Dell PowerEdge R640 4-Bay 3.5"
Configure Your System:
1 Processor Required
Step 1: Choose a series
Both generations drop into the same R640 motherboard, so this is a price and features decision, not a compatibility one.
Intel Xeon Gen 1 (Skylake-SP, 4100/5100/6100/8100 series) is the value play. Mature, plentiful, and the best price per core in the lineup. Memory tops out at 2666 MT/s.
Intel Xeon Gen 2 (Cascade Lake-SP, 4200/5200/6200/8200 series) adds higher clocks per tier, 2933 MT/s memory support on Gold 6200 SKUs, and Optane PMem support on L-series parts. Worth the premium when per-core performance or memory features are load-bearing; otherwise Gen 1 does the same job for less.
Whichever series you pick, populate both sockets. A single-CPU build loses half its memory channels, half its DIMM slots, and a PCIe riser.
Step 2: Choose your CPU
Match the tier to the workload, and let your OS licensing steer the core count.
Running Windows Server 2022 or 2025? Two 8-core Gold CPUs land exactly on the 16-core base license. Go to 10 cores each and you are buying four extra core packs before the server performs any differently. Our pick is the Gold 6134 (8 cores, 3.2 GHz, 130W): high clocks per core, which is what Windows and SQL Server workloads actually want. The Gold 6250 is faster still, but costs many times more for a modest clock gain, and at 185W it requires the high-performance heatsink.
Virtualizing on vSphere? Ignore the 8-core strategy. VMware licenses 16 cores per socket regardless, so buy the cores your VMs need. The Gold 6230 (20 cores, 2.1 GHz, 125W) is the balanced default; the Gold 6230R (26 cores, 150W) is the VDI density play.
Silver covers light-duty and budget builds. Platinum buys maximum core count at a steep premium, and most workloads that need 24+ cores per socket are better served by the Gold 6248 or 6252 at a fraction of the price.
Fair warning on any CPU above 150W: the high-performance heatsink and fan kit are required, not optional. Confirm they are in your build before checkout.
2 Heat Sink Optional
Standard heatsink covers CPUs up to 150W TDP in a normal datacenter environment. That is most of what we quote, including the Gold 6134 and Gold 6230.
High-performance heatsink is required above 150W (Gold 6250, 6248R, 6258R and similar), and also on lower-TDP CPUs in two cases: rear-drive or fully populated NVMe configurations, and ambient inlet temperatures above 30°C.
If you are unsure about your rack's inlet temperature, spec the high-performance heatsink. The cost difference is small and it is a lot cheaper than a thermal event.
3 Memory (RAM) 24 DIMM slots, modules added in sets of 2 Required
There are two benefits. The first is performance: full population at 2 DPC keeps all twelve memory channels working at maximum depth, and on memory-bandwidth-sensitive workloads like VDI and high-density virtualization it consistently outperforms a partial population running at higher clock. The second is cost per gigabyte: 24 smaller modules hit the same capacity for meaningfully less than 12 larger ones, and 768 GB as 24× 32 GB is the best cost per GB in the lineup.
We know the instinct is to leave a few slots open for a later upgrade, and on paper that is sensible. In practice it rarely plays out well. One year from now you are hunting for DIMMs that match the rank and speed of what is already in the box, and a mismatch either clocks the whole server down or refuses to train. The server is in production by then, so adding memory means a maintenance window you have to schedule and justify.
Buy the capacity now, in one matched set. If you outgrow it, the right move is usually another node rather than another DIMM.
384 GB (24× 16 GB) for file, print, domain controller duty through a modest virtualization host or SQL Server instance
768 GB (24× 32 GB) is our default for a general-purpose ESXi or Hyper-V host
1.5 TB (24× 64 GB) for dense consolidation, VDI, or in-memory workloads
Past 1.5 TB you are into 128 GB LRDIMMs, which reach 3 TB but cost substantially more per GB. Most buyers are better served by a second host.
4 RAID Controllers Optional
PERC H730P (2 GB cache, battery-backed) is our default on this chassis and the most common spec we quote. The 2 GB cache is appropriately sized for an 8-drive array on mixed or read-heavy workloads, and it covers the large middle of what this server actually runs.
