Server Batteries

Add CyberPower APCRBC141-CP UPS Replacement Battery Cartridge, New to the cart
Add Cyber Power APCRBC152-CP UPS Replacement Battery Cartridge, New to the cart
Add 878643-001 - NEW - HPE 96W Smart Lithium-ion Battery to the cart
Add 398648-001 - NEW - HPE Server Battery to the cart
Add 660091-001 - NEW - HPE Flash-Backed Write Cache (FBWC) Capacitor Cable Pack w/ 12” Long Cable to the cart
Add 462976-001 - NEW - HPE 4.8v, 650mAh BBWC Battery Module to the cart
Add 70K80 - NEW -Dell H710/H730 Raid Battery Backup to the cart
Add NU209 - NEW - Dell 5i/6i Raid Battery Backup to the cart
Add 587324-001 - NEW - HPE Flash Backed Write Cache Super Capacitor with 24-inch Cable to the cart
Add NWJ48 - NEW - Dell H740/H745/H750/H755/H840 Raid Battery Backup to the cart
The Dirty Connector Problem: MPO Inspection and Cleaning

Contamination is the leading cause of fiber link failure, and one dirty position on an MPO ferrule presents as a...

My Optic Won't Link: Reading DOM and Working the Optical Loss Budget

Before replacing a transceiver, read its DOM. Transmit and receive power split a dead link into three buckets in under...

What Can You Actually Run on a 48GB NVIDIA A40? Measured Model Sizes and Tokens Per Second (2026)

One A40 48GB runs 27–34B models at 23–29 tokens/sec, fits a 70B at Q4 at 12–13.5 tokens/sec, and serves 1,705...

The Dirty Connector Problem: MPO Inspection and Cleaning

Contamination is the leading cause of fiber link failure, and one dirty position on an MPO ferrule presents as a failing module. The inspect-clean-inspect discipline, the right tools, what never to do, and the symptom patterns that should send you to the scope instead of the spares shelf.

My Optic Won't Link: Reading DOM and Working the Optical Loss Budget

Before replacing a transceiver, read its DOM. Transmit and receive power split a dead link into three buckets in under a minute. Per-lane readings, the loss budget arithmetic, and the failures DOM is blind to.

What Can You Actually Run on a 48GB NVIDIA A40? Measured Model Sizes and Tokens Per Second (2026)

One A40 48GB runs 27–34B models at 23–29 tokens/sec, fits a 70B at Q4 at 12–13.5 tokens/sec, and serves 1,705 tokens/sec of aggregate output on an 8B model at 100 concurrent requests. The measured numbers, the KV cache math behind the 72B cliff, and the three things an Ampere card cannot do.

Keith L.