modulation – so many names, so much signal

Hey again! Let’s talk more Wi-Fi – yeah, all that modulat-y stuff. It gets messy, but I’ll keep it chill.

So like, Wi-Fi uses all kinds of tricks to push bits through the air. And those tricks got fancy names: BPSK, QAM, OFDM, DSSS, even infrared (yeah, that beam-y thing). Each one’s got its own job – some are old-school, some are super fast, and some are just weird but clever.

OFDM? It slices your data into tiny streams and slaps them on different frequencies. Neat! DSSS? Spreads your signal like peanut butter over a wide channel – smooth. QAM? Makes your signal stronger *and* shiftier. 64-QAM, 256-QAM, even 1024-QAM... those are like turbo Wi-Fi moves. Then there’s CCK, DBPSK, GFSK – sounds like robot talk, right?

Anyway, all these types of modulation are used in different “PHYs” (that’s short for Physical Layers – the part of Wi-Fi that actually hits the air). Each PHY’s got its own set of moves, from the simple stuff (1 Mbps) to the wow-stuff (gigabit-level speed!).

We’ll break ‘em down, show you who does what, and why some are fast, some are tough, and some are just there to keep the old gear happy.

Let’s dive in, one mod at a time


 

let’s talk ofdm – the big brain behind fast wifi

okay, so now we enter the serious part. like, really. this thing here – OFDM – is the muscle and brain of your wifi, the thing that makes stuff fly fast through airwaves. orthogonal frequency division multiplexing. yeah, long name. don’t stress.

imagine tryin’ to stuff a big highway full of trucks. it clogs. but what if you break that one big road into lots of tiny roads? and send one lil’ truck per lane, all together? that’s what OFDM does. splits your data, sends it smart across many tiny lanes (they call 'em subcarriers), and bam – your internet gets real slick.

so what’s it doin exactly?

first, it works mostly on 5GHz bands, but newer wifi types got it workin on 2.4 and 6GHz too. speeds? how about 6, 12, 54 megabits per sec. depends how you tweak it – more band, more speed. it even has slow modes for tight spaces (like 5MHz lanes).

it uses things called modulations – BPSK, QPSK, 16-QAM, 64-QAM. fancier modulations = more bits per ride. also does error checking with somethin’ called convolutional coding. sounds fancy but it’s like digital spell-check for wifi bits.

there’s structure too, not just chaos

OFDM splits up the signal: preamble (the “hello” part), header (says how big the message is), and payload (actual stuff you send). those short and long training sequences? they help your wifi receiver tune in, like sayin “hey i’m here – now focus”.

it’s also got a thing called Guard Interval – kind of like extra space between cars, to stop crashes when signals reflect off walls and come late. so yeah – it’s prepared for mess.

extra features? oh boy

you got pilots too – no, not flying ones. these are little signal pieces that help with sync. there’s also DSSS-OFDM (a mix of old and new), and OFDMA, where many devices share chunks of the air at the same time like a polite dinner party.

so what’s the point?

OFDM is what keeps your wifi chill, fast, and smart. without it, you’d be in 1999 speeds. it’s like the backstage crew makin’ sure the show runs without the audience even noticin’. and it’s only getting better with new tricks like OFDMA.

how fast is fast? let’s dig into ofdm speed stuff

so yeah – we already said ofdm is like the wifi power engine. but now we zoom in and look at how fast it really goes and what kinda space (frequencies) it needs. turns out, there’s more to it than just “it’s fast.”

where does ofdm live?

mostly in the 5 GHz band – it’s like its home turf. but it also plays in 4.9 GHz, 2.4 GHz (with help), and now the shiny new 6 GHz band too. modern wifi (like 802.11ax) spreads this tech across all the bands, makin’ sure it works no matter where it’s needed.

the secret sauce: headers, scramblers & pilots

first thing a receiver gets? the header – especially that SIGNAL field. it’s always sent slow and strong (bpsk + 1/2 code rate), so everyone hears it. it tells the rest of the packet what’s coming.

then you get the payload – the real data – scrambled, coded, modulated, and pushed onto 48 subcarriers (the other 4 are “pilot” carriers – they help the receiver keep its head straight).

time is everything

each symbol has a set duration – usually 4.0 microseconds for 20 MHz. if you go down to 10 MHz or 5 MHz, that time gets longer. there’s also a guard interval, kind of like a safety zone between symbols so they don’t trip over each other.

