Some people have swapped to a better quality 8 mm inductive prox sensor, but the more accurate ones are usually 12 mm in diameter. Heating the nozzle won’t help much because not much of that heat makes it to the prox sensor. I think parking the prox sensor close to the bed and heating the bed that then indirectly heats the prox sensor is about as good as it gets. Preheating the prox sensor (perhaps with the Start G code I posted above) reduces the amount that the sensor heats while it’s creating the bed mesh. If the prox sensor is the same temperature during the entire bed mesh, the amount the measurements drift with temperature changes is minimized.
It doesnt matter if your bed is warped. A lot of people go on and on about taco beds. This is what the probe fixes and compensates for. Just go mainline klipper, takes an hour if you don’t have any issues, and get the eddy probe. The only issue I had on the mainline upgrade was with Katapult and the solution for that is on the git now.
What is “Katapult” 3d printers vs. what is “Catapult” 3d printers. Please explain THANKING YOU
Many start print macro doing that (preheat bad + nozzle). I so a difference with both instead of just one … But I trust that generating also more ‘noises’ on the cables so from that I am using that, I have more failures during the bed meshing (extra MCU timeout). I hoping that with a better power converter + eddy usb, that will help (but that starting to doing big improvements). The cheapest option with just M12 probe is probably the most common fix.
There is an SV08 Eddy Upgrade Kit is the works, but as of now there is no release date.
Most likely it will be after the release of the SV08 Max.
Read the wiki… Step 7
It could be using the Eddy duo, which has both USB and CAN.
I installed Eddy via Mainline klipper. It is not the paradise as many sell.
I cannot say it is better. If your room temperature is constant. The effort doesn’t worth it.
I have an OCD and I manually adjust the z offset on the first three lines of purge.
The Eddy with USB has some issues, of your distribution has a bad shielded cable you will have connections issues and wlguess what, the printer stops. Eddysl connection issues are a common thing and some people fixed with a new cable, different cable routing, ferrite rings and good luck.
https://github.com/bigtreetech/Eddy/issues/30
I removed Eddy and went with factory probe. Left the cable there. I opened a ticket with Biqu team. They expect a video of all my setup to decide what to do. So far it doesn’t worth the effort.
Heat you bed for a while so that the temperature sensors are kind of stable before you start your print
A good bed mesh is very helpful but using an accurate bed mesh so Klipper can compensate for a 1 mm or more warp across the bed surface works in theory but isn’t a good solution in practice. If it worked well, there would be no need to do the quad gantry level because Klipper could just compensate for the bed that isn’t level while it’s compensating for the bed that isn’t flat.
One problem is Z axis backlash. Klipper may instruct the four Z axis stepper motors to move 0.8 mm across a large print but if there is 0.1 mm of backlash, that would still result in the low spots being 0.1 mm too low. Of course, if the bed only had 0.1 mm of warp across the print surface, 0.1 mm of backlash would still result in the low spots being 0.1 mm too low. Bed mesh compensation is only as good as the Z axis backlash, so bed mesh compensation is limited by backlash even on a relatively flat bed.
Another issue is accuracy. Many of us are making parts that need to fit other parts. Even if bed mesh compensation worked perfectly, the top of the part will be flat and level but the bottom of the part will not be flat and level. It’ll have the warped surface of the bed. Klipper would compensate for the warped first layer and compensate less for subsequent layers on the bottom of the part until that bed compensation is averaged away over several layers until the layers are flat and level for the upper portion of the part.
I must say, I have to agree with others here that this so-called Taco Bed ‘Problem’ (issue) has been blown up out of all proportion. Sadly, the metrics of YouTube encourages certain YouTubers to jump on the Taco Bed Bandwagon, as it’s a great way to get lots of views. Ultimately then, something that in reality is a bit of an annoyance, becomes a Big Issue!
I mean, if printer beds were completely flat, with no height variation of any degree, then there would be no need for Z-probing anyway! It would be a matter of Set It Once And Forget It. For years now I’ve used a Prusa Mk3s and before every print it performs a 9-point Z-probe matrix across the entire bed (it’s not even adaptive!) and it works. As far as I know it uses a PINDA probe, which I think is inductive.. or something! 
I can clearly see that my SV08 bed is not completely flat, I can see that from the height maps that are generated whenever it does its bed-probing… and that is exactly what bed probing is for! It probes (adaptively in the case of the SV08) and then uses that height map to compensate for the variations.
Now, I’ve learnt by now that I have to make sure the first layer is close enough/not too close to the bed, before every print. There’s clearly some inaccuracies with the probe, but I’d rather do that every time a print begins than have failed prints. What I do is:
I give every single print a 5mm skirt around the models on the bed ..around 15mm from the printed objects.. then, once the print has started, via the Tune menu, I slow the Speed down to around 50% and then manually adjust the Z-offset as it puts down the skirt on the bed.. I’ve even feel it with my finger, and manually adjust the Z-offset (via the same Tune menu) until it looks and feels right. Then, when I’m happy, I up the speed back to 100% and let the print carry on.
Yep, it’s a bit of a nuisance, but not half as much of a pain as multiple failed prints would be! I’ve had this printer 3 weeks now, and after having worked out this routine, I’ve honestly not had a single failed print.
Would I like it to probe more accurately so that I don’t have to do this every time? Absolutely!
Am I prepared to brick the whole machine in a bungled attempt to get it to do so?
No way!
I’ve studied all of the ‘Fixes’ in great detail, and it seems to me that they are a collection of great ways to introduce yet more issues.
