yes we can get prices from various places but usually they are close...man so lenovo is taking people to the cleaners???
Lenovo Thinkcentre/ThinkStation Tiny (Project TinyMiniMicro) Reference Thread
- Thread starter Parallax
- Start date
-
- Tags
- lenovo tiny tinyminimicro
Notice: Page may contain affiliate links for which we may earn a small commission through services like Amazon Affiliates or Skimlinks.
From my experience with RAM in new, custom-built systems, Lenovo's prices are above retail but I think that is true for other major OEM PC manufacturers for their custom-built. Their prices for non-custom, pre-built systems tend to be more "normal." If you wait for a sale, you usually get +40% off on custom-built. They just had one for their P3 workstation series.
One thing I really like about Lenovo, if you search the model and "Field Service Manual" or "Hardware Maintenance Manual", you can almost always find the Lenovo technician's hardware maintenance manual (hmm). Here is the "M90q hmm". Usually for Lenovo PCs, and this is true for the M90q per the HMM, if the RAM is NOT soldered on (i.e. upgradeable / replaceable RAM) then it is a CRU (Customer Replaceable Unit). You can upgrade CRUs and it will not void your warranty. (If it is marked FRU, field replaceable unit, in the HMM, then it requires a field technician or authorized shop otherwise it will void your warranty). Given that, if I'm buying for a business I will pay the Lenovo prices. If I'm buying for personal or residential, I usually get the minimum amount of RAM and upgrade with retail RAM. But I still try to purchase during sales which they have constantly.
I just put the 64GB DDR5-4800 SODIMM Kingston kit in a ThinkPad X1 Extreme Gen5 - no problems at all.
One thing nice about that Kingston RAM is that it is CAS Latency 38 at the default JEDEC - you don't have to mess with XMP - which is great for OEM PC that might not have robust overclocking. Datasheet
This is the 4 port 2.5 gb nic that I bought from AE to use for OPNSense on a m720q.
It worked right out of the box and OPNSense detected it. They were available on amazon and ebay. But the price was $30-40 higher. So I took the risk and got from it from Aliexpress. I've had it running in OPNsense for a month now and have had no issues. Performance has been as expected via iperf3 test on my local lan and was able to saturate it to max speeds.
If you do get it from AE, make sure you get the one with the m720q/920q bracket and PCIe riser combo. I got it without the bracket and noticed that the card was a sagging without it and almost touching the main motherboard PCB components. So I had to order the bracket separately later and paid quite a bit more on ebay for just the bracket. If you have a 3d printer, you can try printing one. But I haven't found a model that will work properly. So paid the extra money to order the bracket.
One additional thing I plan on doing is drilling a 20mm diameter hole on the top cover of my m720q and put a mesh to allow for better cooling. As you can see from the AE page I linked above, there is a fan on the card with a decent size heatsink. But don't think there is much room left to let enough air circulate between the m720q cover and the card. Even though I haven't had any heat issues in m720q running as a router - I still want it to have better cooling.
Edit: FWIW - I'm running OPNSense as a baremetal install and NOT in proxmox or any other virtual environment. But I can't see this card having any issues in Proxmox.
Hope that helps.
Last edited:
Many thanks for this info! its exactly what I was after.
Ive seen that card on Ali, my only concern was I wanted passive cooling
How loud is the fan on the card? can it be heard over the CPU fan?
Anyone have system board schematics for m90q gen3? Ordered 30mm x 30mm and 40x40 fans 5v &12v, looking to potentially power them by soldering a JST terminal on CUP fan2 (pictured atrached) but do not want to take any chances before knowing if there will be any implications or issues in doing so.
My questions:
1. Can I solder a terminal on CUP Fan2
2. Can CUP fan2 support 12v or only 5v
3. Any other modifications needed to the system, bios, additional capacitors etc
3. Likelihood I fry the mobo after plugging in the fan and powering on the system
4. Is there a other location I am overlooking that may have a 2 pin terminal already installed that I can use or tap into?
My questions:
1. Can I solder a terminal on CUP Fan2
2. Can CUP fan2 support 12v or only 5v
3. Any other modifications needed to the system, bios, additional capacitors etc
3. Likelihood I fry the mobo after plugging in the fan and powering on the system
4. Is there a other location I am overlooking that may have a 2 pin terminal already installed that I can use or tap into?
Attachments
The fan is surprisingly quiet. It's not even audible when the cover is closed on the m720q.
Are you running bsd / *sense native? I've got running OPNsense on my m720q with a proxmox host, with a dual i226 nic with pcie passthrough, and as well I don't get any temp readout as well on my card.
