Getting a Mirror Bed

tirelessirony
Posts: 10
Joined: Mon Oct 08, 2018 6:22 pm

Getting a Mirror Bed

Post by tirelessirony » Mon Oct 22, 2018 6:05 pm

Hi all, this post contains some info I gathered on replacing the Promega bed glass. Perhaps it'll help you too! My Glass hasn't cracked but it came a bit warped. I cant tell if the repeated heating and cooling worsened it, but flipping it to either side shows an interesting bulge/depression near the center that flipping the glass and adding more clips couldn't solve. If I recall correctly, Discord user Shadowhawk has replaced his with a 1/4" Mirror and his bed mesh looked fantastic (most all green). So that seemed like a good plan.

My main concern was the thickness and how that affected the time it takes to get the bed to heat.

So today I went to my local glass shop and after explaining the issue they were kind enough to furnish me two small, squarish pieces of scrap mirror for testing with the bed heater. One piece 1/8" thickness (~9"x~4"), one piece 1/4" thickness (~10"x~5"). The original glass is 1/16" clear glass the same size as the heater pcb. Due to the mirror pieces being smaller than the original glass and the bed, this test really only shows how quickly/effectively the surface of each heats up - not how well they print OR how well a ~15" piece with rounded corners would heat up. I'll post more data on that once I get it.

Methodology:
I pre-heated the full bed to 60 degrees Centigrade without any glass or mirror and let it stay at temperature for a good length of time to allow heat to radiate throughout all edges and aluminum. Using an IR thermometer, I verified the Temperature across the bed while the web console read a consistent 60-60.1c reading. I noted near the center temperatures of up to 62-66c, with ~55c in the corners tapering to <40c at the very edge of the corners. When switching between glass, I let the bed sit without anything on it for 5-10 minutes before starting measurements with the next piece. The bed is at its lowest position, resting on the Z limit switch for all measurements. No enclosure.

To ensure solid IR thermometer readings on the mirror surface, I cleaned and applied small sections of painters tape to each piece of glass (StickyBedz brand 8" Orange 3D print tape, to be exact). I sat each piece of glass onto the bed, centered as best as possible and used a stop watch to record the surface temperature every 30 seconds. Here's the data expressed in a chart:

Image
I stopped measuring at 25 minutes for the 1/4" mirror, taking another reading at the 35, 40 and 60 minute marks. It reached 60c somewhere between 40-60 minutes. *OEM glass in closer to 1/12" thickness, not 1/16"!

Analysis:
I can see now why M3D chose to go with a 1/12" piece of glass. As my glass shop said, it must be good quality seeing as its so thin and can withstand repeated heating's in excess of 100c! I'm sure you can look around to find a 1/12" mirror, but I doubt it's going to be appreciably better when it comes to warping. Perhaps someone with more experience with glass and mirrors could weigh in on that point...

Regardless, the time it takes to heat up a 1/4" piece of glass was indeed substantially greater than I'd hoped - even if we consider its slightly large dimensions. For both mirrors I suspect getting a full size piece would yield faster heat up times due to more surface area, but it may be larger glass leads to slower heat up times. I ordered the 1/8" mirror after taking these measurements, so I should soon be able to report if the larger 1/8" mirror heats up faster than these small test pieces.

Conclusion:
Without more data, 1/8" Mirror seems like a solid option. It heats to 55c in the same time the glass heats to 60, and I don't mind waiting 10-15 minutes for my bed to heat up completely if it gets rid of the warp. Only time will tell if that added thickness will avoid warping. 1/4" mirror is thick stuff for a heated bed - taking over 40 minutes to get to a full 60c! I hope to get some insulation for the underside of the bed in the near future to try & shave more time off. Perhaps cork or some foam board cut to shape.
Last edited by tirelessirony on Fri Jan 11, 2019 12:35 pm, edited 1 time in total.

jwmueller
Posts: 10
Joined: Tue Sep 11, 2018 4:30 pm

Re: Getting a Mirror Bed

Post by jwmueller » Mon Oct 22, 2018 6:33 pm

Thanks, my bed is 2.25mm from the factory.

I used to using 5mm borosilicate but in this range I was looking at $200. I have watched the glass warp while heating everything without the clips on, I now clip it in 4 spots on the front and back. I keep watching what everyone else is doing before deciding what direction I want to go.

