• State Resolved
  • Locked Locked
  • Replies 2 replies
  • Answers 1 answer
  • Subscribers 62 subscribers
  • Views 877 views
  • Users 0 members are here
Support feedback
Options
Options
Similar topics

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.

TL431B: Reference voltage not in spec

Matthew Findlow
Intellectual 880 points

Part Number: TL431B

Hi 

I want to use TL431BQPK (0.5% SOT89) in a simple solar charger circuit to charge a 12V lead acid battery with a 5W panel (250mA max power/500mA short circuit). The circuit  below doesn't show the diode between Vout (14.4V) and the battery +(13.8-13.9V).  When the battery is charged to 13.9V and Vout is 14.4V, the TL431 shunts 30mA to ground through the 180R which raises Vin to the open circuit voltage of the solar panel, where the only current from the panel is the 30mA being shunted to ground. Simples.  I ordered five  TL431BQPK (the 0.5% type) from RSonline and built a few sample circuits.  Both the transistor and TL431B are SOT89 packages mounted on very large pours, so neither barely get warm to touch.  The circuit works really well, with less than 1V drop across the transistor when charging at its maximum rate of 250mA. (if someone can recommend an LDO that would do a better job, I would appreciate the advice, I'm not clever and need all the help I can get))  

 My problem is that the reference voltage appears outside spec.  With the solar panel hooked up, and no battery attached I measure Vout at 14.21 and Vref at 2.466.  According to the data sheet Vref should be between 2.483 an 2.507, but I'm reading around 2.470 with a reasonable good quality multimeter.  The math checks out, (1 + 10/2.1)x2.466 - (.02)=14.21 so I don't think my multi meter is out of cal'.  So I built another board and it was even worse with Vref at 2.464.  This is causing me stress because I want to use the TL431BQPK on my production boards but now I don't trust them to be within spec.  The data sheet says its specs relate to test conditions with a cathode current (Ika) of 10ma, while I am shunting closer to 30mA, but as I said the IC is mounted on a big pour and barely feels warm to touch. When I lick my little finger and press it on the IC, I'd say the IC case couldn't be more than 40 degrees C.  The recommended operating conditions say up to 100mA, which I am well below, so I would still expect the Vref to be within spec even if my shunt current is greater than the 10mA data sheet test conditions.

What I did notice is, as the sun went behind the clouds and conditions were very dull,  Vref went up to 2.477 and Vout went up with it to 14.28.  This will  be due to a reduction in cathode current.  But even so, the Vref was still below spec.

Does it strike you as odd that I get a Vref of only 2.47, or would you expect this under my circuit conditions?  Do you think I should count on Vref dropping by 0.01V with an Ika of 30mA or do you think it might not be reliably repeatable? I was thinking perhaps RSonline have labelled the part wrongly and have sent me the 2% variety,???  What can I do to renew my faith in the part?

Regards Matt.

  • 0
    TI__Intellectual 1540 points

    Hi Matthew,

    The TL431B is a 0.5% accurate part at 25C, 10mA cathode current, and Vka=Vref. By changing the cathode current and the Vka value, you will be directly effecting your reference voltage. Here is an app note that goes over how these parameters effect the reference voltage. 

    The largest factor here is the change in Vref due to the change in Vka. This will cause a nominal drop in voltage of about 15.5mV, and max drop of about 30mV under your Vka conditions. This plus the 0.5% variation (12.5mV) can lead to an output values as low as 2.453. 

    A simulation in TINA verified my result that the nominal value should be around 2.48V. 

    Regards, Diego

  • 0 in reply to Diego Lewis
    Intellectual 880 points

    Thanks Diego, that explains everything. I'm very grateful for your support. Cheers Matt.

    One thing though.  Can you explain why my Vref and Vout went up when the sun went behind the clouds?  In dull conditions the Solar panels open circuit voltage comes down which means Vin is less and as such Ika is less.  As my Ika(28mA) is greater than Ika(nom) I would have expected a decrease in Vref as my Ika decreased towards the nominal of 10mA , due to Zka.