Methods for dropping 9v down to 5v

[jacobyjd]

So I have a couple situations where I'm using ICs that need 5v power along with ICs that can use 9v. I know the most solid method of achieving 5v from a 9v regulated supply is to use a regulator...what are some other ways of doing this?

I've tried it using a voltage divider, which works ok--I know some of you guys do some pretty wild things with diodes and whatnot--I'd like to have a more comprehensive set of ideas to work with as I customize designs where I run into this obstacle. Thanks!

[SonicVI]

5V zener diode?  what was wrong with the voltage divider?

[~arph]

78L05.


sorry just read you don't want to use a regulator. I would though. Thiss will make sure that you have a stable 5v when the battery voltage goes down.

[Sir H C]

Some questions first:

Current needs?

How consistent the current usage (does it vary a lot)?

How well regulated do you need?

Why not a regulator, you can get ones that look like a little '92 transistor?

What is the driving parameter, cost, board space, reliability, ease of use, etc?

[jacobyjd]

Some questions first:

Current needs?

How consistent the current usage (does it vary a lot)?

How well regulated do you need?

Why not a regulator, you can get ones that look like a little '92 transistor?

What is the driving parameter, cost, board space, reliability, ease of use, etc?

Good questions--really, my question is a general one. I'm really not opposed to any one method, be it a regulator or whatever. I'm just curious as to what options are available in case I have tighter specifications on future builds for things like current, space, etc.

So far, we have:
-regulator--I've used these before, and this would be my preference
-voltage divider--I've used this for testing, but it's definitely not as stable
-zener diode--this I've never tried--I actually don't know a whole lot about zeners, so I'll do some reading.

What other possibilities are there?

[Mark Hammer]

Zeners don't actively attempt to stabilize, though they come in handy in a pinch.  Several tens of thousands of Phase 90 users think it works pretty okay.

Dividers are another way, but they only provide a proportion of what you start out with.  If what you start out with changes (battery drop) then the result drops too.

Series diodes can be used to subtract voltage drops.  So, in EPFM, Craig Anderton has a circuit (CMOS switcher) where he assumes you're starting out with a +/-9v supply, but wants to power a CMOS chip safely, so he sticks three Si diodes in series along each power line to drop +/-9v down to +/-7.5V.  Again, this subtracts a fixed voltage drop from an unknown or perhaps unreliable starting point.  In Anderton's case, the circuit would work even if a pair of 9v batteries had dropped to 8v each, because +/-6.5v is still fine for the CMOS chip.  If you want to drop a known 9V supply down to 5V, though, seems to me that a 78L05 would be smaller and cheaper than eight 1N4001s, and provide greater certainty.

[frequencycentral]

Zeners don't actively attempt to stabilize, though they come in handy in a pinch.  Several tens of thousands of Phase 90 users think it works pretty okay.

Mark, I've wondered why the Phase 45/90 use a diode instead of a resistor for the vref. Is there any reason it couldn't be replaced witha  resistor providing the ratio is correct?

[earthtonesaudio]

Another vote for regulator, but a different type:

LM2937 500mA, low dropout, low quiescent current, not picky about output capacitor, about $2 US.

[Gus]

The low drop out reg as posted above might be your best bet.  3 terminal regs often need >3 VDC in to out for low ripple and good regulation.

"Zeners don't actively attempt to stabilize"?  Simple zener circuits are often shunt regulators they will regulate withing a range of voltages(due to current in the diode) from the lowest speced current until they burn up.  The way you design a simple resistor zener shunt reg is first know the current draw of the circuit then you select the series in resistor to pass a bit more current than the max current needed by the circuit.  You need to know the wattage rating of the zener to make sure it does not burn up.

9 to 5 with a zener would be a resistor to the 9 the other end to the anode of the 4.7VDC 5.1VDC etc zener, cathode to ground often a electro bypass cap is across the zener.  The output is from the anode of the zener and ground.  Something to be mindful of are the current in the zener when it does not have a load across it, this is when max power in the zener happens.

If the 5 VDC load is constant a resistor divider and cap can work OK.  If the load has big current swings a regulator circuit is what you might want  work.

You can build a regulator from tubes, BJTs, fets, opamps etc.

As posted above in the thread for good help you need to post the current needed for the 5 volt section and the available current from the 9volt supply. 

[Mark Hammer]

Zeners don't actively attempt to stabilize, though they come in handy in a pinch.  Several tens of thousands of Phase 90 users think it works pretty okay.

Mark, I've wondered why the Phase 45/90 use a diode instead of a resistor for the vref. Is there any reason it couldn't be replaced witha  resistor providing the ratio is correct?

Well that's the thing, isn't it -  the mixture of relatives and absolutes.  The op-amps may be entirely satisfied with a bias voltage of one half of whatever V+ is, but the JFETs want a specific bias voltage.  So the zener provides a means of identifying a specific range of possible bias voltages for the JFETs and the trimpot takes it the last lap and fine tunes within that range.

By analogy, many of the MN3207-based pedals (and the very reason why the 3207 probably exists) use +9v for the supply, but employ a 5V regulator to drop the supply voltage down for the BBD.  Why?  Because when the bias provided to the BBD is based on an unregulated battery, setting the bias trimpot for the BBD with a fresh 9v battery may only be valid for a little while.  In contrast, if one regulates that 9v down to 5v and powers/biases the BBD using that derived supply, whatever you set the bias to will remain valid until the battery is too damn weak to be of any use.  Again, the op-amps may be quite content with relative biasing, but the BBD will want absolute biasing.

