This thread has been locked.

If you have a related question, please click the "Ask a related question" button in the top right corner. The newly created question will be automatically linked to this question.

  • Resolved

SMD alternative for 1N457 (LM334 current source minimizing tempco)

Prodigy 20 points

Replies: 4

Views: 3848

L.S.

I'm a noob on temperature coefficents of diodes.

The datasheet of LM134/234/334 shows a way of minimizing the tempco by adding a diode in series, the tempco's cancel each other out, at least for most parts.

But the 1N457 is a conventional type, and I cannot even find a single SMD diode where the tempco is specified in the dataheet.

Can anyone help me out? Does a drop in replacement of the 1N457 exist?

Maybe we can persuade Texas Instruments to make an addendum to the datasheet of this ancient (National instruments) part that mentiones an SMD diode? 

Thanks in advance,

Loek

  • MMBD1201 / MMBD1202 (SOT-23-3), or FDLL914A (SOD-80), all from Fairchild Semiconductor are ~similar~, not exact performance.

    The Fairchild dataheet for the FDLL914A has a Typical Performance curve (Page 3, Figure 6) that suggests the forward voltage tempco will be about -1.43mV/°C, with a If of 1mA, and a typical forward voltage of about 615mV at Ta=25°C.

    http://www.fairchildsemi.com/ds/1N/1N914.pdf

    The suggestion in the LM134 datasheet that "... this value should be directly measured ..." is the best way to ensure that you have best information. If the forward voltage of an unknown silicon diode can be measured across a reasonable temperature range then the tempco can be calculated and almost any silicon diode can be used.

    And the standard disclaimer for this type of circuit is included: "...however, the circuit should be built and tested over temperature..."

     

  • In reply to Donald Jones:

    Hi Donald Jones,

    Thank you for your answer. I would never have found the FDLL914A datasheet or would have imagined that the MMBD1201 is similar. My preference would be a SOT23 package, and there is no clue in that datasheet that it is similar (and no tempco graph!). 

    For a (medical) electronics shop like us measuring a temperature value directly is a bit cumbersome: one needs a form of rather precise temperature control. And the statistics on extrapolating the results of a single or a few diodes to have a confidence interval on a whole batch of products is not trivial. Getting values off a datasheet is more in my comfort zone...

    You have at least pointed me to an eye-opener: apparently the tempco is also (exponentially) related to the forward current. My current source will be 0.7uA so my tempco will be around 2mV/degree instead of the 1.43 @ 1mA.

    What we will do is build a test PCB with a MMBD1201 and see if a quick and dirty validation is possible by molding it in epoxy and testing with melting ice water and boiling water.

    Again, thank you very much for the swift response,

    Cheers!

    Loek

     

  • In reply to Loek Gijben:

    All the formula in the datasheet needs to counterbalance the device tempco is some simple inexpensive device with a negative tempco. Generic silicon diode fits the requirement. You could easily use the base-emitter junction of a transistor.

    Yes, forward voltage and tempco are strongly dependent on the forward current.

     " ... Getting values off a datasheet is more in my comfort zone ..."

    Keep in mind that Typical Characteristics curves are just that ...'typical'. Rarely is there any indication provided by the manufacturers’ of how many samples were used to define the typical, or how wide the distribution (6 sigma) might be. But, the typical curve values should get you close.

    “… one needs a form of rather precise temperature control …”

    Not really. You need a precise temperature measurement. I have had customers use a toaster-oven for high temperature testing, and their freezer for low temperature testing. As long as you can create a temperature difference, and measure the temperature difference, whatever it might be, along with the forward voltage at those temperatures then you can work the formula:

    Vf Tempco = ΔVf / ΔT

    Your methodology should be more than adequate.

     

  • In reply to Donald Jones:

    One can use Schottky diode. We obtain decreasing drop (0,35v instead of 0,6v) and TemCo (1...1,2 mV/degree). R2/R1 will be 5 or less.
    Hello from Moscow. Vladimir

This thread has been locked.

If you have a related question, please click the "Ask a related question" button in the top right corner. The newly created question will be automatically linked to this question.