Could some electronics guru please help ? A typical linear voltage regulator (LDO, Standard etc.,) has a feedback loop, with an amplifier, one of whose inputs is a reference voltage. How and where is this reference voltage generated ? Any hints would be helpful.
Most linear voltage regulators incorporate a "band gap" voltage reference
-- Bill Sloman, Nijmegen
Several ways, most common is a "bandgap", with "buried zener" probably a close second. ...Jim Thompson
--
| James E.Thompson, CTO | mens |
| Analog Innovations, Inc. | et |
| Analog/Mixed-Signal ASIC's and Discrete Systems | manus |
| Phoenix, Arizona 85048 Skype: Contacts Only | |
| Voice:(480)460-2350 Fax: Available upon request | Brass Rat |
| E-mail Icon at http://www.analog-innovations.com | 1962 |
Make the world a better place, dismember a Democrat.
Bandgap.
Zener.
Charged capacitor.
John
Some more interesting, albeit generally impractical, ideas:
About the only linear regulator that I know off that uses a zener reference is the venerable uA723, and it pre-dates buried zeners. Some people like the part because it gives a relatively low-noise reference voltage.
There some three-terminal voltage references that do use - much quieter - buried zener diodes, but they tend to be expensive.
-- Bill Sloman, Nijmegen
Weird.
Oh, there's also the jfet equivalent of a bandgap, like in the ADR420 series. Pretty low noise.
John
Linear regulators with internal references use either the so-called band gap scheme or a specially designed buried zener (like the uA723).
e
Let's add a few:
Battery (standard cells)
Forward biased diode (LM334 style) - temperature dependence guaranteed but you can combine several diodes and null that, which is called 'a bandgap reference' because the null occurs at a voltage pinned to the Si bandgap
Saturating inductor (ferroresonant power unit)
Current regulator diode (a JFET with gate-source connected) - add resistor load to make a voltage source.
Generator with mechanical speed regulator on its input motor.
Spark gap (or gas-filled tube equivalent).
Thermopile with known-temperature reservoirs. (hey, it's primitive, but that's how Georg Ohm did some important stuff!)
The general scheme is that any transfer characteristic with a nonlinear character, as an amplifier with saturation or a diode with curvature in its forward or breakdown region, has an operating point (with resistor or other load) that doesn't track with the presumed variable input voltages. Some kind of linear combination of the input voltage and the operating point voltage is a near- constant value.
ere
.....
Drat, I missed one:
Ballast tube (basically a hot-filament lamp, the radiative heat loss is the nonlinear part). That's the regulator element in the old HP 200 oscillators.
Yes, and probably a few more, but not inside a voltage regulator chip. Neon lamp for example.
Moving coil meter movement with photoelectric interrupter feedback!
Varicap in a bridge!
I really like the Intersil thing, a capacitor charged at the factory, to the target voltage, then just left alone.
John
where
'Voltage regulator', though, includes the old Fluke 405 (?) which regulated up to 3 kV and had a 300V gas tube for reference. The original post was about 'a linear regulator', no single-chip requirement stated.
Drat, there's also the corona off a smooth sphere, like on a Van de Graaff... not exactly 'spark gap', it's another candidate. And I suppose the work function of a photocell cathode (with the stopping voltage being the reference, and any spectrally pure exciting source) is another...
where
Victoreen used to make small glass corona-based VR tubes that ran up into the kilovolts. I had a Knight Kit DC-5 MHz triggered-sweep scope that used them in the unblanking circuit.
John
And, of course, the standard volt, which is a bunch of series-connected SQUIDs in a microwave field whose frequency is calculated very accurately.
You could also use a Hall effect device with a known magnetic field applied.
Cheers
Phil Hobbs
where
Radioactive source of betas, retarding field, photomultiplier. Adjust the field until the betas just barely make it through.
A beta battery would, theoretically, charge itself up to that point. Cool, if you need a super-high-Z, hundreds of kilovolts reference.
Hmmm, you could do beta source spectroscopy with the retarder/detector rig. I wonder if that's ever been done.
Surely there must be a few even dumber ideas for voltage references.
John
where
Sure. Potato batteries, for instance. ;)
Cheers
Phil Hobbs
where
Ah, the Irish Standard Volt.
John
where
Not to be confused with Dan Quale's 'Potatoe batteries'.
-- Politicians should only get paid if the budget is balanced, and there is enough left over to pay them.
where
Yep. It's okay in the morning, but gets a bit wobbly after lunch. ;)
Cheers
Phil Hobbs
ElectronDepot website is not affiliated with any of the manufacturers or service providers discussed here. All logos and trade names are the property of their respective owners.