Pin23 to Pin24 seems to works with anything lower than 600 ohms (3x 1.8k in parallel). 900 ohms (2x 1.8k in parallel) doesn't work.
I don't get it then why mine does NOT work at all in the same Conditions.
No idea, I also eliminated one more variable. The outlets I have at my house are primarily ungrounded and I had been testing thus far in one of those. I've now tested both PSU's in an outlet with a safety ground. It made no difference, both PSU's power on and supply the main 12V output.
. Obviously I had the PSU turned back to front so I didn't see that ...
That's a classic.@mideel: OK, just found 1 Issue: the Switch was turned OFF
BUT still 2 x Amber LEDs (AC Fault & DC Fault), Green Status LED Blinking and ... I got 13.25 V Output.
I have both on with the green led blinking.
Yeah. Murphy's Law
.
.Thanks for the Tip about B2 (PS_ON), the Sparring and the Help in Troubleshooting
. I might have eventually figured it out, but I originally spent a couple Hours one Day, then parked this Task for like 2 Weeks until I received some I2C/SPI/UART Programmers since I basically thought that was the only way forward.Thanks
.It's 14 Power Connectors (4 Long + 7 Short + 3 Long) + 24 Signal Connectors.
forum.digikey.com
.
molex "14 Power Circuits" + 24 Signal Circuits connector and sure there is one:
?
) is to blame. .I think I'll end up with a "Peace Was Never An Option" solution, rather than sourcing the proper female connector.
Well I posted what I could find. How we deal with it now it's another Topic.
Nice
.
I'm a bit puzzled because PSUs that are about half as small (thinking Dell/Supermicro 1U Redundant PSUs) can deliver ~ 1kW continuously without Issues I believe.
..I had the lid completely off during the testing and I used the air compressor at some distance, but that shouldn't be a problem because the volume of air coming from an air compressor is much greater than any fan. It did work and cooled it down quite dramatically, but it started to heat back up again just as quickly.I'm a bit puzzled because PSUs that are about half as small (thinking Dell/Supermicro 1U Redundant PSUs) can deliver ~ 1kW continuously without Issues I believe.
I agree that they are typically oversized to have Room to spin up Hard Drives.
But 600W with some Air Flow seems very low.
About Passive Cooling, keep in Mind that Load Current != Filter Current, thus the Capacitors are ALWAYS stressed (a bit more / a bit less), even though the Load Current might be Zero. So you ALWAYS need some Airflow to cool the Capacitors down. And the Filter Choke to a lower extent.
About the Fan Cooling, are you sure the Cooling was done in an uniform Manner, i.e. NOT focused only on one Point ? Did you put the Lid back on during the air Flow Test ? I ask because without a Duct the Air will NOT be forced front to Back and will just escape from the Top.
Other Possibility is of course to use a big Fan (140mm or 200mm), take the Lid OFF, then force all air from Top to Back of the PSU.
The Choice depends of course on how you plan to integrate the PSU into the rest of your System.
I had done a similar thing with my Solar Inverter: when I sealed up all Gaps with some PC Plates & Silicone Sealing Strips, Temperatures at very low Load dropped from ~ 55°C to around 25 °C. You need to have the air properly guided from Front to back. Lid is one Way. But if your Fan is bigger than 1U or if it's not "flush" with the PSU, then you need some Air Duct to force all air through. And of course, you need enough Static Pressure to achieve that Airflow (P/Q Curve for Fans).
Personally I'd probably stack 3 of these Suckers up and try with an Arctic Cooling P12/P14 Max in the Front to get relatively good Airflow & Static Pressure at Human survivable Noise Level
Alternatively each of them with the LID off and an Arctic Cooling P14 Max on top, forcing the air to the Back of the PSU.
You could of course try to drill some extra Holes in the Lid or even at the front of the PSU, especially if you cannot achieve enough Static Pressure with the Fans that you are using right now.
