Zener Diodes / Voltage Regulation

[FyberOptic]

I'm having some confusion regarding zener diodes which I hope somebody
much brighter on the subject can clarify somewhat, or at least point
me in the right direction!

The basics of diodes has always been "current can only flow one way",
and beginners are left to assume this means they'll always do just
that. But obviously that's not the case, since normal diodes will
stop doing such a thing if they're pushed too far, I understand. And
as a result, people take advantage of this effect more reliably with
the zener diode apparently, by making them "fail" at much lower rates,
it would seem.

So while this is possibly out of the scope of asking on a newsgroup,
what sorts of applications could they be used for reliably? I hear
they can be used as regulators, but only if the load is fairly
constant. So, for something like an electronic circuit, with chips
going on and off and constantly changing the load, would this be of
any use there? I've seen them used in situations such as when one is
drawing power from the PC parallel port. I'm assuming this isn't
related to actually getting that power, but keeping it from going too
high.

I'm curious because, despite my lack of knowledge regarding the
details of analog circuits, I have a pretty decent understanding of
digital ones, and find myself making and planning new things in that
realm all the time. I'd like to power some of these sorts of things
off of batteries sometime instead of a 7805 on an ac adapter, and am
interested in the best way to go about it. One project in particular
involves an 8052 microcontroller with an lcd which I'd like to make
portable, for example.


On a side note, how possible would it be to boost the voltage from a
couple of AA's to run 5v logic reliably? I can't really think of how
such a circuit would be made, though I know they surely exist. I
assume it uses capacitors and an oscillation of some kind, but I
dunno. When I think of changing voltages, I mostly just think of
transformers.

Anyhoo, any help or pointers would be greatly appreciated!

 

[Eeyore]

FyberOptic wrote:

Could you perhaps try a post with a couple of straight questions ?

Graham

 

[Michael A. Terrell]

FyberOptic wrote:

Could you perhaps try a post with a couple of straight questions ?

Graham

Dammit, you KNOW that this is yet you
just have to slam the newbies.


--
Service to my country? Been there, Done that, and I've got my DD214 to
prove it.
Member of DAV #85.

Michael A. Terrell
Central Florida

 

[Michael Black]

So while this is possibly out of the scope of asking on a newsgroup,
what sorts of applications could they be used for reliably? I hear
they can be used as regulators, but only if the load is fairly
constant. So, for something like an electronic circuit, with chips
going on and off and constantly changing the load, would this be of
any use there? I've seen them used in situations such as when one is
drawing power from the PC parallel port. I'm assuming this isn't
related to actually getting that power, but keeping it from going too
high.

In the old days, voltage regulation wasn't common. There were a few
things that could make use of it, and you'd see voltage regulation
applied to that stage rather than to the whole equipment. Only exotic
lab equipment would have fully regulated power supplies.

So in the tube days, that voltage regulation would be done with VR tubes,
ie Voltage Regulation tubes. There'd be a current limiting resistor
from the power supply, and then the two terminal tubes would kick in
when the voltage went too high. Change the load, and you'd need to
change the current limiting resistor.

Transistors came along, and so did zener diodes. They were like
those VR tubes, except you could now get zeners that would
regulate at low voltage.

IN the early days of transistors, you would see those zener diodes
used basically like those VR tubes, regulating specific stages
rather than whole power supplies. They were lousy for general
supply use, since they tended to be low current devices and you
had to adjust the current limiting resistor as the load changed.

But they disappeared mostly after a decade. People made the
transisition to solid state devices, and realized things weren't
quite like the days of tubes. Tubes used such low current at
high voltage that the power supplies were pretty high impedance output.

But solid state needed high current at low voltages, and the old
high impedance output power supplies didn't really work so well. You'd
see people adding and adding large value electrolytics to the output
of the rectifier(s) in the solid state supplies, trying to get that
low impedance output. And it was never a complete success.

