RE: wifi vs ethernet speeds
wow, any slower and I'd be thinking I was on 56k dialup where 6min for 1MB was great! haha Oh wait now I'm showing my age!
Yes, in 2023, I'm kind of surprised that was considered good enough to ship....
I wish I could get that speed, I just transferred a 8.6Mb gcode to the printer and it took 16 Minutes! so about 2 Mins per Mbyte.
With the included usb stick was 4x slower than that.
This is with the hard wired ethernet not wifi.
RE: wifi vs ethernet speeds
oh that's good to know. I hope they've scale tested that prusaconnect service... my sincere hope is that they've used gcp/aws/something to offer regional points of access as opposed to everyone's printers having to talk back directly to Prague. I'm glad it can be configured to be offline until they figure things out.
It seems to be something to do with PrusaConnect, I have now got PrusaLink working offline and uploading prints to the printer is reasonably fast i.e. a few seconds rather than minutes.
RE: wifi vs ethernet speeds
This right here! Use hard-wired ethernet if at all possible. you don't want your wifi-router getting rebooted by your isp for firmware update while you've got a multi-day print going...
Prusa printers are not affected by that. They never print over LAN or over cloud. g-code file is fully uploaded to the printer (over LAN or over cloud) and only then print starts.
you don't want your wifi-router getting rebooted by your isp for firmware update while you've got a multi-day print going...
No it's much worse than that.
Wifi, in towns, is becoming unuseable.
People experiencing any sort of wireless issue generally blame the wifi router and demand better, more powerful wifi hotspots - which are invariably worse.
My town's chamber of commerce has installed free wifi throughout the town centre. One of the council members asked me to test it as there had been very few users and several complaints. So I took a diagnostic kit and logged in - or rather, tried to.
They had installed several hotspots and boosted the power to increase coverage - big mistake. My unit could 'see' several competing access points and after I had tried several until I sucessfully logged into one it naturally stayed locked to it - when I walked down the street it suddenly stopped working but remained locked with a strong signal. I was now far enough away that any incoming signal could not be replied to in the same phase at the carrier frequency, I was blocked by physics and the speed of light.
I was in a South London coffee shop a little while ago which advertised free wifi - so I tried logging in and was astonished how slow it ran. Then I scanned and found over 180 access points within range.
Several years ago I gave presentation at a games convention in a small Yorksire town, Hebdon Bridge. I had already checked the hall wifi and determined that it could handle about 20 connections max so relied instead on the attendees mobile 'phones - I gave a link to my games server for a simultaneous demonstration and ... it failed. My server logs showed barely a blip but in the pub later locals were moaning about how their 'phones all lost connection earlier. About 80 - 120 people had tried to use the accessible masts at the same time (my fault) and crashed the local cells.
Wifi and related wireless are polite networking.
It doesn't matter how powerful, within reason, the signal is.
A station, base or mobile, wishing to send a packet first listens for other traffic on the same channel, if it hears some it backs off for a random period, initially just a few microseconds and tries again. After each successive failure it is supposed to back off for a longer time (not all do) and so on until the channel is free. The station it is talking to does the same when replying - but as it is in a different place it is within range of different, overlapping set of stations. The more powerful the station the greater radius it covers and the more competitors it sees.
More powerful wifi does make some sense in the countryside where there are no competing stations and you want to log in from the garden but in towns you want the least powerful transmission that *just* covers your desired area and so 'sees' fewer competitors.
Each time you add a wifi node you slow all the traffic within range by a small amount.
And then you attach the wifi aerial to a chunk of earthed metal with moving parts. Each time it moves the reflection pattern changes; possibly altering the phase, and the deadspots shift.
Whenever possible install a cable, your wire has no competitors you do not control. It will always be faster and more reliable.
Cheerio,
RE: wifi vs ethernet speeds
I guess I'm older because I had a 300 baud, and this MK4 is bringing back memories. I've had mine only a few days and have been pleased overall, but this transfer rate is embarrassing. My RAISE3D E3 is ancient by comparison and transfers files in seconds, and my Bambu printers leave this thing in the dust.
wow, any slower and I'd be thinking I was on 56k dialup where 6min for 1MB was great! haha Oh wait now I'm showing my age!
Yes, in 2023, I'm kind of surprised that was considered good enough to ship....
I wish I could get that speed, I just transferred a 8.6Mb gcode to the printer and it took 16 Minutes! so about 2 Mins per Mbyte.
With the included usb stick was 4x slower than that.
This is with the hard wired ethernet not wifi.
RE: wifi vs ethernet speeds
yeah.... I wasn't going to admit to anything older than 56k... but my commodore 128 had a 300bps modem... back in the old bbs days haha life was simpler then.
I guess I'm older because I had a 300 baud, and this MK4 is bringing back memories. I've had mine only a few days and have been pleased overall, but this transfer rate is embarrassing. My RAISE3D E3 is ancient by comparison and transfers files in seconds, and my Bambu printers leave this thing in the dust.
wow, any slower and I'd be thinking I was on 56k dialup where 6min for 1MB was great! haha Oh wait now I'm showing my age!
Yes, in 2023, I'm kind of surprised that was considered good enough to ship....
I wish I could get that speed, I just transferred a 8.6Mb gcode to the printer and it took 16 Minutes! so about 2 Mins per Mbyte.
