View Full Version : HXTAL NYL-1 Epoxy Metal-Glass Bonding, opinions?

Matthew Leung
04-09-2012, 02:17 PM
Hi all,

Is anyone here familiar enough with HXTAL NYL-1 to know about its capabilities of bonding glass to 303 Stainless Steel for structural purposes? Not too much documentation available online on successes.

Please see attached pic for reference on my project:
1) Black Frames with Metal pucks on top are the final legs for the table.
2) All unpolished lumber is part of the scaffolding
3) Glass which is not shown here is 72" x 42", where the 72" spans to the middle point of each of the chairs.

I am building a dining table (as a hobby for self) where a large piece of glass (42"x78"x0.5") is supported by 2 frames each with 2 metal pucks (3.5" diameter each) that will hold the glass up. The sole joint for each frame will be the glue joint between the glass and the stainless pucks. The original plan is to use HXTAL NYL-1 for glue-up after proper treatments (talc for cleaning glass, amino saline onto glass, alcohol on stainless and then torch to sweat the stainless).

Idea for Glueup:
I have constructed a rig with 4 hydraulic jacks on a scaffolding that holds the glass (see picture). The idea is to secure the 2 frame-puck pieces in place, apply the HXTAL epoxy, and lower the glass onto the frame-puck constructs whilst maintaining a consistent glue fillet. Originally I had the idea of using 1mm glass standoffs of 1/2" x 1/2" x 1mm to maintain and ensure a uniform glue fillet across all 4 pucks to the glass. However after talking with Bob at Hisglassworks, swapped out the idea with pre-curing "drops" of glue onto the metal puck as standoffs, since his idea is much better and eliminates glass-to-glass in the joint.

Note again that the only thing keeping the glass and frame from toppling like dominos is the glass-metal glue joint, thus it is essential that it is reliable.

Bob had mentioned thermal expansion being a potential problem, but this table will be kept at 72F-78F. Just in case though, the glue joint will be at least 1mm thick. The table will likely lead a fairly sedentary life, but needs to obviously be able to safely handle someone bumping into it and leaning on it but nothing extraordinary. Bob also suggested considering 3M's DP105, however, I am not comfortable with its short 5 minute working time.

Because my principal concern is actually the strength of the metal bond (not so much the thermal expansion at the moment), I have bought a bar of 303 SS and plan to bond two metal segments together and do destructive testing after full cure to see what the yield strength is.

My question is based on your experiences and intuition, if HXTAL is acceptable for this application. Has anyone bonded metal to glass with HXTAL in furniture or on some large scale? Any feedback is greatly appreciated.

Regards and thanks!

I understand that any recommendations expressed here are for educational purposes only.

Scott Novota
04-09-2012, 02:47 PM
I have never used this but figured it might be of use to you if you have not seen it already. It seems to be addressing the same application.


Marty Kremer
04-09-2012, 10:35 PM
I think you're asking a lot from a glue joint.
The glass alone will weigh 135 pounds. Add any lateral force (someone pushing away from the table? someone stumbling against it?) and I wouldn't want to be sitting there.
I'd either glue across the entire frame (no pucks) or drill the glass (before tempering) for a mechanical hold or-- best-- run a cross brace at the top from one frame to the other.

Dan Ellis
04-09-2012, 10:40 PM
I used to work for a company making glass awards and we once had issues with making glass to stainless bonds using UV glues. We'd glue up the parts and often in a few days the glass was checked under the metal. the glass was 2"x2"x12" square columns and the stainless pieces were 3/16" x 1" disks.

Talking with the Dymax guys we figured in was either the different rates of expansion between the glass and metal as the temps changed or the adhesive was contracting as it cured and since the glass was more brittle then the metal it cracked. Temperatures were only the differences between days and nights during the winter. We never did find out for sure what caused it.

We did a bunch using a quick cure epoxy but I forget which one and we did a bunch with very slightly flexible UV adhesive from Dymax, Both worked fine but all the bonds were a lot smaller than what you're doing. The flexible adhesive isn't like a silicone type glue, it was only flexible microscopically, just enough to compensate for expansion or vibration.

Rick Wilton
04-09-2012, 10:54 PM
There are plenty of tables constructed like that they are generally glued with uv glues however not hxtal. C.R. Laurence sells all kinds of glues, and hardware to do things like this.

Pete VanderLaan
04-10-2012, 02:43 AM
I saw a joint once done with metal seal using hxtal and the piece did break but not at the glue joint. It was heavier than what you describe. Don't use
UV or silicon.