One nice thing with the thicker beds is that you get a longer cooldown period which can benefit some really large prints.

tirelessirony
Posts: 10
Joined: Mon Oct 08, 2018 6:22 pm

Re: Getting a Mirror Bed

Post by tirelessirony » Tue Oct 23, 2018 1:45 pm

Wow, good catch. the OEM glass is 2.25mm, closer to 1/12 of an inch. I can't be sure what the best replacement solution is, but given time I'm sure the community will reach consensus.

User avatar
Shep
Posts: 62
Joined: Tue Sep 11, 2018 4:18 pm
Location: Cleveland, OH
Discord Handle: @Shep
Owned M3D Printers: Original Micro, Pro, ProMega, Crane
Contact:

Re: Getting a Mirror Bed

Post by Shep » Wed Oct 24, 2018 3:26 am

So would the results be the same for regular window glass? The commonly-held (in Discord anyhow) notion for the window glass (not mirror) thickness is:

2/16" (1/8") is too thin, prone to warping/shattering
3/16" is just right, being strong enough to hold while being thin enough to heat well
4/16" (1/4") is too thick, plenty strong, but suffers from temp issues

I do wonder if 100 C would be another test point? See how far it can go, give it up to about two hours to reach temp and if it doesn't, kill it.

Also curious if there's any convection going on here. The 1/4" piece covers the full bed, even so far as with curved corners, while the other two pieces are smaller sizes -- is it possible the parts that aren't covering the bed are heating the air and wrapping around to the edges/top? Only way to find out is with a size matching the 1/4" glass (though Leon and ShadowHawk measured 16 3/8" square, while I measured a near-exact 16 1/4" square -- your mileage may vary)

StickyBedz brand 8" Orange 3D print tape... I gotta look that stuff up, I knew I wasn't crazy when I said I swore by blue tape!

I would also wonder if third-party sources can enlighten us on the abilities of glass at different thicknesses re:strength, flex, etc. Surely someone has some seriously scientific setups for this by now, glass has been around a long time and so has the window industry, even before they did the whole vacuum-seal thing.

BTW side note, glass like this is considered "float glass", i.e. it's not meant to be clamped down firmly when heated, That's why you'll often see 3D printed clips which make you question how their evidently springy nature can work. It's supposed to have some flex built in, and that's why window frames use foam around the edges to boot -- insulates and alleviates strain. Binder clips are extremely rigid and don't really flex well at all. ShadowHawk has a design somewhere of his (sadly not on the GitHub and I don't know his Thingiverse either) that is rather springy that would work well here, and should be fairly easy to adapt also.

User avatar
Kenny66
Site Admin
Posts: 13
Joined: Wed Sep 05, 2018 11:43 am

Re: Getting a Mirror Bed

Post by Kenny66 » Thu Oct 25, 2018 11:39 am

At the moment I have the M3D glass, which luckily mine is pretty flat, a 1/4" mirror and just yesterday bought 1/8" mirror.

The quality of mirror you buy I think makes a difference. 1/4" mirror is I think to heavy for the stock Promega bed. A bed crash with that much weight will certainly break something.

I've seen a couple of people now with warped glass from M3D, and one person has mentioned it's not borosilicate. Whatever it is it's apparently pretty sturdy stuff. Someone posted a pic of theirs on the floor after the print head hit their print and shoved the whole thing out the front of the printer. The glass didn't chip or break.

That being said, I'm opting for 1/8" mirror from the local glass shop. It's got a little heft to it, but not to heavy and it should heat up fairly quick. The guy at the glass shop told me it should do fine heating up to 100-110C without issues.

The very important part to any of this is ALWAYS heat the entire bed. Leave your left power switch in the ON position. Any of the glass beds will break if heated unevenly, and heating the middle while leaving the outside cool is asking for a shatter.
Kenny66

User avatar
Shep
Posts: 62
Joined: Tue Sep 11, 2018 4:18 pm
Location: Cleveland, OH
Discord Handle: @Shep
Owned M3D Printers: Original Micro, Pro, ProMega, Crane
Contact:

Re: Getting a Mirror Bed

Post by Shep » Thu Oct 25, 2018 4:53 pm

Wonder how beds that aren't glass would do as far as max temp as well as time to 60 C. PEI Sheets, BuildTak, realistically the whole point of glass is you can print directly on it, but for guys like me who swear by blue tape we just want something on there. If glass is a limiting factor -- and I bought 1/4" because it was $11 each -- I'll glady ditch it.

tirelessirony
Posts: 10
Joined: Mon Oct 08, 2018 6:22 pm

Re: Getting a Mirror Bed

Post by tirelessirony » Sat Oct 27, 2018 7:59 pm

I got my 1/8th Mirror Bed and am pretty happy with its flatness so far. Mine priced at <$20 cut, corners rounded and edges sanded. No clue about its quality or manufacture, other than to say it looks nice and has a black backing on it.