[km-r]

buck converter 

[Sir H C]

Zeners don't actively attempt to stabilize, though they come in handy in a pinch.  Several tens of thousands of Phase 90 users think it works pretty okay.

Mark, I've wondered why the Phase 45/90 use a diode instead of a resistor for the vref. Is there any reason it couldn't be replaced witha  resistor providing the ratio is correct?

Also the zener method is better with a low impedance circuit coming off it, as the resistive divider has the divide ratio whacked by whatever current the circuit is using.  So you set the voltage to 5v out, but if the circuit uses enough current, it will pull this down further.  Also need a cap there to hold it steady, lowering the AC impedance for either method.

[earthtonesaudio]

An alternative to the zener+resistor approach is the buffered zener.  You can use a big (but not too big) resistor to limit zener current (zeners are noisy at high currents) but then buffer the zener voltage with an emitter follower or something. 

You could use a constant current source to feed the zener, for less noise than a resistor.  But by then you've got like 6 components and still not as good of performance as a regulator...

[Gus]

Yes you can build all kinds of circuits for a regulator but until the first poster posts the requirements the thread will be full of noise.

To do this right, to start, you need
current being used by the 5 volt section
Current the 9volt section can supply
Size wanted
Space the reg is in
Cost
how well regulated it is needed to be

Then you can select the circuit type.

Google power supply design read some app books.

[jacobyjd]

Yes you can build all kinds of circuits for a regulator but until the first poster posts the requirements the thread will be full of noise.

Right. As I mentioned above, I don't have a specific application in mind--I was just looking into the possibilities that are available out there. It's hard to learn how to properly apply something so far-reaching as power supply design in a niche as small as stompboxes when one doesn't already have a broad knowledge of where to look. The amount of reading out there is fantastic and full of great knowledge, but it would take me weeks to satisfy my curiosity.

In contrast, a question on the forum to summarize everyone's thoughts, ideas, and commonly-used applications for this specific niche of electronics takes a matter of minutes.

Thanks, everyone, for your responses--I definitely have some new info (to me) to crunch on for the next couple of days.

[David]

It won't get you 5V, but there is a post here:  http://www.diystompboxes.com/smfforum/index.php?topic=18596.msg109649#msg109649
that will produce rock-solid 4.5V using a 386 audio amp.

[~arph]

It won't get you 5V, but there is a post here:  http://www.diystompboxes.com/smfforum/index.php?topic=18596.msg109649#msg109649
that will produce rock-solid 4.5V using a 386 audio amp.

..not unless you have a rock solid 9v too.. 

[R.G.]

There are only a few generic ways to do this.
1. Voltage divider; as you've found, unless the current provided out of the divider is less tnan 10% and preferably less than 1% of the current through the divider alone, the load causes the divider voltage to vary. That is, it has poor regulation (that is, voltage stability) with changes in load. It also has no ability to keep a constant output voltage with changes in input voltage, so the line regulation is absymal. Voltage dividers can only be practical for low current loads.
2. Series regulators. There is some series element which lets a little of the higher voltage through, and actively samples the output voltage, trying to keep it to a constant level. This breaks further into linear and switching series regulators. A linear regulator uses some active element, such as a transistor, FET, or variable resistor to lower the output voltage, and changes the amount of current it lets through to keep the output voltage constant. A switching regulator chops the incoming voltage into rectangular bites, and lets through only wide enough bites to keep the output voltage constant.
3. Shunt regulators. Shunt regulators need a series element to drop the incoming voltage across. For linear shunt regulators, this is a resistor, for switching shunt regulators, this is an inductor. The way a shunt regulator works is that the series element is sized to let through the maximum current the load will ever need. The active part of the regulator then "eats" the difference between the maximum current and the current the load needs to keep the voltage constant.

Zener diodes are shunt regulators. They have the quirk that after the voltage across them passes the zener voltage, they will let any amount of current pass, right up to and past where they burn up from the sheer power of the current times the zener voltage. This is where that limiting series element comes in. A zener has no other way to stop eating current, and zeners can and will burn up given any chance by the series element to do so.

That's really all there are.
Voltage dividers are impractical for all but the tiniest loads.
Series linear regulators are easy, cheap, reliable, self protecting and high performance.
Series switching regulators are harder, more expensive (although getting cheaper), modestly self protecting, even higher performance and great makers of RF noise.
Zener regulators work for small loads, but will burn out on you in any real power providing application unless you really know what you're doing. Active linear IC shunt regulators exist, but they're primarily for voltage references. Active switching shunt regulators work fine, but cost more, are harder to design, yada, yada, yada.

The three terminal regulator is the sweet spot for power regulation unless you really know what you're doing and have valid technical reasons for doing something else. If you don't meet that last criteria, stay with three terminal regulator chips.

[Sir H C]

And in terms of cost, unless you are making thousands of these, the 3 terminal regulator cost is not that much.

[jacobyjd]

Thanks, RG, that is exactly what I was looking for.

The low-current application for voltage dividers matches exactly with the results I've gotten from experimenting on my breadboard.

Part of the reason for asking about this is that, given my small collection of parts, regulators aren't something I always keep on hand, so in a situation where I don't have one on hand, but still want to continue to pursue results, knowing other ways to 'fake' it (maybe not in the most efficient or precise manner) in a pinch is invaluable.