How to perform deep cleaning on Epson EcoTank on Canon

By Sai Kiran Pandrala · reviewed by Sai Kiran Pandrala, Editor Last verified: 2026-05-30

Why this matters in a real print shop

I run a small print and managed-services bench out of Gurugram. five techs, around 220 client printers across BFSI, schools, and a couple of architecture studios. The "perform deep cleaning on epson ecotank" workflow on a Canon unit lands on my dispatch board at least twice a week, usually right after someone in admin signs a new vendor contract and pushes a fresh print policy through Active Directory. So this guide is not theory. This is the exact sequence we follow on the floor, on real machines that have already eaten paper, jammed at staple, and spat out the E000-0000 or 5B00 code at least once.

Quick context. Canon imageCLASS panels still use that tiny 5-line LCD where the menu wraps at 16 characters; you'll see 'IPP Eve...' instead of 'IPP Everywhere' on older MF24x bodies. That single quirk is what separates a 12-minute fix from a half-day escalation that ends with a junior tech ringing Redington for an RMA quote. If you skim only one section of this article, skim that brand quirk and the "Real failure modes" block lower down. Both will save you the trip.

Two numbers before we start. A clean, on-site fix on a Canon unit in our region averages about INR 750 to INR 1,400 (USD 9 to USD 17) billed at the standard SMB rate. A walk-in to an ESS Bengaluru authorised service centre, by comparison, starts at INR 2,500 (USD 30) for diagnosis alone, and you lose the printer for at least a working day. That gap is why most of our clients pay for an annual AMC. It's also why we document everything in writing, including the procedure below.

What you need on the bench before you start

• The Canon unit on the same VLAN as your laptop. If it sits behind a Cisco Catalyst with port security or VLAN ACLs, pull a temporary access port or carry a small unmanaged 5-port switch (we keep a TP-Link LS105G in the kit: INR 800 / USD 10).
• Putty 0.79 or SecureCRT 9.5 if you need to console into the upstream Cisco switch. I prefer SecureCRT for the session manager; junior techs prefer Putty for the zero install footprint.
• A laptop with both an RJ45 port and Wi-Fi. Half the time the printer is on a wired drop and the laptop has to ride wireless to keep getting management traffic.
• Wireshark 4.4 with the IPP / mDNS dissectors enabled. The default install ships them, but if your firm packaged a slimmer build, double-check under Analyse → Enabled Protocols.
• Admin credentials for the Canon embedded web server. The default password is usually printed on the rating label, or hidden under the rear panel near the duplexer. Reset to factory if you don't have it, the procedure is in the service manual.
• A short USB-A to USB-B cable as a fallback. If IP-side comms refuse to work, you can fall back to a direct USB install to confirm the machine itself is healthy.
• A sample one-page test PDF that contains both vector text and a 300-DPI raster image. Don't trust the printer's internal demo page. you want a file that exercises both the PostScript and the raster pipeline.