PERC H740P (8 GB NV cache, battery-backed) is the production storage pick for write-intensive or transactional workloads: SQL Server, heavy logging, anything where write latency is load-bearing. The step up from the H730P is small money for four times the cache. If you are unsure which side of the line your workload falls on, this is the safe answer.
PERC H730 (1 GB cache, battery-backed) is the 13th-gen carryover that still works in this chassis. Viable when budget is the constraint and write performance is not, but the H730P is a small step up for double the cache. We rarely quote the H730 when the buyer sees both prices side by side.
PERC H330 (no cache) is entry-tier hardware RAID for light workloads. No cache means write performance depends entirely on the drives. Fine for a domain controller or file server; wrong for anything transactional.
S140 software RAID is dev/test only, and SATA drives only. Not a production recommendation.
One configuration this list does not cover: if you are running vSAN, Storage Spaces Direct, or Ceph, you do not want hardware RAID at all. Those stacks need direct drive access through an HBA330. Call us at 1-800-778-1545 and we will quote it.
5 Storage Drives Select up to 4 drives (0/4 Slots Used) Optional
This is the one R640 where spinning disk is the point, not the compromise. The 4-Bay 3.5" exists for capacity per dollar: bulk file storage, backup targets, archive volumes, log retention, surveillance. Large-capacity 3.5" drives deliver cost per terabyte that no SSD matches, and for the sequential, capacity-driven workloads that belong on this chassis, spinning disk is the right call. If your workload is random-IO-heavy (databases, virtualization with local storage), stop here and look at the 8-Bay or 10-Bay 2.5" configurations instead; four spindles cannot deliver the IOPS those workloads need.
Buy new. We steer you away from refurbished on drives specifically. With only four bays, a single drive failure puts you one more failure from data loss during the rebuild, and rebuilds on large-capacity drives take a long time. Do not start that clock on a drive with unknown hours.
The RAID math matters more here than anywhere else in the family: RAID 6 over RAID 5, no exceptions on large-capacity spinning disk. A RAID 5 rebuild on big drives runs long enough that a second failure mid-rebuild is a real risk, and RAID 5 does not survive one.
Keep the OS off your four bays: a Dell BOSS card, available in the Add Ons section below, gives you mirrored M.2 boot drives and leaves all four bays for the capacity you bought this chassis to hold.
6 Remote Access Required
The dividing line is the remote console. iDRAC9 Enterprise gives you full remote KVM and virtual media, meaning you can reach the server's screen, mount an ISO, and rebuild an OS from anywhere, even when the operating system is down. Express does not include the remote console, so any problem below the OS level means walking to the rack.
Our recommendation: Enterprise for any server in a datacenter, colo, or remote site. The first time a host hangs at 2 AM and you fix it from your desk instead of driving in, the license has paid for itself. It also adds automated firmware updates, two-factor authentication, and secure erase for compliance environments.
Express is the right call only when the server sits down the hall and someone can physically reach it without ceremony: a single office server, a lab box, a bench unit.
If you are deploying 5 or more units, take Enterprise on all of them. Fleet management without remote console access does not work in practice.
7 Power Supply 2x included in build Required
2x 495W covers light builds: Silver CPUs, partial memory, a few drives. Peak draw on that configuration runs around 260W, which leaves comfortable headroom.
2x 750W is our default for most builds on this chassis. A balanced configuration (Gold 6230, full memory, 8 SSDs) peaks around 460W, and even a dense VDI build stays within this unit's envelope. When in doubt, this is the answer.
2x 1100W is required territory for top-bin builds: 205W CPUs with full memory and drives peak around 720W, and any GPU configuration belongs here at minimum.
2x 1600W is the headroom play for multi-GPU inference builds or anyone who wants maximum margin.
Two rules of thumb from our deployments. Size for roughly 50 percent load at your expected peak: a PSU running at half capacity runs at peak efficiency, runs cooler, and lasts longer than one working near its limit. And when your build sits between two tiers, size up; the price step is small. Watch the estimated TDP counter at the bottom of the page as you add components to see where your build lands.
One planning note: if a GPU might be added later, buy the 1100W or 1600W now. PSU swaps are hot-swappable and painless, but ordering the right unit the first time is cheaper than a second shipment.
