Binning is a manufacturing term usually reserved for processors, although many micro-chips are binned. To oversimplify, processors are designed around a specific core design and this single core will be used throughout the entire model line.
Models aren’t merely sorted by speed, but also by cache depending on the manufacturer and packaging (not box package). For example, an extreme Intel Core X6800 contains the same core as its less costly E6300 brethren except the less costly model has just half the cache, with the other half disabled on purpose. In a perfect world (for the manufacturer and for the Purist Overclocker) every CPU produced at the Fab would be capable of meeting and surpassing the performance requirements of the flagship or fastest model.
Since it costs approximately 2-billion dollars to build a Fab which is usually intended to manufacture a single core design, it wouldn’t make financial sense to build a Fab for every model designed off the same template (core architecture). And it’s also not reasonable to expect the consumer will desire, need, or have the funds for $1000 CPU.
The answer then is binning and packaging which creates products at a performance price point for all. What has occurred with Intel’s Conroe processor is that it was one of their best production runs ever, this is evident in the fact almost every core off the line was capable of running far beyond the flagship model’s intended speed without a need to increase Vcore. CPU makers always make sure to design in some overhead or “speed ceiling” as a longevity buffer. In a less successful CPU production run Overclocker’s may only get a few hundred MHz if their lucky and they will have to raise Vcore just to get that with additional heat output. An example of this can be found in some of the less successful Intel Pentium 4 Northwood’s in which slower models were packaged to run 1.525V while faster models required 1.550V vcore. The Overclocker lives for processors like Conroe in which the Allendale’s (Conroe with less cache and lower multiplier) overclocked so high their speed compensated for less L2 cache in performance.
Did binning happen on a smaller scale with the Etasis EPAP-750 aka Silverstone Zeus ST75ZF? Can a PSU be binned? From the photo above it does seem possible at least some parts of a PSU designed to meet certain specifications can perform beyond those specs, we constantly see this in the "Peak" value. With only a 100W difference between the 750 and 850 models, it is not only logical but cost effective the same circuit board could be used for several models. So long as the design is the same, four 12V-Rails for example, then a difference of 100W +/- could very well share some basic parts. What originally piqued my interest was the obvious main board discrepancy. Several reviewers commenting on this suggested Etasis or Silverstone had purportedly "upgraded" parts and the stickers placed on these were evidence of this. Below the main transformer bears a sticker identifying it as
EPAP-850 part.
Once fully disassembled ambiguities arise in the discord between the EPAP-850 sticker and more permanent printing on the transformer side identifying the part number
EPAP-750 (below).
When examining the ST85ZF in an attempt to find any part not bearing the EPAP-750 identification it seems I may have confirmed this was an EPAP-750. Every transformer on the power supply are labeled EPAP-750, such as the PFC transformer off the primary PCB.
Next the independent voltage regulation transformers supplying VRM3V and VRM5V also bear these same numbers. To reiterate it is entirely possible these parts could be "tweaked;" however, this does seem extraordinary
The circuit board running along the DC-output voltage to PC bears the EPAP-750 part number as seen in the photos below.
A close-up reveals the part number.And as indicated on the previous page, the main PWM or Voltage Regulator Module bears the same part number.While part labels do not necessarily negate an 850W power rating nor prevent the PSU in keeping with its specifications as listed by Silverstone for their
Zeus ST85ZF, it does beg the question: “Would it be advantageous to purchase the ST750ZF instead?”
At the very least without a visual side by side comparison of Etasis EPAP-750 / EPAP-850 or Silverstone ST85ZF / ST75ZF; any conclusion would be based on conjecture. I have compared photos and part descriptions from what I believe to be every review on the Internet pertaining to the ST75ZF, ST85ZF, EPAP-750 and EPAP-850, although none have explored this issue in-depth.
Perhaps most compelling would be to carefully compare the two photos below, the left thumbnail exemplifies the secondary PCB borrowed from
Overclock3D Etasis ET-750 review (pg.5) and on the right a photo of our ST85ZF taken from the same viewpoint (right thumbnail).
Onto testing -->
99% of reviews are wrong sounded is a bit odd. It depends on the type of audience; 99% are electrical engineers (by formal training or otherwise) and PC enthusiasts; leaving 1% reader who asks that give me a power supply that fits well in my case and support my system without over kill in paying for something I don't need is a strange statement.
By the way, the article serves me well