Variable Voltage Sources without dedicated ICs

Thread Starter

clark.donaldson

Joined Aug 19, 2019
26
Hi all -
I recently came across a design that specifies a 2.8v reference to bias a circuit.
I've traditionally used programmable voltage regulators, but I'd like to build the circuit and wait for yet another parts order.
I have some 3.3v zeners and it seems that putting a resistor between the cathode and V+ allows me to adjust the voltage to the value I need.
What I'm wondering is how "bad" is this? I think I read voltage dividers have a 5% accuracy and basic zener shunts have a 4% accuracy... Am I looking at a 4.5% accuracy?
Are there small, simple improvements to make? Would a trimpot in place of the resistor be better or worse?
 

TeeKay6

Joined Apr 20, 2019
464
Hi all -
I recently came across a design that specifies a 2.8v reference to bias a circuit.
I've traditionally used programmable voltage regulators, but I'd like to build the circuit and wait for yet another parts order.
I have some 3.3v zeners and it seems that putting a resistor between the cathode and V+ allows me to adjust the voltage to the value I need.
What I'm wondering is how "bad" is this? I think I read voltage dividers have a 5% accuracy and basic zener shunts have a 4% accuracy... Am I looking at a 4.5% accuracy?
Are there small, simple improvements to make? Would a trimpot in place of the resistor be better or worse?
You will not achieve a good voltage reference using zeners below about 4.7V. You are getting 2.8V with a 3.3V zener because you are not providing enough current for good zener action. With less than the specified current a zener diode operates much like an ordinary resistor. With a low enough current you can get any low voltage but it will not be regulated at all. To get a stable 2.8V reference you must use an IC, not a zener. There are many voltage reference ICs and voltage regulator ICs in the market that can produce a stable 2.8V output.

As for tolerances, zener diodes commonly are sold with ±5% tolerances, although some are available with ±2% tolerance. It is important to understand that the zener voltage must be measured at a specified current if you expect the voltage to be within the tolerance. If less current flows through the zener, the voltage will drop; if more current flows, the voltage will rise; the manufacturer specifies the voltage to be within tolerance only at a specific current (see the datasheet for the device). Low voltage zener diodes are not very good voltage regulators; higher voltage zeners (e.g. >8V) have a sharper "knee"--that is the operating point where voltage regulation begins; below that point regulation is very poor.

Voltage dividers, comprised of resistors, can have any tolerance you wish; resistors can be easily purchased in tolerances as close as ±0.1% or as loose as ±20%. There is no single "standard" tolerance.

If you publish the schematic for the circuit you wish to use, someone here can recommend how to create the 2.8V.
 

crutschow

Joined Mar 14, 2008
23,544
If you want to use a zener, then it should be at least 5.1V for reasonable stability.
You can use a voltage divider to then get the desired voltage.
How much current do you need?

Better would be to use an inexpensive programmable shunt references (programmable zeners) such as the TL431 and the TLV431.
The TL431 can be adjusted down to 2.5V and the TLV431 down to 1.24V.
With 1% fixed resistors the standard TLV431 can have a set voltage accuracy of about 3.5%.
If you want it closer, you can use a potentiometer to tweak the voltage.
Those references have sufficient temperature stability for most requirements (0.5% typical over the 0-70°C range for the TLV431).

Below is the LTspice simulation of a TLV431 configured to generate 2.8V.
The TLV431 adjusts its current until the Ref voltage from the R1-R2 divider equals 1.24V.
So you can adjust to voltage for any value you want from 1.24V up to its 6V maximum (with a greater than 6V supply of course).

upload_2019-10-10_0-50-40.png
 

Thread Starter

clark.donaldson

Joined Aug 19, 2019
26
Alright - it sounds like everyone agrees that I need to just order a programmable shunt.
I just want to clarify a few things:
1. Traditionally, people have connected a 5.1v zener diode as a shunt - and used a subsequent voltage divider - to get a voltage source. I understand this is not ideal - but I'm looking at a need for about 6mA, so I don't know how ruggedly it needs to be built.

2. Can the TLV431 be connected to produce a negative voltage reference also?
 

crutschow

Joined Mar 14, 2008
23,544
I'm looking at a need for about 6mA, so I don't know how ruggedly it needs to be built.
Not sure what you mean by "ruggedly".
It just needs to be designed to deliver at least 6mA.
But if you want that amount of current from a voltage divider, then it likely needs a buffer amp on the output, otherwise the divider resistor values get very small.
Can the TLV431 be connected to produce a negative voltage reference also?
Sure. You just turn it upside-down. See below.
I increased the bias current through R3 to 10mA so it can deliver at least 8mA output current without losing regulation.
upload_2019-10-10_8-46-29.png
 

Thread Starter

clark.donaldson

Joined Aug 19, 2019
26
Thanks for the clarification - I'm still a little worried the TLV431 won't work, though - its going to be built off of a 15v/3a supply.

Not sure what you mean by "ruggedly".
It just needs to be designed to deliver at least 6mA.
As for this, I guess I just meant that the application is probably ok with a little fluctuation around that value. I'd be OK with 90% accuracy.
 

crutschow

Joined Mar 14, 2008
23,544
I'm still a little worried the TLV431 won't work, though - its going to be built off of a 15v/3a supply.
It should work fine with that voltage.
You just must keep the adjusted output (TLV431 cathode) voltage below 6V.

And increase the value of R3 to give the desired bias current [(15V-2.8V)/10mA = 1.21k for 10mA as shown below].
upload_2019-10-10_11-41-46.png
 

ci139

Joined Jul 11, 2016
821
upside-down
? applies to what part exactly please note the protective reverse diode at fig.12 , chapter 8.2 , Fig.1 , . . .
False alert = Your battery symbol has [+] where the [-] is been output = a weird battery that has its Cathode named [+] ...
... e.g. -- denoting : " It's positively a Cathode " ??
actually the rechargeable Ni-Cad and some non-rechargeable Alkalines can develope a small voltage reverse charge !!! -- in a series pack
__________

Falstad Custom (not too good for above 200 or 400Hz) Shunt Ref
 
Last edited:

crutschow

Joined Mar 14, 2008
23,544
so it's fig. 2 ? what's the difference (conditionally it is and in figure 12 as by the representing symbol of the TLV431)
I have no idea what your question is. o_O
The Zener diode symbol with the 3rd connection on the side is the TLV431 symbol.
I thought that was apparent.
 
Top