8 Network Cards Required
Speed first. Quad 1 GbE (Intel i350 or Broadcom 5720) is fine for domain controllers, file servers, and branch boxes where the pipe genuinely will not be the bottleneck. For a virtualization host it is a false economy: vMotion, backup traffic, and storage traffic will saturate 1 GbE and you will feel it. The 2x 10 Gb + 2x 1 Gb combo cards are the sweet spot for most builds: 10 Gb for production traffic, 1 Gb ports left over for management and backup networks. Step to quad 10 Gb (Intel X710, X550, or Broadcom 57840S) when you are separating production, vMotion, backup, and management onto their own links, which is the right architecture for any serious cluster.
Then port type, and this is the one that catches people. SFP+ ports take fiber or DAC cables and talk to datacenter-grade switches. The -T models (X540, X550) are 10GBASE-T copper and use ordinary RJ45 cabling. Neither is better; they must simply match your switch. Order SFP+ against a copper-only switch and the server racks up with nothing to plug into. Check your switch before you check this box.
Fair warning that applies to every option here: a 10 Gb card does not deliver 10 Gb unless the switch, cabling, and the device on the other end keep up. The card is the cheapest link in that chain, which is exactly why we say buy the 10 Gb card and let the rest of the network grow into it.
Need 25 GbE for all-flash SAN traffic, or additional PCIe network cards beyond the NDC? Both are in the Add Ons section below, or call 1-800-778-1545 and we will spec the whole path.
9 Operating System Required
Windows Server 2022 is the version to buy today; 2019 belongs on a new build only when an application vendor requires it. Standard licensing is per core with a 16-core base, which is why our processor guidance steers Windows buyers to two 8-core CPUs: the 16-Core license covers that build exactly. Past 16 total cores, you need the 24-Core option. On CALs, count your users and round up.
Fair warning on Evaluation Editions: free and fully functional for 180 days, then they expire. Right for labs and proof-of-concept, wrong for production. If the server is going into service, buy the license now.
VMware ESXi 8.0 is the one to deploy; 6.7 and 7.0 are past end of support and only belong on hosts joining an existing cluster. Licensing is yours to bring.
Proxmox VE is what we increasingly quote for buyers done with VMware licensing costs: open-source KVM virtualization with clustering and live migration, no per-core math. The R640 is a well-proven Proxmox platform.
Ubuntu Server LTS covers Linux application servers, containers, and web workloads with five years of free updates.
No OS is there for teams imaging from their own deployment infrastructure.
Not sure which fits? Tell us the workload at quote time, or call 1-800-778-1545.
WServer Warranty
Dell PowerEdge R640 4-Bay 3.5" Drives [14th Gen] Detailed Review
The R640 4-Bay 3.5" is the refurbished LFF capacity outlier of the R640 family. Every other R640 variant is built around 2.5" SFF drives optimized for density; this configuration swaps in four large-format 3.5" hot-swap bays in the same 1U chassis. The result is a platform that prioritizes raw storage capacity per bay over drive count: four 18 TB NL-SAS drives yields 72 TB raw in a 1U footprint, the capacity equivalent of many 2U storage servers in a single rack unit.
This chassis is for a specific procurement scenario: high-capacity spinning disk in a 1U form factor where four bays is enough and 2U is not an option. Edge computing nodes with local archive requirements, branch office servers that handle compute and bulk local storage in one unit, remote backup targets where capacity-per-rack-unit matters, and log aggregation or archive nodes where IOPS is not the constraint. If you need more than four bays, the R740xd 12-Bay 3.5" or R740xd2 24-Bay 3.5" 2U platforms are the right call. If you need SSD primary storage or NVMe, one of the 2.5" R640 variants is the better fit.
To configure a build, call 1-800-778-1545 or use the quote form below. Every refurbished unit ships under our 180-day warranty with 12+ hour burn-in testing, and volume pricing starts at 5 units.
When 4-Bay 3.5" Is the Right Choice
The 4-Bay LFF chassis earns its place when one of these patterns applies: edge computing or branch-office deployments where 1U density is a hard requirement and the workload needs meaningful local capacity, remote backup targets where 72+ TB raw in 1U beats stepping up to 2U in dense colo or remote sites, log aggregation or archive nodes where sequential write throughput on spinning disk is sufficient and capacity is the design constraint, and capacity-focused file servers in environments where the 2U upgrade is not justified by the workload size.