Code:
root@OPNsense:~ # ifconfig -vvvm igc0
igc0: flags=8863<UP,BROADCAST,RUNNING,SIMPLEX,MULTICAST> metric 0 mtu 1500
options=4e0072b<RXCSUM,TXCSUM,VLAN_MTU,JUMBO_MTU,TSO4,TSO6,LRO,RXCSUM_IPV6,TXCSUM_IPV6,NOMAP>
capabilities=4f43fbb<RXCSUM,TXCSUM,VLAN_MTU,VLAN_HWTAGGING,JUMBO_MTU,VLAN_HWCSUM,TSO4,TSO6,LRO,WOL_UCAST,WOL_MCAST,WOL_MAGIC,VLAN_HWTSO,NETMAP,RXCSUM_IPV6,TXCSUM_IPV6,NOMAP>
ether 02:76:c6:xx:xx:xx
media: Ethernet autoselect (2500Base-T <full-duplex>)
status: active
supported media:
media autoselect
media 2500Base-T
media 1000baseT
media 1000baseT mediaopt full-duplex
media 100baseTX mediaopt full-duplex
media 100baseTX
media 10baseT/UTP mediaopt full-duplex
media 10baseT/UTP
nd6 options=29<PERFORMNUD,IFDISABLED,AUTO_LINKLOCAL>your output is similar to mine. Found this post which the output includes temp. I like to be able to check remotely.
Link
I made a post over on Reddit about my new cluster setup, but figured I'd post it here as well, since it is very specific to this thread.
Finally got around to setting up my new cluster. My current one is 3 m720q running Proxmox and using their local storage or connected via NFS to my UNRAID server. Suffice to say, the bandwidth performance isn't great. The 1g link is heavily saturated, so doing VM migrations or even booting VMs over NFS is pretty painful.
Ordered a Mellanox SX6036 on eBay for $100 shipped and a couple of Connect X3 cards for about $15 shipped each. So, for less than $200 (with the QSFP cables) I have a 40gbe network going.
The hosts are m720q, 32GB of RAM, a 2TB Intel NVMe on the bottom NVMe port, a 256GB Hynix in the WiFi M.2 slot (for the OS), Connect X3 cards, a 3D printable fan shroud I designed to cool off the cards, and a baffle I designed to allow a clean look on the back side of the server with a notch to allow a USB cables to pass through the back.
Fan Shroud:
www.printables.com
Bracket:
www.printables.com
Running CEPH on the Proxmox cluster with a "RAID5" erasure coded pool, I was able to achieve some pretty impressive speeds!
Pretty dang fast for using storage over the network
This is with only one port used, I'm not sure if these little units can really benefit from using the second port in a LAG setup, but I'm willing to try!
Not sure if I can get that 32gbps up to 40, but it'd be nice (9000 mtu)
With the erasure coded pool, I get about 4TB of usable space, redundancy, and incredible speed. The nice thing is, the more nodes I add to the cluster, the faster the storage gets! Now I wont have to wait minutes for a VM to start up, or worse, a migration taking a LONG LONG time, since all the nodes already have the needed data, it's practically instant.
Just wanted to share!
Power Usage:
Single Node at Idle with a Windows VM = 17w
Single Node with a maxed out network test = 56w
Single Node with a maxed out CPU (No network) = 50w
Switch at idle = 46w
Switch with 100gbps blasting through it... 46w. Lol.
Finally got around to setting up my new cluster. My current one is 3 m720q running Proxmox and using their local storage or connected via NFS to my UNRAID server. Suffice to say, the bandwidth performance isn't great. The 1g link is heavily saturated, so doing VM migrations or even booting VMs over NFS is pretty painful.
Ordered a Mellanox SX6036 on eBay for $100 shipped and a couple of Connect X3 cards for about $15 shipped each. So, for less than $200 (with the QSFP cables) I have a 40gbe network going.
The hosts are m720q, 32GB of RAM, a 2TB Intel NVMe on the bottom NVMe port, a 256GB Hynix in the WiFi M.2 slot (for the OS), Connect X3 cards, a 3D printable fan shroud I designed to cool off the cards, and a baffle I designed to allow a clean look on the back side of the server with a notch to allow a USB cables to pass through the back.
Fan Shroud:
Printables
www.printables.com
Bracket:
Printables
www.printables.com
Running CEPH on the Proxmox cluster with a "RAID5" erasure coded pool, I was able to achieve some pretty impressive speeds!
Pretty dang fast for using storage over the network
This is with only one port used, I'm not sure if these little units can really benefit from using the second port in a LAG setup, but I'm willing to try!
Not sure if I can get that 32gbps up to 40, but it'd be nice (9000 mtu)
With the erasure coded pool, I get about 4TB of usable space, redundancy, and incredible speed. The nice thing is, the more nodes I add to the cluster, the faster the storage gets! Now I wont have to wait minutes for a VM to start up, or worse, a migration taking a LONG LONG time, since all the nodes already have the needed data, it's practically instant.
Just wanted to share!
Power Usage:
Single Node at Idle with a Windows VM = 17w
Single Node with a maxed out network test = 56w
Single Node with a maxed out CPU (No network) = 50w
Switch at idle = 46w
Switch with 100gbps blasting through it... 46w. Lol.