Mark Wilson
04-10-2012, 06:58 AM
i would use E6000 craft glue. i use it to bond glass to just about everything. i use it to bond glass to my outdoor glass art...it is outside 24/7..365...you can find it at most craft stores and also walmart.

Tom Fuhrman
04-10-2012, 08:54 AM
The glass on my dining room table, 135-150lb. is held on with suction cups. we bought it at Ikea and there are 1000's of them out there like it. had it 10 years with no problem with it coming loose. KISS method, keep it simple stupid.
you can also check out the adhesives from Tangent industries,they have a new website that gives all the specs on their adhesives and will help if possible, call them.

Lawrence Ruskin
04-10-2012, 10:48 AM
What Marty said about glue joints.

Who knows if that glue joint is going to last 20 years.

Tom Fuhrman
04-10-2012, 11:50 AM
What Marty said about glue joints.

Who knows if that glue joint is going to last 20 years.

I'm quite certain that I won't last another 20 years.

Dave Hilty
04-10-2012, 12:11 PM
I have used Loctite 352 UV/Heat/Activator cure with Loctite 7387 Activator for many glass to metal bonds and so far none have failed. I have experience several failures of glass to metal bonds when using the E6000 glue. I think Marty's suggestion of spreading the bond across the rail and not just pucks is good.
My other suggestion is to turn the thing upside down, especially if you are attempting the puck only route. Put the glass on the floor/carpet and then place the activator-prepared pucks attached to the rails (activator takes about 8-10 minutes to be ready to take the adhesive into contact) on the glass after smearing the glass target area with the adhesive. This will insure more surface area of two pucks at a time forming the bond rather than trying to even up 4 pucks across that span. Clean up any excess adhesive/activator around the pucks with Acetone before it cures. Once you have the thing adequately cured, you have the advantage of mechanical leverage to really lever against the rails sideways (if you have the nerve) to test the strength of the bond. My experience with this glue is that if you go too far, the glass will break out and not the bond.
I'll leave it to you to figure out how many guys you need to turn the table back upright when you are satisfied with the cure/bond. Ultimate leveling would come from leveling feet I assume you are placing at the four corners of the rails. I have to assume it is tempered glass of course, so that when you are practising your judo defense technique and you throw your "Cato" over your head on top of the table, it busts up nicely.
Have you gotten an engineer's go-ahead and loading analysis of the table span of .5" tempered?

Pete VanderLaan
04-11-2012, 11:48 AM
My criteria for trusting a glue joint is "would you sleep under it?"

I would never sleep under a joint from E6000 or any UV glue and that's based on catastrophic failures of the joints at unexpected intervals.

I think that a way I used to see Michael Cohn deal with Hxtal in that circumstance was to place some very small pieces of slide covering glass ( as in Microscope preparations) on the pucks you describe, then lay down the hxtal. The prep glass will keep the glue from squeezing out of the joint. It you let the Hxtal set sufficiently enough to not squeeze out, It's a bitch to keep from trapping bubbles that will loom large as you stare at them.

Matthew Leung
04-11-2012, 09:05 PM
Thank you all for weighing in on the issue. I am leaning towards trying out a HXTAL with the microscope slides as separators to force a robust thickness on the glue fillet.

One possibility instead of using glass in the glue joint (since glass-glass-metal may be ideal) is to prepare some HXTAL left out to thicken and apply small dabs to the glass itself where it will be mounted to the metal pucks. Once it sets to touch, it can be shaved down to the required fillet glue joint thickness. I would apply to the metal, but the metal needs to be torched to drive out the residual moisture or the bond won't be good (and can't torch metal with cured epoxy on it obviously!).

I'm pretty comfortable with HXTAL's famous strength when bonded to glass. If the "legs" I was gluing to was also glass, I would not be concerned. However I haven't heard of its merits with metal, so I'm going to run some failure experiments with Stainless 303 flat bar. Once it arrives, I will saw up two sample pieces and glue the metal to metal and then try to evaluate the force required to bring it to failure.

The glass weighs close to 200 pounds and railings would certainly make it more secure, but I'm afraid adding railings would defeat the purpose of the contemporary design.

I'll keep you guys updated.

Pete VanderLaan: From the test pieces that I have run with HXTAL, I have not seen notable bubbling in the glue fillet. It helps to lower the pieces so the epoxy contacts both surfaces, then raise the surface again, reflood the glue joint with additional epoxy, and lower the surface finally and fully for curing. Removal of tape protected surfaces about 12h into curing seems to work great. I'd be curious to hear more about these heavier adhesive preparations you described. Do you have any more info?