What does not belong on this chassis: random-I/O-heavy workloads (databases, virtualization, VDI all need SFF SSD or NVMe), deployments requiring more than 4 drive bays of LFF capacity (use the 2U R740xd or R740xd2), and any workload where rebuild time during a large-capacity-drive failure cannot be tolerated. We will tell you directly at quote time when one of those constraints applies and the LFF chassis is not the right fit.
Storage - 4 LFF Bays (the Defining Characteristic)
Four 3.5" hot-swap drive bays on a SAS/SATA backplane. This is where this configuration diverges entirely from the 2.5" R640 variants. 3.5" drives give access to capacities that simply do not exist in 2.5" form factor:
- NL-SAS HDDs up to 20 TB: Near-line SAS drives deliver the highest capacity available in spinning disk. Four 18 TB drives yields 72 TB raw, four 20 TB drives yields 80 TB raw. Dual-port connectivity for redundant path access. Sequential throughput is excellent; random IOPS are modest (typically 100 to 200 IOPS per drive). The right call for archive, backup, and sequential-read workloads.
- SATA HDDs up to 20 TB: Lower cost than NL-SAS at the same capacity. Single-port vs NL-SAS dual-port, lower sustained throughput, less suitable for multi-host shared-storage access patterns. Appropriate for backup targets and local archive where SAS dual-port redundancy is not a requirement.
- 3.5" SATA SSDs: Available in enterprise grade up to 8 TB. Unusual for this chassis. If SSD performance is the requirement, the 2.5" chassis variants are the practical choice. The option exists for specific cases where high-capacity SSD in LFF format is needed.
BOSS module for boot is mandatory on this chassis: With only four front bays available, dedicating one to a boot drive is an expensive trade. The BOSS module (dual mirrored M.2 SSDs on a dedicated PCIe card) keeps the OS off the front bays and preserves all four for data. We include BOSS as a default on every LFF build we configure; it is not optional in any serious deployment.
Capacity planning note: Four bays with RAID 6 (the configuration we recommend for data protection on large-capacity spinning disk) leaves you with approximately 2 drives of usable capacity, or 36 TB usable with 18 TB drives. RAID 10 gives 2 drives usable with better performance but the same usable capacity. RAID 5 is technically supported but we do not quote it for large-capacity spinning disk arrays: rebuild times on 18+ TB drives are measured in days, during which a second failure is catastrophically likely.
Storage Controllers
Same Dell PERC controller family as the rest of the R640 lineup. The 4-bay LFF workload profile (large sequential writes, RAID 6 protected, sustained-read on retrieval) shapes the controller choice:
- PERC H740P (8 GB NV cache, battery-backed): Our recommendation for any configuration with meaningful write workload or production data. Battery backup is particularly important on large-capacity spinning disk arrays where rebuild operations put sustained stress on the controller and drives simultaneously. The 8 GB cache size is well-matched to a 4-drive LFF array and helps absorb the parity calculations RAID 6 requires.
- PERC H730P (2 GB cache, battery-backed): Adequate for read-dominant workloads such as backup targets, archive retrieval, and sequential-read applications where peak write throughput is not the constraint. The 2 GB cache is workable on a 4-drive array though tighter than the H740P under sustained write load.
- PERC H730 (1 GB cache, battery-backed): The 13th-gen-era controller Dell maintained Mini-PERC slot compatibility for on 14th gen. Appears on the secondary market frequently as a carryover from prior deployments. Viable on this chassis on read-dominant LFF workloads where write throughput is light: cache size is small for a 12-TB-plus drive array but the workload pattern of an archive or backup target tolerates it. Quote when budget is the constraint; otherwise the H730P is a small step up for a meaningful cache size increase, and the H740P is the right answer on production data with mixed write load.
- HBA330 (pass-through): For software-defined storage or backup applications that manage drives directly (Veeam, Veritas, certain ZFS-based stacks). Many backup applications explicitly prefer direct drive access over hardware RAID for snapshot integrity reasons.
- PERC H330 (no cache) and S140 (software RAID): Light-workload only. Not recommended for production data on large-capacity spinning disk.
The controller mounts in a dedicated internal slot, so the full PCIe slot count remains available for networking and any add-in cards regardless of controller selection.
