Nope. But I'm if you're talking about the meanwell 12V output, i'll leave the ends open ended so you can wire your PICO to it.
Nope. But I'm if you're talking about the meanwell 12V output, i'll leave the ends open ended so you can wire your PICO to it.
So I've been playing in Sketchup for a while now trying to figure out how I would like to layout a Mean Well EPP-200 build and the headache has always been actively cooling the PSU. One option would be to just lay all the components flat (Node 202/RVZ-01 style) and put a fan above the PSU, but I wanted to try something different. The 2 options below both have (probably major) issues but I though I'd get some feedback. Both designs rely on a Zalman CNPS2X size cooler but the CPU is going to be a 51/35W part depending.
Layout 1:
The main issue here is that the PSU won't get enough airflow, the PCIe riser will block air from exiting that side of the case, so I suspect it wouldn't be able to output 200 W.
Layout 2:
This has the benefit of good PSU cooling at the cost of CPU cooling. It also has the benefit of allowing a rigid riser which would help hold up the GPU.
In both cases the front panel could be vented but that's something I'd ideally avoid.
Alternatively there's secret option 3, use the EPP-300 instead which has a passive 200W output, kind of cheating but may be easier.
In this case the increased width op the PSU allows for a larger CPU cooler (the box here is Cryorig C7 sized)
All of this is a very round about way of saying I keep talking myself back and forth between the EPP-200 and EPP-300. The 200 would be more interesting from an experimentation standpoint, but the 300 seems more likely to work as intended.
All input greatly appreciated
Layout 1:
The main issue here is that the PSU won't get enough airflow, the PCIe riser will block air from exiting that side of the case, so I suspect it wouldn't be able to output 200 W.
Layout 2:
This has the benefit of good PSU cooling at the cost of CPU cooling. It also has the benefit of allowing a rigid riser which would help hold up the GPU.
In both cases the front panel could be vented but that's something I'd ideally avoid.
Alternatively there's secret option 3, use the EPP-300 instead which has a passive 200W output, kind of cheating but may be easier.
In this case the increased width op the PSU allows for a larger CPU cooler (the box here is Cryorig C7 sized)
All of this is a very round about way of saying I keep talking myself back and forth between the EPP-200 and EPP-300. The 200 would be more interesting from an experimentation standpoint, but the 300 seems more likely to work as intended.
All input greatly appreciated
Last edited:
This is a common mistake people make, to embed Imgur pics:
either paste the non-file extension link by itself into the forum post editor:
Copy the direct image link and then use the "Image" button in the forum post editor to insert it:
Or copy/paste the "BBCode (Forums)" link straight into the forum post editor:
Code:
http://imgur.com/zYyTtNWCopy the direct image link and then use the "Image" button in the forum post editor to insert it:
Code:
http://i.imgur.com/zYyTtNWg.pngOr copy/paste the "BBCode (Forums)" link straight into the forum post editor:
Code:
[img]http://i.imgur.com/zYyTtNWg.png[/img]
Last edited:
So I've been playing in Sketchup for a while now trying to figure out how I would like to layout a Mean Well EPP-200 build and the headache has always been actively cooling the PSU. One option would be to just lay all the components flat (Node 202/RVZ-01 style) and put a fan above the PSU, but I wanted to try something different. The 2 options below both have (probably major) issues but I though I'd get some feedback. Both designs rely on a Zalman CNPS2X size cooler but the CPU is going to be a 51/35W part depending.
Layout 1:
The main issue here is that the PSU won't get enough airflow, the PCIe riser will block air from exiting that side of the case, so I suspect it wouldn't be able to output 200 W.
Layout 2:
This has the benefit of good PSU cooling at the cost of CPU cooling. It also has the benefit of allowing a rigid riser which would help hold up the GPU.
In both cases the front panel could be vented but that's something I'd ideally avoid.
Alternatively there's secret option 3, use the EPP-300 instead which has a passive 200W output, kind of cheating but may be easier.