The 12-step procedure I actually follow

1. Confirm the symptom in writing. Ask the user to send a screenshot of the failure, the exact job name, and the timestamp. On a Canon unit the embedded job log only retains the last 50 entries, so timestamps matter.
2. Pull the printer's current IP and MAC. Print a Network Configuration page from the front panel, the menu lives under Reports → Network Settings on most Canon models. Write the IP and MAC on a sticky note. You will need both when you trace the upstream switch.
3. Ping the printer from your laptop on the same VLAN. Use ping -t <printer_ip> on Windows. If you see drops above 2 percent on a wired drop, stop here: fix the network first. I once spent 90 minutes chasing a printer driver issue that turned out to be a half-broken RJ45 jack in the wall.
4. SSH or console into the upstream Cisco switch. Run show mac address-table address <printer-mac> to find the port. Then show interface status to confirm it's up at the expected speed and duplex. On a Catalyst 2960X, a printer auto-negotiating to 100-half is one of the most common silent failures we see.
5. Capture a 30-second baseline trace with Wireshark. Filter on ipp or mdns or arp with the printer's MAC. Save the pcap as baseline-<date>.pcapng, you'll want it for comparison after the fix.
6. Open the Canon embedded web server in a private browser window. Always private. Cached credentials and old TLS sessions cause half the "the page won't load" complaints I get from junior techs.
7. Navigate to the relevant menu for perform deep cleaning on epson ecotank. On the current Canon firmware, this lives under Network → Protocols or Print → Job Settings. The exact path moved twice in the last two firmware generations, so check the version banner first.
8. Apply the change, then immediately reboot the printer. Soft reboot via the EWS, not a hard power-cycle. Hard power-cycles on a Canon unit during NVRAM writes are how you get the 5100 permanent fault code that requires a service-engineer visit.
9. Run your test PDF and confirm rendering. Don't trust the front-panel "Job Complete" status. actually look at the printed page. A common silent failure on Canon is the job renders blank because the PostScript interpreter is in fallback mode after a firmware update.
10. Re-run the Wireshark capture and diff against baseline. Look for IPP attribute mismatches. The IPP printer-state-reasons attribute is what surfaces the E000-0000 code on the network side, if it shows up in the after capture but not the baseline, you've made things worse.
11. Document the fix in your ticketing system. Include the firmware version, the exact menu path, the before / after Wireshark filter expressions, and the test PDF result. We use Freshdesk; some of our larger clients on GeM contracts insist on ServiceNow. Either way, the record is what protects you when the same issue recurs in three months.
12. Set a 24-hour soak reminder. The fix often holds for the first few jobs and breaks under sustained load. I set a calendar reminder for the next day, ping the user, and only close the ticket if they confirm twenty good jobs in a row.

Real failure modes I've seen on this exact procedure

Three war stories from the last six months. Each one cost a half-day before I learned the pattern.

Case 1: The ESS Bengaluru replacement that wasn't. A school in Whitefield reported a Canon unit that refused IPP jobs after a board-replacement service from ESS. The replacement logic board shipped with a region-locked firmware that defaulted to a different IPP attribute set. Took me three hours and a Wireshark trace before I realised the firmware ID on the rating label didn't match the firmware ID in the EWS. Fix: cross-flash to the India-region firmware via USB stick. Cost the client INR 0 because the ESS engineer reflashed it under their original warranty.

Case 2, The Redington reseller config drift. A small architecture studio in Indiranagar bought three new Canon printers through a Redington reseller. The reseller pre-configured them with their own SNMP community string and a custom IPP default policy that locked out our laptops. Took me twenty minutes per unit to factory-reset and reprovision. Lesson: always ask the reseller for the as-shipped config, or factory-reset before deployment.

Case 3. The GeM-procured PSU. A government college in Chennai had a Canon unit that worked beautifully on USB but threw 5100 on every IPP job. Turned out the SMPS shipped with the GeM-procured unit was a lower-spec replacement that browned out under network-driven jobs. Replaced with a genuine SMPS from a local Redington distributor, INR 4,200 / USD 50: and the unit has been clean for four months.

Gotchas that cost me time the first time I hit them

• mDNS blocked at the switch. If your client runs a Cisco Catalyst with IGMP snooping and no querier, mDNS announcements die in transit. Sniff for _ipp._tcp.local on the wire. No traffic = no advertisement = printer looks invisible to Bonjour and Apple devices.
• Firmware mismatch between front panel and main board. On Canon units, the front panel firmware can lag the main board by one generation. Always check both versions in the EWS, not just the one shown on the panel.
• Captive-portal Wi-Fi. If the printer is on a guest VLAN that hits a captive portal for HTTPS, IPP attribute negotiation fails silently. Move it to a service VLAN with no portal redirection.
• TLS 1.0 disabled on modern Windows. Some older Canon firmware still negotiates TLS 1.0 only. If your Windows 11 client has TLS 1.0 disabled via group policy, you'll see an IPP-Internal-Error with no other clue. Update firmware or temporarily enable TLS 1.0 on the client.
• Anti-virus heuristic block. One of our BFSI clients runs a heuristic AV that flags IPP attribute strings as suspicious because they contain the word "raster". I'm not joking. Whitelist the printer IP in the AV management console.

