I just noticed I misread the DDR3 specifications, it says:
Supports Single Channel DDR3 & DDR3L 800/1066/1333 (1.25V)
I thought it was all only 1.25 volt, but it also supports regular DDR3 which is 1.5 volt but Biostar also specified it wrong because DDR3L is 1.35 volt.
DDR3 = 1.5 volt
DDR3L = 1.35 volt (Low Voltage)
DDR3U = 1.25 volt (Ultra Low Voltage)
So the CPU temperature probably has nothing to do with the RAM.
Your experience is similar to mine with an AMD A8-5455 (1.7 GHz, 19 W). It came with a similar sized heatsink, plus 40 mm fan. When the fan died, it would shut down after ~ 5 minutes. Replacing the fan got back to normal operation. You could try fitting a fan.
Does the heatsink quickly get hot to touch ? If not, then the heatsink paste might be missing or ineffective.
[quote=âPascal Herget, post:24, topic:10085, full:true, username:eth1â])
Maybe I place an old 40x40 mm fan somewhere.
[/quote]
A UEFI having graphical display is likely to continuously display CPU temp. This is before an OS takes charge and attempts to reduce CPU frequency & heat, so the CPU will run at nominal frequency & heat generation. Try monitoring that for 15 minutes or so.
Although 77 C is within the CPUâs limits somewhere closer to 50 C would be desirable.
A 40 mm fan would probably suffice for your 10 W CPU. I went one step further with my 19 W CPU and installed a 50 mm fan from a Pentium 3. Needed some metal-bashing to fit a thin & shallow Aluminium âtubâ to the top of the heatsink.
Itâs working quite well for being unworkable. RED is connecting to a cable router.
I also tried putting on a 40x40 mm fan from an old Intel Pentium II CPU but the aluminium heatsink is actually only 35x35 mm, I attached it with only 2 screws but it was just rattling too loud. So I let it be.
I would have to replace the whole cooler. The one from A68N-5600E looks like it would fit.
Unfortunately I also didnât know that these SFX PSUs get really hot because of their compactness, could also be improved.
Maybe it is âworkingâ somehow.
But putting both interfaces ( red0 and green0 ) into the same network ( 192.168.0/24 ) is an error by design and doesât make much sense.
One functionality of IPFire is the routing of networks. To accomplish the security functions ( firewall, IPS, ⌠) data transfer between LAN ( green0 ) and WAN ( red0 ) must be forced to go through IPFire.
That it the advantage of the âtubâ. Put a 35 diameter hole in it and screw it flat to at least two corners of the heatsink
Sounds like you are using an economy, inefficient SFX PSU. Itâs difficult to find an SFX rated at less than 250 W and those are much less efficient at about 20 W for your system. I have abandoned SFX/ATX PSU for systems less than about 40 W. A pico-ATX PSU plus external power supply is more efficient plus keeps heat out of the PC case
I think there is some misunderstanding, the cable router is only connected to the RED WAN, it is not connected to the network directly, so it must always pass the firewall.
I actually wanted to use an external power supply before with a smaller case before I noticed that the case had no room for the additional PCIe card.
I donât think so. You have two possible NICs for the same network. Why should the data stream be sent through the firewall? IPFire works just as a switch in this case.
And the switch just connects the two trunks of your LAN.
OK, I overlooked this line. But why did it properly boot to the setup after the second run?
Now at home I canât confirm this, I can neither ping, traceroute the IP or access the web interface of the cable router. Shouldnât this be possible if IPFire acted like a switch?
Unfortunately I also cannot add a pass firewall rule for the cable router to prove itâs working because I get a warning "Source and destination IP addresses are from the same subnet. " But itâs actually working as I expected it to.
The DHCP is also complaining:
17:16:04 dhcpd: Wrote 1 leases to leases file.
17:16:04 dhcpd: Multiple interfaces match the same subnet: red0 green0
17:16:04 dhcpd: Multiple interfaces match the same shared network: red0 green0
17:16:04 dhcpd: Server starting service.
But I donât see why, the cable router is obviously behind the firewall.
I think it isnât worth to explore the situation with two different networks ( red0 green0 ) being the same.
IPFire is designed to differentiate WAN and LAN networks. There may be various side effects not handled properly by the system.
Just use different networks. BTW, isnât the WAN network (192.168.0.0/24) defined by the cable routerâs/ISPâs DHCP server?
It has a static IP (192.168.0.1) on the LAN side which is the WAN side on the IPFire. It has DHCP server for LAN clients. I could also set the IPFire WAN to be DHCP client, that should also work. And of course the cable router gets an IPv6 address on its own WAN side (unfortunately DS-Lite).
Ipfire is not designed to be used the way you have it.
You will have double NAT, but that should be fine
Red = 192.168.0.1 fixed is fine
Green = 192.168.10.*