Corsair HX 850 80+Gold

mi_schmid schrieb:
Hallo Zusammen,

ich bin Mitarbeiter von Corsair und möchte nur etwas zum Thema Schutzschaltungen beitragen.
In diesem Dokument sind alle Schutzschaltungen der verschiedenen Corsair Netzteile aufgelistet.

Schön, interessiert nur nicht.

Denn hier gibts 'nen Review vom HX850 v2 und dort vom GS800 V2.
Auf der Seite 'a Look inside' (sehr weit unten, vor den Bildern von der Rückseite) ist ein Weltrend WT7502 Protection IC zu erkennen, der nur 8beinchen hat. Gut, OVP und UVP scheint der Chip wohl doch für +12V zu können (+12V = VCC), das ändert aber nichts an der Tatsache, dass OCP auf +12V fehlt. UVP löst auch erst bei etwa 2-3V unter dem Sollwert aus...
Ergo: im Kurzschlussfall wird wohl am ehesten die OPP auslösen, wenn überhaupt...

OK, bei einem Single Rail Schweißgerät ist das eh egal, siehst du hier eindrucksvoll

Und bevor du ankommst: Das MUSS so sein, bei einem Single Rail Netzteil. Denn dieser Zustand ist ja gerade DER VORTEIL von Single Rail.

Buttom Line: Die aktuellen Corsair Produkte sind überwiegend auf einen geringen Preis ausgelegt und nicht besonders gut, was die Seite der Schutzschaltungen betrifft. Das ist sogar 'ne Frechheit, für den Preis, den ihr für euer Zeugs verlangt.
 
Zuletzt bearbeitet:
Stefan Payne schrieb:

Sorry, I have to reply in English.

Stefan, on the "lack of OCP and OTP" concern, it seems you like to assume a lot and understand little.

True, the WT7502 only supports OVP and UVP, but these PSUs still have OCP and OTP. Since the PSU uses DC to DC, you can use the PWM controller for OCP since the PSU is essentially one giant +12V PSU. Of course, each DC to DC board has an AP7159 IC that supports OCP for the non-primary rails. And any PSU that has a thermostatically controlled fan (pretty much ALL of them these days, right?) has OTP. When the fan controller IC detects a critical temperature, it tells the WT7502 to latch off.

Hope this clears things up.

Jonny
 
Zuletzt bearbeitet von einem Moderator:
Nutzt man zur Sicherung der +12-Volt-Leitung die Überlast-Sicherung des PWM-Controllers, hat dies zwei Nachteile:
1. Das Auslöselimit liegt immer recht hoch, OCP ist damit nicht so wirksam wie bei Multi-Rail-Netzteilen. Gerade im Leistungsbereich 800 Watt macht mcih das etwas nervös.
2. Möglicherweise löst die Schutzschaltung minimal (zeitlich) später aus.

OCP ist damit wohl schon an Board, da hat Jonny mal wieder Recht.
 
Philipus II schrieb:
Zitierenregel beachten.

Nee, das ist OPP! Das darfst du jetzt nicht mit OCP verwechseln. OPP ist primärseitig implementiert (üblicherweise im PFC Kreis), OCP Sekundärseitig. Im Falle eines Fehlers schaltet OCP einfach schneller ab als OPP, bei dem erst noch nachgeregelt wird. Und auch der OPP Trip Point ist z.T. relativ hoch. Manchmal auch nicht vorhanden (einige HEC Geräte, bei denen man 400W Geräte mit 900W belasten kann)

Gut, bei einem Single Rail Schweißgerät der 800W ist das eh egal, wenns da 'nen kurzen aufm Board/auf der Grafikkarte gibt, fackelt die eh erst mal ab...

Und dass man OTP über den Regelkreis des Lüfters implementiert, halte ich mal für ein Gerücht. Ohne Beweis glaub ich das jetzt erst mal nicht. Aber Jonny kann ja gern ein entsprechendes Nezteil raus schicken, dann kann das ja mal wer testen ;-)
 
Zuletzt bearbeitet von einem Moderator:
Die Auslösezeiten muss ich nachschaun. Oder bist du fleißig. Ich gehe davon aus, dass es wie geschrieben etwas länger dauert.
 
Ja, eben genau das ist ja auch der Punkt...

Bei OCP gibts 'nen hohen Strom, das Netzteil schaltet ab -> nur noch die im Gerät enthaltene Ladung muss abgebaut werden
Bei OPP gibts 'nen hohen Strom, das Netzteil regelt nach, es merkt, dass die Leistung 'nen bisserl hoch ist und schaltet irgendwann mal ab.

Ist natürlich die Frage, wie groß der Unterschied ist. Aber dazu kannst dir ja mal 'nen Dark Power nehmen, um das mal zu testen ^^
 
I'd also like to know how they rigged the wire catching fire in that video. We tried to duplicate the test on the Chroma at Corsir and couldn't get the wire to heat up. The PSU always shut down.

As you know, the thinner the wire and the longer the wire, the more resistance. Despite this, we used a 22g wire that was 16 feet long and the PSU still tripped off within milliseconds with 300W, 600W and 1200W loads put on them.

Bottom line... there's much ado about nothing... especially when talking about real world usage.
 
Ich bin dem mal nachgegangen. Meine Industriekontakte gehen davon aus, dass eine Integration über den PWM Chip nicht notwendigerweise langsamer agiert, und wenn, dann in der Größenordnung von Millisekunden. Eigene Messungen habe ich aber nicht.
 
You're both correct. OPP can be used for OCP on a PSU that uses DC to DC for the non-primary rails. 100% of the PSU's capability is available on the +12V, so it's just as reliable. And any delay between catching an "overload" on the DC side versus the AC side is a mere matter of milliseconds. You'll note that PSUs that do NOT use DC to DC (like CX600 and below, for example) actually implement an IC that does support OCP. That's because one can not reliably use OPP to monitor over current conditions.

On another note....

We (at Corsair) also tried to simulate the "burning wire" video that was posted. I'd like to know how long and what gauge that wire is. As you know, the thinner the wire and the longer the wire, the more resistance. We sued a 22g wire that was 20 feet long and plugged it into +12V and ground on an AX1200. The PSU immediately hut down given a number of different loads programmed onto the Chroma. We eventually gave up because the wire never even got hot.

Now, I'm not saying it's impossible. I've worked tech support long before working in the PSU industry and have seen people pinch wires in chassis side panels, creating a short and causing the wire to heat up and the insulation to melt off. But when trying to duplicate the "magic trick" in the video, we couldn't do it. Was there a ceramic resistor off camera that helped heat up the wire?
 
Zuletzt bearbeitet:
Hey jonnyGURU!
I'm Simon who did the short circuit experiment in the video discussed above. Sorry for my late reply. I've been on vacation, however, here is my eagerly-awaited answer :)

I caused the short circuit with a 'common household' wire made of copper. I didn't measure its length before, but I would say the wire was about 1,5m (~5 feet) long. The diameter without isolation was about 0,15mm (~ 0,07 inch). I'm sorry, that I can't give you more explicit information about the wire, but as you can see in the video it was enough to circumvent the protection circuits of the AX1200 (it was an early version of the AX1200 - got it right after the release).
 
Zurück
Oben