K8200: use of mirror on heatbed

Hi there,

Noticed there are some problems regarding the flatness of the heat bed. One of the solutions to this problem is the use of a 20x20cm mirror on top of the heat bed.

I do have some questions regarding the use of this mirror as there are some opposite solutions to the matter.

1- How to place the mirror onto the bed? Back or front side ‘up’?

2- Do you still have to use tape on the mirror to improve the ‘stickiness’? Which tape should be used? Kapton? Blue painters tape? … or maybe some other solutions?

3- What’s the best way to mount it on the heat bed? Double sided tape seems to be difficult when mirror needs to be removed later.

4- Which settings have to be adjusted in the software due to the extra use of the mirror. Temperature? Speed? …

Looking forward to the answers to my questions… Thanks in advance.

PS: purchased my mirrors @ Brico (Belgium) - set of 4 pieces < 10€.

Hi BigDee,

the best way to mount the mirror is to use binder clips (that are these things: en.wikipedia.org/wiki/Binder_clip). Just clip them onto each corner but be careful that the print head does not crash into them when homing!

Regarding the orientation of the mirror I would say reflective side up. I myself don’t use a mirror but a piece of glass that I found lying around in the garage. I also don’t use any kind of tape or anything. Just clean the glass before printing and the PLA sticks fine. Bed temperature is set to 60°C. Everything else is standard settings. Works well for me!

  1. the backside is on the bed, (Frontside is up)
  2. For PLA you need no tape or coating (you print on the blank glass surface of the mirror).
  3. the clips i think ar the best
  4. same settings but it takes longer to heat up

my five cent to your questions:

  1. i use my mirror with the mirror side up.

  2. currently i am printing only with PLA. The only thing to do is "have a clean, fat free mirror. I cleaned my mirror one time with Acetone. After that i use the scraper for ceran cooking fields to clean the mirror between prints. The heat bed must be heated (i use 50 degree Celcius). On the warmed mirror the stickiness of PLA is perfect. Important is the distance between mirror and nozzle. To less blocks the filament and produces “blobs” of filament now and then. To much distance let the filament “lay around”. With the correct distance the filament is slightly pressed onto the mirror and sticks perfect.
    After cooling down to room temperature after the print the printed goods “pops” of the bed very easy.

  3. i clam the mirror with paper clips onto the heat bed. Currently i use two strong clips on the front and the back side of the bed (y-Axis). Beware that the nozzle can collide with the clips. That did no harm to my printer, but it could (actually i am thinking about a different, smaller mounting option). Because my bed is quite warped (from the very beginning), i use small stripes of paper between the bed and the mirror. This way the mirror gets not bended and is perfectly flat. I know that my technique is really bad in terms of heat flow from the bed to the mirror, but with PLA it works fine.
    The heating of the bed takes from 17 degree celcius to the targeted 50 degree about 10 to 15 minutes. I can see from the temperature curve that the power supply for the heat bed is on its limits. The stronger power supply for the heat bed is already ordered and will enable faster heating and higher temperatures (e.g. for printing ABS, Nylon, etc.) We will see.

  4. no settings adjusted for the mirror. Just the z-Axis end stop has to be appropriate.

Greetings from Hamburg

P.s. I bought my mirror at the swedish house.

P.p.s. My modifications to the printer are so far a replace z-motor-holder (inspired by edirol) http://www.thingiverse.com/thing:144813 and a fan duct from eytec http://www.thingiverse.com/thing:160252 .
I will add a X-Axis Belt-Tensioner from Stephan http://www.thingiverse.com/thing:166031

I use a Mirrorplate.

  1. I scrached the mirrorside with scotchbrite - Mirrorside up ( at 50 °C sticky, at 25°C Parts jump off)
  2. On the backside i used thermal compound (cheepest one, not the expensive one used for computer cpu´s)
  3. On the 4 edges i used normal “hotglue” - The Heatbed used at 60°C, the Hotglue starts melting at 150°C
    The glue not between the bed and the mirror - just the edge of the mirror to prevent sliding
  4. Before Printing, if the bed reach 50°C i clean it with Isopropanol or a Cleanpad for glasses.

You can find all of my modifications at Thingiverse:
A picture of one edge of my heatbead


I have been using the mirror for over a week now, and must say results are great.
Also noticed that it is an absolute must to spend some extra time on levelling the mirror and adjusting the Z-height ‘home’-position.

PLA sticks like glue with heat bed @ 60°C and extruder @ 190°C. When finished, just wait until heat bed is back at room temperature and part will pop off with no extra force.

I recommend it to all.