Processors
CPU options: Dual 1st Generation Intel Xeon Scalable (Skylake-SP) or 2nd Generation Intel Xeon Scalable (Cascade Lake-SP), socket LGA 3647 on the Intel C620-series chipset. Skylake and Cascade Lake are drop-in compatible. Up to 28 cores per CPU. The platform vocabulary matches the rest of the R640 family; the workload profile is what differs.
Our SKU recommendations on this chassis: Right-sizing compute to workload matters more on this chassis than on the 2.5" variants. Pure backup-target or archive workloads do not need top-bin CPUs; the drives are the bottleneck, not the CPU. Intel Xeon Silver 4214R (12 cores, 2.4 GHz, 100W) or Silver 4216 (16 cores, 2.1 GHz, 100W) are our most common specs for backup-target and archive builds. Gold 5218 (16 cores, 2.3 GHz, 125W) is the right step up for edge nodes running compute alongside the local storage tier (branch office file plus application server, edge analytics with local archive). Higher core counts (Gold 6230 and above) are appropriate only when the node runs meaningful compute workloads alongside the storage serving role.
Heatsink requirement still applies: Any CPU above 150W TDP requires Dell's high-performance heatsink kit and high-performance fan kit. Most LFF builds do not need it because the workload typically calls for Silver or low-end Gold CPUs. When the build does include a top-bin CPU (a misallocation worth flagging at quote time), the kits are mandatory regardless of chassis variant.
Single-socket warning: A single-CPU LFF build is supported and is sometimes the right answer for pure backup-target or branch-office archive nodes where dual-socket is overkill. With one CPU populated only 12 of the 24 DIMM slots are accessible and half the PCIe lanes are inactive. For genuine single-socket workloads (low-throughput backup, edge archive with light compute), this is acceptable. For nodes running compute alongside storage, dual-socket is the right call.
Memory
Architecture: 24 DDR4 DIMM slots, 12 per CPU across 6 channels at 2 DIMMs per channel. Same Purley 6-channel layout as the rest of the family. Partial population is more defensible on this chassis than on the SFF variants because the most common LFF workloads (backup target, archive, branch-office storage) do not consume the bandwidth that full population delivers.
Supported DIMM types:
- RDIMM: Standard enterprise choice. Up to 64 GB per DIMM, 1.5 TB total at full population. Most LFF builds size between 64 GB and 256 GB, well below the RDIMM ceiling.
- LRDIMM: Up to 128 GB per DIMM, 3 TB total. Rarely the right answer on this chassis; the LFF workload profile does not justify the LRDIMM premium.
- Intel Optane Persistent Memory (PMem): Cascade Lake L-series CPUs only. Not a typical LFF chassis workload pattern; if Optane is in the design, the chassis choice probably should not be the 4-Bay LFF.
- NVDIMM-N: Niche; not applicable on typical LFF workloads.
Memory sizing by workload: Pure backup target with Veeam or similar: 64 to 128 GB. Branch-office file plus application server: 128 to 256 GB. Edge node with compute alongside storage: 256 to 512 GB. Calculate memory against the actual workload, not the chassis maximum. The full-population speed-step penalty (DDR4-2666 at 2 DPC vs 2933 at 1 DPC on Gold 6200 / 5222) matters less here than on the compute-first chassis variants because the workloads are not memory-bandwidth-sensitive.
Mixing rules: Match ranks, capacity, and timing within a channel. We do not quote mixed configurations for production.
Networking and PCIe Expansion
NDC options: Spinning disk sequential throughput tops out well below 10 GbE saturation on a 4-drive array. The networking requirement on this chassis is more about access pattern than raw bandwidth:
- 4x 1 GbE: Functional for genuinely low-throughput backup or file-serving workloads at remote sites where 1 GbE is the available WAN. We do not love recommending 1 GbE in 2026, but it is appropriate in genuinely bandwidth-constrained remote contexts.
- 2x 10 GbE SFP+ + 2x 1 GbE: The baseline for most edge and branch deployments. 10 GbE for the data path, 1 GbE for management. The most common NDC on this chassis.
- 4x 10 GbE SFP+: For nodes connected to a 10 GbE storage fabric or carrying meaningful network traffic alongside the storage role.
- 2x 25 GbE SFP28: Overprovisioned for most LFF workloads. Quote on request but typically a sign that the network was sized for a different chassis class.