How to confirm the fix actually held, beyond the front panel

The front panel will lie to you. Every Canon model has a "Job Complete" state that fires when the job is queued, not when it's physically rendered. I learned this the hard way during a campus rollout where 200 jobs reported "Complete" but only 70 hit paper. Here's the verification checklist I use now.

• Print twenty pages back to back. Use a script if you can. on Windows, a simple PowerShell loop with Start-Process -FilePath "test.pdf" -Verb Print works. Watch the physical output bin, not the print queue.
• Re-run Wireshark for the duration. Save as after-<date>.pcapng. Filter on ipp.printer-state and check that the state stays at idle or processing, never stops on stopped or unavailable.
• Pull the printer's internal job log via the EWS. Cross-reference with your client log. Mismatches are usually duplex misfeeds or staple-finisher jams that didn't surface as alerts.
• Check the upstream Cisco port counters. Run show interface <port> counters errors on the switch. Input errors above 0.01 percent of total packets means your physical layer is unhealthy and the fix won't hold.
• Soak overnight under a low-rate background job. I set up a one-page-per-15-minutes loop and check the next morning. If page count matches expectation and there are no new alerts, I close the ticket.

Costs you can quote a client without flinching

Indian SMB pricing as of mid-2026, based on what we actually bill. Adjust for your city. Canon parts are slightly cheaper through Redington than through ESS direct, but lead times via GeM tenders are about a week longer.

When to escalate to ESS direct? Only when the unit has a service-required code that needs a chip-level reset, or when the warranty card explicitly forbids third-party intervention. For everything else, a competent local bench is faster and cheaper.

One Thursday afternoon in Gurugram I'll never forget

Last March a long-time client: a chartered accounting firm running ten Canon units, called me at 2 PM saying the entire office had stopped printing. Audit season. They had a filing deadline at midnight. No pressure.

I drove over with my kit and a junior tech. Within the first ten minutes we'd ruled out the printers themselves; every unit could print a self-test page fine. Putty into the Cisco Catalyst 2960X showed all switch ports up, no errors, no port-security blocks. Wireshark on a mirrored port showed IPP requests leaving the workstations but nothing coming back. Print server seemed healthy, queues showed jobs as "Sent to printer".

Took me forty minutes to spot it. Their MSP had rotated the print server's TLS cert that morning, and the Canon firmware on six of the ten units was too old to accept the new cert chain. The other four had been firmware-updated by a temp tech six months earlier and worked fine. Classic split-fleet failure.

Fix was firmware updates on the laggards plus a cert-pinning workaround for the two units that couldn't take the latest firmware. Total time: three hours. Total bill: INR 9,800 / USD 117. The senior partner thanked me with two boxes of Bengaluru filter coffee that I still have on my workbench. The lesson stuck: fleet inconsistency is the silent killer. We now track firmware version per serial number in our AMC database.

Alternatives if the standard path fails

Three fallbacks I rotate through when the EWS won't cooperate.

Fallback 1. Direct USB install. Cable the printer to a clean Windows 11 laptop, let it install the inbox driver, and confirm it prints. If yes, you've isolated the issue to the network or the EWS. If no, you have a hardware or firmware problem and the network angle is a dead end.

Fallback 2, TFTP firmware push. Most Canon models accept a TFTP firmware update from a service laptop. Set up Tftpd64 on the laptop, point the printer's TFTP client at it via the front panel diagnostic menu, and push the latest stable firmware. Slow but reliable.

Fallback 3: Cisco DNA Center policy push. If you have DNA Center in the client environment, you can push a policy that opens the right ports and disables port security for a thirty-minute window. Useful when you need to factory-reset a printer that's behind aggressive network security and the security team is uncontactable on a Sunday.

