diy 2.5v voltage source

[ME]

Is it possible to  get a stable 2.5v for the control voltage for an ocxo using a tl431?.

[Smokey]

https://www.ti.com/lit/ds/symlink/tl431.pdf

[Andy Chee]

Yes, you can generate a stable 2.5V using a TL431.

Be aware, the TL431 is a shunt regulator.  This differs from the LM317 and 7805, 78xx regulators which are series regulators. 

[ME]

Can anyone give me a rough drawing of a schematic to do it?.

[RoGeorge]

Take the 'Simplified Schematic' on the first page from the datasheet linked by Smokey.
If you apply +5V at 'Input', the resistor should be 220\$\Omega\$ for best performance (to get roughly 10mA through TL431).

V_KA will be 2.5V, your reference voltage.

[Gyro]

If it's for a precise application like an OCXO Vadj, you probably want something more temperature stable than a TL431. There are many 2.5V low Tc Bandgap references available, eg, LT1009, REF03 etc.

[ME]

its because i have some 431 components to hand,i dont want to order yet,thier was a fella built a freq ref with them using a cti 10Mhz ocxo, I cant find the thread tho.

[TimFox]

The TI data sheet (and the sheets for the other precision units cited) specify the stability and temperature performance of the devices.

[floobydust]

The LT1009 is a 2.5V shunt regulator, it comes with these low cost 10MHz OCXO boards.
You could put in a TL431 with no problem, aside from the value of capacitor C6 across it. Some makes of TL431 will oscillate depending on that value, so I would first check their datasheet i.e. pg. 18 fig. 6-18 . There's only 3 pins to worry about. TL431 datasheet. Same pinout in TO-92.

I'm not sure how much of an effect ref 100ppm/°C vs 15ppm/°C will have when the module control voltage spec is +/-2 ppm from 0-4V.
Quick napkin sketch 10°C change of a TL431 is 2.5mV shift, for 0.0025ppm or 0.025Hz but I might be wrong.

[ME]

any chance of posting a napkin scketch on how to wire it up so i get 2.5v out,ie a rough schematic??.

[TimFox]

For a TL431, connecting Ref and “cathode” (positive terminal) directly together forms a 2.5 V Zener diode.
Other voltages > 2.5 V require extra resistors. 
All this is in the data sheet, better than a napkin.

[floobydust]

any chance of posting a napkin scketch on how to wire it up so i get 2.5v out,ie a rough schematic??.

Same circuit as you or 807 posted in the other thread.
TL431 Anode to GND, K or Cathode is +2.5V (has R4 3k3 resistor to +5V), jumper REF to K.

[Andy Chee]

any chance of posting a napkin scketch on how to wire it up so i get 2.5v out,ie a rough schematic??.

Rough schematic is available on page 26

https://www.ti.com/lit/ds/symlink/tl431.pdf

[Picuino]

Add a potentiometer for fine regulation.

[Picuino]

Another cheap and simple solution: MCP1501T-25E/SN

https://eu.mouser.com/ProductDetail/Microchip-Technology/MCP1501T-25E-SN?qs=BA62vJVifGq6P0twjIFVRA%3D%3D

https://ww1.microchip.com/downloads/aemDocuments/documents/APID/ProductDocuments/DataSheets/MCP1501-Data-Sheet-DS20005474.pdf

[Refrigerator]

Connecting the reference pin to the cathode of the TL431 and adding a shunt resistor will make the simplest voltage reference.
But if you want more current you can add a single NPN transistor and now you will have what is basically a 2.5V voltage regulator.

[Konkedout]

Connecting the reference pin to the cathode of the TL431 and adding a shunt resistor will make the simplest voltage reference.
But if you want more current you can add a single NPN transistor and now you will have what is basically a 2.5V voltage regulator.

I think the pass transistor is a good idea, but I would be a bit worried about oscillation.  I have added some additional parts to help prevent that but my values are "guesstimates".  You may need to experiment on the bench.  I do not completely trust simulation to guarantee stability.  Different versions of the TL431 are different!  Please see my schematic image.

For Q1; 2N3904 is OK for lower current up to 50-100 mA.  2N4401 works well if the current is higher.  There is a lot of overlap between these two transistor types.
R1=1K is good for 8V-12V input; you may need to lower it for lower V_INPUT.
R2, R3, and C1 are feedback loop compensation.
R4 and C2 are some output bypass.  The TI TL431 datasheet describes using resistance in series with output capacitors to insure stability.  Adding Q1 makes instability more likely so my warning of possible instability ought not be ignored.
R5 is a minimum load. If your actual load current can ever get down below 1-2 mA you may need this.  R5 = 1K may be better but that will burn more power.

[floobydust]

The load requirement for the VCOCXO control voltage pin is very low about 10uA. So I don't see a need for the NPN buffer transistor, the TL431 etc. circuit alone has enough current available.
Also, I haven't confirmed if the control voltage pin sources any current, which is something else to consider but a dummy load resistor looks after this in the added buffer transistor circuit posted.

[ME]

I want this to provide vref for a OSC5A2B02 module before it gets trimmed with a multiturn preset pot.