Why does 100Mbps broadband only download at ~10MB/s? Am I being scammed?
No, you're not being scammed—it's a unit confusion issue. When ISPs say 100M broadband, they mean 100Mbps (megabits per second, lowercase b), while download software shows MB/s (megabytes per second, uppercase B). Since 1Byte=8bit, theoretical peak = 100÷8 = 12.5MB/s. After subtracting protocol overhead, network loss, and router/NIC bottlenecks, a stable speed of ~10–11MB/s is normal. Similarly: 500Mbps≈62.5MB/s, 1000Mbps (gigabit)≈125MB/s.
How do you tell bps, Bps, kbps, Mbps, Gbps apart?
The core difference is b vs B: lowercase b = bit (smallest information unit, 0 or 1), uppercase B = Byte = 8bit. So 1Bps = 8bps. Rate unit prefixes: 1kbps = 1000bps (decimal, telecom standard), 1KiB/s = 1024B/s (binary, computer storage). Note: data transfer rates almost universally use decimal (1000 base), while file sizes use binary (1024 base). Rule of thumb: ISP Mbps ÷ 8 = download MB/s.
When should I use decimal (1000) vs binary (1024)?
Simple rule: network bandwidth/serial/USB/PCIe/fiber rates — use decimal (1000 base, SI standard); hard drive/file copy speed/memory bandwidth — you may see either, check device documentation. Telecom almost exclusively uses decimal (Ethernet 1000BASE-T = 1Gbps = 10^9bps); internal computer data transfer may mix both. In practice, judge strictly by documented unit notation (case of k/K/M/G, presence of i): kbps=1000bps, Kibps=1024bps, Mbps=10^6bps, MiB/s=2^20B/s.
What are theoretical speeds for 5G, gigabit fiber, USB 3.0?
Common interface/network theoretical peaks (decimal): Gigabit fiber = 1000Mbps ≈ 125MB/s; 5G (Sub-6GHz) downlink ≈ 1Gbps ≈ 125MB/s, mmWave 5G can reach 10Gbps; USB 2.0 = 480Mbps (actual ~30–40MB/s); USB 3.0 = 5Gbps (actual ~400MB/s); USB 3.1 Gen2 = 10Gbps; USB 3.2 = 20Gbps; Thunderbolt 3/4 = 40Gbps; SATA3 = 6Gbps (actual ~550MB/s); PCIe 4.0 x16 ≈ 32GB/s (bidirectional); Wi-Fi 6 (AX) ≈ 9.6Gbps; Wi-Fi 7 ≈ 46Gbps (theoretical). Actual throughput is typically 50%–80% of theoretical.
How long to download a 10GB game on 100Mbps broadband?
Estimation method: 10GB file = 10×1024×8 = 81920Mb; 100Mbps = 100Mb per second; theoretical time = 81920÷100 = 819 seconds ≈ 14 minutes. But actual speed is affected by server bandwidth, resource popularity, peak-time congestion, and routing hops—allowing 20–30 minutes is realistic. On 20Mbps (typical home) it takes ~70 minutes; on gigabit fiber (1000Mbps) theoretically about 1.5 minutes.
Are baud rate and bit rate (bps) the same thing?
Not exactly—this is a common point of confusion in serial/modem contexts. Baud rate is the number of signal changes (symbols) per second; bit rate is the number of bits transmitted per second. In binary baseband transmission (like RS-232 serial), 1 symbol carries 1bit, so baud rate = bit rate (e.g., 9600 baud ≈ 9600bps). But with higher-order modulation (like QAM256/QPSK), 1 symbol can carry multiple bits (e.g., 256QAM = 8 bits per symbol), where bit rate = baud rate × log2(modulation order). In most engineering contexts, '9600 baud' effectively means 9600bps.
How much upload bandwidth do I need for live streaming/video conferencing?
Common video service bandwidth reference (down/up in bps): 1080P streaming push: ~4–8Mbps upload; 4K streaming push: ~15–25Mbps upload; Zoom/Teams 1080P: ~2–3Mbps up/down; YouTube/Twitch/niconico 1080P viewing: ~5–8Mbps download; 4K streaming: ~15–25Mbps download; 8K streaming: ~50–100Mbps download. Note: live streaming and video conferencing demand high upload bandwidth; home broadband upload is typically only 1/10–1/5 of download (e.g., 100M broadband may only have 10–30Mbps upload).
Why does Wi-Fi speed test only show a few hundred Mbps on gigabit broadband?
Bottlenecks can occur at multiple points: (1) Wi-Fi standard: Wi-Fi 5 (AC) single-antenna max 433Mbps; Wi-Fi 6 (AX) needed to approach gigabit; (2) Band interference: 2.4GHz band is congested, actual speed only 50–150Mbps—use 5GHz band; (3) Router ports: some older routers have only 100Mbps WAN ports; (4) Ethernet cable: Cat5e supports gigabit, Cat5 does not; (5) Client NIC: old phones/laptops may not support; (6) Encryption/handshake overhead: actual TCP throughput is ~60%–70% of physical rate. Recommendation: gigabit router + Cat5e/Cat6 cable + 5GHz band + Wi-Fi 6 clients.
How many MB/s is 1Gbps? How to convert?
1Gbps = 1000Mbps = 1000÷8 = 125MB/s (decimal). Using binary (1MiB=1024KiB) it's about 119MiB/s. When people say 'gigabit downloads at 125 megs' they mean 125MB/s. HDD sequential writes are ~100–200MB/s, SATA3 SSDs ~550MB/s, NVMe PCIe SSDs can reach 3000–7000MB/s, so gigabit network usually isn't bottlenecked by storage.
How long to transmit one byte at serial baud rates 9600/115200?
Serial commonly uses 8N1 format (8 data bits + 1 start bit + 1 stop bit = 10 bits per byte): At 9600bps, 1 byte takes 10÷9600 ≈ 1.04ms; at 115200bps, 1 byte ≈ 86.8μs. Modbus RTU and industrial protocols commonly use 9600; debugging/firmware updates commonly use 115200. Note: with 7E1 (7 data + even parity + 1 stop = 10 bits/byte) it's also 10 bits per byte, but only 7 bits are payload. Calculate based on total frame bits, not simply ÷8.
Is there a case difference between k/K and M/G in bandwidth units?
Yes, strict differences: lowercase k = 10^3 = 1000 (decimal), uppercase K in strict usage means 2^10 = 1024 (binary); but in practice many technical documents mix them. IEC recommends Ki/Mi/Gi for binary (KiB=1024B) and k/M/G for decimal (kB=1000B), but marketing materials for network bandwidth almost universally use lowercase k/M/G at 1000. Linux/ifconfig shows KB/s at 1024; Windows Task Manager shows Mbps at 1000. When confused: network bandwidth → calculate at 1000; file size → calculate at 1024.
Why use this tool? Can't I just divide by 8?
The simple Mbps→MB/s ÷8 estimate works for quick decimal math, but manual calculation easily goes wrong when: k/K is mixed (1000 vs 1024), Gbps/Tbps large orders of magnitude are involved, serial baud rates include framing bits, different scenarios (comms/storage/video bitrates) use different bases, or you need multiple unit comparisons at once. This tool converts all 18 units in one go, supports 1000/1024 switching, gives precise results, and avoids manual calculation errors.