In this case the increased width op the PSU allows for a larger CPU cooler (the box here is Cryorig C7 sized)
All of this is a very round about way of saying I keep talking myself back and forth between the EPP-200 and EPP-300. The 200 would be more interesting from an experimentation standpoint, but the 300 seems more likely to work as intended.
All input greatly appreciated
I've been debating a very build to yours, though with a lower end GPU.
The 1060 + i5t should draw about 160W, so you don't nearly the full output of the PSU, however it is not quite natural convection rated, so you need some airflow. Based on the output ratings, ambient temperature doesn't affect the output nearly as much as moving the air from the heatsinks, so I suspect as long as we get SOME airflow, we'd be good to go.
My layout uses two 92mm fan, either slim, or fullsize, to supply air to the CPU and move air around the meanwell heatsink. Since it is just a 35W CPU, you don't need a whole lot of air, just enough to keep it cool. A 35W CPU only needs about the airflow of the CNPS2X to keep it cool.
Meanwell PSU can be front mounted, or top side mounted. Most compact would be to occupy the space next to the GPU. Front mounted options allow you to use a 300W for only the cost of 5mm more clearance.
Here is an example of what I'm talking about:
I didn't draw up the front mounted layout, but I think you got the idea.
Planning a low power build with a EPP-200 as well (yet, I would need to wait for spending on some parts). But with a 1050 Ti. The GPU will be placed parallel and right next to the PSU. Time will tell if the GPU's fans blowing exhaust air will negatively affect the PSU or not.
Most 1050ti has a front/rear exhaust flow, so if you vent it properly, it should just be directed outside the case.
Be careful about choosing the riser. I've been reading some of the riser threads and not all of them can supply 75W. However, if you find a small 1050ti with a 6 pin, that might save you some hassle there.
Alternatively, using a low profile 1050ti will only incur a .3-.5L increase.
Never thought I'd actually be looking for a 1050Ti with a 6 pin lol.
The HD-Plex riser is only 15CM and has been out of stock for quite a while. Maybe it could work for a 1050TI LP card, but for a normal card you need around 250CM to invert the card so that the fans are facing the same was as the CPU cooler. For exhausting down it would work fine. If you want to go under the MB and to the left you need 300~mm.
I know HDPlex is working on a longer version of the riser, however.
Load test with G4560 (i3-6100 equivalent) + RX 460. Ran for 20 minutes. No issues. Used a cheapo PDCB from Aliexpress ($12), cut out the CPU 4 pin so it ran straight from the PSU. The panels around the PSU did get warm, about 90F. CPU stayed cool and quiet with my set up. The GPU however, not sure due to my set up of 120mm fan + deshroud, or just that it is a cheap aluminum block, did have to ramp the fan up to 40-50% to stay at 80C.
Conclusion is that you don't need a high end PDCB to run a 75W GPU + 50W CPU.
I will have an infrared gun soon so I'll give some readings of the heatsink.
Conclusion is that you don't need a high end PDCB to run a 75W GPU + 50W CPU.
I will have an infrared gun soon so I'll give some readings of the heatsink.
Great results! When you say you cut out the CPU 4 pin do you mean you ran without it connected at all (don't think this should work) or with it direct from the Meanwell? I've been looking the past couple of days to figure out what power distribution is divided between the ATX and the 4 pin with a view to doing this.
I ran it straight from the Meanwell PSU. I basically just cut it, combined the 4 wires to 2 16AWG wires, and wired it direct the PSU.
If you are running the EPP-200-12, you will have 6 total pinouts from the PSU, 3 positive, 3 com. You should just run 2 pins to the PDCB, and 2 (or 4 pins) to the 4pin CPU. Should be fairly simple. I'll provide pictures of the EPP200-12 when I receive it.
Here it is! The EPP-200-12! This thing is tiny cute. Would be so useful in builds.
What do you mean? The EPP-300 is used which is one of the higher wattage units. Or do you mean the PDCB?
What do you mean? The EPP-300 is used which is one of the higher wattage units. Or do you mean the PDCB?