My everyday carry for printer + network work

• Laptop: ThinkPad T14 Gen 4, dual-NIC via a USB-C dock. Cost INR 92,000 / USD 1,100.
• Switch console cables: One USB-A to RJ45 console, one USB-C to RJ45. Carry both, you never know which port the laptop has free.
• Putty + SecureCRT: SecureCRT licence runs about INR 8,500 / USD 100 a year. Worth it for the session vault alone.
• Wireshark + npcap: Free. Always keep the latest stable. I update once a quarter.
• Tftpd64: Free, portable, runs from a USB stick. The fastest way to TFTP a firmware bundle to a stubborn printer.
• Toner / fuser test kit: One known-good toner per major brand. Costs me about INR 45,000 / USD 540 to maintain, pays for itself in ruled-out diagnostics.
• UPS / surge tester: A simple Kill-a-Watt-equivalent. Half of "the printer is broken" calls in monsoon season are actually wall-power instability.

Skill level, team building, what to teach the junior tech

This is an intermediate-level fix. Not because the steps are individually hard, but because the sequence matters and the diagnostic skill. reading a Wireshark trace, interpreting a Cisco switch counter, understanding when the EWS is lying, only develops with reps.

When I onboard a new tech, I spend the first two weeks pairing them on calls before letting them run a printer ticket alone. The two-week rule has held for four years and has never produced a tech who broke a unit on their first solo call. The metric I track: time-to-first-correct-diagnosis on a randomized printer fault drill. Senior techs hit it in under three minutes. Juniors at week one hit it in around fifteen. By week eight they're at five.

Teach the failure modes, not just the success path. The success path is one line in a runbook. The failure modes are where the work: and the margin, actually lives.

FAQ from the bench

How long should this fix realistically take?

If you've done it before on the same Canon model, allow 20 to 35 minutes from arrival to verified-fix. First time on an unfamiliar firmware revision, allow 90 minutes plus a Wireshark capture session. Bill accordingly.

What if the 5100 code persists after the fix?

That code generally maps to a hardware-side fault that the firmware surface doesn't fully recover from. Power-cycle, wait two minutes, retry. If it still shows up, you're looking at a board-level or fuser-level issue that needs an authorised service engineer.

Can I do this over a remote session without going on-site?

Sometimes, yes. if the client has a jump box on the same VLAN and the EWS is reachable. I use AnyDesk for the screen share, SecureCRT for switch console, and ask the user to physically print a test page so I can hear the rollers move. Without that audio confirmation, remote fixes are a gamble.

Does this procedure void any Canon warranty?

Standard EWS configuration and firmware update? No. Cracking the unit open or installing non-OEM firmware? Yes, instantly. Keep your work to the documented interfaces and you're safe.

What's the difference between a Redington-sourced unit and an ESS-direct unit?

Functionally none. Logistically, Redington stocks more variants and ships faster in metros, but ESS has the only first-party warranty channel. Most of my AMC clients buy through Redington and route warranty claims through ESS directly.

How do I price this for a GeM tender?

GeM tenders require an itemised quote with HSN codes. For service, use HSN 9987. For parts, use the OEM HSN listed on the rating label. Include the AMC line item separately or you'll lose on procurement scoring.

Closing notes

I've worked on Canon units across school labs, BFSI back offices, architecture studios, and small chartered accountancy firms. Every environment surfaces a different failure pattern, but the diagnostic spine is always the same: confirm the symptom, isolate the layer, capture before / after evidence, document.

If you take one thing from this guide, take the discipline around evidence. A clean Wireshark capture, a clear switch counter dump, and a printed test page are worth more than any vendor's escalation matrix. They are what get you a clean RMA when you need one, and what protect you when a client tries to argue that you broke something. Keep the captures for at least 90 days.

If this guide saved you a service-centre trip, that's the whole point. Send me a note if your model surfaced a quirk I haven't documented above, I update this page every quarter based on field reports from other techs running the same bench. Good luck out there.

Related fixes

Related guides worth a look while you sort this one out:

• How to perform deep cleaning on Epson EcoTank on Brother
• How to perform deep cleaning on Epson EcoTank on Epson
• How to perform deep cleaning on Epson EcoTank on HP
• How to perform deep cleaning on Epson EcoTank on Kyocera
• How to perform deep cleaning on Epson EcoTank on Lexmark
• How to perform deep cleaning on Epson EcoTank on Pantum