As far as voltage reference stability is concerned, how good is a zener?
For example, I can use a 5.1V zener in series with a 820Ω resistor. I need 9V output, so that should give about (9-5.1)/820 = 4.76mA of current through the zener.
How stable will this setup be?
Noise will be quite easily filtered, I just want to know how it will typically perform.
Why use this setup? It can't get any more simple or cheap. But I am looking for output variations of 1-2% max. Will this suffice?
For example, I can use a 5.1V zener in series with a 820Ω resistor. I need 9V output, so that should give about (9-5.1)/820 = 4.76mA of current through the zener.
How stable will this setup be?
Noise will be quite easily filtered, I just want to know how it will typically perform. Why use this setup? It can't get any more simple or cheap. But I am looking for output variations of 1-2% max. Will this suffice?
LM329 have active shunt regulator circuitry lowering the (DC-1kHz) dynamic impedance of the internal buried Zener to < 1 Ohm at any current within its operating spec
true Zener's are crap - have poor dynamic Z, broad "knee" require lots of current
above ~6 Vz rated "Zener diodes" Avalanch mode takes over - gives sharper knee, better Z at lower curernts
read some data sheets - look for tutorials or app notes if you don't understand the specs impact on circuit performance
true Zener's are crap - have poor dynamic Z, broad "knee" require lots of current
above ~6 Vz rated "Zener diodes" Avalanch mode takes over - gives sharper knee, better Z at lower curernts
read some data sheets - look for tutorials or app notes if you don't understand the specs impact on circuit performance
Last edited:
Prior postings here many years ago concluded that true zeners were very noisy, and that as long as you were far removed from the true zener voltage of ca 5.6V then it didn't matter much if you used zeners or LED's. Since then I suspect but I don't know this to be fact, that band gaps have taken over completely in the semiconductor industry. Don't know how good they are.
Thank you for your answers!
DUG, I haven't bought anything yet, but seems like I will. I have an electronic store close to me that implements orders from there if I want so. Do you think that ordering directly will come cheap?
EDIT: LM336 seems good and cheap. Has anyone had any experience with it?
DUG, I haven't bought anything yet, but seems like I will.
I have an electronic store close to me that implements orders from there if I want so. Do you think that ordering directly will come cheap?EDIT: LM336 seems good and cheap. Has anyone had any experience with it?
Last edited:
Thank you for your answers!
DUG, I haven't bought anything yet, but seems like I will.
...
I really have no idea how much this stuff costs in Greece.
Use the Digi-key part search and narrow down the selection with what features you want.
How do we interpret the initial accuracy spec found in voltage reference IC's?
I mean, say an LM336 of 5V has a 1% initial accuracy. Does this mean that:
(1) it will be 1% away from the 5V, say 4.95V, but always that value
(2) or that the reference voltage will be always susceptible to changing within this 1% margin? That means on Monday it is 4.95V, on Saturday it can be 4.98?
That said, I ask since I want to decide between an LM329 and LM336 to build a Sulzer regulator. The LM329 has better noise specs, the LM336 best accuracy. But if the reference meets the assupmtion (1), I should choose LM329 for its better noise (I will trim the output to the desired value). But if (2) is valid, then LM336 and moire filtering is best, I assume.
I mean, say an LM336 of 5V has a 1% initial accuracy. Does this mean that:
(1) it will be 1% away from the 5V, say 4.95V, but always that value
(2) or that the reference voltage will be always susceptible to changing within this 1% margin? That means on Monday it is 4.95V, on Saturday it can be 4.98?
That said, I ask since I want to decide between an LM329 and LM336 to build a Sulzer regulator. The LM329 has better noise specs, the LM336 best accuracy. But if the reference meets the assupmtion (1), I should choose LM329 for its better noise (I will trim the output to the desired value). But if (2) is valid, then LM336 and moire filtering is best, I assume.
the initial value is typically only modified by tempco, line, load regulation, drift which are spec'd on better refs
buried Zener long term drift is expected to be very good
thermal cycling can add or relieve mechanical stress - which does affect the V of reference chips - real instrumentation geeks do stress relief thermal cycles to reduce this higher rate aging/drift
it is hard to imagine audio electronics applications where typical Vref aging/drift would be an issue
buried Zener long term drift is expected to be very good
thermal cycling can add or relieve mechanical stress - which does affect the V of reference chips - real instrumentation geeks do stress relief thermal cycles to reduce this higher rate aging/drift
it is hard to imagine audio electronics applications where typical Vref aging/drift would be an issue
Those ADI parts do look decent for low frequency noise and of course drift. Although it's not so much of a concern as it is readily amenable to filtering, it's interesting that the midband voltage noise density is about the same as an On Semi TL431, a highish 48nV/sq rt Hz. Of course the TL431 is bandgap-ref based.
Something I have not seen but must exist: the use of stacked TL431 and/or mixing the outputs, for a square-root-of-N noise reduction. Board area and power consumption would be a concern for some applications but presumably not a big deal for diy.
Brad
Something I have not seen but must exist: the use of stacked TL431 and/or mixing the outputs, for a square-root-of-N noise reduction. Board area and power consumption would be a concern for some applications but presumably not a big deal for diy.
Brad
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.