PCIe expansion: Up to 3 PCIe Gen3 slots depending on riser configuration. The 4-Bay LFF preserves the full riser budget structurally. Common builds: external SAS HBA for connecting to a JBOD shelf (extending the storage tier past the 4-bay limit), Fibre Channel HBA for SAN-attached secondary storage, or a single NIC for a separated management network. Multi-card builds are uncommon on this chassis; the workload mix typically does not need them.
GPU Support
GPU support is supported by the chassis (up to 3 single-width low-profile NVIDIA T4 cards or a single FPGA) but is uncommon on LFF workloads. If the deployment is edge analytics with a local GPU plus an archive tier on the four LFF bays, the configuration works cleanly: the workload mix is not the typical case but the platform supports it. For any GPU compute beyond single-card inference, look at the R740 family. For workloads that pair LFF capacity with active GPU compute (rare; usually one of those two needs is on the wrong chassis), the R740xd 2U platform is the better fit.
Management - iDRAC9 Generation
iDRAC9 Enterprise is especially important for edge: When the node is 500 miles from your datacenter team, remote KVM, virtual media, and predictive analytics are worth meaningfully more than they are on co-located hardware. Do not deploy a remote LFF node without out-of-band management. iDRAC9 Express is acceptable only on co-located builds where physical access to the console is straightforward.
Security baseline: Silicon Root of Trust anchors firmware verification in immutable silicon. System Lockdown mode prevents unauthorized firmware changes after deployment. TPM 2.0 module supported and recommended; compliance frameworks (NIST 800-171, CMMC, FedRAMP, HIPAA, PCI DSS) do not have geographic exceptions for edge nodes. Branch-office and remote-site servers carrying production data need the same security baseline as the central datacenter.
Lifecycle Controller and OpenManage Enterprise: Same Dell management plane as the rest of the R640 family. For distributed edge deployments, OpenManage Enterprise's centralized firmware compliance and configuration drift detection across remote sites is the operational win; the homogeneous fleet profile of distributed branch nodes makes drift detection meaningful.
Power and Cooling
3.5" HDDs draw more power than 2.5" SSDs, and spin-up current on large drives is significantly higher than steady-state draw. PSU sizing for this chassis:
- Light (Silver CPUs, partial RAM, 2 HDDs): 2x 495W Platinum, peak draw approximately 270W
- Balanced (Gold 5218, full RAM, 4 NL-SAS HDDs): 2x 750W Platinum, peak draw approximately 450W
- Edge node with compute (Gold 6230, full RAM, 4 HDDs, single accelerator): 2x 1100W Platinum, peak draw approximately 620W
Spin-up current consideration: Large-capacity NL-SAS and SATA drives draw significantly more current at spin-up than steady state. Staggered spin-up is managed by the RAID controller and BIOS, which handles this for a single unit cleanly. For multi-unit deployments on shared PDUs, account for spin-up surge in rack power sizing. A rack of LFF servers spinning up simultaneously after a power event can trip PDU breakers. Our team includes this calculation as part of every multi-unit LFF quote.
Thermal note: LFF chassis depth is typically slightly longer than the SFF variants (approximately 750 to 790mm) to accommodate the 3.5" drive form factor. Airflow design is similar to other 1U R640 variants; standard fan configuration is sufficient for the typical LFF workload thermal profile. Eight hot-plug redundant fans standard. ASHRAE A3 (40C) extended ambient support is achievable with the high-performance fan kit but uncommon on edge deployments where ambient is usually closer to A2.
Physical Specs & Platform Notes
- Form factor: 1U rack server. 42.8mm H x 434mm W x 750 to 790mm D depending on bezel and cable management options. Slightly deeper than the SFF variants to accommodate the 3.5" drive form factor. Standard 19-inch rack mount with Dell ReadyRails II. Confirm rail kit clearance in shallow racks before order, particularly in branch-office cabinets that may not be standard datacenter depth.
- PCIe expansion: Up to 3 PCIe Gen3 slots across the supported riser configurations. Multi-card builds are uncommon on this chassis; the workload mix typically does not need them.