Pete VanderLaan
04-12-2012, 02:16 AM
Matthew when I referred to the slide covers, they are seriously thin if you know what I'm referring to. It's not the slide, it's the cover. I would be inclined to first glue the slide cover to one of the two surfaces using about one gram of hxtal for each slide cover. Use three covers. Let that set up for 24 hours, then mix your second batch and figure on about a gram per square inch. After you mix it, microwave to heat it just about 100F and let the bubbles come out for about twenty minutes, then pour it. Admire your mess. You will still need a dental pick to remove the remaining seeds and you really want to remove them Think really bright lights to see the joints and the bubbles. All the bubbles look bigger once the joint gets made but given your design, bubbles will really detract from the effect you want.

The failure I saw was on Stainless to Glass and the glass broke about an inch and a half from the joint. It had been subjected to 45F below zero. It was a major crunchy fracture. Note again, the joint was fine. It's actually stronger than the glass.

The surfaces need to be as flat as can be had. Clean up with acetone but beyond cleaning up the slop, wait 24 hours to do fine cleaning. It will be timing issues. Using an exacto will clear it out the best but first use brand new blades and a really light hand. I would practice this first more than once. It is a timing issue more than anything. Once the joint is clean, let it stay warm, not hot for three days and you will be good to go.

Wear an organic vapor respirator when you are applying the stuff despite what anyone tells you to the contrary. Wear latex gloves and roll all of your waste material up in the gloves when you finish up, despite what anyone tells you.to the contrary. I put all that stuff in a zip lock bag and seal it. I use a lot of Hxtal. I do also sell it but probably in bigger quantities than you need. I do have glue sets that Jon Kuhn makes and uses that are a lot smaller and it is great stuff as well. I don't use that amino isotonic stuff at all. I do make up a xylene laced (.05%) isopropyl alcohol mix to clean the joints before gluing. Don't ever apply windex to a glue joint prep.

I can't think of a lot else although there's always more.

Matthew Leung
04-13-2012, 02:56 PM
Thank you for sharing Pete.

I have the SS303 bar in so I will do some testing hopefully next week.

At first I thought the aminosaline preparation was a little hokey (sounds that way right?), but at least based on other people's microscopic anaylsis data (https://www.xiameter.com/en/ExploreSilicones/Documents/Silane%20Chemistry-2a-95-719-01-F2.pdf), the treated joints appear to have a much more smooth glass-to-epoxy interface compared to without. The pictures sold me on using a silane coupling agent - not that I know much about this topic, just thought I'd share.

Pete, I see why you were saying that any bubbles in the glue will get really big. When I spoke of microscope slides, I was aiming for a 1mm glue fillet joint, and you are referring to more like a 0.3mm glue fillet joint. I would like to hear what you or anyone else thinks about a 1mm or greater glue joint:

I was pondering whether to even make it thicker and go for 2mm, since the cured product seems to be extremely resilient, unless I am overlooking something, it would not hurt to use a thick joint. Going the direction of thinner fillets, I've been warned that the thermal expansion poses an increasingly greater risk as the glue joint is made thinner and thinner because HXTAL does not give and bonds very strongly to glass. I was thinking of using a >=1mm thick joint to let HXTAL take up some of that stress, but more so (since temp should be stable) have it take up more impact stress, i.e. should someone bump into it or lean on it (it is after all a table!), since it has much more elasticity than glass or metal parent materials.

Any thoughts?

Pete VanderLaan
04-14-2012, 06:48 AM
Whenever I have tried thick joints I have had the adhesive actually pull away from the glass and cause checking. I did that filling bored holes in Shott F2. As to the treatment I have been working with Hxtal for about 20 years and used the stuff you refer to. It didn't seem to make a difference in the Long run. I don't recall Herbert Hillary ever mentioning it and we used to talk a lot. I think that came from Doug at Conservation Materials and you would need a long memory for that.

Matthew Leung
04-14-2012, 03:52 PM
Herbert Hillary!! Isn't he the inventor of HXTAL? :)

Well that is certainly something I would not want happening! What thickness did you consider as thick in your projects?

I've been running tests with 1mm fillet joints spaced with 1mm microscope slides. They do not seem to exhibit pulling away from the joint itself unless I apply overpressure, lift the adhesive and let it resettle (because this tends to alleviate any remaining bubbles), and then reapply pressure (but a different amount or loaded somewhat otherwise differently). Then the glue will be missing on some edges, creating a lens-like effect. However, I was able to remedy this by reflooding the joint (allowing it to flow freely over the edge of the simulated puck during the resettling phase.

