Weight as a Characteristic of Power Supply Reliability - Revisited

[bill_bright]

Apologies first:

I would like to apologize to doctorgonzo, David M, Alaron, Staren, kstatefan40, thefultonhow, PCMech Forum members, and finally glc. The other day, I opened a thread to criticize something I had read elsewhere on PCMech, and did so with a tone that was too harsh and not very professional .

My goal was, and always is, to use my many years of experience as an electronics/computer/IT hardware technician by being an advocate for the end-user and to ensure consumers have the correct information needed to make the correct purchasing decisions, and to have an enjoyable and safe computing experience for themselves and their families. However, the approach I took immediately put everyone on the defensive and it became a totally unproductive thread. The direction that thread went was totally my fault, and again, I apologize and hope that can be put behind us so we can get off to a better start this time around.
***

Down to business: Weight as a Characteristic of Power Supply Reliability - Revisited

There are a lot of folks, even experts, convinced that weight is a sure sign of reliability and output power in an computer power supply. Several years ago, that was true. But that is no longer the case.

Please hear me out.

I think we can all safely assume there have been some advancements in technology in power supplies over the last 11 years since the ATX form factor standards were established, and took over from the Baby-AT standard. Many changes and advancements were mandated by the EU, FCC, FormFactor.org Standards Committee, and consumer demands on the industry itself. Advances made in lightweight materials, manufacturing techniques, EM Interference control, quality, reliability, power (wattage), and last but by far, not least, efficiency have been significant. Advances in particular, over the last 5 years, are notable due to the explosion in CPU and GPU performance gains and energy demands.

Note too that the gaming industry is a HUGE driving force in PC advances, and not just in performance. Many gamers are also into LAN parties where luggable, lightweight but sturdy, and reliable cases from the likes of Lian-Li and others are in demand. This demand extends to components within the case too, including a demand for lighter power supplies, while at the same time, ensuring these supplies are able to support today's dual core CPUs and multiple power hungry video cards, while staying cool and in many cases, quiet too.

I agree that light weight might be an indicator of cheap parts. But with today's technologies, it may very well be an indicator of advanced technologies and the use of state-of-the-art parts too! A single lightweight integrated circuit (IC) may be used where several discrete components were used just a couple years ago. It is amazing to think that there are over 125 million transistors in the Intel P4 Prescott CPU cores today. But understand that there were "only" 55 Million in the P4 Northwood cores of just a few years ago. Today's top video cards can easily have 300+ million transistors! It should not be hard to imagine, similar - though not on that scale certainly - but similar miniaturization in PS technology has occurred too. Especially by the makers with the high-end offerings. But there is just no way to determine that without looking inside the power supply case. You can't do it by a set of scales.

Certainly, if weight were legitimately a factor, generic PS makers with no scruples would just add dead weight to their products, just to make the sale - and then make a point of it on the packaging.

Some experts may point to the Directron.org study as evidence of weight being a characteristic of quality in power supplies. Or they may point to this HighPowerSupply FAQ on how to choose a power supply. Please note that the Directron study was conducted nearly 6 years ago in December of 2000, and the HighPowerSupply FAQ answer on weight is word-for-word (as in plagiarized?) from the Directron study.

To support my position, please note the below examples of several current name-brand (that is, PS makers with a reputation for fine power supplies) and low-budget brands. You will see the brand-name quality power supplies weigh less and/or have more power than the low-budget brands. That is contrary to the Directron study, clearly illustrating that weight can not be used to judge the supply's reliability or power when comparing like rated supplies.

I note the name brand Antec TP II 550 weighs in at 2.3kg while relatively unknown StarTech's smaller PS Startech 480 weighs in at 2.4kg. The ThermalTake 550 comes in at a measly 2.14kg while the Kingwin 450 comes in at 5lbs or 2.3kg.

You can see from those examples that if a newbie walked into a PC shop, with no prior knowledge of power supplies, and picked up 2 like rated supplies and did a balancing act to determine which was heavier, he would likely pick an inferior supply. That would be a mistake.

Using the same Directron.org study as above, I refer you to Directron's own updated Table 1 from that study. Scroll past the first list from the year 2000 to "Power Supplies (300W-600W)" section. This second list was complied recently in January 2006. Please review those power supplies and see if you can make a definitive correlation between weight and quality. Note, for example, the Enermax NoiseTaker 600watt EG7001AX weighs 5.1 lbs, while the lower power Enhance ENS-0246B 460watt supply weighs more at 5.45 lbs! I think you will find as I did, that you cannot conclude one way or the other as the weights are all over the field for a given power rating.

The Extreme Overclocking.com article, Powering Your PC: Truths and Misconceptions, from May of 2005, is an excellent, in depth article on power supplies. Note the section concerning power supply weight on Page 12.

If you can trust the Government, you might be interested in the EnergyStar.gov report on power supplies. Please note on page 5 where it reports,

"Highly efficient power supplies tend to be smaller, lighter in weight, and more convenient. They operate at cooler temperatures, contain fewer parts, and are likely to result in greater product reliability."

[Cricket]

Since you're bringing this topic up again you may want to revisit this warning:

Originally Posted by glc There is no excuse for you jumping in here the way you did with your flamethrower all cranked up. Do it again and I promise you that you will be banned instantly. I stand behind the tip as published and intended 100%.

Tread lightly.

Cricket

[bill_bright]

Cricket, I am not certain it was necessary to bring that up. I admitted that I had opened a previous thread poorly, so it would seem obvious to all reading that I was not hiding anything - I feel they could have easily found it themselves if curious. I really see no point in bringing it to the forefront other than to stir things up. But I guess you had your reasons that perhaps I just don't understand.

