How to hold a specific voltage

Hi
I want to make a bench supply with fixed voltage of perhaps 12VDC and 15VDC.
Doing so, I am thinking about having a high enough main supply of like 13-14V or 30V?, and then using either a zener diode or voltage regulator to create the correct Vout on the bench supply.

Are there any reasons to chose a dedicated voltage regulator like LM78XX instead of zener diodes?
Can zener diodes be found, which are capable of handling 1A?

Current! The current will determine what the best choice would be. Give this link a read as it explains it well. Based on your post a simple regulator like a 7812 or or 7815 should do fine up to about an amp with adequate heat sink. When using regulators like the 78XX series remember to consider "drop out" voltage. You want the source voltage a few volts above the regulated voltage. Work from the regulator data sheet(s) and get familiar with how the regulators in this family actually work. My suggestion would be using regulators between the options you mention.

Ron

Things like bench supplies are best bought (and cheaper) than making one on your own, nothing you can home brew without detailed knowledge of electronics will be inferior to a unit you could buy for half the price (in both time and materials)

But i do have detailed knowledge of electronics, just mostly on a theoretical level.
A bench PSU in real life, takes a lot of practical experience, which I lag, so I ask here )
And i would think that it would be cheaper to build one yourself, a PSU doesnt need many parts afterall.

Thanks for links.

And i would think that it would be cheaper to build one yourself, a PSU doesnt need many parts afterall.

Click to expand...

You would be wrong at least if you wanted a GOOD regulated supply.

I want a psu with fixed 12V, 15V and 1 or 2 regulated outputs in the scale 2-30V
But you may link to your cheaper product.

I have access to most basic components on my school, so I would probably only need a transformer and a box/casing

What are your regulation requirements? There are two terms, line regulation which is affected by the difference in input/output voltages, and load regulation, which is the regulation of the supply based on the load's current, if you can't specify the line/load requirements then you can't even start designing the supply. The set voltage and the design current and range are secondary to the actual regulation requirements. Many modern bench supplies are linear at fractions of a percent of their set voltage/current draw

I fail to see how Line/load regulation should be an issue here because of 2 reasons:

1 I would expect the input voltage to be relatively stable, if im wrong please tell me why im wrong )
2 If i get a good voltage regulator, and taken that i dont use it to power a brushless motor demanding too much current, is this really a problem?

As said earlier, my lack of real world experience may cause my expectations to be...naive

lotus21 said:

But i do have detailed knowledge of electronics, just mostly on a theoretical level.
A bench PSU in real life, takes a lot of practical experience, which I lag, so I ask here )
And i would think that it would be cheaper to build one yourself, a PSU doesnt need many parts afterall.

Thanks for links.

Click to expand...

Way wrong. You need a chassis and good heatsinks, you will need power transformers and filter caps and panel meters to read the voltage. I built my bench supplies but mostly just to learn and make them "stackable" and also able to pararllel outputs.

1) Yes, you are wrong. Via a transformer for one the voltage will depend on the load. If you pulse the supply frequently it will lose voltage regulation because of the swings in current (load regulation) if the compliance (ability of the regulator you attach to it to regulate the output voltage) isn't within the input swing from load currents you will lose regulation.

2) A 'good voltage regulator' is HIGHLY application dependent. You assume too much, in the real world a good solid bench power suppoly has to be able to hold a very precise voltage (down close to mv's) AND still provide that regulation for it's supply range (sub 1amp to 2-3-4etc..+ amps.)

The line and load regulation on a typical ATX power supply is only a few percent of the given values, and they contain hundreds of components. A quality bench supply will have many hundreds of components in part because of the increased regulation requirements and and application independent components (pulse current, thermal compensation etc..)

I would like to qualify this post on one basis. You said you have a good theoretical understand of power supplies, which obviously you do not because I do even though I've never built one (aside from a small 100ma-2ma constant current supply)

I have however tested both high quality linear and wall wart style supplies, so I have no practical understanding of the actual realization of power supply and stability issues.