Pete VanderLaan
04-14-2012, 05:13 PM
Indeed the same Herbert Hillary. Its now made by his son Nathan and sold through a few distributors of which I am one but I tend to deal with OCD users in volume.

There could have been a problem with pouring it in a hole. I don't know Every indication I have seen is that the epoxy is stronger than the glass. The less part two you use (up to a point) the stronger the joint.I think in epoxies a good rule of them is that the faster it sets the weaker it is.Don't heat it above 104F. Use tons of light when you're picking bubbles.

Greg Vriethoff
04-14-2012, 06:51 PM
Don't ever apply windex to a glue joint prep.
Windex should never be allowed in any kind of glass making studio. I see fusers and kilncasters using windex to prep materials before firing, and this is a mistake as well.

Rosanna Gusler
04-14-2012, 08:18 PM
Windex should never be allowed in any kind of glass making studio. I see fusers and kilncasters using windex to prep materials before firing, and this is a mistake as well.ditto. rosanna

Pete VanderLaan
04-15-2012, 01:11 AM
I use it to clean finished work. I think the. 05% xylene in 90% Isopropyl Alcohol to be a good cleaning prep with a tooth to it but I'm interested in what other people use.

Rick Sherbert
04-24-2012, 01:57 PM
I've heard the same about Windex, but I've never heard why. Someone care to share?


Pete VanderLaan
04-24-2012, 03:08 PM
It has wax in it.

Matthew Leung
06-02-2012, 03:54 PM
Hey guys, I need some further advice.

The test for glue strength of HXTAL with glass to metal substrates was a very bad failure. I am very glad that I did not assume that the strength would be sufficient, as later the table might have killed or dismembered someone. The failure was very sudden and occurred at the metal to HXTAL joint, very cleanly, suggesting that HXTAL bonds very inferiorly to 303 Stainless Steel.

For those who may wonder, the following was my procedure/design:

1) 1/2" untempered glass, small sheet about size of letter size paper
2) 3/8" 303 Stainless steel bar, gluing surface about 2.5" x 2.5", the rest of the bar overhanging part of the glass. The overhanging part had 2 holes drilled for simulating lever forces with a long wood 5' central fence post (joint failed before even doing this test...)
3) HXTAL, applied as 4 small dabs (1mm high) to regulate HXTAL fillet joint thickness and ample amount to fill the joint of that thickness (1mm), applied after 5 hours of letting epoxy set in pot.

1) Glass prepared by (a) cleaning with fine powdered talc (b) cleaning thoroughly with 99.99% isopropyl alcohol (c) amino silane + alcohol solution from HisGlassWorks applied and let dry.
2) HXTAL that had already sat for 5 hours to become viscous was applied as 4 small nubs on the glass (used to regulate HXTAL fillet joint), uniformly spaced under where metal will be mated to glass, nubs were 1mm in height.
3) Set for 5 days.
4) Metal bonding surface was prepared by sanding down with steel wool and then (since that didn't work really well) fine mesh abrasive sand paper 600-1000 grit.
5) Metal bonding surface was then rubbed with alcohol thoroughly, dried.
6) Used torch to sweat the metal, ensuring all moisture is dissipated.
7) Once, cooled to slightly warm, not hot, HXTAL was applied to the glass with the nubs and the metal was pressed against the nubs.
8) A weight of several pounds was placed on top of the metal for it to set.
9) Joint was allowed to sit undisturbed for 2 weeks, well beyond standard curing time.

Long rubber mallet was used to pound on the steel, with the glass on a protected surface.
-OKAY: Pounding PARALLEL to the glue joint, the joint was very stable.
-FAILURE: A moderate tap in the perpendicular direction to the glue joint on the metal that is extended past the glass (the drilled portion where the wood fence post would have been screwed in), resulted in IMMEDIATE and sudden failure of the joint, and the piece of metal flying off the glass.

Inspection of Joint Failure
1) Everything is 100% except for the metal to HXTAL interface. It completely failed as if there was insufficient bonding action between the two.

This is the same result that I achieved when gluing to polycarbonate. I was later, however, told by HisGlassWorks that HXTAL does not like polycarbonate. I thought that was fine, as long as HXTAL bonds sufficiently well to metal to withstand reasonable human weight strains on a table glass ~200 lbs. I am finding that this is CLEARLY not the case.

Please all who contributed previously, or any one that would have insight in this matter:

1) Is the joint supposed to be so weak with Metal to HXTAL? In reviewing my preparations, is it due to something being insufficient? I performed the preparations as I have been told verbatim by most experts.

