LM79L: Do you have temperature coefficient information?

Hi Teritama,

While our voltage references have a specified ppm/'C term, our LDOs typically do not. 

However, we do still have characterized temperature performance for many of our devices (see LM79L below)

This isn't the same as a min/max spec for ppm/'C, but gauging from this graph, the -5Vout version of the LM79L may exhibit -100ppm/'C which is outside your spec. 

The other concern is that it doesn't seem that we sell a fixed output version of the LM79L that fits your output requirements.

Is there another acceptable definition of temperature dependence I can use to search through our LDO portfolio (total accuracy across a given temperature range, etc) or must it be a ppm/'C temperature coefficient?

Best,

Gregory Thompson

Hi Gregory-san,

Thank you for your answer. I understand that typically TI LDOs do not specify ppm/℃.

I would suggest a shunt regulator solution because the customer's requirement is for very low current (under 10 mA).

Also, since the customer requires a fixed output to avoid the influence of the thermal characteristics of an external resistor divider, I think the configuration using LM337—where Vout and ADJ are directly connected to set the output voltage to Vref (typically −1.2 V)—can be proposed as a viable solution.

Please your opinion about above.

Best regards,

teritama

Hi Teritama-san,

Unfortunately, it doesn't appear that we have dedicated shunt regulator for negative voltages but there may be a configuration that makes it possible. I don't believe we have collateral around using an LDO in a shunt configuration, the concern is that most of our LDOs don't expose the positive terminal of the error amplifier required to regulate a shunt.

Fixing an adjustable part to its reference voltage the is a good idea. You can set the output to the fixed voltage reference on any of our adjustable LDOs without any external resistors. For floating LDOs (like the LM337) you connect the ADJ pin directly to ground, but for non-floating LDOs (most of our portfolio) you wire the FB directly to the output like you suggested. This is a viable and proven method of setting output. Many of our characteristic curves are taken in this configuration.

I will note that in my own experience with temperature testing, the ratio of two feedback resistors of the same type and package is very well matched across temperature, so it does not alter accuracy very much (as the ratio is the thing that affects Vout vs VFB).

Best,

Gregory Thompson

Hi Gregory-san,

Thank you for your comments.

Regarding the LM337-N or LM137, the datasheet specifies a Temperature Coefficient of 50 ppm/°C.
Could you please clarify the test conditions under which this Temperature Coefficient is defined?

We are considering a configuration where the ADJ pin is connected directly to ground to output the reference voltage, and would like to confirm if this approach is appropriate.

Best regards,

teritama

Hi Teritama-san,

1)

The 50 ppm/'C figure comes from dividing the Thermal Stability term by the temperature range it is defined over (for the LM337-N). 0.6%/(125'C-0'C) = 0.0048%/'C = 48ppm/'C which rounds to 50ppm/'C.

And is subject to the Electrical Characteristics table conditions (from above the EC table, seen below) and the Recommended Operating Conditions where not defined by the EC table or parameter conditions.

A more detailed graph of temperature dependence can be found below:

2)

Directly grounding the ADJ pin will work, this is appropriate. Adapting the front page application circuit:

Best,

Gregory Thompson