ofdm’s cooler cousins

  • DSSS-OFDM: old and new shake hands. a hybrid mode to play nice with older systems. same rates as regular ofdm.
  • OFDMA: the latest jam – split the spectrum into slices, give each device its own piece. multiple users, less waitin’. way more efficient.

and yeah, you can even use different power levels for different chunks in OFDMA. fancy stuff.

ofdm adapts. it scales. it’s the swiss army knife of wifi. whether you need fast speeds, tight lanes, old hardware support or lots of users at once – it’s got the moves.

let’s talk about ofdm, but the nerdy version

alright, so here’s the deal. OFDM ain’t just some fancy name — it’s the base layer, the bones, the thing that makes wifi fly in the air. it started out as the 5 GHz champ, but now it’s kind of everywhere, including 2.4 GHz and 6 GHz, all thanks to newer wifi standards like 802.11ax.

how fast? depends on the lane width

you wanna go fast? or save some space? that’s what channel width is for. and OFDM gives you options:

  • 20 MHz: you get 6 to 54 Mb/s. symbol time = 4.0 µs, slot time = 9 µs. that’s the classic setup.
  • 10 MHz (half-clocked): slower highway, goes from 3 to 27 Mb/s. symbol time doubles to 8.0 µs, slot time 13 µs.
  • 5 MHz (quarter-clocked): tiny lane. only 1.5 to 13.5 Mb/s. symbol time is 16.0 µs, slot time 21 µs. slow but useful.

and for the modern HE PHY (that’s high-efficiency wifi), you’ve got 20 MHz as must-have. 40 and 80 MHz are required too, but only in 5/6 GHz. the big ones — 160 and 80+80 MHz — they’re fancy add-ons.

modulation? it's all about the bits per symbol

OFDM splits your data into tones (they call ’em subcarriers). 52 of ‘em. and then you slap modulation on top:

  • BPSK – basic, one bit at a time
  • QPSK – two bits, four phases
  • 16-QAM – four bits per symbol
  • 64-QAM – six bits, and you better have a clean signal
  • 256-QAM and 1024-QAM – for HE PHY, where things get crazy fast

the SIGNAL field? always BPSK with code rate 1/2 – so every STA gets the message loud and clear, no matter how old it is.

and yeah, the bits are turned into complex numbers (constellations), and normalized with something called Kmod – so your signal doesn’t blow up. values like 1, 1/√2, 1/√10... all to keep things steady.

pilot subcarriers = little helpers

4 of the 52 tones (-21, -7, 7, 21) are pilots. they help the receiver keep track of timing and noise. these little guys are BPSK too, scrambled so they don’t make weird spikes on the spectrum.

dc... what?

DC is bad here. so OFDM skips subcarrier 0 (right in the middle) to avoid it. cleaner signal that way.

bonus round: dcm for picky use cases

some new HE packets can use DCM (dual carrier modulation). it spreads bits across two tones, like backup copies. but only for MCS 0, 1, 3, 4 – and only if you’ve got 1 or 2 streams. no MIMO here.

error correction – 'cause real life is noisy

bits get messed up, so OFDM does coding:

  • Convolutional Coding: basic one – code rates 1/2, 2/3, 3/4. works up to 4 streams.
  • LDPC: the fancier one. needed for wide bandwidths or high MCS. HE stations doing big things need this.
  • Scrambling: yep, all bits go through a scrambler (polynomial math magic) to avoid long runs of 0s or 1s. makes things more robust.

the SIGNAL field, tho? that part stays unscrambled. it’s gotta be read clearly first.

okay, but what does all that mean?

it means OFDM adjusts. it’s flexible. you can slow it down to fit narrow bands, or juice it up to go crazy fast. it picks what mod scheme fits the moment. and it even helps clean up your data as it goes.

stick around – next we dive into how OFDM structures its packets, symbols, and guards. yeah, even wifi needs personal space.


DSSS-OFDM

what's dsss-ofdm? kinda old but kinda clever

okay so here's the scene: OFDM is that slick thing in wifi that splits up your data across tons of little channels. but back in the day, we also had DSSS (Direct Sequence Spread Spectrum), which was simpler but slower. then DSSS-OFDM came along and mashed ’em together. kind of like a remix – legacy vibes, modern beats.

why even bother with it?

DSSS-OFDM was cooked up to help older wifi gear (like 802.11b stuff) talk with newer OFDM-powered ones, without needing extra coordination. it shows up in the 2007 version of the standard and yeah – it’s marked deprecated now, but still worth knowing.

speeds? it's the usual suspects

with DSSS-OFDM turned on, you’re lookin’ at these data rates:

  • 6, 9, 12, 18, 24, 36, 48, and 54 Mb/s

same ones as ERP-OFDM. it builds on top of the old DSSS speeds like 1, 2, 5.5, and 11 Mb/s. so it’s backwards-compatible and faster.

how's the signal structured?

a DSSS-OFDM frame (PPDU) has 3 parts: preamble, header, and data. the preamble and header use classic DSSS-style modulations (DBPSK), so old devices can still recognize it.

the SIGNAL field is always set to “3 Mb/s” (hex 1E). this tells old gear to chill and let the new gear take over for a sec. the actual data gets modulated using OFDM with BPSK/QPSK/16-QAM/64-QAM – you know the drill.

data part: the real deal

the data part is straight-up OFDM style – same as clause 17 in the standard. here’s what happens inside:

  1. long sync sequence: OFDM uses it to lock onto the signal
  2. ofdm SIGNAL field: says how fast the data will come and how long it is
  3. data section: the actual bits, encoded and scrambled, then dropped on 52 subcarriers
  4. 6 µs signal extension: kind of like a breather for the receiver, helps meet timing rules like SIFS

special sauce: phase handoff

now here’s the spicy part – when switching from DSSS to OFDM inside one frame, the signal phase needs to stay smooth. no sudden jumps. so the first OFDM symbol is rotated 45 degrees clockwise compared to the last DSSS one. this keeps demodulators happy.

error correction? of course.

as always, noisy air means you need forward error correction. DSSS-OFDM uses the same stuff as classic OFDM:

  • convolutional coding: with 1/2, 2/3 or 3/4 code rates, using 64-state encoders (k=7)
  • scrambling: so you don’t get weird repeating bit patterns – it uses the G(z) = z⁻⁷ + z⁻⁴ + 1 polynomial

anything else cool?

yeah – DSSS-OFDM lets receivers do linear distortion mitigation across the whole frame without resetting. no re-sync needed when switching from the DSSS part to the OFDM part. and pilot tones (-21, -7, 7, 21) help clean up phase noise and frequency offset.

is it still used?

not really. it’s deprecated now. but back in the early wifi evolution days, DSSS-OFDM was a pretty clever way to bridge the old and new worlds without breaking stuff. you might still bump into it in legacy hardware or older specs.


OFDMA

yo what's ofdma anyway?

so like, OFDMA stands for Orthogonal Frequency Division Multiple Access – it's basically a group-sharing version of OFDM. instead of one user hogging all subcarriers, we slice 'em into chunks called RUs (resource units) and hand 'em out to different users. more sharing, more efficiency.

how it's diff from classic ofdm

in plain OFDM, one user gets all the subcarriers. with OFDMA, we break 'em up and let multiple users talk at once. every PPDU (that's the data packet thingy) can switch up how those chunks are handed out.

subcarriers & resource units (rus)

  • data subcarriers: carry the good stuff (actual data)
  • pilot subcarriers: keep the signal clean from noise and phase drifts – especially those on -21, -7, 7, 21

btw, power levels can be tweaked for different RUs too – like more juice where needed.

modes and use cases

OFDMA’s used in High Efficiency (HE) PHY – think 802.11ax. it works both ways: downlink (AP to users) and uplink (users to AP). bunch of modes here:

  • single stream to one user (SISO)
  • multi-stream to one user (SU-MIMO)
  • multi-stream to many users (MU-MIMO)

plus there's this cool trick called UORA – uplink OFDMA random access – where STAs just jump in and transmit on special random-access RUs. only works if they got the right flag flipped (that OFDMA RA support bit).

bandwidth stuff

so yeah, OFDMA works across all kinds of widths:

  • 20 MHz: must-have for everyone
  • 40 + 80 MHz: required if you're not 20 MHz-only
  • 160 & 80+80 MHz: totally optional vibes

and spatial reuse lets folks cram more into same bands by managing power and overlap smartly.

formats – how a packet looks

OFDMA HE PPDUs come in all shapes depending on width:

  • 20 MHz with RU < 242 tones
  • 40 MHz with RU < 484
  • 80 MHz with RU < 996
  • 160 MHz or 80+80 MHz with RU < 2×996

they got all these headers like L-STF, L-LTF, HE-SIG-A/B and more – basically helps your wifi chip know what’s up.

trigger frames & multiuser uplink

if an AP wants to make users talk at the same time, it sends a trigger frame that tells each STA what RU and stream to use. it’s like a traffic cop with a whistle. no chaos, just organized noise.

coexistence and overlap tricks

APs can choose to accept super narrow 26-tone transmissions from overlapping BSSs without thinking it's radar noise. if a device ghosts a trigger frame, the AP might skip sending it another 26-tone RU next time. basically: behave or you’re benched.

extra goodies

  • tx accuracy test: checks that empty subcarriers don’t mess up the overall signal (yep, even the quiet parts count)
  • LDPC tone mapping: if you’re using LDPC coding, there’s a special mapper for matching tones to code. it doesn’t kick in for older BCC modulations though

why it matters

OFDMA = smoother traffic, more efficient air time use, and lower latency. it’s what makes 802.11ax actually “high efficiency.” without it, everyone’s just yelling one at a time. with it, it’s more like coordinated group chat over the air.


UORA... say what?

UORA stands for Uplink OFDMA-based Random Access. It's like giving Wi-Fi devices a way to shout “me first!” without stepping on each other. Totally part of that fancy Wi-Fi 6 (802.11ax) vibe.

why tho?

UORA lets non-AP stations (your phones, laptops, etc.) randomly grab airtime in little blocks called RUs (resource units). Makes things faster by cutting down protocol chat and boosting the real data speed.

how does it even work?

  • Optional stuff: Devices don’t gotta do UORA, but if they do, they flip a capability flag (yep, `dot11OFDMARandomAccessOptionImplemented` = true).
  • The AP (router dude): sends out Trigger Frames with special RUs just for random access. Could use BQRP, BSRP, or just basic ones.
  • The STA (your gadget): listens for these and jumps in when ready — if it supports it.

RUs and all that jazz

RUs are chunks of frequency. The AP can mark a few of them as “free for all” (RA-RUs). Then it’s like musical chairs — devices back off a random time and try their luck.

OCW and OBO? sounds like cereal

  • OCW: “OFDMA Contention Window.” Bigger window = more backoff wait options. Starts small, grows after fails.
  • OBO: “OFDMA Backoff Counter.” Starts at some random value, ticks down each time a new frame shows up.
  • If OBO hits 0 and enough free RUs are in the air, the device jumps in and sends its data!
  • If not? It waits. Ticks down more. Shrugs.

retry dance

Didn’t make it? The device doubles its OCW and rolls a new OBO. Tries again later. Still fails? Keeps going till OCWmax is reached, then just chills at that level until reset.

extra flavor

  • Not associated? Devices not yet linked to the AP can still sniff those special RA-RUs and figure out who’s who (BSS coloring and all).
  • Power saving: With TWT (target wake time), devices can nap and wake up only when they expect good RU giveaways. Save battery, win Wi-Fi.
  • NAV rules: Devices follow some access rules (NAV), depending on whether they know the AP or not.

SISO

yo what’s siso?

alright, SISO stands for Single Input, Single Output. it’s the chillest setup in wireless – one antenna sending, one antenna receiving. no fancy tricks. just plain ‘one guy talks to one other guy’ kinda deal.

siso in ofdma – simple but solid

even in the world of OFDMA where everyone’s sharing the air, sometimes you just wanna keep it simple. that’s where SISO kicks in. inside a resource unit (RU), you can have a single stream going to one user. no extra spatial streams, no juggling.

where it fits

in every PPDU (that’s your Wi-Fi data packet), you can run SISO. it’s perfect when you don’t need MU-MIMO or multiple streams. it just gets the job done. clean. basic. reliable.

why use it?

less hardware, less complexity, more compatibility. most older or simpler Wi-Fi devices roll with SISO. also good for low-bandwidth jobs or when signal conditions ain’t that great.

so yeah – SISO might not sound sexy, but it keeps your Wi-Fi talking when all else fails.


SU-MIMO

yo what’s SU-MIMO?

so SU-MIMO is short for Single-User Multiple Input, Multiple Output. it’s like giving your Wi-Fi extra arms. you got many antennas at the sender, and many at the receiver – and all of 'em work together to shoot multiple data streams to one single device. yeah, one user, multiple streams. boom.

how it’s different from siso?

well, with SISO you got one stream, one antenna – classic old-school style. but SU-MIMO? nah, it’s advanced. it slices your data into parallel spatial streams and beams it to just one lucky user – all at once. that’s what they call spatial multiplexing.

su-mimo inside ofdma

inside OFDMA (ya know, where many users share different chunks of air), SU-MIMO can happen inside a single resource unit (RU). like, the RU could carry one stream (SISO), or multiple (SU-MIMO), or even be shared across users (MU-MIMO). SU-MIMO means all streams go to one person, just faster.

beam me up – beamforming

SU-MIMO uses beamforming – kinda like pointing all your signal guns straight at your buddy. the sender (called beamformer) figures out where to aim using feedback from the receiver (beamformee). that feedback’s like: “yo, tilt ϕ left, twist ψ a bit right”. super precise. better signal. better speed.

who can play this game?

HT and HE stations with multiple antennas can do SU-MIMO. a HE STA says “i can beamform” by setting a flag in its capability field. Access Points with 4+ antennas must support it. receivers say “yeah i can catch beams” with a different flag. some gotta support it, some it’s optional.

what’s the ppdu deal?

there’s a few PPDU formats just for this: HE SU PPDU and HE ER SU PPDU (extended range stuff). the amount of training symbols (HE-LTF) depends on how many streams (NSTS) you got. more streams = more training.

modulation + coding

the signal can be BPSK, QPSK, 16-QAM, or 64-QAM – depends on the rate. if LDPC coding is used (that’s low density parity check, fancy error-fixing stuff), then a tone mapper gets applied to all the streams in the RU. this ensures the bits get spread out real smart.

SU-MIMO makes your Wi-Fi smarter and faster by sending multiple data rivers at once to one receiver. like running a 4-lane highway straight into your laptop. 


MU-MIMO

yo what’s mu-mimo?

okay so MU-MIMO means Multi-User Multiple Input, Multiple Output. it’s a way for Wi-Fi to talk to multiple peeps at the same time. same frequency, different users, boom – everyone's happy and downloading.

how’s it different?

SISO talks to one user with one stream. SU-MIMO talks to one user with many streams. but MU-MIMO? it’s the boss move – talk to many users at the same time using spatial multiplexing. more antennas, more action.

downlink mu-mimo – AP takes the lead

in the old days (like 802.11ac), only the access point (AP) could do MU-MIMO downlink. AP beams out to 2, 3, or 4 users in one go, each on their own spatial stream. kinda like a DJ spinning different tracks to different people at once.

uplink mu-mimo – now clients talk too

with 802.11ax (that’s HE PHY), now even non-AP stations (STAs) can yell back at the AP at the same time. AP sends a trigger frame, clients shout back using HE TB PPDUs. and they can mix OFDMA + MU-MIMO if they wanna.

mu-mimo meets ofdma

yep, they work together. inside a PPDU, some RUs (resource units) can run SISO, some SU-MIMO, and some full-blown MU-MIMO. you just need big enough RUs (like 106 tones and up). parts of the band can even mix it up. wild.

how many users and streams?

HE PHY lets MU-MIMO hit up to 8 users per RU, and each can get up to 4 spatial streams. but total streams can’t go over 8 in one go. keep it balanced, ya know?

beam me up, again

MU-MIMO uses beamforming too, like SU-MIMO. AP (the beamformer) figures out where to send what using feedback from each user. it lines up signals so they don’t trip over each other. clever math involved. feedback = precision.

ppdu formats and bits

DL MU uses HE MU PPDU. UL MU uses HE TB PPDU. there’s a field called HE-SIG-B that tells who gets what, how many streams, and how to pack the PSDU so everything finishes on time. teamwork, but in binary.

who needs to support what?

APs with 4+ streams must support DL MU-MIMO. STAs that do UL MU-MIMO on partial RUs gotta tell the AP in their capability field. oh, and all HE APs must support plain DL OFDMA (aka no MU-MIMO). fair’s fair.

MU-MIMO = many antennas + smart beamforming + parallel users. more speed, more efficiency, more Wi-Fi for everyone.