2) Any opinions on DP105? I don't like its fast set time, it will be hard for me to work with, and there will be other overhead costs like the dispenser which I do not have, but if I cannot improve the HXTAL glue joint, it is simply too dangerous to use for this purpose (glass table, see original post).

3) Glass to HXTAL seems so strong it is unbelievable. However, metal to HXTAL is horrible, from my experiments so far. Any suggestions on improving HXTAL bond to metal, or is this pretty much a case of trying to use something for an application it was not designed for?

Thank you.

Matthew Leung

Pete VanderLaan
06-02-2012, 04:17 PM
You shouldn't let Hxtal get above 104F Matthew. It's really easy to do. Whether that was instrumental in the failure, I can't say. I do know when John Nickerson glued aluminum to glass with the stuff, that when the joint broke it was a break in the glass, not the actual glue joint.

I have broken joints glass to glass when putting it in tension. I think mechanical connectors are worthy of consideration when you start talking liability. I go back to my observation "Would you sleep under it?"

Rick Sherbert
06-03-2012, 01:53 PM
I know that the Hang Your Glass folks had issues with joint failure until they started doing an extensive cleaning of the metal pieces at the factory. Now, no failures that i know of. They were'nt using Hxtal and I don't know what cleaning involved, but cleaning the pieces yourself just didn't work. This is all a long way of saying you might consider the microscopic cleanliness of the metal.

Matthew Leung
06-03-2012, 03:08 PM
Hey all,

Apologies, I should have been more clear, when I said slightly warm, I meant almost room temperature. I meant to convey that it hadn't been torched, say, like a few hours before glue up.

I wonder if cleanliness of the metal is the issue, and if I had not taken off a sufficient layer of the stainless steel for sufficient bonding. The SS303 arrived with some kind of oxidization, not really rust, but just dulling and discoloration. I scraped most of it off via sanding, but certainly not till there was a solid metallic sheen.

I observed a "dirty" translucent look to the failure site on the HXTAL epoxy side, even though the metal still feels smooth, the HXTAL epoxy side feels rather rough and abrasive textured. Maybe the epoxy bonded to some oxidization and also some bare metal, but ultimately it wasn't enough to take a hard blow from a rubber hammer.


-From the feedback I have received and my inspection of the failure, I think that likely the test piece of metal was not cleaned enough down to the bare metal? What do you all think?

-As a general prep question: Is it better to have an unfinished (smooth to touch but not mirror finish) or a more finished (almost mirror finish) metal substrate for bonding to something like HXTAL?

-Rick, do you know someone I could speak with from Hang Your Glass that might have some insight? Any info would be useful. Would a freshly turned or machined surface not clean enough? (maybe machine oils or something?)

-Any other alternative glue suggestions reputedly maybe close to HXTAL but great on metal with set time of >10-15 minutes?

Meanwhile, I will prepare a new metal piece that is more clean and hope for better performance. The 1 week cure time is a disadvantage to running these tests quickly!

Matthew Leung

Greg Vriethoff
06-03-2012, 03:09 PM
The question of whether or not to trust a glue joint by sleeping under it is a non-starter for me as I would never sleep under anything supported by glue. By their nature failure is always an option with adhesives.

By what you have described (and shown) you are asking a lot of the materials. The surface area of the bond(s) between the glass and metal is very small relative to the overall mass and weight of the finished piece. Regardless of how "super-duper" we all know HXTAL is, my opinion is that even if you do everything "right" with the materials, what you are attempting goes beyond the products' abilities.

The other point I'd like to make is that the stainless may just not like the chemical reaction of the epoxy. This may be similar to trying to use silicone on aluminum. Aluminum doesn't react well with the acetic acid produced by the silicone when it's curing (just a thought).

For safety's sake, I think you should consider a mechanical solution.

Pete VanderLaan
06-03-2012, 03:20 PM
Some materials are strong in compression and in tension. Quartering Shear is something else again. Mechanical connections rock.

Matthew Leung
06-04-2012, 10:58 AM
Thank you for everyone's input.

I think if I can make it so that the bond failure is at least partly within the epoxy instead of 100% between the metal to HXTAL, I would be okay with sleeping under it provided that I place low friction standoffs along the base to reduce shock and the lever action of the floor against the glass (i.e. felt, teflon, or delrin).

I am not a pioneer in this kind of this exact table design, this concept has been done before, I just don't know how and am trying to figure out. For reference, the idea came from a table that is discontinued by Modloft. See link.


