAN330, etc.
Leo

In this case, you still didn't say why operating it at 3.0V, as I suggested, would be a problem.

The TCAN330 should be perfectly "happy" with 3V.

8.3 Recommended Operating Conditions
MIN NOM MAX UNIT
VCC Supply voltage 3 3.6 V

It might marginally change the max data rate you can achieve in given conditions, which in turn would  be a problem only if you run it at the edge.

But hey. This is your pick.

[not1xor1]

Since the ground current is not that well controlled (typ. 5mA max 10mA) it may require some select-on-test components in the ground leg to adjust the output voltage to an acceptable tolerance range.

This comes 40-50 years after the old LM78XX and its quiescent current is 50-100µA.
In any case quiescent current varies with the load (according to the datasheet), but a schottky would manage that and even compensate the output voltage drop caused by the load.

[not1xor1]

I use parts that won't be happy at 3.0V level like TCAN330, etc.
Leo

In this case, you still didn't say why operating it at 3.0V, as I suggested, would be a problem.

operating voltage range is 3-3.6
a silicon diode like 1N400x would likely provide a drop of 0.5 V or less @100µA