DL MU-MIMO

yo what’s dl mu-mimo?

DL MU-MIMO (Downlink Multi-User Multiple Input, Multiple Output) means your Wi-Fi router (the AP) blasts out different data streams to multiple devices at the same time over the same frequency. like group texting, but smarter and faster.

the idea and how we got here

unlike SU-MIMO (one-to-one stream action), DL MU-MIMO is one-to-many. started out in the 802.11ac (VHT) days, but got way better with 802.11ax (HE PHY). back in the day it was all about downlink only – uplink came later with HE.

how it works

  • AP builds multiple A-MPDUs, each for a different device (STA).
  • sends 'em all together as separate spatial streams in a single PPDU.
  • each STA picks out its stream and decodes it. boom – less waiting, more throughput.

streams and users

in HE PHY: up to 8 users, max 4 streams each, but total can’t go over 8 spatial streams across all users. older VHT setups capped out at 4 users/streams total.

beam me down, scotty

AP does beamforming magic – figures out how to steer signals with its antennas. needs feedback from the STAs (called Vk) based on sounding packets (HE NDPs). then it tweaks the Qk steering matrix to deliver clean signals to everyone at once.

works with ofdma too

  • Full bandwidth DL MU-MIMO: whole channel used for MU-MIMO.
  • Partial bandwidth DL MU-MIMO: MU-MIMO only on parts of the channel (specific RUs).
  • can even mix & match – some RUs do MU-MIMO, some don’t.

protocol stuff

the AP uses group IDs to manage who gets what. each RU must end its data at the same time – synchronized like a marching band but for bits.

why it rocks

more data to more devices in less time. users get faster downloads and the network breathes easier. can boost per-user throughput by 4x compared to old-school SU-MIMO.

what’s needed?

  • AP: if it can do 4+ streams, it must support full bandwidth DL MU-MIMO.
  • STA: must support receiving DL MU-MIMO with at least 4 total spatial streams across all users.
  • they also advertise this in their HE Capabilities – look for flags like Partial Bandwidth DL MU-MIMO.

so yeah – DL MU-MIMO’s like the AP saying “i got plenty of bandwidth, everybody gets a piece.” and everyone wins.


UL MU-MIMO

yo what’s ul mu-mimo?

UL MU-MIMO stands for Uplink Multi-User Multiple Input, Multiple Output. in normal words: bunch of devices talk to the same Wi-Fi router at the same time using the same channel but on different spatial streams. magic? nah, just math + antennas.

what makes it special?

it’s like SISO = one stream, SU-MIMO = many streams to one user. but UL MU-MIMO = many users sending many streams to one AP all at once. it’s all about that spatial multiplexing life.

how does it start?

wasn’t a thing before 802.11ax (HE PHY). now, the AP sends out a trigger frame, says "yo you, you, and you – send me your stuff now." the clients send using HE TB PPDU format. synchronized chaos, but it works.

teamed up with ofdma?

you bet. UL MU-MIMO works with or without OFDMA. if it’s mixed, they call it partial bandwidth UL MU-MIMO. if it uses the whole band? that’s full bandwidth. RU (resource unit) needs to be at least 106 tones if it’s not full band. size matters.

how many streams per person?

each user can send up to 4 spatial streams, if their device can handle it. but all together? total max is 8 streams for UL MU-MIMO. gotta keep things in check.

beam me... from the client?

yup, kinda. the AP can tell the clients what LTF mode to use: single stream pilot or masked HE-LTF, depending on how many HE-LTF symbols (1x, 2x, 4x). even in uplink, beamforming coordination’s a thing.

who supports this?

for clients (non-AP HE STAs), it’s optional. but if they wanna do UL MU-MIMO, they gotta say so in their HE PHY Capabilities. if they do full bandwidth, they set dot11HEFullBWULMUMIMOImplemented. if they also do partial, they set dot11HEPartialBWULMUMIMOImplemented. APs gotta check these flags before triggering anything.

UL MU-MIMO lets lots of devices talk back to the Wi-Fi router all at once. less waiting, more uploading, smoother Zoom calls for everyone.


Guard Interval

what’s a guard interval (GI)?

