Silverstone Zeus ST85ZF 850W Power Supply Review

Cases & PSU/Power Supplies by KeithSuppe @ 2007-01-11

Silverstone Technology Ltd. is a relative newcomer to the PC-world having been founded in 2003. As a manufacturer of power supplies, CPU-coolers and enclosures they aggressively sought to compete with the very best in these product categories. Over the last year they have come into their own with the release of the well received Temjin TJ06 aluminum case and Zeus line of PSUs. Silverstone is now a relatively familiar name and we have the flagship of the Zeus series, the ST85ZF.

  • prev
  • next

Circuit Topology, Secondary PCB

In a typical PSU primary and secondary stages are usually divided on a single PCB by the main transformers which are usually placed at the center of the PCB. In order to keep the ST85ZF compact for standard ATX cases Etasis chose a dual PCB design in which secondary power stages along with the main transformer are mounted on their own PCB. This includes voltage regulation and DC output to the PC.

If you look carefully at the main transformer you can just make out the part number EPAP-750 printed on the side, yet on top we have a haphazardly placed sticker indicating the part number “EPAP-850 B”. This discrepancy will be the topic of our next section on the next page.

Madshrimps (c)

Madshrimps (c) Madshrimps (c) Madshrimps (c)

The secondary board is rather congested, however; given the number of circuits Etasis has done an exceptional job of organization. As stated above the secondary PCB also houses the main transformer, rectification stages and DC output stages to the PC. In addition to those mentioned the ST85ZF features “true” independent voltage regulation for the ultimate in rail stability.

Madshrimps (c)

True independent voltage regulation is rare. Even among high end power supplies the best we may see are dual Mag Amp output regulation circuitry in which there is a “shared” regulation among the rails. Therefore if one rail is affected the others may be as well.

On the secondary PCB independent regulation stages occupy three daughter boards (or "daughter in law" boards given the number in the entire PSU). Seen in the photo below from left to right, the main stage card features a potentiometer, (this does not adjust rail output) while the two remaining cards regulate 5V and 3.3V legs respectively. In the thumbnails below we have a close-up of the twin transformers feeding the 5V and 3.3V independent regulation and a second view with labels.

Madshrimps (c)

Madshrimps (c) Madshrimps (c)

While independent voltage regulators have their advantages most PSU manufacturers do not find this method cost effective. Fortunately Etasis/Silverstone chose stability over cost. In the close-up below the first independent voltage regulator card features a potentiometer which PCModding-Malaysia identified as; "...adjusting the protection threshold for the power supply." This initial circuit is based on a Texas Instruments UCC3895DW (left thumbnail) advanced phase shift PWM controller, in the next thumbnail the card is identified by the part number EPAP750-CB4. The next two thumbnails exemplify printing found on the secondary mainboard for each voltage regulation stage.

Madshrimps (c)

Madshrimps (c) Madshrimps (c) Madshrimps (c) Madshrimps (c)

There have been questions as to why Etasis/Silverstone chose true independent voltage regulation. In my opinion Etasis/Silverstone wanted to ensure at the very least the 3.3V and 5V rails would not cause fluctuation on the 12V rails. While I don’t know of a PSU which actually places independent regulation on each 12V rail as well, the intent here is for the 12V rails remain steadfast regardless of load (ideally). Building a multi-rail PSU correctly is difficult enough when one begins to understand the contradictions in the 12V Power Supply Guidelines.

In this case there was no need for ad hoc solutions such as "Rail Fusion" in which an overload on any 12V line result's in a combined rail voltage to compensate. This is counterintuitive to the intentions behind multi-rail circuits and while desperate times call for desperate measures, ultimately the overload is on the consumer with so much "techno-hyperbole". The ST85ZF isolates all 12V rails at their source off the PCB, which is evident in the photo below and something I haven’t seen on any other multiple 12V rail PSU beyond the 12V1 and 12V2 level.

Madshrimps (c)

Turning the PCB over we see the 12V rails separated into distinct solder point sections including 12V-Rails associated with baseboard (CPU) connectors and PCIe, connectors.

Madshrimps (c)

Circuitry markings -->
  • prev
  • next
Comment from Sidney @ 2007/01/11
From Hardware Secrets; Why 99% of Power Supply Reviews Are Wrong. I'm sure that article will have a substantial impact on most PSU reviews you have read.
I've read the article the day it came out, and found the statement rather inaccurate. A Chroma station, measuring amp and or wattage or voltage by replicating the manufacturer QC station process, and even taking IQC (incoming QC on purchased components) are for the engineering professional in electrical field. When I test drive a car, I only need to know how it responses, sitting/driving position, steering inputs, acceleration (if it is important to me), gas mileage (if it is labled for good mileage consumption), price point and the "look", etc, etc ...... One thing I don't mind not knowing is all the engine, chassis, suspension, gear-box design and specifications (although some of us do want to know because we are very good at it).

