and some transformers ex Antek, the sec voltage given is at NO load with 115V/230V mains. in this case (254/230) x 38 = 42Vac unloaded. (42Vac - 1.2V) x 1.414 =57.7VDC
5.3VDC under the 63VDC rating at worst case ever.
5.3VDC under the 63VDC rating at worst case ever.
Here's a novel idea, why not just measure the output and put all the guessing to rest. I guess it you just sit at the computer all day and don't actually build anything then you have no choice than to speculate and fantasized but the easiest and safest way would just to measure and make sure you know what voltage your transformer is putting out unloaded. If it is 38vac, you are going to be plenty safe with 63V caps. The mains here in sunny California vary as much as 4vac. I have see as low as 118v and as high as 122V. On a transformer with 38Vac secondaries that will result in less that a 2V variance since the secondaries are less than half the primary.
I've got to agree with you on this. You're in California where supply is notoriously bad. Here in Ontario I measure 124 volts day and night for 10 years. Hydro obviously selected the wrong tap on the transformer for my house but the most fluctuation I have ever seen is .2 volts. Pretty much every cap I've checked out has a surge rating for fluctuating supply anyway. I'd worry more about the temperature degrading them that voltage spikes.
Depending on capacitor's quality for caps with Ur below 315V Us is 15% to 25% higher (Chemicon) and Ut is even much higher but for transients only. Surge voltages can be withstand only for reperive shot time intervals and this again somewhat varies depending on manufacturers approach. And then nominal Ur and actual Ur are not exactly the same. Everything deepends whether long life is valued most or budget as there is a substantial price difference between 63V, 80V and especially 100V caps.
cheers,
cheers,
right Terry,
in my other house in the province, my mains are 240 volts all day long, i lived beside a pole mounted distribution transformer ...
That is indeed novel. For some reason, you appear to resent the notion that design engineering aims to remove guesswork. Twice you have stated that you have never witnessed more than 122Vac maximum voltage on your mains. So what does your measurement really mean? That the utility company supplying your region is obliged to maintain the supply at your measured levels and will guarantee to do so ad infinitum? This is, of course, pure conjecture. That aside, you remain under the false impression that I have stated categorically that a 38Vac transformer would not be suitable for a 63V capacitor. In reality, I have only invited theoretical validation as you will confirm if you actually read and digested my earlier posts.
An incorrect guess.
This offers little confidence to others who may not have specified or purchased their transformers and does not take into account mains variations. It may well not be an issue, but a simple calculation takes seconds.
“If it is...”. The ultimate question. So, if there is any concern about the working voltage rating of supply capacitors, including incoming spikes which can affect their longevity - to do the job properly and stop all this guessing – obtain the official mains tolerances for your location, calculate then match this with your empirical evidence.
i always prepare for the worst case scenario...in my amps i compute for a worst case line voltage of 250volts in a line that normally cruise along 230 volts...
an 80% of the cap working voltage is what i aim for as much as possible....
an 80% of the cap working voltage is what i aim for as much as possible....
This is the sensible way. Assess the risks and design accordingly.
Again very sensible.
Using only 80% of max working voltage for the operating conditions that most often occur leaves that spare 20% (125% of the normal operating voltage) for spike tolerance and still not exceeding the capacitor rating.
This results in very few spike incidents exceeding the cap rating and increases the life of the capacitor considerably.
For those interested In the US, ANSI C84.1 is the standard. Essentially the nominal voltage is 120 +/-5%, 126V max. http://www.pge.com/includes/docs/pd...ergystatus/powerquality/voltage_tolerance.pdf
Safety margins are good, but I think sometimes we go overboard building bulletproof amps that will last forever. With some understanding of the failure modes and proper close fusing of the mains I think we can reduce the safety margins on occasion.
The ease of changing out amps in the DIYAudio Store enclosures means I probably won't be leaving a particular amp and power supply in use very long. Do I really need to design an amp for 20 year life? Of course you are free to use whatever safety margins you prefer.
Safety margins are good, but I think sometimes we go overboard building bulletproof amps that will last forever. With some understanding of the failure modes and proper close fusing of the mains I think we can reduce the safety margins on occasion.
The ease of changing out amps in the DIYAudio Store enclosures means I probably won't be leaving a particular amp and power supply in use very long. Do I really need to design an amp for 20 year life? Of course you are free to use whatever safety margins you prefer.
That is useful information, thank you.
With the exception of electrical (or indeed personal) safety, where minimum safety guidelines should be strictly adhered to, the home-constructor has the freedom to incorporate as much, or as little safety-margin / redundancy as they see fit in their creations. Some enthusiasts may seek to create an ultra-reliable amplifier, close the lid and forget about it for years, while others may prefer to tinker, frequently upgrade or tune various parameters as they become aware of ways to improve what they have. There is little point in vastly over-engineering an amplifier to meet near-military grade specification levels when most of its life is spent in the owner's living room on a shelf or rack in a domestic environment.
Sharing knowledge and technical skills is hugely advantageous in this respect and being armed with the techniques and methodology to permit fellow enthusiasts to decide exactly how much “engineering” they may require can potentially save them hours in terms of frustration and research.
We all have different motives!
Thank you, you said it better than I did. The discussion was developing a confrontational tone that I was hoping to defuse.
Daniel,
I don't know where your x 1.45 multiplier comes from? If you have a 35-0-35 Vac transformer (with primary voltage matching your mains voltage), across both rails:
[(35+35) * 1.414] - 1V = 97.98V. Divide this by two and your rails will be 49VDC. This is well within your capacitor voltage rating even taking into account transfomer voltage regulation. Why do you include a diode and paralleled resistor in each rail?
There is some logic behind that design, and the rest is just what Dan thinks sounds best.
What for you is the best PSU?
What for you is the best PSU?
in order to know the answer, we first must define "best psu" what does it mean?
to me a psu with low rail sag at full load is a capable one, but best? i dunno...
to me a psu with low rail sag at full load is a capable one, but best? i dunno...
How many folks really play their amp at full load or even close? Maybe a 50W amp but even then? Maybe with very inefficient speakers.
For an unregulated supply, there isn't a lot of scope for variation. Personally, I prefer (where possible) a transformer with good regulation, as few diodes as possible to perform the necessary rectification and adequate reservoir / filter capacitance. LED indication of rail state can be very beneficial, i.e. "are the capacitors charged" and minimal bleed to reduce heat and wastage. Other than that, a low-impedance PCB layout that is optimised for current handling if not using point-to-point wiring. Did I miss anything?