If you use high RPM Fans (> 2500 rpm), be careful with your Fingers, put some Fan Grills on to avoid getting them torn apart
I guess for the Air Compressor you used the "standard" (very small) Nozzle, didn't you ?
Sure, we are DIY-ing here
..Sure, because the System Integrator (Dell) had the Task of ensuring that there is sufficient Airflow going through this Power Supply at all Times.
docs.broadcom.com
..Or a Duct / Ventilation Channel. You do NOT necessarily need a Small Fan. You need Airflow and Static Pressure.
www.thingiverse.com
I'd trust the Thermocouple more than a Thermal Camera on a non-black (and potentially quite reflective, those Parts are silver-ish) Surface.
Input current (A) | Output current (A) | Input Power (W) | Output power (W) | Power factor | Efficiency | Dissipated power (W) |
0.5 | 1 | 72 | 12.4 | 0.63 | 0.17 | 59.6 |
0.51 | 2 | 76 | 24.8 | 0.64 | 0.33 | 51.2 |
0.53 | 3 | 84 | 37.2 | 0.67 | 0.44 | 46.8 |
0.54 | 4 | 87 | 49.5 | 0.68 | 0.57 | 37.5 |
0.57 | 5 | 96 | 62 | 0.72 | 0.65 | 34 |
0.61 | 6 | 109 | 74.4 | 0.76 | 0.68 | 34.6 |
0.66 | 7 | 122 | 86.8 | 0.79 | 0.71 | 35.2 |
0.7 | 8 | 135 | 99.2 | 0.82 | 0.73 | 35.8 |
0.75 | 9 | 148 | 111.5 | 0.84 | 0.75 | 36.5 |
0.8 | 10 | 162 | 123.9 | 0.86 | 0.76 | 38.1 |
0.85 | 11 | 175 | 136 | 0.87 | 0.78 | 39 |
0.9 | 12 | 188 | 148 | 0.89 | 0.79 | 40 |
0.95 | 13 | 201 | 161 | 0.9 | 0.80 | 40 |
1.01 | 14 | 215 | 173 | 0.91 | 0.80 | 42 |
That makes NO SENSE to me in Normal Operation. See the Bottom of the Post for more abnormal Scenario.
.Well that's roughly 3A Input Current, for a 10A Input Nominal Current (~ ish). Of course they don't design PFC to be optimal at 0.01% Load
.
Again, that doesn't make much sense to me.
Back there ? The PFC is in the Front ...
Well, you have to blow Air through the Case, not on the Case. Without a Duct you'd have zero Chance.
At which Input Current Load is that ?
Again, weird. I'd say the Switching Frequency is going UP at low Load, which is the Opposite of what I'd expect.
Output Stage with no Airflow ? For ~ 30% Nominal Power I'd say it's not too Bad
.If you have a Scope it would be interesting to investigate the PFC Circuit as to why it behaves like this and what changes between very low Load and 7A Load.
.
I had the thermocouples mixed up (had trouble with my less than good thermometer, it kept reading nonsense due to the switching noise). The heatsink that gets hot is the one on the right side of that. Which should be part of the switching stage and not the PFC.