You'd start to see zener diodes used to supply a constant voltage,
rather than real power, so they'd be used with transistors that would
pass the actual current. And then you'd see supplies where there
was feedback, so no matter what the load on the supply, the voltage
was constant. Gradually these came into force, so fewer and fewer
solid state supplies had anything but a regulated supply. But they
weren't regulated because the devices had an absolute need
for an exact and constant voltage, but because the regulating state
allowed for that low impedance output that the devices wanted.

It was also a lot easier to make up a solid state regulator than
a tube type regulator, since large numbers of transistors would
fit in the space of a single vacuum tube.

Then a few years later, the discrete regulators basically disappeared.
IC regulators became common, especially 3 terminal regulators. It
became so easy to use them, it was hardly worth not using them. No
need to figure out the current limiting resistor, and no need to adjust
it every time the load changed. It was great.

You now rarely see zeners. They are relegated to where a low current
voltage source is needed. And where the drain of a 3 terminal regulator
is prohibitive.

Michael

 

[John Popelish]

Here's a little challenge for the OP. Given that

Q=CV
C(series) = 1/(1/C1 + 1/C2) (alternatively C1C2/(C1+C2))
C(parallel) = C1 + C2
Qt(series) = Q(C1) = Q(C2)
Qt(Parallel) = Q(C1) + Q(C2)

Then take two capacitors with a ratio of 10:1 (say 1uF and 0.1uF). Hook
them up in parallel and charge them to some voltage (say 5V)

Disconnect them and reconnect in series (a thought experiment, really)
and calculate the total voltage across the pair. You might be surprised
at the result.

Is the challenge to find the error in these equations?

 

[PeteS]

John Popelish wrote:

I can't really think of how
such a circuit would be made, though I know they surely exist. I
assume it uses capacitors and an oscillation of some kind, but I
dunno. When I think of changing voltages, I mostly just think of
transformers.

There are ways to charge a pair of capacitors, in parallel, with the
battery voltage, and then reconnect them, so that they are in series,
doubling the battery voltage, and connect that voltage to an output
storage capacitor to act as a power source while you go back and repeat
the cycle. The key words for this sort of thing are [charge pump].

Here's a little challenge for the OP. Given that

Q=CV
C(series) = 1/(1/C1 + 1/C2) (alternatively C1C2/(C1+C2))
C(parallel) = C1 + C2
Qt(series) = Q(C1) = Q(C2)
Qt(Parallel) = Q(C1) + Q(C2)

Then take two capacitors with a ratio of 10:1 (say 1uF and 0.1uF). Hook
them up in parallel and charge them to some voltage (say 5V)

Disconnect them and reconnect in series (a thought experiment, really)
and calculate the total voltage across the pair. You might be surprised
at the result.

Cheers

PeteS

 

[John Popelish]

Snort

Do tell. Let's see

Q = CV. I think we can agree on that. It was figured out a long time
before we were born.

Needs some interpretation. Q is the charge transfered by a
change in voltage V.

Series capacitance; reciprocal addition is the same (for two devices) as
product over sum.

True but not terribly applicable to placing two charged
capacitors in series and measuring their open circuit voltage.

Parallel capacitance is simply the addition of the capacitances.

True, and useful in determining how much total charge it
will take to produce a given change in voltage across the
parallel pair.

The charge on any number of caps in series of any capacitance is identical
The total charge on any caps in parallel is the sum of charges.

What does that have to do with the open circuit voltage of
an arbitrary string of precharged capacitors?

I don't think I said anything different above.

The question is applicability of those somewhat simplified
equations to this problem.

I assert that if you charge a 1 uF cap and and a a 1 nF
cap, each to 12 volts and place them in series, the total
voltage across the pair can be either zero volts or 24
volts, depending on the relative orientations. How do your
generalities show this to be true or false?

 

[Charlie Siegrist]

The basics of diodes has always been "current can only flow one way",

That's a bad assumption, and not the "basics" of diodes at all. Maybe a
Cliff Notes version.

