Mains Voltage variation within North America?

[eyAyXGhF]

I just read the choice of power supplies... So let me get this straight, you're supplying 15 volt linear regulators from a 12 volt halfwave rectifier and you expect it to work at all? The instant you apply any load to that the linear regulators will stop working because the ripple voltage will be massively bellow it's dropout voltage especially from a half wave rectifier! Even LDO's wouldn't work properly in this situation you're working three volts under your regulator voltage at it's rated load! Might as well skip the regulators altogether cause they're not serving a purpose on anything other than a fraction of the supplies load at some perfect supply voltage.

Hey Sceadwian, I know what you're saying but let me try to explain myself. Although I admit to being a newbie at this.

I'm using a 12VAC-1000mA wallwart and drawing 250mA from it. Because I'm not drawing full load from the wallwart, it is outputting slightly higher voltage since it's rated 12V for full load. In my case (with our nice 122.5VAC mains supply), it is actually outputting 14.2V. This is RMS, so the peak to peak is actually about 40V (14.2 x 2 x 1.414) and measures exactly that on the scope. After diode drops on the halfwave rectifier and filter capacitors, I measure 19V and 18.5V going into the regulators with about 1V of ripple. According to the datasheets, the LM317 and LM337 require a voltage drop of about 1.6V at the load I'm putting on them, making this design totally acceptable I thought.

What I didn't account for was that a few of my customers are getting supplied with way less than the 122.5VAC I experience here. I never knew it varied that much, but thanks to this thread.. I know better.

I can't move to CT transformers as you suggested as much as I'd like to, because I'm selling these things and cannot afford any CE or UL certification (see https://www.electro-tech-online.com...tuff-that-plugs-into-mains-120-240vac.123506/).

So far the only solution I'm really seeing is moving up to a 14VAC or even 16VAC wallwart, but I'm open to any other solutions that will still let me sell these things legally.

 

[Sceadwian]

I'm using a 12VAC-1000mA wallwart and drawing 250mA from it. Because I'm not drawing full load from the wallwart, it is outputting slightly higher voltage since it's rated 12V for full load.

This is a poor design choice. You NEVER EVER work AT your design goal for peak power, you have to engineer the overhead into the product.

What I didn't account for was that a few of my customers are getting supplied with way less than the 122.5VAC I experience here. I never knew it varied that much, but thanks to this thread.. I know better.

This is basic engineering 101. You don't engineer for what you expect, you engineer for two ranges, low and high, and you make sure that even if it goes towards low or goes towards high you add a little bit more of leeway... This is exactly why I don't like 'working standards' because you'll never know about them unless you're working with them, and usually this way (a product that does not work) Where as a written standard that is followed will warn you against this type of thing.

CE or UL certification is pointless, if you know you're selling a safe device then sell it. The problem is you didn't account for one variation in the circuit design already so you may not know what is or is not safe. CE and UL certification is for one thing and one thing only, insurance claims, there is nothing whatsoever preventing you from selling something that has no UL or CE certification if you want to aside from the legal culpability if you do something that injures or kills someone or otherwise damages properties you will be held to blame.

The only thing using these generic wall-warts will do for you is place the blame on the company that produces them, mind you better be sure those CE/UL certifications are real, Chinese nockoff products will place that rubber stamp on unsafe stuff all the time.

 

[ronv]

My guess is you don't want to throw away a bunch of transformers. If so you could maybe move to some low dropout regulators like the LM2990 and LM2940. That would pick you up avout a volt or more. In a panic you could also use a bigger cap with the same lead spacing if you have room to reduce the ripple. Between the 2 you might pick up enough to "slip by" until you can get a higher output supply.

 

[Sceadwian]

A volt or more won't help that he's 3 volts out from the start on the low end....

This is a quintessential example of poor engineering (no offense intended)

How about full bridge rectifiers from a properly chosen supply? That will actually solve the problem.

You can not engineer for best case scenarios, you engineer from worst case, and if you don't know the worst case, that's on you.

 

[ronv]

That would be true if he was 3 volts out, but he may not be. Since we don't know the low end where he needs to run the LDO would give him margin (2%) at 110VAC in if his numbers are correct. A 3300 Ufd. cap would buy him another few percent. Better diodes might also add a couple of points.

 

[Sceadwian]

For ripple, on a half bridge?

BAD ENGINEERING

 

[ronv]

For ripple, on a half bridge?

BAD ENGINEERING

But maybe good economics.

 

[Sceadwian]

Only in countries and industries that have no culpability laws.

Everything is free!!! Until you have to pay.... Sometimes, what it costs to make something is not worth what it costs to certify it, so you go cheap.

UL/CE basic engineering stems from these simple rules.

If you KNOW what you're doing... No problem!