Only minor is the size of my mirror. I bought a standard 20cmx20cm in the shop. Using clips to mount it on the bed gives you a smaller ‘usable’ size as you have to avoid ‘hitting’ the clips with the nozzle of the extruder. Will have one custom made size 21,5cmx21,5cm.

Regards & Big Fun


I just sanded my heatbed a bit, this sticks really well too, is the stickiness (and the perfect smoohness) the only advantage of the mirror on heatbed ?

I also use a mirror. the surface is now perfectly flat and sticky (even sanded the heat bed didn’t stick enough : I had to use blue tape - and was not flat).
I also used thermal compound and glued the corners with hotglue. Heating time is almost the same as without the mirror.
To adjust the heatbed I’ve put 4 springs between the heatbed and the aluminium frame. It has another advantage : if by accident the noozle hits the mirror it won’t break as the springs will sink a bit.


installing a mirror was one of the first tweaks to my printer, since the original heatbed had a notable curvature right from the start, making it unsuitable for direct printing. After having a look at the mirror tiles from the swedish furniture store (like Matthias mentioned), I decided to ask a local glass shop - square 21 cm, 4 mm thick, trimmed edges for less than 5 €.

  • Heating is slower. Most days, when multiple print jobs are pending, I don’t let the bed cool down between prints to save time (50 °C, PLA mostly).
  • Heat flow to the lower side was quite high. Now it’s reduced a bit by installing a 4 mm cork tile between the alu baseplate and the cardbord spacer.
  • Fixing was done with a stripe of kapton tape on each corner - and that’s all. I took out the screws originally used for fixing the PCB, so mirror and PCB can have good thermal contact. Since the kapton is thin, there’s no real danger of collision with the nozzle (as long as it doesn’t scratch the surface).
  • Cleaning involves (in ascending order, as needed): a scraper (like those for ceramic cooktops) for sticky plastic; a dry paper (as “soft” as a usual paper towel) for polishing; sometimes a drop of water in case of fingerprints etc.; and as a last resort for degreasing isopropanol.
  • Although the heat transfer seems to be slow, it’s not as bad. Advantage: the surface temperature is quite close to the setpoint (applied thermocouple shows only a few tenth of a degree deviation). Disadvantage: switching on the fan to 45% or above nearly immediately reduces the temperature, which in turn forces the PID to constantly apply power to the heatbed. I don’t know if this is also the case without the mirror, since I could only use the printer after installing it.
  • The table is considerably heavier in this configuration, so I reduced vibrational issues in my prints by setting maximal accelerations.
  • Next plan is to get a slightly larger (21.5 cm), but thinner (3 mm or less) mirror to reduce weight.


Can you explain what changes you made to the acceleration settings and why you choose for any particular settings?


Hi Robert,

I set acceleration (in Slic3r’s speed settings) to 120 - 150 mm/s². Default (i.e. the reset value) is 500 mm/s², which corresponds to my firmware; 1000 would be as provided by Velleman - see Configuration.h, line 228 here.
Speeds are generally set to 50 - 60 mm/s, exceptions being perimeters, top solid infill (for surface quality), and gap infill, all slower. And travel, which is of course set higher, at 200 mm/s.

With default settings some of my prints had problems. One example: in a print, upper layers, there was a short, ca. 20 mm long zig-zag infill. The same infill had already been present in some lower layers without problems, but for some reason at that point the carriage would built up vibration, locking the x-bearings and thus loosing steps. Reproducible, each time same point of failure. Otherwise, the x-axis is running smoothly, just the increased weight and high center of mass in my heatbed config seem to cause trouble in the drivechain.

Apart from such critical failures, setting the acceleration also had the nice effect of producing better infill. E.g. for rectlinear infill: when the nozzle started infilling, the first few lines were often broken, as if the material did not immediately flow at the right rate. With acceleration, those (mostly very short) first lines are slowed down enough to prevent gaps, while allowing the printer to operate at the set infill speed when long lines are printed.

There are some quality issues you can get rid of by setting a reduced acceleration, others will appear. So this stuff is not without tradeoffs, unfortunately. In my case, I’d rather use it than frequently face lost steps and skew.

Above values may of course be far away from the optimum, for your configuration and also for mine. I started low and increased the values until the printer was fast enough for my taste (at least at the moment). All those tests were done using PLA, in case it matters.

Happy experimenting!


Thank you fo sharing this with us.
I will apply his in finetuning my settings.


I use a piece of 8mm plexiglas without Any tape.
It sticks as hell. :wink:
Hope it helps.

Greets, passi.