1) I hope not the voltage out of the transformer would depend on the load? I do believe the voltage of lets say a: (24V-0-24V) transformer might be (26V-0-26V) without load, but with load it 'should' hold at (24V-0-24V) Ofcourse I will keep below the specifications on the transformer.

2) Voltage regulators alone can not keep voltage to lets say +- 50mV? I thought they could.

And i have looked inside my ATX PSU, but those 2 shouldnt be compared easily, since i can understand a computer would need much smaller tolerances than my bench psu, and the ATX psu delivers a lot of power and has a lot of outgoing connections.
So - same same, but different.

If all you want is a fixed 12 and 15 volts not to exceed 1 amp on either line all you really need is a 12.6 volt output transformer rated at about 3 amps. Rather than using the traditional LM78XX series of regulators you could use what is known as a low dropout version like the LM2940 series which can be found here. If you want to use the LM78XX series regulators then find a transformer with an 18 volt secondary similar to this one. Come off the transformer with a full wave bridge, even a 4 amp 50 volt bridge from Radio Shack would be fine. Then I would parallel a pair of 2200 uF 35 (or greater) volt rated capacitors. Place a pair of .33 uF caps right at the input to each regulator. On the regulator outputs place a pair of .1 uF and 10 uF capacitors.

Given a choice I would use the 18 volt secondary transformer. You are building a basic fixed voltage bench power supply. Anyway, that is all I would do.

Now if you want something adjustable then I would use a pair of LM317 regulators as can be found here. Circuits are shown in the data sheet.

Ron

Thanks
One thing Im wondering about, is the capacitors, calculating the values for the caps using to smooth out the full wave diode bridge is no problem.
But how do I estimate 'buffer' capacitors like the ones you mentioned on the input/output leg on the regulator?
How do you get those values you mentioned?

since i can understand a computer would need much smaller tolerances than my bench psu

Click to expand...

Why would you think that?
A bench PSU for any kind of analytical understanding would require higher precision. Even CPU/GPU's have on board regulators AT THE CHIP to avoid line/load regulation issues with a centralized power supply that can't be solved any other way.

Reloadron, how are we into the 15th post on a power supply design without finite definable measure of line OR load regulation?

lotus21 said:

Hi
I want to make a bench supply with fixed voltage of perhaps 12VDC and 15VDC.
Doing so, I am thinking about having a high enough main supply of like 13-14V or 30V?, and then using either a zener diode or voltage regulator to create the correct Vout on the bench supply.

Are there any reasons to chose a dedicated voltage regulator like LM78XX instead of zener diodes?
Can zener diodes be found, which are capable of handling 1A?

Click to expand...

Sceadwian,

That was the original post. Now my take and granted only my take is the original poster wants little more than a fixed bench type power supply that will deliver 12 & 15 volts with a maximum current not to exceed an amp. There is no mention of precision regulation or output response times to adjust for loads or rapidly changing loads. There is no mention of a laboratory grade precision power supply? Certainly nothing in the original post that would imply that is what the OP is wanting to build. For a basic fixed 12 & 15 volt power supply I saw no reason to drag in output transient response times for loads or ripple factors. I provided some basic information and suggestions on what I thought the OP wanted.

Even if the OP were to build or buy a precision lab grade bench power supply the average home electronics enthusiast sorely lacks the small mountain of test equipment required to actually test such a supply.

lotus21:

1) I hope not the voltage out of the transformer would depend on the load? I do believe the voltage of lets say a: (24V-0-24V) transformer might be (26V-0-26V) without load, but with load it 'should' hold at (24V-0-24V) Ofcourse I will keep below the specifications on the transformer.

2) Voltage regulators alone can not keep voltage to lets say +- 50mV? I thought they could.

And i have looked inside my ATX PSU, but those 2 shouldnt be compared easily, since i can understand a computer would need much smaller tolerances than my bench psu, and the ATX psu delivers a lot of power and has a lot of outgoing connections.
So - same same, but different.

Click to expand...

The voltage out of a simple step down transformer like we are talking about should remain pretty constant as long as the rated load for the transformer is not exceeded. The transformer is a ratio device so voltage out is a function of voltage in. The transformers current out is a function of the wire gauge used in winding it. Granted there is more to it but in the best interest of keeping it simple, that is about it. If I have a transformer rated for 120 VAC RMS input voltage and a 12 VAC RMS output voltage I can pretty much assume a 10:1 turns ratio. Mains voltage is not a fixed value. 120 VAC could be 115 VAC or 125 VAC. This morning mine is around 123 VAC at the outlet beside me. In the case of my example if the input voltage was 125 volts there would be a proportional change on the output and it would be 12.5 volts.

As to an ATX PSU? Actually the voltage output regulation is not all that great. For example the 12 volt rail is only regulated to +/- 5% so an output anywhere between 11.4 and 12.6 is acceptable. The transient load slew rate is = 1.0 A/μs. So if I step 5 amp loads on it it should return to whatever its voltage was within 5 uS. The design spec calls for 50% of rated load so if I have an ATX PSU rated for 20 amps on a 12 volt rail and I apply a 10 amp load it should respond within 10 uS and be in specifications. The real critical regulation is a function of the motherboard for the uP and chip-sets etc.

So it all comes down to what the end user wants from a power supply.

Ron

Great I am learning already )
And ATX PSU's are sloppy at best.

So first things first, keeping the transformer within limits, and making sure that the voltage regulator input is at a sufficient level, then I can make a fixed output on my PSU at 12V and 15V.
Sorry if I asking something i could look up myself in datasheets, but what precision can i expect from regulators?

Now the regulated outputs, what designs can you recommend? (precision I would prefer to stay within +-5% - hopefully better)

What I want my psu to do is, powering RC-servos, a microcontroller and sensors attached to it. All in all, not anything that would draw much current, the reason I want a more 'powerful' PSU is that I want to be able to use it in future development.

In the best interest of keeping this simple:

Load regulation is the capability to maintain a constant voltage (or current) level on the output channel of a power supply despite changes in load.

Line regulation is the capability to maintain a constant output voltage level on the output channel of a power supply despite changes to the input voltage level. Line regulation is expressed as percent of change in the output voltage relative to the change in the input line voltage. This measurement is normally taken under conditions of low power dissipation to reduce the effect of average chip temperature and is typically represented as a percentage of the output voltage.

Click to expand...

The above quoted material was stolen from the Wikipedia while they weren't looking.

Sorry if I asking something i could look up myself in datasheets, but what precision can i expect from regulators?

Click to expand...

Ask all the questions you wish to ask, you may not get the answers you want but you will get answers.

The expected precision or better put output uncertainty of a regulator is a function of many things, for example temperatures. This is why the Tron God invented data sheets.

When designing or building a circuit the individual components are chosen for their intended applications. Enter a pile of data sheets. So what uncertainty can you expect from regulators? Depends on the regulator used which is selected based on the uncertainty you need.

When the ATX PSU design guide was written the needed uncertainties were based on the intended application of the PSU. The system application dictated the PSU parameters and specifications.

Yes, if I need a current of 3 amps I would not begin my design with a transformer only rated for 2 amps, I would likely look for a transformer rated at 4 or 5 amps to allow headroom and so it goes.

Ron

lotus21 said:

2) Voltage regulators alone can not keep voltage to lets say +- 50mV? I thought they could.

Click to expand...

Read the specs of the regulator. The typical LM7824 has an output voltage spec in the 3% ballpark as I recall. They also spec load and line regulation. The specs are on the data sheet.

Reloadron said:

If all you want is a fixed 12 and 15 volts not to exceed 1 amp on either line all you really need is a 12.6 volt output transformer rated at about 3 amps.

Click to expand...

12.6 VAC would probably be a little low to build a 15V out reg. You would need at least 18V ballpark input to the reg to stay in regulation and you have to allow for a couple of volts of ripple on the input rail.