LT3081 - nice feature upgrades to LT3080

[konfu]

Hi folks,

some weeks ago I found some new product being released by Linear Technology - the LT3081.

Basically it's the same like LT3080 but with more features: current control via an extra pin, current sense output and a pin to give a current equivalent to the die junction temperature. Nice features.

Sounds like a great upgrade to Dave's LT3080 power supply...


I now struggle a bit to understand how the current limiting function works.

To set the current limit LT uses a resistor between the current-set pin and the output pin of the LT3081. So for example: 0.5A output current limit is done by 1.65k between these pins.


I am now thinking whether I can control the current limit the same way we do the voltage: apply some variable voltage to the Iset pin. But I guess it's not that easy.

So how does this current limit set feature works internally? It's not about voltage being applied to the Iset pin as the resistor would change it's delivered voltage if the output voltage is being changed. Same applies to the current flowing through the resistor.

There is no relation to the output voltage if I read the datasheet correctly. So the Iset resistor is not delivering either a current or a voltage to the Iset pin. I think it could be like this: there might be some internal resistor setting the current limit (if Iset is being left floating, there is some internal current limit of 2A). The Iset resistor is a parallel resistor to the internal one thus changing the current set resistor value.

But how would I then control the current limit with some MCU? I could of course attach some nice SPI/I2C settable potentiometer. But these potis usually only accept some smaller voltage at their resistor network pins. Since one end of the variable resistor would be attached to the Output pin of the LT3081 I would need to use something like MCP41HV51 to cope with higher voltages.

Anyone any idea how this current limit function works and how to control it with a MCU?

[firewalker]

Have they addressed the issues with LT3080?

https://www.eevblog.com/forum/projects/lt3080-wierdness-dave%27s-power-supply-%28eev224%29-gone-mad/

Alexander.

[Bored@Work]

Feck, you know what they have as application example in the datasheet? Arduino Monitored Supply. Don't they know any shame at Linear?    The world is going to hell in a handbasket. I smell a wave of Arduino-powered "lab" power supplies coming up.

Have they addressed the issues with LT3080?

From the datasheet:

The LT3081’s IN pin withstands ±40V voltages with respect
to the OUT and SET pins. Reverse current flow, if OUT is
greater than IN, is less than 1mA (typically under 100µA),
protecting the LT3081 and sensitive loads.

Clamping diodes and 400 [Ohm] limiting resistors protect the
LT3081’s SET pin relative to the OUT pin voltage. These
protection components typically only carry current under
transient overload conditions. These devices are sized to
handle ±10V differential voltages and ±25mA crosspin
current flow without concern. Relative to these applica-
tion concerns, note the following two scenarios. The first
scenario employs a noise-reducing SET pin bypass ca-
pacitor while OUT is instantaneously shorted to GND. The
second scenario follows improper shutdown techniques
in which the SET pin is reset to GND quickly while OUT
is held up by a large output capacitance with light load.

I guess "without concern" followed by "application concerns" is marketing speak for

Will blow up if ±10V differential voltages or ±25mA current between SET and OUT are exceeded. 

 And they call that a "rugged" IC.

[Legit-Design]

Am I missing something or what is LT8612 ?



I can find

• http://www.linear.com/product/LT8610
• http://www.linear.com/product/LT8611
• http://www.linear.com/product/LT8614

but no LT8612, are they also going to release that?
http://cds.linear.com/docs/en/solutions-manual/Linear_Automotive_Brochure_web.pdf
On page 7 we have:

[konfu]

The schematic shown as "typical application" is NOT an Arduino controlled but just an Arduino monitored setup.

And I really don't get why they use a LT3092 programmable current source. They could just feed voltage into the SET pin to control the output voltage, couldn't they? As far as I can see the SET pin just goes into a comparator + input internally. So it tries to make the SET pin voltage equal to the OUT pin voltage. No big deal - just the same like LT3080 here.


But still: I don't understand how to control the current limit with a MCU.

LT8612 is just some sync step down converter. It's the same technique Dave used for the "tracking pre-regulator" to get the power loss via the LT3081 down. Maybe just a typo or a new version of 8610/8611 not yet released ;-)

[kizzap]

Just wondering, has anyone had a detailed look at the datasheet, and worked out if it is possible to control the maximum current output by driving the Ilim pin with an op-amp in some way? I'm trying to wrap my head around it, but I just cannot decipher how the Rlim circuitry works. Knowing my luck, it is going to be something obvious, but my lack of knowledge is really holding me back here.