My first concern was the pressure from my binder clips. I tend to use between 4-8 depending on how the bed mesh comes out. Clipping it wasn't a big problem, including at ~100c (I stopped at 100c). If it was any thicker I wouldn't be able to use this size binder clip, and I believe the next size up has a lip which won't fit under the spacing of the bed. So, you can use commonly available binder clips with an 1/8" mirror [they're tight!], but any bigger and you'll need something else (such as the printable 1/4" mirror clips Shadowhawk posted).

I took my time with this mirror, not sure of how it'd react to heat. First, I sat the new mirror on the bed, clipped it in the corners, and heated it up from room temp to 60c. I let it sit at that temperature for over 3 hours, paying attention to the any warps or spacing changes around the edges. Everything looked great! No warping, chipping, cracking. So I ran through the macros, adjusting my machines z endstop height for the ~1mm or so loss of Z height. No problem there, onto bed meshing. Here's the Bed Mesh with the 1/8" Mirror:

Image
Ignore the bad corners - it's the best I can get skipping the belts for now and was more level than but consistent with my prior Meshes.

The metrol probe confirmed what I was seeing. It is extremely flat - even at temperature. Unlike the methodology I used last time, this time I had placed the full sized mirror on the cold bed and let it heat up with the PCB. At the exact moment the Web Console reached 60c, my IR thermometer read 55c. Approximately 1:30-2:00 minutes later, I got a 60c reading from the bed. This represents a more real world usage scenario and it is well within an acceptable time tolerance for my use.

After the 3 hours, I got consistent readings between 55-61 near the center on the bed. Satisfied it wasn't going to crack at 60c, I set the heater to 100c in the Web Console and let it heat up. It took a considerable amount of time - granted I have no frame of reference for this as I had never exceeded 70c prior! I let it do its thing for another 3 hours. I moved some of the clips around, added a few more on while at 100c (7 total) and didn't notice any issues other than it was too hot to comfortably touch.

Ultimately, I let the mirror cool completely to room temperature, removed it, and heated the uncovered bed to 60c. Then I used the same methodology as was outlined in my original post to generate data on how quickly the full sized 1/12" glass and 1/8" mirror Beds heated up. Here's the data expressed in another graph:

Image

According to the data, it takes an additional 30-90 seconds to heat the thicker, 1/8" mirror. Again, an acceptable amount of added time for most purposes. Using the larger, shaped piece of mirror did yield much faster results - heat ups in 5 minutes vs 10 minutes. That would likely hold true for 1/4" mirror as well, but only those who've gotten one can truly say. What is clear is the increased surface area matters - either by increasing heat transfer or by holding in and insulating more of the heat. I don't think it would be unreasonable to assume only an added 2-10 (7-15 total vs 40-60) minutes for a full size 1/4" mirror bed, but that's just a guess. I hope this added information can help someone.

I am happy with my 1/8" Mirror so far, will post updates after more use.

sohjsolwin
Posts: 6
Joined: Mon Oct 01, 2018 9:25 am

Re: Getting a Mirror Bed

Post by sohjsolwin » Tue Nov 20, 2018 9:16 am

I picked up this mirror https://www.target.com/p/square-framele ... A-51672241 from Target a few weeks back and it's been working well. You lose a little print volume, but it until I can get to a window show to get a custom one, it fits well enough and is reasonably priced.

MyronW
Posts: 4
Joined: Sun Oct 07, 2018 9:34 pm
Location: Valley Center, CA
Discord Handle: MyronW
Owned M3D Printers: Original Micro, Micro+, ProMega

Re: Getting a Mirror Bed

Post by MyronW » Mon Nov 26, 2018 11:06 pm

A lot of owners have given up on the M3D-supplied glass bed. If you recall, that piece of thin glass was 99 bucks, and I just can't keep wasting money on this machine so I was determined to get it right. This is my bed after leveling and shimming.