...the zener diode apparently [will] "fail" at much lower rates...

Start understanding better by using the proper term, which is not "fail,"
it is "breakdown."

So while this is possibly out of the scope of asking on a newsgroup,

Not at all, your question is exactly what this newsgroup is for.

what sorts of applications could they be used for reliably? I hear
they can be used as regulators, but only if the load is fairly
constant...

Actually, as long as the load and the supply are reliable within a fairly
large range, the zener can be used pretty safely, as long as the regulation
does not have to be tight. Zeners have quite a large +/- tolerance range,
on the order of 15%.

I'd like to power some of these sorts of things
off of batteries sometime instead of a 7805 [or] an ac adapter, and am
interested in the best way to go about it.

The best way to go about it is to choose a battery value that most closely
matches the need of the load. Any voltage regulation using a zener will
waste power from the battery, and there are many, many batteries available
on the market to pretty much match any requirement.

On a side note, how possible would it be to boost the voltage from a
couple of AA's to run 5v logic reliably?

It's best done, as above, by choosing the battery voltage you need. Any
voltage conversion, up or down, again, will require some possibly
significant amount of power just for the conversion. Since getting 5V with
a battery of AA cells is not practical without conversion, due to the
voltage variation of differing types, you want to find a marketed 5V rated
battery.

To choose the right one, figure out the amperage requirement of your
project, then decide how long you want to run the project reliably. You
mentioned wanting to use AA batteries, which are in the range of 2000-2500
mAh for NiMH rechargeables. Say you are going to use 1A for 4 hours.
That's 4 amp-hours. Multiply by 125% and you arrive at the requirement for
a 5 amp-hour (5000mAh) battery, which is twice the typical capacity of the
NiMH AA. Now find a 5000mAh 5V (typically 4.8V, 1.2V per cell) battery
pack, which can be had for around $30 US.

When I think of changing voltages, I mostly just think of
transformers.

Which, of course, don't work well with DC. Hence the prevalence of AC in
power transmission.

 

[Charlie Siegrist]

What does that have to do with the open circuit voltage of
an arbitrary string of precharged capacitors?

That's the question of the day. I'm thinking the answer depends highly
upon unstated conditions, and then after the stating of said conditions,
lots of sammy-the-snake squiggly lines and cuss words such as "dv/dt" and
"logarithmic graph paper." Either that, or magic.

 

[PeteS]

Is the challenge to find the error in these equations?

Snort

Do tell. Let's see

Q = CV. I think we can agree on that. It was figured out a long time
before we were born.
Series capacitance; reciprocal addition is the same (for two devices) as
product over sum.
Parallel capacitance is simply the addition of the capacitances.
The charge on any number of caps in series of any capacitance is identical
The total charge on any caps in parallel is the sum of charges.

I don't think I said anything different above.


Cheers

PeteS

 

[Michael A. Terrell]

| Here is a basic tutorial for the buck and boost converters:
| www.national.com/appinfo/power/files/f5.pdf

This pdf says a * has an efficiency of 80% @12v input.
Then if I replace my 1990 Honda engine's 4 resistors (
each 5.6 ohm ) which cut voltage ( max 12.9 v ) to fuel
injectors ( solenoid @ 2.1 ohm ) with a *, can I avoid
these 4 resistors' wastage of current ( 15.71 w ) by using
a * which presumably will use just 5.89w x ( 20% ÷ 80%
) = 1.47w ?
Was any * available in 1990, to Honda or car owners ?
This engine's ECU was made by NEC ( japan ), could
NEC include a * in this ECU, so only max 3.51v is fed
to these 4 injectors ( w-o any need for resistors ) ?

[Image]

This is not a binaries newsgroup. You can post it to
by attaching it as a GIF, JPEG,
or other common image file. Then, post a message here giving the name
of the post. If you post a binary file here, most people will never see
it.


--
Service to my country? Been there, Done that, and I've got my DD214 to
prove it.
Member of DAV #85.