Input current (A) | Output current (A) | Input Power (W) | Output power A (W) | Output power B (W) | Output power Total | Power factor | Efficiency | Dissipated power (W) |
0.5 | 1 | 72 | 12.4 | | 12.4 | 0.63 | 0.17 | 59.6 |
0.51 | 2 | 76 | 24.8 | | 24.8 | 0.64 | 0.33 | 51.2 |
0.53 | 3 | 84 | 37.2 | | 37.2 | 0.67 | 0.44 | 46.8 |
0.54 | 4 | 87 | 49.5 | | 49.5 | 0.68 | 0.57 | 37.5 |
0.57 | 5 | 96 | 62 | | 62 | 0.72 | 0.65 | 34 |
0.61 | 6 | 109 | 74.4 | | 74.4 | 0.76 | 0.68 | 34.6 |
0.66 | 7 | 122 | 86.8 | | 86.8 | 0.79 | 0.71 | 35.2 |
0.7 | 8 | 135 | 99.2 | | 99.2 | 0.82 | 0.73 | 35.8 |
0.75 | 9 | 148 | 111.5 | | 111.5 | 0.84 | 0.75 | 36.5 |
0.8 | 10 | 162 | 123.9 | | 123.9 | 0.86 | 0.76 | 38.1 |
0.85 | 11 | 175 | 136 | | 136 | 0.87 | 0.78 | 39 |
0.9 | 12 | 188 | 148 | | 148 | 0.89 | 0.79 | 40 |
0.95 | 13 | 201 | 161 | | 161 | 0.9 | 0.80 | 40 |
1.01 | 14 | 215 | 173 | | 173 | 0.91 | 0.80 | 42 |
1.05 | 15 | 227 | 173 | 12.38 | 185.38 | 0.91 | 0.82 | 41.62 |
1.1 | 16 | 241 | 173 | 24.76 | 197.76 | 0.92 | 0.82 | 43.24 |
1.16 | 17 | 254 | 173 | 37.13 | 210.13 | 0.93 | 0.83 | 43.87 |
1.21 | 18 | 268 | 173 | 49.49 | 222.49 | 0.93 | 0.83 | 45.51 |
1.26 | 19 | 281 | 173 | 61.84 | 234.84 | 0.94 | 0.84 | 46.16 |
1.32 | 20 | 294 | 173 | 74.18 | 247.18 | 0.94 | 0.84 | 46.82 |
1.38 | 21 | 309 | 173 | 86.51 | 259.51 | 0.95 | 0.84 | 49.49 |
1.43 | 22 | 322 | 173 | 98.83 | 271.83 | 0.95 | 0.84 | 50.17 |
1.49 | 23 | 336 | 173 | 111.15 | 284.15 | 0.95 | 0.85 | 51.85 |
1.55 | 24 | 350 | 173 | 123.46 | 296.46 | 0.96 | 0.85 | 53.54 |
1.61 | 25 | 363 | 173 | 135.74 | 308.74 | 0.96 | 0.85 | 54.26 |
1.67 | 26 | 378 | 173 | 147.99 | 320.99 | 0.96 | 0.85 | 57.01 |
1.72 | 27 | 391 | 173 | 160.26 | 333.26 | 0.96 | 0.85 | 57.74 |
1.78 | 28 | 405 | 173 | 172.5 | 345.5 | 0.96 | 0.85 | 59.5 |
1.84 | 29 | 418 | 173 | 184.75 | 357.75 | 0.97 | 0.86 | 60.25 |
1.9 | 30 | 432 | 173 | 196.98 | 369.98 | 0.97 | 0.86 | 62.02 |
1.95 | 31 | 446 | 173 | 209.18 | 382.18 | 0.97 | 0.86 | 63.82 |
2.01 | 32 | 460 | 173 | 221.38 | 394.38 | 0.97 | 0.86 | 65.62 |
2.01 | 33 | 457 | 173 | 233.56 | 406.56 | 0.97 | 0.89 | 50.44 |
2.76 | 46 | 635 | 173 | 390.97 | 563.97 | 0.98 | 0.89 | 71.03 |
The PFC IS switching. I believe they use Active PFC (Boost Converter) on such a high Power Module.
Again I'm not sure I'd trust the Thermal Camera on Silver color & reflective Surfaces.
.It's not the reflective surface that the thermal camera is measuring, it's the black case of the SMD component (there's a gap between the SMD component and the metal case on top).A thermocouple gives roughly the same reading, so the reading is accurate enough. Also the front case has a sticker on it, which gives a good enough surface to read the temperature.
No idea, but it's those components that are getting the hottest when you start pulling current. Which makes sense from a thermal design perspective that you'd put those at the front where the most airflow is.
Actually it could very well be the Power MOSFETs in SO-8 or 8-SOIC Package of ~20A-50A each: