Fax-Native SIP Trunking
The SIP Trunk That Actually Supports Fax
Most SIP trunks treat fax as an afterthought. T38Fax built its entire network around it. In-house T.38 stack. ECM on every call. Dialogic-certified.
Why Fax Breaks the Moment You Move to SIP
You made the switch. Everything worked — until someone tried to send a fax.
Pages failed mid-transmission. The receiving machine picked up but nothing came through. Or fax calls dropped after the T.38 re-INVITE. Your SIP provider told you to lower the baud rate, disable ECM, and try again. Sometimes it worked. Then it didn’t.
Here’s what’s actually happening: your SIP provider doesn’t run T.38 infrastructure. They run voice infrastructure. When your fax equipment initiates a T.38 call, your provider’s system passes it upstream — to a carrier they don’t control, whose T.38 implementation is inconsistent, whose ECM configuration you’ll never see, and whose routing changes call to call.
Your fax problem isn’t your ATA. It isn’t your fax server. It isn’t the T.38 protocol. It’s the carrier.
Most SIP trunk providers weren’t built for fax. They were built for voice, and they added T.38 pass-through because customers asked for it. Pass-through means your fax traffic terminates on someone else’s network — an upstream provider your SIP trunk vendor doesn’t control, can’t modify, and can’t diagnose when it breaks.
T38Fax was built the other way. We started with fax.
What a SIP Trunk Needs to Support Fax
Not all T.38 support is equal. Three things determine whether fax over SIP actually works — and most providers fail at least one.
An in-house T.38 stack, not pass-through. The difference between a carrier that implements T.38 and one that passes it through is the difference between accountability and a shrug. Carriers that own their T.38 implementation can diagnose failures — they can see every step of the protocol. Carriers that pass upstream can’t. Ask any potential provider: where does your T.38 conversion happen? If the answer involves “relay,” “upstream carrier,” or “our network partner,” you’re looking at pass-through.
ECM enabled by default. Error Correction Mode divides each fax page into frames, checks each frame for transmission errors, and retransmits any frame that arrives corrupted. Without ECM, a page can arrive with missing or garbled sections and neither side knows — the call completes, a corrupted document lands in the tray, and there’s no indication anything went wrong. Most SIP carriers disable ECM by default to reduce processing load. Check whether ECM is on or off by default before signing up. “ECM is optional” means “we disabled it.”
The ability to diagnose a failure. Ask any prospective carrier: if a fax fails, can you tell me exactly why it failed — which T.30 phase, which ECM frame? If the answer is “try lowering your baud rate and turning off ECM,” they cannot diagnose the failure. That’s a guess. A carrier that understands T.38 at the protocol level can read a trace and give a specific answer.
T38Fax meets all three. We built our T.38 stack in partnership with Dialogic — the company behind the SR140, the fax processing engine embedded in OpenText RightFax, HylaFAX Enterprise, GFI FaxMaker, and XMedius. ECM runs on every call, no exceptions. And when something fails, we diagnose with T.30 traces, not trial and error.
Built for Fax. Not for Voice with Fax Bolted On.
Consistent T.38
We built our own T.38 stack and place our media gateways inside carrier networks. Your fax equipment talks to our gateways on every call — not an upstream provider we don't control. No variable routing. No mystery failures. The same call sent twice takes the same path and gets the same handling.
ECM Error Correction
Most carriers disable ECM to save processing power. We insist on it. Without ECM, fax pages can arrive corrupted with neither side knowing. The word "facsimile" means exact copy — and ECM is how you deliver one. It's on by default, on every call, with no option to disable it.
Security & Compliance
SOC 2 certified. HIPAA compliant. Optional IPSec VPN tunnels at no additional charge. We never store fax content — T.38 is real-time, so there's nothing to retain. When your compliance team asks about fax infrastructure, you'll have answers that hold up.
Premium Support
Fax industry engineers who diagnose with T.30 analysis, not superstition. We understand ECM partial-page retransmission, UDPTL redundancy settings, and why your fax failed at page 47 of 200. Unlimited technical support is included — not a paid tier.
Works With Your Existing Equipment
The most common question isn’t whether T38Fax works — it’s whether T38Fax works with what you already have. The answer is almost always yes.
Power-T.38 is a SIP-native service. If your platform supports standard SIP trunking, connecting to T38Fax is straightforward. We’ve tested and documented configurations for the most widely deployed fax servers, softswitches, and ATA hardware.