With the included usb stick was 4x slower than that.
This is with the hard wired ethernet not wifi.
RE: wifi vs ethernet speeds
the day i go my printer (last week) i spent 2 hours with chat support trying to send a file from Prusa Slicer to the MK4 via prusalink / prusaconnect
Turns out the issue is extremely slow wifi and no status update on the Connect side besides telling me file has not been uploaded (we tried debugging why to finally realize it because it had not finished downloading yet... when it finally did (1h later) the file actually did print so my setting and api key etc where all good.
On the attach screenshot this is all you see on Connect -- no progress bar showing that it is in fact uploading (at snail speed) but knowing that would have been nice -- only a pink error screen showing file not uploaded.
This was brought up to the developers from whom I've not heard back from yet. Support told me that is something we should see fixed in the a firmware update.
I gave up trying to use the built in wifi and this connect software -- i simply plugged in my Raspberry Pi 3 i had working for my MK3 (USB A on the Pi to USB C port on the printer), changed a few setting like the name and bed size in Octoprint and after 10 min i was back using my great workflow of Octoprint / Octo everywhere / and Polymer on my iPhone oh and uploading same file that way via wifi from PrisaSlicer takes 5 sec not 1 hour. go figure!
Only issue with Octoprint and MK4 is that the MK4 unlike my mini, does not switch the the green Octopi screen where you can for example speed the print. I've red in other thread that basically any feature that alters the gcode is not currently possible from Octoprint on the MK4 (another thing i hope the next release of firmware will resolve)
Cheer from Montreal, Canada
RE: wifi vs ethernet speeds
So where is the file saved upon transfer - do you need to have a USB drive inserted so the file can be saved there? In essence its a transfer from computer to USB drive connected to printer?
This right here! Use hard-wired ethernet if at all possible. you don't want your wifi-router getting rebooted by your isp for firmware update while you've got a multi-day print going...
Prusa printers are not affected by that. They never print over LAN or over cloud. g-code file is fully uploaded to the printer (over LAN or over cloud) and only then print starts.
--> MK4 - MK4S - MINI+ - Accelerometer Guide <--
RE:
I assume for the Mk4 it's the same as the XL and Mini, so yes, Prusalink requires a USB stick to be inserted.
Formerly known on this forum as @fuchsr -- until all hell broke loose with the forum software...
RE: wifi vs ethernet speeds
Unfortunately the ESP8266 even if it is fast if used on its internal memory, uses serial port to communicate (what it receives or must send on WiFi) with the ST Mini or MK4 board, and the reliable serial speed allows a network speed around ~100KB/s MAX.
I have a look at the schematic of the Mini and it has several STM dedicated line connected directly to the Ethernet chip, and it doesn't use any serial pin internally; so it must be fast.
RE: wifi vs ethernet speeds
Could someone retest wifi and wired ethernet with new firmware?
https://github.com/prusa3d/Prusa-Firmware-Buddy/releases/tag/v4.6.1
"Developers have optimized file transfers between Prusa Connect and the printer resulting in higher transfer speed."
RE: wifi vs ethernet speeds
So here is Wifi off of my AC6-Pro
10.43mb file sent in 74 seconds
140.94k/sec
Channel1 (2.4 GHz, 20 MHz)
Signal-40 dBm
AP/Client Signal Balance Good
Standard WiFi 4
MIMO Configuration 1x1
Rx Rate 43.3 Mbps
Tx Rate 58.5 Mbps
Even though 43Mbps is slow as shit, its like 5mb/sec, but if we are maxing out the serial on an ESP at ~150k then hello octoprint. What were they thinking?
Could someone retest wifi and wired ethernet with new firmware?
https://github.com/prusa3d/Prusa-Firmware-Buddy/releases/tag/v4.6.1
"Developers have optimized file transfers between Prusa Connect and the printer resulting in higher transfer speed."
RE: wifi vs ethernet speeds
"10.43mb file sent in 74 seconds" So the 40+meg file I just uploaded to by MK3/OctoPrint in a few seconds would take 5 minutes, WOW!! In my eyes that makes the new "Ethernet and Wi-fi connection" and most likely the "Prusa Connect" useless.
With that news I am rethinking my Mk4 order!
If I do keep it, I will have to plan on putting OctoPrint on the MK4 as well and just disable Prusa's WiFi.
RE: wifi vs ethernet speeds
If you already have a Pi with Octoprint and can't get another one for a reasonable price now you can always put another Octoprint instance on the first Pi. Chris Riley has several how-tos for this.
RE: wifi vs ethernet speeds
The hardware choice and design is 100% dumb.
Unfortunately, I read somewhere that they spent a bunch of software resources to fully rewrite the ESP firmware. That pretty much means they have invested so much ego into both the hardware and software design, that they will never admit it's a dumb design and change it.
RE: wifi vs ethernet speeds
Or the Orange Pi 5 is available fairly reasonable. I hear it runs Octoprint very well.
If you already have a Pi with Octoprint and can't get another one for a reasonable price now you can always put another Octoprint instance on the first Pi. Chris Riley has several how-tos for this.
RE: wifi vs ethernet speeds
Another option are second hand RPIs. Where I live, there’re plenty of 3B/3B+ in good condition at reasonable prices ( 40-50 euro). I got one for my Voron and probably I’ll get another for running Octoprint on my MK3S+.