Guard Interval (aka TGUARD) is like a lil’ time gap between Wi-Fi signal chunks (OFDM symbols) to stop them from messin’ with each other. Used in OFDM, like in 802.11ax and pals.

why it’s there

Wi-Fi signals bounce all over the place (walls, metal, your cat), and bits can arrive late. That causes intersymbol interference (ISI) — overlapping symbols, big decoding mess. So, the fix? We slap a copy of the end of the signal at the front — called a cyclic prefix. It’s like givin’ the signal a running start.

symbol timing breakdown

One full OFDM symbol is made of:

  • TDFT: the useful data time
  • TGI: the guard time

So: TSYM = TDFT + TGI

guard times in HE PHY (802.11ax)

Depending on what part of the Wi-Fi frame we’re in, guard times change:

  • TGI,Pre-HE: 0.8 µs (used in pre-HE stuff like RL-SIG, HE-SIG-A)
  • TGI,L-LTF: 1.6 µs (for L-LTF field)
  • TGI1,Data: 0.8 µs — normal GI for data
  • TGI2,Data: 1.6 µs — double GI, extra safe
  • TGI4,Data: 3.2 µs — four times GI, super safe but slower

And yeah, data and HE-LTF fields use the same guard interval value.

don’t mix it up with guard tones

Heads up — guard interval is time-based. Guard tones are frequency-based — empty subcarriers on the sides to keep Wi-Fi neat and clean. Different purpose, different world.


Spartial Streams

spatial streams? what’s that?

So yeah, when your Wi-Fi talks with multiple antennas, it don’t just shout louder — it splits up the data into chunks called spatial streams. Each stream rides a different antenna path. More streams = more speed (theoretically).

DL MU-MIMO vibes

Downlink MU-MIMO (multi-user, multiple antennas, crazy fast) means the AP (the Wi-Fi box) can beam multiple streams to multiple devices at the same time. Everyone gets their own stream or two. Nobody’s waiting in line anymore.

how many streams we talkin?

In the latest Wi-Fi 6 stuff (HE PHY aka 802.11ax), the AP can toss out up to 8 spatial streams total — max 4 per user, across up to 8 devices sharing the same RU (resource unit).

beamforming magic

To make sure streams don’t collide mid-air, APs use beamforming — like Jedi mind-tricks but with math. Devices send back channel info, AP shapes the stream just right. This includes:

  • steering matrix (Qk)
  • channel feedback (Vk)
  • cyclic shifts to prevent accidental stream overlaps

devices gotta declare their stream game

Not every Wi-Fi gadget can handle lots of streams. They all declare their max in special fields:

  • HT devices: ‘Tx Max Spatial Streams’ field (0=1 stream, 3=4 streams)
  • VHT/S1G devices: look at ‘Rx NSS’ field (1 to 8 streams)
  • HE devices: signal it with ‘Operating Mode’ and ‘HE-MCS & NSS Set’

Note: APs that can do 4+ streams must support full-on DL MU-MIMO. Non-APs must be able to receive it.

trigger frames and stream assignments

In coordinated sessions (like OFDMA), the AP sends a trigger frame that says: “yo device X, you get 2 streams starting at position 3.” That’s SS Allocation.

extra fun stuff

  • Midambles: help when stuff is moving — keep signal estimation fresh (up to 4 streams)
  • STBC: for extra reliability, used with just 1 stream (if no DCM)
  • Spatial stream underutilization: basically when your AP has 8 streams but your device can only take 2. The rest just sit there, bored.

smokin’ heads and ofdm waves

when too many bits ride the air, heads fry and ofdm’s still chillin’ like math jazz in the sky.


If you got the feeling now like “Yeah, I totally get Wi-Fi!” – then uh... nope. That was just OFDM Stuff, my friend. There’s a whole bookshelf waitin’ for you if you wanna be a real Wi-Fi geek. CWNA, CWAP, CWDP, CWSP, CWISA... yeah, it’s a ride. Buckle up.

Just a quick FYI:
This article’s got no tables or fancy graphics – on purpose. It’s built that way so screen readers and text-to-speech tools don’t freak out. Keepin’ it clean for the accessibility crew.

Heads up, Wi-Fi nerds:
This whole guide was put together using the CWNP books CWNA, CWAP, CWSP, CWDP, CWISA and the IEEE Std 802.11-2020 All the deep-dive stuff about OFDM, 802.11 weirdness, and packet wrangling comes straight outta those.