Fax Server Software
- OpenText RightFax
- HylaFAX Enterprise
- GFI FaxMaker
- XMedius
- Biscom
- Softlinx ReplixFax
- Esker VSIFax
Softswitches & IP-PBXs
- Asterisk / FreePBX
- FreeSWITCH / FusionPBX
- 3CX
- Cisco CUCM
- Cisco BroadWorks
- Avaya
- NetSapiens
- Metaswitch
- Grandstream UCM Series
ATAs & Gateways
- Grandstream HT801, HT802, HT812, HT814, HT818
- Grandstream GXW4004, GXW4008
- Cisco ATA 191-MPP, ATA 192-MPP
- AudioCodes MediaPack 1xx
- Obihai OBi200, OBi202, OBi302
- Patton SmartNode
The full certified device list — with tested configurations and step-by-step setup guides — is in our Knowledge Base. If your device or platform isn’t listed, bring it to the 30-day free trial. If it supports T.38, it should work — and our support team will get there with you.
Running RightFax, HylaFAX Enterprise, GFI FaxMaker, or XMedius? Power-T.38 was engineered around the Dialogic SR140 — the same T.38 engine your fax server uses. Deep interoperability, not coincidental compatibility.
Enterprise Fax Servers →Want the full technical picture? How our T.38 stack is architected, what ECM actually does at the frame level, and why gateway placement inside carrier networks matters more than anything else.
Our Solution: Power-T.38 →Frequently Asked Questions
- Yes — if the SIP trunk is built for it. The T.38 protocol exists specifically to make fax over IP reliable, but it requires a carrier that has implemented it correctly: in-house T.38 stack, ECM enabled by default, and gateways placed inside carrier networks rather than routing through upstream providers. Most SIP trunk providers treat fax as a secondary concern and offer T.38 pass-through rather than a native T.38 implementation. T38Fax is a dedicated fax carrier — our entire network was built around T.38, and fax isn’t optional or bolt-on. It’s the only thing we do.
- The root cause is almost always the carrier, not your fax equipment or fax server. Most SIP providers route T.38 traffic upstream to a carrier they don’t control — their T.38 ‘support’ is really pass-through. That upstream carrier may have inconsistent T.38 handling, ECM disabled to reduce processing load, or variable routing that changes call to call. ECM (Error Correction Mode) is particularly critical: without it, a page can arrive corrupted and neither sender nor receiver knows. The result is fax that works sometimes, fails mysteriously other times, and produces no useful diagnostic when it breaks — because nobody in the chain actually owns the stack.
- T.38 is the ITU standard protocol designed specifically for fax over IP. It converts fax tones into UDPTL — a dedicated data protocol with built-in packet redundancy that makes it resilient to the packet loss and jitter normal on IP networks. G.711 fax over SIP encodes fax tones as audio and sends them as voice packets; it works on short, clean network paths but degrades quickly under real-world conditions. When a carrier claims T.38 support but delivers G.711 passthrough, you’re getting the less reliable option without being told. T38Fax runs native T.38 on every fax call — G.711 passthrough is not part of our network.
- Three things matter most. First, whether the provider runs their own T.38 implementation or resells carrier pass-through — if they can’t tell you how their T.38 stack is architected, it’s pass-through. Second, whether ECM is enabled by default: a provider that disables ECM is choosing lower processing cost over document accuracy. Third, whether they can actually diagnose a failure — if the answer to a failed fax call is ’lower your baud rate and turn off ECM,’ they don’t understand the protocol. T38Fax built its T.38 stack in partnership with Dialogic, enables ECM on every call, and diagnoses failures with T.30 protocol traces. We can tell you exactly why a specific call failed.
- Ask your provider two questions: where does your T.38 conversion happen, and what upstream carrier handles the T.38 leg? If the answer involves words like ‘relay,’ ‘upstream carrier,’ or ‘our network partner,’ you’re looking at pass-through — your fax quality depends entirely on a carrier your provider doesn’t control. You can also run an ECM test: T38Fax offers a free ECM self-test at foip.t38fax.com/ecm-self-test that tells you in minutes whether your current SIP trunk properly supports T.38 with ECM enabled. Most customers discover their current provider fails the test.
- You can use T38Fax for fax while keeping your existing voice SIP trunk for voice calls — that’s the most common configuration. T38Fax is a fax-specialized carrier, not a general-purpose voice provider. Most customers point their fax equipment directly at our network while their IP-PBX continues using their existing voice SIP provider. Some customers run all traffic through T38Fax, and that works too. Our setup guides cover both configurations for Asterisk/FreePBX, 3CX, FreeSWITCH/FusionPBX, Cisco, and other common platforms.