Topic: Overclock/Overvolt the Avalon6 with a DPS-1200FB Made Simple (Read 3963 times)

Update on the HSTNS-P11 alternative to the DPS-1200FB:  The link I shared above shows the wrong location for the voltage pot.  The actual location is up and to the left (about 3/8" to the right of the MC33274ADG OP AMP).  Mine adjusts from 11.588V to 12.427V.

I don't have the DPS-1200, but I do have a similar HP supply, 490594-001 (HSTNS-P11 I believe):

I'm going to try to voltage adjust one of these this weekend.  I found this post online:
http://www.rcgroups.com/forums/showpost.php?p=27851172&postcount=2602

It looks as easy as the DPS-1200 to adjust.  The fourth picture shows where the adjustment pot is supposed to be and has a decent range for an Avalon6 (11.7V-12.8V).

I have run three of these 490594-001 supplies on SP20s for a year now.  They have the same ratings as the DPS-1200 (900W@120V, 1200W@200-240V).  I do not know if they will take the same abuse as kilo17 gives the DPS-1200 (getting the 1200W rating out of 120V instead of 240V) since I use them on 240V.  Unloaded (open) they read 12.3V by multimeter and loaded they read 12.00V exact.  They are VERY quiet and surprisingly efficient (I've calculated mine to be providing 93% efficiency when running a SP20).  The only downside is that while a simple resistor mod can be used to start the supply like with other HP models, the 490594-001 will not auto start after a power outage with that resistor mod.  The supply needs a short time delay after power returns before it will accept the PS_ON signal.  I got around this by using the 5v standby output pin 32 to charge a 100uF capacitor via a 470ohm resistor and tied that to a base of an NPN transistor, collector to the PS_ON pin, emitter to ground (edit: my memory was mistaken on what I did with the C/E pins a year ago; reviewing photos I took of the mod it appears I have the collector on pin 33 and the emitter on pin 36)  (RC charge time delays the transistor switching, which in turn activates the supply).  It has worked perfectly.  No, I didn't do a breakout board-- I went cheap bodge style with the parts soldered together hanging off the card edge.

I'll be getting my first Avalon6 when the group buy delivers, so I'm hoping the voltage adjustment on this model is successful since I already have four of them (3 live and one spare).  I plan to get the highest allowed (within spec) level (12.2V) instead of exactly 12V.

I see some very nice numbers here, but i also see some pretty high temps. What temp would be considered overheating, for the Avalon 6?

PSU's still strong after a couple of weeks

Hey Kilo,

How's the voltage change going so far?  Has it been stable for a few days?  What is the best voltage?  The Dell 750W PSU are the ones on Holybitcoins.com including the breakout boards n cables. Is your HP 1200W PSU very quiet?  My HP 1000W PSU is extremely quiet but run a bit hotter than the Dell.  The Dell is LOUD as heck but with the board, I can adjust the fan speed to make it run hotter but quieter.


http://holybitcoin.com/product/dell-750w-server-psu-breakout-adapter-board/

P.S I'm still scratching my head as to why they say that PSU can only do 900w on 110V but your proved otherwise as even non OC, the Avalon uses more than 900W.

Post a pic of your PSU- I want to make sure it is the same as the ones I have (750's) - I think I have the adjustments for it also

You guys are making me envious!!!  Still in 3.3TH LALA land........so awesome to see consistent 3.8 to 3.9TH using 12.55V.  Definitely going to try that once I get the new Finksy board.

That is why when you backed off from 12.79 to 12.55 the psu cooled as it drew less amps to do the 12.55 volts.

The Avalon six seems to run very well in the 12.55 volt area.

I did get the psu as high as 12.83 volts.  But the hash rate moved very little and the power use needed was much more.

The avalons remind me of gpu mining the top maxed over lock is wasteful.

The sp-20 was also like that.

I have found 12.55 volts is really good for the Avalon sixes.

That is a true statement at "stock" settings but by increasing the voltage it is increasing the wattage I do believe  (watts=volts * amps) on the 12 volt side.  On the wall side it would need to increase the amps since the voltage remains stable at 120

I have backed the voltage down a tad to about 12.3 (by software) and 12.5 on the PSU and achieve the same hash results and the PSU is much cooler.

Here are a different miner and PSU.
Picture:
1 - Miner before overvolting - Avalon #3
2 - Miner running "stock"
3 - After overvaulting PSU
4 - Plug measurement
5 - Miner running overvalued
6 - Results after overvolting - Avalon #3

based on his numbers it must do more maybe 1100.  1100 clean watts would give about  3500gh.

How is the Avalon getting enough watts? The PSU on 120v maxes at 900w right?

I am using 1 of them, I plan on doing it to another tonight and will take pictures of it running.  Also, it will be a good follow up to verify the results.

are you using 2 of these per avalon then?  I dont see how at 900w on 120v you can run one avalon with increased voltage to boot

That is correct.

Mine are all currently running on 120v, including the one that I bumped up the voltage on.

Looks like J4bberwock's common slot breakout board, available on Scott's website: http://holybitcoin.com/product/hp-1200-watt-common-slot-psu-bundle/

This is a really great article and discussion. One thing I think that was missing from the outset is that this will only work with an input voltage of 220V. It won't work on 120V (nominal) wall power.

I say this based solely on the label which suggests it won't supply enough DC power when powered from a typical 120V outlet.

Which breakout board are you using on this DPS-1200FB power supply?

that is certainly a possibility, although most of them wont even attempt to start. they just flashe red and stays at 12.4 or 12.5v. sometimes i wonder if its heat related. If i drop the voltage down to 12.1 or so 50% of the time it will take off on its own. the other 50% i have to unplug the psu, start the miner all the way till hashing and then raise the voltage. And it is certainly different miner to miner. I've tried moving psu's around and trying different things. most of it boils down to differences in the miners more than the psu's. i've got 20 on 3 different types of psu's, but none higher than 1300w, although all running at 240v.  Never any problems while running, most have been up for 6-7 days now since last power issue. It just takes me an hr or two to get them dialed in after such a failure. And mostly getting each miner set up to the voltage it runs best at overclocked without going past where its efficient (which can be a very small window).