By the way, great job on the analysis of mirror thermal performance. Your report is quite professional. Well done!
M3D Glass Plate Mesh.JPG

tirelessirony
Posts: 10
Joined: Mon Oct 08, 2018 6:22 pm

Re: Getting a Mirror Bed

Post by tirelessirony » Thu Jan 24, 2019 2:03 am

I got my ¼” Mirror Bed and am so far similarly pleased with its flatness. Also ~$20 cut, corners rounded and edges sanded – I had a trace made of the OEM glass, so it’s just ever so slightly oversized, which seems ideal. This time I ordered three!

No Clue about their origins, but I was told ¼” mirror like this – or at least this mirror in particular – is tempered. Before testing the bed, I tried various sizes of binder clips but found what other users had already reported – there really isn’t a commonly available size that works for thicker mirror (mine measures 5.98mm thick). I printed Shadowhawks ¼” Mirror clips (link) which slide on nicely.

Before getting to how long they take to heat up, I heated the 1/4" mirror to 60c and let it sit for nearly two hours – during which time I re-zeroed and meshed the bed. No issues, chipping cracking or warping. So I set the printer to 100c and let it run for a few hours more. I returned to find it hovered between 99.6-100.0. I’m not exactly sure how long it took, or how well the bed would maintain that heat raised up or the effects of using an enclosure, but it was able to reach and maintain ~100c.

Bed Mesh Flatness:
Before posting the data on the time it takes to heat each different thickness bed up, here are some screenshots of the best bed meshes I achieved using an upgraded z-system featuring a metrol probe, an optical endstop and my rail foot tensioners as well as these flat bottomed C4 sliders:

1/12" M3D Glass:

OMEglassBestBedMesh.jpg

1/8" Full sized Mirror:

1-8mirrorBestBedMesh2.jpg

1/4" Full Sized Mirror:

1-4mirrorBestBedMesh2.jpg

These clearly show the issues using M3D's $100 glass bed. Again, as I posted up above, it's not bad glass - mine was just not flat glass. Also, I have been using the 1/8" glass almost exclusively now for months, and even after countless heat cycles, the 1/8" still performs very similarly to 1/4" when it comes to flatness.

---

IR Temperature Testing
Using the same methodology laid out in the first post, I prepped by heating the uncovered bed to 60c in the DWC and let it permeate throughout the aluminum of the frame for well over an hour and verified the temperature across the bed. I cleaned and prepped each of the three glass bed surfaces and applied a small section of painters tape (StickyBedz brand 8" Orange 3D print tape) to the center. I then placed them one at a time on the bed and took measurements using my IR thermometer every 30 seconds until they reached 60c. I allowed 15 minutes or so after removing a glass bed to allow the PCB to fully heat back up and normalize before measuring the next. Here's that data expressed in a chart:

3dPrintBedGlassMirrorComparisonChart.png

Analysis:
The 1/8" Mirror and M3D glass had results consistent with my original measuring - 1/8" mirror takes about 60-30 seconds longer to reach 60c than M3D's glass bed. I had speculated it would take between 7-15 minutes to fully heat a 1/4" mirror based on earlier observations - and that was a safe hedging. 1/4" takes twice as long to heat up as M3D's 1/12" glass, and about three minutes longer than 1/8" mirror. Nine minutes in total using this methodology.

For thoroughness, I placed a cold mirror onto a completely cold Promega heatbed and set it to 60c and watched with a stopwatch until the DWC read 60c. Readings in the DWC went from 24c to ~40c in two minutes, got to 50c in four minutes thirty, and hit 60c at seven minutes thirty. Although, I could not find any point on the bed that read 60.0 with my IR thermometer until eight minutes thirty. This represents more real world use and shows the 1/4" mirror bed takes less than 10 minutes to heat up.

Conclusion:
You cannot go wrong with a mirror. Having two Promega's I bought one 1/8", but three 1/4" mirrors. I did this primarily because there has been discussion on the discord about how the additional weight is thought to help with Z-hop accuracy - perhaps with mesh bed adjustments as well. The trade off is an increase in bed heat times and loss of some extra print height (my Promega with a 1/4" mirror has 370mm or ~14.5" print height available).

While the heat up time doubling over M3D's glass isn't insignificant, it's far from damning. 1/8" Mirror will serve just fine and has a few benefits - it heats up slightly faster and can use commonly available binder clips. There are also 3/16" mirrors and glass available, and I would expect it to perform similarly well. I'm glad to have both sizes, but based on the first layer adhesion I've gotten using the 1/4" mirror, I would highly recommend the 1/4" mirror option using the stock or a similarly modified 80/20-slider based Promega Z-system.

I hope this report can aid others in making an informed choice. Cheers!

Post Reply