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
Comment from EsaT @ 2007/01/19
I just checked article more closely and I think you are wrong about 12V rails being really separate.

Basing to these pics looks much like at least 1 and 2 are "jumpered" (those thick metal jumpers) from same trace.
And basing to bottom side of PCB also 3 and 4 are propably similarly connected.

In fact looks like all 12V rails are directly connected to each other after those "jumpers", PCB traces seem to connect to each others on left edge of pic.

Originally Posted by lazyman View Post
When I test drive a car, I only need to know how it responses...

99% of reviews are wrong sounded is a bit odd.
So you test drive car using just gear 1?
Or more precisely without any driving and seeing just that engine works at idle?
Comment from jmke @ 2007/01/19
rather testing driving the car on race roads, normal roads, traffic jams, long distances, short distances, parking, booth space, easy maintenance, wheel removal kit, seat arrangements, mirror placement, etc etc

you don't need to know how the engine works to make an evaluation of how the car drives

knowing how the engine works can help to explain why the car drives as it does, but it's not a necessary ingredient to make a good evaluation of the car, of course, cars and PSU are not quite a like, and comparo here might not be appropriate
Comment from Sidney @ 2007/01/19
If it is automatic trans; most people don't just put it in "D"; drive away
If it is humming along quietly, most drivers don't care how it works
I am referring to majority of drivers.
Comment from EsaT @ 2007/01/19
Originally Posted by lazyman View Post
I am referring to majority of drivers.
But most people propably wouldn't trust test drive of car done by using only gear 1, that's what that majority of PSU reviews with barely 1/3 of load specified in ads of PSU are...

On the other hand this specie might have once again digged deeper than previously reached level of rock bottom.
Comment from Sidney @ 2007/01/20
You don't buy a car just by reading test report I hope
Comment from jmke @ 2007/01/20
it does help make the decision a lot, testing driving PSU is hard
Comment from Sidney @ 2007/01/20
My initial post was related to the 99% reviews being wrong. Here is my take-
1) A car is rated to have 500HP rear or front wheel power; top speed 180MPH. There will be manufacturing variables, and a dyno test may or may not prove 100% accurate. Now comes the 180mph test, again there are many variables plus having a track to do the test against normal traffic. A test drive reviewed at 120 MPH without the track to top out is not 100% wrong. It merely ran out of track on this part of the review.
2) Since PSU is not regulated by any government (unlike cars) rating except UL for safety. It is highly dependable on the manufacturer's ablility to hold certain professional standards such as ISO which I strongly recommend consumers to look into.
3) The use of any test equipment not regularly certified does not mean much to me.

Lastly, PSU tested even at 100% spec on bench using Chroma station at a given time does not gurantee real life performance against weather condition and irregular or variable input voltage (from the wall or motherboard). Look for solid and reputable manufacturer or name brand; reading test reports include cable length, flexibility, size to fit most or just certain cases, and noise level; to say 99% of reviews are wrong because they might not test the max output is "odd".

We could go another step to look into component specs +/- 3%? Or requesting BOM (bills of material) and supplier listings?
Comment from Liquid3D @ 2007/01/25
I actually re-evaluatewd that article myself. I also found several points which were inconsistent. I wasn't trying so much to discredit our or any other methodologies as I was trying to challange those sites whom feel their tests are beyond reproach. Many of the top test sites with the very best hardware use voltage regulation as a standard test.

When you think of it, since all PSU's have voltage regulation it's not as if their testing the PSU's "power reserve" their actually measuring (if it's possible) the unit's ability to compensate for Rail fluctuation. Without "true" independent voltage regulation fluctuation will be regulated, often at the xpense of theo others Rails. Therefore measuring rail stability may only show one or two phenomenon: First - time it takes to compensate (get the rail back to "spec". Second - wether or not PSU is able to compensate without any affecting other rails or circuits in the design. The paper states the latter is almost impossible because one would need to monitor all rails simultaneously.

IMHO it's not so much the rail in question is immedately regulated, but more importantly this regulation is proprietary as is the current source behind it.

For exmaple if 12V3 dips and it is compensated for, does this now effect 12V1, 12V2 and 12V4? While independent voltage regulation may avoid any impact on other rails, much more important to me is where that power comes from?

The jumpers you refer do not seem to connect any of the 12V rails on the face of it, but I'll follow the traces to be sure if you like?