Michael A. Terrell
Central Florida

 

[neon]

you ask questions and i will try to answer zeners are diodes doped to breakdown to whatever range is required you may use them as regualtor as long you do not slip beyond the soft knee of the regulator or zener, unfortunately thy are not used much nowdays there are better devices and at real low voltage are realy lousy. Transistor Vbe breakdowbn is much sharper BUT you cannot predict the voltage breakdown. all devices are controlled by one factor HEAT

 

[default]

| Here is a basic tutorial for the buck and boost converters:
| www.national.com/appinfo/power/files/f5.pdf

This pdf says a * has an efficiency of 80% @12v input.
Then if I replace my 1990 Honda engine's 4 resistors (
each 5.6 ohm ) which cut voltage ( max 12.9 v ) to fuel
injectors ( solenoid @ 2.1 ohm ) with a *, can I avoid
these 4 resistors' wastage of current ( 15.71 w ) by using
a * which presumably will use just 5.89w x ( 20% ÷ 80%
) = 1.47w ?
Was any * available in 1990, to Honda or car owners ?
This engine's ECU was made by NEC ( japan ), could
NEC include a * in this ECU, so only max 3.51v is fed
to these 4 injectors ( w-o any need for resistors ) ?

The four resistors would only "waste" 15 watts if they were pulling
current all the time. The fuel injectors only open very briefly to
allow a spray of fuel in.

The resistors may also serve another function - like modify the inrush
current or help absorb flyback current.

15 watts is a drop in the bucket when it comes to auto efficiency -
even if the injectors stayed open 100% of the time. Synthetic motor
oil, intelligent driving, tire pressure, etc., will all make a greater
impact.

 

[quietguy]

Just thought I'd mention that Jaycar (and others I presume) sell a device
that does this - ie a portable USB supply

David

 

[neon]

I'm having some confusion regarding zener diodes which I hope somebody
much brighter on the subject can clarify somewhat, or at least point
me in the right direction!

The basics of diodes has always been "current can only flow one way",
and beginners are left to assume this means they'll always do just
that. But obviously that's not the case, since normal diodes will
stop doing such a thing if they're pushed too far, I understand. And
as a result, people take advantage of this effect more reliably with
the zener diode apparently, by making them "fail" at much lower rates,
it would seem.

So while this is possibly out of the scope of asking on a newsgroup,
what sorts of applications could they be used for reliably? I hear
they can be used as regulators, but only if the load is fairly
constant. So, for something like an electronic circuit, with chips
going on and off and constantly changing the load, would this be of
any use there? I've seen them used in situations such as when one is
drawing power from the PC parallel port. I'm assuming this isn't
related to actually getting that power, but keeping it from going too
high.

I'm curious because, despite my lack of knowledge regarding the
details of analog circuits, I have a pretty decent understanding of
digital ones, and find myself making and planning new things in that
realm all the time. I'd like to power some of these sorts of things
off of batteries sometime instead of a 7805 on an ac adapter, and am
interested in the best way to go about it. One project in particular
involves an 8052 microcontroller with an lcd which I'd like to make
portable, for example.


On a side note, how possible would it be to boost the voltage from a
couple of AA's to run 5v logic reliably? I can't really think of how
such a circuit would be made, though I know they surely exist. I
assume it uses capacitors and an oscillation of some kind, but I
dunno. When I think of changing voltages, I mostly just think of
transformers.

Anyhoo, any help or pointers would be greatly appreciated!

One way is not exacty right they do conduct both way one is forward current the other is saturation current breakdown. if the current is not limited because of the power dissipation of the device blweeyyyy. HEAT blows them up. zeners below 6v have a very lousy knee and zeners above 6v have a dedency to drift with temp if it is in the same package configuration. it is hard to find 3 zeners in the same family number to be within 1% of the specified voltage. at one time that was all we had nowdays the sky is the limit and they are not more use much. and as voltage controllers they must carry al load curent to regulate