Had it been still available for purchase, I would have just bought it, and perhaps now looking back been willing to pay double or triple the cost.

Honestly, on the frustrating days I think it would have been a better option to just buy a nice looking retail contemporary dining table and not have to deal with this kind of hokey reverse engineering...

However a lot of effort, time, and money has gone into this, I am 90% there, with the leg frame glueup, priming, painting, puck machining/tapping, drilling and mounting pucks to legs, getting the glass, building the pneumatic scaffolding, etc. I understand the safety concerns and it is a concern for me too, however, at this point, I would like to continue to improve adhesive techniques and compounds in my tests to find a candidate that, through failure tests, shows to be sufficiently strong for normal dining and gentle use.

The dimensions of my table are very similar to the manufacturer's original specifications for the Modloft Howard Glass table. I have also seen some other designs in furniture stores that go as far as using individual legs for dining tables (see attached picture).

Do any of you all know what kind of glass-to-metal adhesive and techniques these manufacturers may have used for the aforementioned products?


Jon Myers
06-04-2012, 06:26 PM
Do you know why that table was discontinued? Maybe they had the same problems...

Matthew Leung
06-04-2012, 09:58 PM
For the Modloft table, I have not heard of any reasons for the table to be discontinued, but to be equally fair, I have not heard of any lawsuits arising from that table.

Here is another design even taking bond strength to a further extent using 4 individual and unsupported table legs, which has received good reviews and is still in stock:


I have found some other furniture companies that employ this kind of metal-to-glass bonding, taking it to a further extent as well:


I would assume and hope that they have used a method/adhesive/process that yields a bond strength that is safe for consumers. If it is not HXTAL that is commercially used, I am willing to try whatever it is that they use, so if anyone has an idea, please chime in.

Allan Gott
06-04-2012, 10:06 PM
That would likely be a mechanical joint. The tabletop is drilled to receive a bolt which is mounted in the cap or leg. The cap turns down or onto the bolt, pressure for security is spread by the surface area contacted by the cap and support plate beneath the glass.

Matthew Leung
06-04-2012, 11:31 PM
I have certainly seen this variation of drilled through glass and a metal cap to sandwich the glass. However, I have also seen the other variety where all metal is located below the glass table surface.

Not sure if it is the 100% exact same model or not as the dCor link (http://www.csnofficefurniture.com/dCOR-design-107842-ZMN1780.html), but I had previously posted a picture in post #30 taken in person of a similar metal-glass bonded design with the same stitching and bonding construction. Inspecting their glue joint, it was extremely thin and transparent, but that is all that could be gleaned. Heh should have brought my trusty rubber mallet!

I am trying to go this route because for better or for worse there are retail instances of this concept.


Rick Wilton
06-04-2012, 11:47 PM
It's a UV Glue. TRUST ME I've done it. We made tables just like this with UV GLUE.

Cr Laurence sells everything you need.



Matthew Leung
06-05-2012, 08:05 AM
I just completed another test piece with HXTAL that was sanded down to a slightly mirrored finish such that no oxidized or contaminated metal would be present. I guess it's another 7 days!

Thanks Rick for your emphatic input, I recall now seeing your post early on about UV glue, but I had been then skeptical. Digging deeper, I now see that there are glues claiming specifically to target the metal-to-glass bonding application.

There is no HXTAL literature making these metal-to-glass claims, so it appropriate to also perform parallel tests with UV adhesive, to ascertain which will be more robust for my particular application.

I was hoping to get a few questions answered:

1) There was previous talk in the thread about random/spontaneous failures of UV adhesives. Could a contributing factor be that UV glue sets up hard and is a little brittle to abrupt shock? (HXTAL would seem to also be relatively brittle.) If properly set, is there evidence that this would occur to a UV adhesive structural member in an indoor, air-conditioned environment?

2) What are your recommendations for UV glues, that have datasheets?
I see that C.R. Laurence has a UV604 (formerly UV682) glue that is especially well adapted for metal to glass bonding.

3) I will need the glue viscosity to hold a slight fillet of ~1mm due to table tolerances. I can add spacer in the form of a pre-cured nub of epoxy to ward against squeeze-out, or something no problem. Is that OK?

4) I have not worked with UV cured adhesives. Closest I've ever been is probably at the dentist's office and I certainly hope this experience will be more pleasant! Any insight on technique?

So far, I have read that for curing, must avoid tension by curing whole joint at a time. Degreasing glass/metal and sweating metal less than 5 minutes beforehand, obviously. I'll continue researching, but a run down of the most common pitfalls would be greatly appreciated!