The thing that is confusing me, (and is probably the hint in the operation of the circuit) is that the current limit is set by a resistor, which is independent of voltage, over the entire output voltage range.

If anyone is able to shed any light on it, it would be greatly appreciated.

-kizzap

[Bored@Work]

Just wondering, has anyone had a detailed look at the datasheet, and worked out if it is possible to control the maximum current output by driving the Ilim pin with an op-amp in some way?

I wouldn't bother, because:

* The limiter varies +/-15% over temperature. I.e. it is unusable for a "lab" power supply.

* The limiter cuts the output voltage off, it doesn't regulate it down. This is not what you want in a "lab" supply.

* The information in the datasheet is confusing and contradicting. If Linear can't be bothered to get their shit together, why should you wade through that shit?

For example, already the schematic on the first page, "Typical Application" does not make sense. The IC has an internal fixed limit at 2A, independent of ILIM. But why do they have RILIM in that schematic set to 6.04 kOhm for a 2.24A limit? The internal limit would always trigger first, the external limit does not make sense. Especially not when the circuit is supposed to show a 1.5A  supply. Taking the recommended 20% variation into account, it would make sense to set RLIM for 1.8A, i.e. to 4.9 kOhm.  But they show 6.04 kOhm 

Or take the first circuit at page 21. Two ICs paralleled, for a 3A output. A single 3k current limit resistor shared by both ICs. How on earth did they derive that value? 3k would be a 1A limit for a single IC. Depending on the magic Linear is using for the current limit setting it would make sense that a 3k resistor, shared by two ICs, sets a limit of 0.5A, 1A or 2A. But 3A? 

Whatever Linear is doing in that IC they really don't tell in the datasheet. I wouldn't touch that IC because of the rubbish information and because of its SET/OUT brokenness.

[c4757p]

What a bizarre datasheet. It looks very hastily thrown together.

[Bored@Work]

What a bizarre datasheet. It looks very hastily thrown together.

Yes, it is surely missing the hand of Jim Williams.

[kizzap]

Well, fair enough. back to the venerable lm317 then I guess...

Strange how a series of chips can be so flawed to the point that no-one wants to use them.

-kizzap

[EEVblog]

But still: I don't understand how to control the current limit with a MCU.

You can't, it's not designed for that. Possible to use an EEPOT for example, but they don't go to high voltages.
Similar to many other switching regulators that have resistor programmable current limits.
There may be some tricky way to do it, but the spec is 15% tolerance over the tamp range, so it's pretty horrible. Not usable for a lab supply current adjust.

[Tooms]

In the LT Journal january 2014 there is 8 pages about the LT3081 and that it is now fixed and better then the LT3080

http://cds.linear.com/docs/en/lt-journal/LTJournal-V23N4-01-df-LT3081-RobertDobkin.pdf


/Tooms

[David Hess]

From the datasheet:

The LT3081’s IN pin withstands ±40V voltages with respect
to the OUT and SET pins. Reverse current flow, if OUT is
greater than IN, is less than 1mA (typically under 100µA),
protecting the LT3081 and sensitive loads.

Clamping diodes and 400 [Ohm] limiting resistors protect the
LT3081’s SET pin relative to the OUT pin voltage. These
protection components typically only carry current under
transient overload conditions. These devices are sized to
handle ±10V differential voltages and ±25mA crosspin
current flow without concern. Relative to these applica-
tion concerns, note the following two scenarios. The first
scenario employs a noise-reducing SET pin bypass ca-
pacitor while OUT is instantaneously shorted to GND. The
second scenario follows improper shutdown techniques
in which the SET pin is reset to GND quickly while OUT
is held up by a large output capacitance with light load.

I guess "without concern" followed by "application concerns" is marketing speak for

Will blow up if ±10V differential voltages or ±25mA current between SET and OUT are exceeded. 

 And they call that a "rugged" IC.

The typical LM317 and many similar 3 terminal regulators have the same output short to ground while the adjustment pin is bypassed problem.  The solution there is to add a diode between the adjustment and output pins to discharge the adjustment pin capacitance if the output is shorted.