- Parts availability: Strong. The 4-Bay LFF backplane is one of the less common R640 variants in the secondary market, but Dell parts coverage remains active and refurbished units are readily available. PERC controllers, NDC cards, riser kits, fan modules, and PSUs are the same as the rest of the R640 family. Large-capacity NL-SAS drives are widely available; we assess remaining drive life via SMART data on every refurbished drive before inclusion in a configuration.
- Accessories we recommend: Dell LCD bezel (P/N 521RX security bezel, 7M3F1 LCD bezel without security, 9NN24 with security; confirm part at quote time against your chassis revision), Dell ReadyRails II sliding rail kit, and the Dell cable management arm (CMA). The CMA matters especially on edge deployments where the local hands servicing the unit may not be your team and pulling the chassis cleanly is the only way to access internal components.
- Platform notes: Boot must use BOSS on this chassis (dedicating one of four bays to OS is too expensive). CPU hot-plug is not supported. Drive bays are hot-swap but rebuild times on 18+ TB drives are measured in days, so plan for a degraded array as the steady state during any failure. RAID 5 is not safe at this drive capacity; RAID 6 or RAID 10 is the floor for production data.
Our Assessment
Where it excels: Remote backup targets where 72+ TB raw in 1U beats stepping up to 2U in dense colo or remote sites; Veeam repository nodes, Veritas backup targets, and rsync-style archive endpoints land here cleanly. Branch office file and application servers that combine moderate compute with bulk local storage in a single 1U. Edge computing nodes running local analytics over an archive of operational data (manufacturing telemetry, retail transaction logs, distributed sensor data). Log aggregation endpoints in distributed environments. Archive nodes where retrieval is occasional and capacity-per-rack-unit is the procurement priority.
Where to look instead: If you need more than 4 LFF bays, the R740xd 12-Bay 3.5" or R740xd2 24-Bay 3.5" 2U platforms deliver 3x to 6x the bay count at minimal additional rack space cost, with proper LFF airflow design. If you need SSD primary storage in 1U, the 8-Bay 2.5" or 10-Bay Standard are the correct configurations. If you need NVMe, the 10-Bay NVMe is the NVMe-first variant. If your workload is random-I/O-heavy (database, virtualization, VDI), this chassis is the wrong answer regardless of capacity needs; LFF spinning disk delivers 100 to 200 IOPS per drive, which is not enough for those workloads.
Bottom line: The 4-Bay 3.5" is a specialty pick. It earns its place when 1U is a hard constraint, capacity matters more than IOPS, and 4 bays is enough to carry the workload. For backup targets, branch-office capacity nodes, and edge archive deployments, this is the right chassis. For anything that needs more bays, more performance, or random-I/O response, look elsewhere. We will not quote this chassis when the workload mismatch is obvious; we would rather steer the customer to the right configuration than ship hardware that disappoints in production.
Where the R640 Fits in 2026
The R640 family is 2 to 3 generations behind current Dell production (R650 15th gen / R660 16th gen). The 10-Bay Standard page covers the generational ladder and support status in full. 4-Bay LFF-specifically: the LFF design point is increasingly rare on newer Dell 1U platforms because the storage industry has moved capacity workloads to either 2U high-bay-count chassis (R750xd, R760xd) or dedicated object storage platforms. The R640 4-Bay remains a strong cost-performance pick for the specific 1U LFF use case in 2026, particularly for distributed edge and branch-office deployments where 14th gen fleet standardization keeps procurement on this platform. For new greenfield deployments, the conversation about whether the right answer is "more 1U LFF nodes" or "fewer 2U LFF nodes" is worth having at quote time.
Honest Limitations
- Only four drive bays. Capacity-per-bay is high with 3.5" drives, but if your design requires 6, 8, or 12 bays of LFF storage, you have already outgrown this chassis. Step up to the R740xd 12-Bay or R740xd2 24-Bay 2U platforms.
- LFF spinning disk is slow vs SFF SSD. 3.5" spinning disk delivers 100 to 200 IOPS per drive, orders of magnitude below SSD. For random-I/O-heavy workloads (databases, virtualization, VDI), the 2.5" R640 variants are the correct choice. The LFF chassis is purpose-built for capacity, not IOPS.
- RAID 5 is not safe on large-capacity LFF. Rebuild times on 18 to 20 TB drives stretch into days. The probability of a second drive failure during a rebuild is non-trivial. We will not quote RAID 5 for large-capacity spinning disk arrays. RAID 6 or RAID 10 is the floor for production data on this chassis.