Rick Wilton
06-05-2012, 10:16 AM
all the specs are on the CRL website for each glue. I don't understand the need for a 1mm fillet, that may introduce a whole host of new problems. The uv glues are designed to have glue gap of 1-2 mils (thousands of an inch) not 1mm. There are high, medium and low viscosity glues. the low viscosity ones have a capillary action that you shouldn't need so a medium or high would be best Yes the 604 is a good choice). I, like others here feel safer with a mechanical fastener but there are thousands of tables, showcases and even glass stair cases made with no mechanical fasteners. CR Laurence isn't going to sell a 28" long table leg that gets glued onto a piece of glass without additional support if it's a liability issue. CRL is by far the largest supplier of everything (cold glass related) in the glass industry so they know their stuff. call them they are very helpful.

Pete VanderLaan
06-05-2012, 11:38 AM
In the past CRL was helpful. I don't find that to be true anymore.

Tom Fuhrman
06-05-2012, 12:16 PM
Crl has compromised some of their helpfulness as they have gotten much larger in scale. As companies get larger it's difficult to keep the customer service as good as when they were much smaller.
Mathew: I suggest you contact the following. Gary Grosclaude, Tangent Industries, Winsted, CT. 860-738-7449. He is the owner of Tangent Industries who has been making commercial adhesives and Uv adhesives for industry for the last 20 years. He is one of the best polymer chemists of adhesives that i know of. I have found him very helpful on numerous projects I've had. They are not atuned to working with small quantities though, but he is very knowledgeable on adhesives and has supplied them to many other companies and might be supplying CRL as he has done a lot of business with glass companies in Europe for many years. He can give you specific specs on all the items he manufactures and has numerous products that allow minute amounts of flexibility so they may withstand the hammer test you are using. He has UV setting adhesives as well as 2 part activator cure formulations which do not require mixing and cure at room temps. Tell him I told you to call him. They make several hundred different formulations for adhering different materials together.

Dan Ellis
06-05-2012, 10:34 PM
A year or so ago I called Tangent and they referred me to CRL for the small quantity I was looking for. I got the impression they make CRL's adhesives but I could be wrong about that.

At the place I used to work at we used adhesives made by Dymax, they also had a formulation for a minutely flexible adhesive. It didn't wiggle or move around, just flexible enough to withstand a little shock. We glued metal to glass all the time there but nothing as big as your table.

Greg Vriethoff
06-05-2012, 11:59 PM
Some materials are strong in compression and in tension.
May need to call Jamie Carpenter about that. Maybe Ed too (no relation).

Mechanical connections rock.
Whether they rock or not is not the question, but just how much do they rock?

Do they go to eleven?

Joe Deanda
06-10-2012, 09:34 PM
Iv been manufacturing austrian crystal earrings for almost 20 years. Iv been through the adhesive battle and found one that has worked for over 10 years with no breakage on maybeeeee 10,000 pr of earrings Loctite E-30CL its a 2 part epoxy developed for glass to metal bonds. Mc Master Carr about 12 bucks and you'll need to gun to dispense it for about 20 bucks mix well wait a few minutes and mix again. about 30 min working time and 24 hr cure at room temp. No UV light needed.

Dave Hilty
06-11-2012, 01:45 PM

I don't understand why my earlier post or Joe's latest post regarding Loctite products has been ignored in the continued discussion on your glass to stainless table glue-up. The Technical Data Sheets for both products are available for review by doing a search on either product name. The PSI and shear strength stats seem to me to be more than enough to handle your application. We both have tested these products in use as have other glass folks I talk to.

The project is not rocket science given the already settled product availability and user experience.

Pete VanderLaan
06-11-2012, 05:43 PM
I would certainly agree that if you need flex, hxtal doesn't flex. Its claim to fame is optical clarity.

Matthew Leung
06-12-2012, 02:18 AM
Dave and Joe:
Early on in the discussion, I knew little about what adhesives were available, and I was misled by the critical remarks about UV glue and other adhesives and did not give them a fair examination. My apologies, I know it's nothing new to glass experts, but I am diving into this because of just this table project, so information is at times unclear to me. Thanks for your patience and continued input.

In response to Rick's question about why I need a thicker glue joint:
It is because the tolerances of my table structures are reasonable, but it is a large object and the tolerances are just simply not to the nearest thousandth of an inch, therefore settling for 1mm glue joint for some fudge factor, I need some kind of adhesive whose viscosity isn't too low and runny. I know this presents a problem with low viscosity UV glues, but something of a higher viscosity might work.