I do think it’s a real shame that we’ve paid for something which clearly has a real issue, but, really, it’s really quite easy to live with, and now I’m over that hump in the road this really is a great printer!
Incidentally… the default profile (Orca Slicer) used some absolutely insane acceleration settings. Taming those made a huge difference to the quality of the finished prints AND kept the printer itself in just one place on my bench!
Best to all,
Tim
2 Likes
Agree. 
The Prusa MK3 doesn’t have 1.1 mm of warp across its bed. My SV08 does. Bed warping is easier to avoid on a smaller bed and it’s less of a problem on a smaller bed.
We shouldn’t need to manually tune every first layer to have a successful print, and manually tuning the Z height only corrects inaccuracies in Z offset. It can put you in the middle of the bed warping but it doesn’t compensate for the warping. It’s also not a viable option for production. It’s a nuisance for occasional prints but anyone trying to run production would either fix the bed or sell the printer to a hobbyist and buy a printer that can produce a reliable first layer without user calibration and prayer for each print. The technology is readily available and modern printers now produce reliable first layers. This isn’t 2015. 3D printing is all grown up now. We don’t use paper to manually level the bed and we shouldn’t need to manually adjust every first layer on every print.
Some SV08 printers had a little bed warping and some had a lot. If you have a bed that’s within 0.5 mm of being flat across the entire bed surface you probably won’t have much of a problem but it’s a real pain for those of us with more than 1 mm of bed warping.
I’ve had my SV08 for five months. I love it but the first layer issues are sabotaging its outstanding potential. I’ve been procrastinating, hoping that Sovol would step up and address this deficiency. It sounds like they will be selling us an Eddy probe upgrade and I’ll gladly buy that, but I’ll still need to do something to fix the warping bed. I removed the bed and took a look at shimming it to minimize warping when heated soon after I received the printer but didn’t finish that lengthy trial and error shimming process. I’ll probably do that rather than spending the same amount of time and $400 on an aftermarket bed.
I’m looking forward to the “perfect first layer” that Sovol advertised for the Zero when it arrives, hopefully soon!
But today’s task is replacing the MicroSwiss hot end that was destroyed by a baseball sized blob of death after I watched the first 20 layers, went to sleep and awoke to find the extruder cover hurled across the printer, a lump of molten plastic where the hot end should be, and the heater wires sparking… all because the SV08 can’t make a reliable first layer.
Sounds good, thank you. I will add that code!
1 Like
May you share your profiles please ? (maybe in another topic) I agree with that, we have right to be angry when Sovol made stupid promises and do not respect them… but at the end, if we ignore the false promises, what we have is not bad.
@mistral Really, I wouldn’t want to do that in case my profile messed up someone else’s machine, but I’m including some screenshots of the Speeds section of my Orca Slicer SV08 profile here, which should demonstrate exactly how I’ve set that section. I hope that helps.
As I’ve said, I’m in the school of Slow and Steady when it comes to printing, and that’s very much the way these settings are… having said that, the quality of the prints I’m getting is wonderful now, and it’s no longer like having a caged Tasmanian Devil at the end of the workshop! With the default profile this thing was screaming and pretty much ripping itself apart.
Oh, the other great thing I did was replace the fan behine the panel on the base of the printer with a Noctua equivalent, and that made the world of difference to noise levels.
Hope this is useful.
Edit: Incidentally, I should mention that I print exclusively with PETG.
Best,
Tim.
2 Likes
I also agree 100% with this.
And to add to this, I see the most “problems” arise with people directly modifying the printer with different hot-end and/or other upgrades and changing all the parameters and macro’s (without a specific cause but just because some video says it needs it) and complaining about the printer quality afterwards.
I’ve had two of these printers with the default setup for some months now and in my opinion they works like a treat for the money I paid for them. With some set-up and maintenance every now and then we get very good results on both the one at home as the one shared (and abused) on the makerspace
P.S. Comparing it to a Prusa is like comparing a chinese micro car to a volkswagen and complaining that the cheap micro car doesn’t have all the features and rigidity the volkswagen has.
1 Like
Exactly! The only thing I did to mine is I had to replace the cooling fan so I went with a bigger part cooling fan and port. It cools from the back & sides instead of the front. Is it any better? I don’t think so. I like being able to see the nozzle, that’s the only advantage. I’m going to change the port again to one that only cools from the back because I don’t like having both sides covering the part so I can’t grab a string or blob that occasionally happens. Other than that it runs great for what I paid for it.
1 Like
You’re probably not using the eddy-ng code then. I had problems with the original eddy code as well, especially when I tried using a different bed I got off aliexpress.
The eddy-ng code reduces the issues with temperature compensation that you’re referring to. And even if it doesn’t, you can have different calibrations.
Mine’s been damn near perfect for the last month. I’ve been running multiple jobs a day for weeks and haven’t had any fails.
Taco bead is easily fixed with a few layers of printer paper. Just undo the screws on the sides, lift it, and slide under a few sheets. Do a z-offset calibration, and a bedmesh. Repeat until you are happy. And don’t tighten the bolts to hard, just finger tight pluss half a turn on each bolt.
The warp is from the lack of support under the plastic frame the bed assy is placed.
It’s not a perfect solution, but will help quite a lot.
I know most ppl use some type of tape, but I don’t like the tape glue going everywhere. And the paper sheets are quicker to work with.
After this “mod” your first layer problems should be manageable. Mine is at a 0.120mm diff and could be less with a little tinkering, but I’m happy with it as is.
My bed was actually so off that when I printed parts at the center of the bed, the pre print mesh lowered the nozzle so much that it scraped the edge of the build plate moving to the heat/purge position.