- Boot drive must use BOSS. With only four bays, dedicating one to OS boot is too expensive. The BOSS module is mandatory on every serious LFF build.
- Spin-up current matters at scale. A rack of LFF servers spinning up simultaneously after a power event can trip PDU breakers. Staggered spin-up handles single-unit cases; datacenter PDU sizing must account for the surge across multiple chassis.
- Refurbished spinning disk has finite life. NL-SAS and SATA HDDs have measurable hours and reallocated-sector counts that we assess on every refurbished drive via SMART data. Drives at the end of useful life are replaced or disclosed and priced accordingly. Spinning disk ages differently than SSD; you should know what you are buying.
- PCIe Gen3, not Gen4. The R640 predates PCIe Gen4. For workloads where per-slot bandwidth matters, the R650 or R660 are the better long-term call.
- 14th gen, not current production. Dell's current 1U production platform is the R660. The R640 represents strong refurbished value in 2026 but is not new hardware.
Workload Fit
| This server is right for | Consider alternatives for |
|---|---|
| Edge computing with bulk local storage | More than 4 drive bays needed (R740xd / R740xd2) |
| Remote backup targets (72+ TB raw in 1U) | SSD primary storage workloads |
| Branch office file and application servers | NVMe performance requirements |
| Log aggregation and archive nodes | High-IOPS random I/O workloads |
| Capacity-focused 1U deployments | High-density compute (8-Bay 2.5" more appropriate) |
| Veeam repository / Veritas backup endpoints | Database hosts, virtualization clusters |
Where to Look Instead
- Need more than 4 LFF bays? The R740xd 12-Bay 3.5" or R740xd2 24-Bay 3.5" 2U platforms deliver 3x to 6x the bay count at minimal additional rack space cost. The 4-Bay LFF is the right chassis only when 1U is a hard requirement.
- Need SSD primary storage in 1U? The R640 8-Bay 2.5" or R640 10-Bay 2.5" Standard Chassis are the correct configurations.
- Need NVMe in 1U? The R640 10-Bay 2.5" NVMe is the NVMe-first variant.
- Need maximum drive count in 1U? The R640 10-Bay + RFB brings the total to 12 hot-swap SFF bays in 1U (10 front + 2 rear).
- Pre-validated vSAN HCI node? The R640 VxRail 10-Bay is the vSAN-certified version for VxRail cluster expansion.
- HPE-side 1U LFF equivalent? The HPE ProLiant DL360 Gen9 4-Bay 3.5" is the closest HPE 1U LFF analog (the DL360 family carried a 4-bay LFF chassis across multiple generations). For 14th-gen-equivalent HPE LFF capacity, the DL380 Gen10 12-Bay 3.5" is the 2U capacity-focused alternative on the Purley platform.
- Step up to PCIe Gen4 or DDR5? The Dell PowerEdge R650 (15th gen) or Dell PowerEdge R660 (16th gen) bring forward-generation features at appropriate price premiums.
- Step down to 13th gen LFF for budget? The Dell PowerEdge R430 4-Bay 3.5" is the 13th-gen-era predecessor 1U LFF chassis at a lower price point for budget-constrained edge and backup-target builds.
Ready to Configure?
LFF configurations benefit from a capacity and RAID-level discussion before quoting. The right RAID level for large spinning disk has real implications for usable capacity, rebuild time, and data protection. Tell us your target capacity (TB usable, not raw), workload type (backup target, archive, edge compute plus storage, branch file server), drive endurance preference (NL-SAS vs SATA), CPU sizing relative to workload (most LFF builds run Silver CPUs cleanly), NDC choice, and quantity. Our account team returns a fully validated configuration with formal pricing within 24 hours, including RAID-level sizing math, spin-up current calculation for multi-unit deployments, and confirmed drive remaining-life assessment via SMART data on the refurbished drives we ship. Every refurbished unit ships with the Wholesale Servers 180-day warranty and 12+ hour burn-in testing, and volume pricing starts at 5 units. Call 1-800-778-1545 or use the quote form below.
Save Your Design
Click the Add to Quote button at the bottom of your screen to save your design as a draft order for future reference and to check for discounts, lead time, and availability. Most servers ship within 1-3 days.