Yes, that definitely is small. I would think it will take up 35mm of height once you get it installed with standoffs.
Also, that is some good news to hear about cheap PCDBs not being a problem with the 75W GPU setup. I'm planning to get a Pico-Box Z1 direct plug unit to hook it up to a EPP-200 and was expecting the DC board to be the weak link to the setup. I will try the same pinout that you used with your PCDB.
Also, that is some good news to hear about cheap PCDBs not being a problem with the 75W GPU setup. I'm planning to get a Pico-Box Z1 direct plug unit to hook it up to a EPP-200 and was expecting the DC board to be the weak link to the setup. I will try the same pinout that you used with your PCDB.
Plug-in DC modules' 'goodness' is more important if you're using a 19V supply. If you're using 12V, then those modules are almost entirely acting as a passthrough for 12V, and only need to deal with 3.3V and 5V loads (which are pretty tiny for modern PCs), so is only handling the conversion a few tens of watts at most itself. But at 19V, the plug-in modules needs to do the work of converting all the power the machine is using.
Conversely, if you're using 12V then you're in theory a little more vulnerable to the foibles of your 12V AC-DC solution. If it can't handle peak current, there's nothing the plug-in module can do about it. At 19V the presence of a DC-DC converter means some current spikes will be 'smoothed out' to below the level that would trip out the AC-DC solution.
But in practice, I suspect 19V is more vulnerable to peak transient loads. 19V AC-DC bricks are almost universally laptop PSUs, so not intended to handle power spikes at all. 12V AC-DC supplies in general seem to be designed to be more tolerant of a wide range of load behaviors, because they're not aimed at charging laptops but instead powering equipment directly. The tradeoff is that laptop bricks are mas-produced so can be had cheaply, whereas 12V AC-DC supplies are less common and generally aimed at a market with higher margins, so more expensive.
tl;dr:
19V: Cheap bricks, more expensive plug-in modules, more heat generated by plug-in module, vulnerable to power spikes.
12V: More expensive AC-DC solutions, cheap (er) plug-in modules, less heat generated by plug-in modules, potentially less vulnerable to power spikes (down to your chosen AC-DC solution to handle).
Conversely, if you're using 12V then you're in theory a little more vulnerable to the foibles of your 12V AC-DC solution. If it can't handle peak current, there's nothing the plug-in module can do about it. At 19V the presence of a DC-DC converter means some current spikes will be 'smoothed out' to below the level that would trip out the AC-DC solution.
But in practice, I suspect 19V is more vulnerable to peak transient loads. 19V AC-DC bricks are almost universally laptop PSUs, so not intended to handle power spikes at all. 12V AC-DC supplies in general seem to be designed to be more tolerant of a wide range of load behaviors, because they're not aimed at charging laptops but instead powering equipment directly. The tradeoff is that laptop bricks are mas-produced so can be had cheaply, whereas 12V AC-DC supplies are less common and generally aimed at a market with higher margins, so more expensive.
tl;dr:
19V: Cheap bricks, more expensive plug-in modules, more heat generated by plug-in module, vulnerable to power spikes.
12V: More expensive AC-DC solutions, cheap (er) plug-in modules, less heat generated by plug-in modules, potentially less vulnerable to power spikes (down to your chosen AC-DC solution to handle).
Good writeup explaining the pros and cons. I'm sure you've seen it, but for those that haven't Johnny Guru seemed to conclude that the 19V bricks were much better than 12V (efficiency wise) on the random assortment he tested (probably being due to OEMs using them and having actual QC). I can also conclude from personal experience that commercial 19V adapters have great short circuit protection when you accidentally short out the power jack on your case chassis using an HDPLEX 300W :V
Last edited:
As promised, at running about 120W, the PSU heatsink is around 56C, the board is 45C.
This is inside my case, after running the G4560 and RX 460 for 15 minutes in unigene valley. Ambient temperature is 25C. The power supply is not directly within airflow.
Last edited:
Similar threads