 

[MrAl]

I don't trust working standards, too many different versions of them, and I find it especially criminal that the entire US power grid relies on a working standard. I've always found the divergence between groups such as electricians and commercial/industrial electrical engineers and electronics to be night and day, totally different mindsets.


That's self evident, transients and externally injected noise are possible in many situations, this however can be designed into the written standard with broad guidelines and caveats where appropriate. But I think things like expected line voltage at the junction box to a home should have a set standard that is clearly defined.


I just read the choice of power supplies... So let me get this straight, you're supplying 15 volt linear regulators from a 12 volt halfwave rectifier and you expect it to work at all? The instant you apply any load to that the linear regulators will stop working because the ripple voltage will be massively bellow it's dropout voltage especially from a half wave rectifier! Even LDO's wouldn't work properly in this situation you're working three volts under your regulator voltage at it's rated load! Might as well skip the regulators altogether cause they're not serving a purpose on anything other than a fraction of the supplies load at some perfect supply voltage. Granted this is efficient if it's getting the right supply voltage linear regulators need headroom and there's no way to avoid that statement.

Even the 14-0-14 won't help, you're still designing for 1 full volt under the linear regulators design specs, it just can't work! Just to properly run a 15volt LDO from (generally drop out is about 1 volt above VCC on an LDO or there abouts) you would need a 16-0-16 volt supply. And to account for source supply voltage variation 17-0-17 would be better, otherwise those regulators don't have anything to work with.

Hi Scead and eyAy,


Yeah it is nice to have standards written down, but we dont always have that. It would take some griping i guess.

But line conditioners are not made to work down to 90vac input just because of line transients.

Since i am just finding out now that the application is for a 15vdc 250ma regulator, then the transformer can not be as low as 14vac and be expected to work right. If the input to the regulator chip needs 17.5vdc, then the lowest the ripple can dip is just that, and that means a much higher nominal voltage will be needed. This is where the specs of 105 and 140 (more at 105 than 140 though) come into play. If the design is made to work down as low as 105vac input to the transformer, then it will work in most places in the continental US. Granted though that there could still be special cases that need a line booster/conditioner. That's the nature of the grid.

So with this in mind, a transformer with an 18v secondary might work out pretty nicely with a 3000uf capacitor for filtering (half wave). This should handle low line without a problem. We can go a little lower than that perhaps but remember there are some other things like transformer resistance to think about too which causes some peaks clipping.
The drawback to using such a high ac voltage is that high line will mean more power dissipation in the regulator chip. This has to be taken into account too for the heatsink design and air flow management. That's the nature of the beast this time the beast being the linear regulator instead of a switcher.

Usually half wave is only used when one side of the transformer secondary has to be grounded or in very low current applications. Full wave is of course better. You cant really call half wave bad engineering because it may in fact fit the application better. I have to agree though that here it is probably not a good idea unless it really has to be that way for some reason.

What else we dont know is what filter cap is intended to be used, what value in microfarads. 3000uf is probably a good idea, but it would be interesting to find out what was being used before this thread was started.

 

[ChrisP58]

The assumption that your product is safe just because you are using a UL listed transformer or power supply is absolutely wrong. It will NOT protect you from litigation if someone is hurt from using your device. You CANNOT claim your product is listed/recognized unless your entire, finished product, has been certified by a NRTL. **

Your could build a product containing 100% UL recognized components, and it could still cause a fire, or present a shock hazard.

Safety needs to be carefully engineered into every part of a product. This MUST include understanding what can go wrong when (not if) it is used outside of the expected operating conditions.

Even if you don't cary through with getting your product certified, you should at least read, understand, and evaluate the product yourself, against whatever safety standard that would apply to the product.

Unfortunately, just getting a copy of the standards can cost many hundreds of dollars.

**
NRTL (Nationally Recognized Testing Lab) Not only UL. Other north american labs include CSA and ETL. I know there are others too. And there are many testing labs in other parts of the world that can do the certification.

The 'CE' mark can be self certified, but your neck is on the line if it can be demonstrated that your product causes damage or injury because you fudged the test results or skipped some of the tests. Using an independant test labe is still best.

 

[audioguru]

In Canada, our electricity is designed TO WORK PROPERLY.
If it is ever only 102V then it is a BROWNOUT where the voltage is reduced on purpose because the system cannot handle the load. But our system is designed to handle the load and is upgraded if it ever comes close to overload.
We now have Smart Meters so we are billed at a higher rate during daytime and on work days. So many people do laundry and other electrical chores in the evening which spreads out the load.

Today I measure 121.4V (my Fluke meter is accurate) and everybody has their air conditioners running. In winter I measure 121.0V when everybody burns millions of Christmas Tree bulbs and some people use electricity to heat their homes. In spring and Fall I measure 122.0V when electricity is not used much.

We have backups for our backups (but not in Quebec) so we rarely have outages. The last outage many years ago started in America and most of North America followed since we are tied together.

 

[Sceadwian]

audioguru, here's a quick statistic for you. The US has 10 times the population of Canada, lets not even talk about heavy industry differences in the consumption of electricity.

California alone (one state) is very close to the population of your entire country!

One day you might realize that we have different problems, and direct comparison is impossible... The fact that the US electrical grid works at all simply amazes me!

So in your pride for your country, keep in mind the scale in comparison to the outside world.

 

[audioguru]

We manufacture cars, stoves and refrigerators in Canada too you know. Those manufacturers use lots of electricity.
The millions of "trailer" homes in the US don't have heating, air conditioning or laundry. If they even have electricity then it plays their TV.

 

[RCinFLA]

105 vac to 130 vac

Power companies can get away with just about anything but they fear claims of appliance failures that can be proven to be contributed by their line voltage deviance.

 

[tcmtech]

The millions of "trailer" homes in the US don't have heating, air conditioning or laundry. If they even have electricity then it plays their TV.

I am going assume you are either joking or else you have gone completely senile at minimum. :confused

My trailer house was built in 1974 and it came factory equipped with a 100 amp 120/240 electrical service, both hot and cold running water, a toilet, two sinks, a bathtub with shower, a washing machine and dryer, a refrigerator, central heat and conditioning, two outside doors, three bedrooms, and at least one window in every room and its not one of the high end ones for it day either!

The TV is a 60' HD LCD as well but thats new.

As far as my electrical power goes I have a pretty steady 118 - 123 normal range here with very few outages other than storm blinks now and then.

(There now I am not off topic and wont get scrubbed now either.)

 

[Mr RB]

In Canada, our electricity is designed TO WORK PROPERLY.
If it is ever only 102V then it is a BROWNOUT where the voltage is reduced on purpose ...
...

AG, I take it you have never worked in power distribution? In any country there are a percentage of the population who live near the ends of the distribution channels. People at the ends of country roads etc, and by the time the mains gets to them it can be significantly lower than the "spec". Also some people live close to substations where the mains is significantly higher, like where i used to live.

If he is selling a product to a braod spectrum of users it will be guaranteed that a percentage will have mains power than can be too high or too low.

Like I said before it will be a matter of economics, whethter it's better to sell a sloppy product that will fail for X percent of users and then cost time and $$ in refunds, or whether it's better economics to spend the $$ and make a highly reliable product from the start.

 

[Nigel Goodwin]

audioguru, here's a quick statistic for you. The US has 10 times the population of Canada, lets not even talk about heavy industry differences in the consumption of electricity.

California alone (one state) is very close to the population of your entire country!

One day you might realize that we have different problems, and direct comparison is impossible... The fact that the US electrical grid works at all simply amazes me!

So in your pride for your country, keep in mind the scale in comparison to the outside world.

I don't see how the scale makes any difference? - if you have ten times the population you build ten times the power stations, and have ten times the infrastructure.

As for Mr RB's point about 'end of line' and 'close to substation', surely such areas are simply fed from appropriate taps on sub-station transformers? - there's no excuse for not providing decent voltages at both extremes.

 

[audioguru]

Telephone systems have the same extremes.
I worked with telephone systems and tele-conferencing systems for a while. I thought that the frequency response of a telephone network was 300Hz to 3kHz until I measured the frequency response from one line at work to another line at work. 3kHz was way down at -12dB!

I called Bell and asked why the response was so muffled and they told me that their 3kHz spec limit was -7.5dB to the exchange which is -15dB for a telephone call end to end. That is why so many people on the telephone repeat over and over, "What did you say?"
Of course people very close to the exchange have much better response at 3kHz.

I made and sold a telephone transmit equalizer that boosted 3.5kHz +12dB. (Don't tell Bell). Every time I demo'd one it was sold because it made the muddy and muffled conference sound crisp and clear.

Now they send very high frequency television signals over telephone lines. 'ow doo day doo dat?

 

[unclejed613]

i've worked for companies that built prototype and production test equipment, and not only were the parts such as transformers and power supplies required to be UL approved parts (and this is mostly a concern on the primary or line side of the equipment), but construction practices were also very important. if you build the equipment properly, and use the proper testing procedures, you may not have the certification, but you did do everything right.

 

[Nigel Goodwin]

i've worked for companies that built prototype and production test equipment, and not only were the parts such as transformers and power supplies required to be UL approved parts (and this is mostly a concern on the primary or line side of the equipment), but construction practices were also very important. if you build the equipment properly, and use the proper testing procedures, you may not have the certification, but you did do everything right.

Yes, the problem here isn't really the poor mains, it's the abysmally poor design in the first place - it was never designed to work.