Please note I made it a point to not criticize anyone, or stick needles into old wounds. I think it is sad you chose to do so in light of the fact that I opened with an sincere apology to all those involved. Noting you were not one of them, I truly don't understand your motive. But, if it made you feel better then that's okay as it does not change the science or the fact that advancements in technology has a tendency to change the status quo. My motive is to help people have current information so they can make good purchasing decisions and a safe computing experience.

If you would like to discuss any of the points I made concerning power supplies, I will be happy to do so. If I am to be further chastised for something I have already apologized for, then I see no reason to make any further contributions.

[glc]

Easy there, both of you. Cricket, next time please use the report post link instead of taking matters into your own hands. If you want to be a moderator the invitation is still open, but until you are, please leave these issues to the current moderation staff, there are issues discussed in the staff's private forum that you are not privy to. Bill, please understand that there will be repercussions such as this from the original thread in spite of your gracious apology, and all we can do is try to move on and be productive.

[bill_bright]

Originally Posted by glc Easy there, both of you. Cricket, next time please use the report post link instead of taking matters into your own hands. If you want to be a moderator the invitation is still open, but until you are, please leave these issues to the current moderation staff, there are issues discussed in the staff's private forum that you are not privy to. Bill, please understand that there will be repercussions such as this from the original thread in spite of your gracious apology, and all we can do is try to move on and be productive.

glc - thanks for your understanding. I do realize feelings may still be sensitive, that I understand and accept - hence my apology and my reasoning for not trying to sneak anything by anyone. I am afraid I don't understand any reason for any repercussions for a mistake I have already taken responsibility for, but then I don't always live up to my name . I just know it does no good to poke a stick at a fire you are trying to put out.

Thanks again.

[David M]

Apology accepted Bill.

I still stand by my premise that although you can get excellent electronics that are getting lighter all the time, the transformer, which makes up the single heaviest component of a PSU, has not changed significantly in design since the modern home computer was invented. Therefore, weight does matter but it certainly is not the only factor in determining a good PSU. One cannot reduce the amount of copper and iron in a transformer and expect that transformer to have the same capacity, reliability or efficiency.

There has to a be a minimum power to weight ratio of a transformer below which it is a physical impossibility to push more watts through the transformer. I have no idea what that ratio is though. Just working around transformers with my job, I know you are not going to push 600 watts through a transformer that weighs only a few ounces. The best transformers are heavy little beasts with a significant iron core...period.

Perhaps the discussion should end with weight is one of many factors that can determine capacity but is not a determinant because of all the other factors?

[doctorgonzo]

Originally Posted by David M Perhaps the discussion should end with weight is one of many factors that can determine capacity but is not a determinant because of all the other factors?

I agree with that. It's just one possible factor. I certainly don't purchase PSUs based on weight; I follow the advice of people here. But weight is still something that you can take into consideration when you are faced with an unknown PSU.

[bill_bright]

Originally Posted by David M Apology accepted Bill. I still stand by my premise that although you can get excellent electronics that are getting lighter all the time, the transformer, which makes up the single heaviest component of a PSU, has not changed significantly in design since the modern home computer was invented. Therefore, weight does matter but it certainly is not the only factor in determining a good PSU. One cannot reduce the amount of copper and iron in a transformer and expect that transformer to have the same capacity, reliability or efficiency. Perhaps the discussion should end with weight is one of many factors that can determine capacity but is not a determinant because of all the other factors?

Thanks David.

To your comment, and doctorgonzo's endorsement, for ending the discussion, I beg to differ for one main reason - you both are trying to end it based on a premise I am NOT making. I am not talking capacity. My whole point is quality and reliability - as indicated in the subject title. Capacity/Power have nothing to do with quality and reliability of a power supply and so I ask that you separate those thoughts for the purpose of this discussion. I do accept that weight is a big factor in determining the power handling capabilities of a transformer, and will elaborate on that in a bit. And I also accept that a underpowered power supply affects overall computer reliability. But the assumption here is that we are buying a power supply with a rated power appropriate for the computer at hand. I appreciate your understanding and cooperation on this.

***
Here's the flaw in the argument. First, there are many more components inside a power supply than just a transformer, and the metal housing itself, which blocks the view of the insides preventing a visual inspection and clue to the quality, makes up a percentage of the total weight - so I don't think it is fair to base an entire position on just the transformer.

BUT - for the sake of this discussion, let's talk about just the transformer then.

How can you determine the quality or the reliability of any transformer - especially one that you cannot see inside a closed housing, just by its weight? You can't! You might be able to compare and get some idea of the power handling capability of two transformers if you could hold them in your hands, but inside the case, you have no way of determining the precision of the windings, the purity of the copper, the quality of the construction/assembly, or how much of the weight is copper and how much is iron - or how much is part of the mounting assembly. Transformers in generic PSs may be used in many applications and not just in computer power supplies used to produce 3.3VDV, 5VDC, and 12VDC. Generic makers could easily pull a universal transformer off the shelf and only use taps for those voltages for ATX power supplies, and leave many additional (and weight adding) windings for other voltages unused. There is no way to determine that without a visual inspection of the transformer.

When talking two generic no-name power supplies with equal ratings, connectors, etc., I accept that you might be able to get an idea of which supply has the "greater" power capability by its weight. But that's it. You cannot tell which power supply is made better, which uses higher quality components, which is more efficient, which is more stable, which produces cleaner power or which has a greater reliability, by the weight. There are just too many other factors which negate weight as a factor of quality and reliability. And based on the examples I gave from PC Mech's own list, you can't even rely on weight to determine power.

I am sorry guys - it is not my opinion, it is just the way it is.