All: I would like to know what the working viscosity is like for Loctite E-30CL? The datasheet says "low viscosity", but yet the viscosity values for the two parts show cps similar to molasses for the resin and honey for the hardener, which doesn't sound low. Ideally I want something that has quite a bit of viscosity to it to keep the thick glue joint filled.

Also on 2nd HXTAL test, it has cured and I've managed to attach a wood fence post to it and smack the end of that with a sledge hammer and it still has not broken, yet. I think the failure from the first time was because I did not sand down to enough fresh metal, even though it might have looked 'clean'. Newbie mistake I guess.


Pete VanderLaan
06-12-2012, 04:28 AM
Matthew, when I've seen joints fail, substantive temperature swings were also noticed. Whenever I've seen UV failures, it has been catastrophic and complete detachment. With Hxtal, I've seen delamination visibly but nothing falling off. It's been unsightly, which in my work is really bad.

I have never thought that Hxtal worked as well when allowed to become tacky. I don't have a scientific basis for why that is, I just know that demlamination occurred far more frequently in joints made with partially set epoxy.

Your second Hxtal test is more what I would have expected.

Tom Fuhrman
06-12-2012, 07:50 AM
when I order my adhesives from Tangent I can specify the viscosity of the batch they make for me. different applications require different adhesives. It may seem simple but adhesives are "rocket science". Talk to the guys gluing new knee joints into place and the guys engineering it if you don't believe me. get samples if possible and do lots of testing to establish the best thing for your application. It may take a few months and and few hundred $ or more. R&D can be time consuming and costly if done properly. If you don't want to invest the time and $ then take what you get. Every application can be unique and requires different materials and technique. I've seen production work at one factory location and when transferred to another take months to get it to work properly. Does glass formulation ring a bell.

Dave Hilty
06-12-2012, 10:20 AM
Tom, My statement re. "rocket science" was in regard to this project in particular, not the world of adhesives. Mathew is approaching this project with as much care as could be expected and his test it first methods are spot on.

My statement was meant to underline the fact that several adhesive products have been developed specifically for glass to metal bonds including the two Loctite products. Between the Data sheets, a call to the company, whether Tangent or others suggested, and Mathew's testing protocol, I'm sure the solution is at hand.

Pete VanderLaan
06-12-2012, 12:10 PM
I'm sure the solution is at hand.
That won't stop the varying opinions Dave. It's scarcely beaten into the ground, much less to death.

Matthew Leung
06-20-2012, 10:07 PM
Hey guys, I managed to lug the test pieces to the gym and make some strength tests with large weights. I have included pictures for ease of understanding what the test entailed.

I have some questions, but allow me to explain briefly the test:


1) As shown in picture 1, I have a letter paper size piece of 1/2" annealed glass glued to a 1/4" stainless 304 bar, adhered via 7+ days cured HXTAL epoxy. The technique used is as previously described and advised (sand to bare metal, sweat metal, clean 99.99% alcohol, aminosilane).
2) A 5 foot tall fence post was screwed to the SS304 metal bar, with screw holes close to the glass edge. I know it is not a perfect scenario and it can flex, etc...more on that later.
3) Since some of the epoxy pooled on the edge and sides of the glass when setting and it was not cleaned, I used an Xacto and cut the epoxy down to the glass on all sides of the joint - so that it would not provide an unfair adhesive advantage that might not present in the real table.


1) The test piece was laid in a horizontal position with exercise equipment to hold the fence post level to the ground. Imagine image 1, and then toppled over to the right and supported by something.
2) Weights were placed on top of the fence post (some padding added on the ground in case the whole thing fell over...which was a possibility). I added weights in 45 pound increments, by dropping them on top of each other at a height of 1 feet. I added 6 of these for a total of 270 pounds.
3) Subsequently with 270 pounds, I held onto a machine and dropped my weight on top of the weights for a total of 450 pounds. (I weigh about 150 pounds.) It did not break like I was expecting it to, so I jumped up and down for a little bit, until I realized that I was about to topple the whole thing over. I know, not a very professional test, but weight is weight.
4) I inspected the joint, there was no signs of delamination, cracks, nada.
5) Now here's the surprising part. I took off all the weights, and I was suspicious but satisfied that it was quite strong. Now imagine image 1 with fence post structure tilted 35% clockwise from vertical. I accelerated it into the ground glass edge first, while it was a few feet above the ground (into industrial carpet, no padding), with about the force that would be used when shoveling loose dirt.
6) There is tense "bonk" sound and the joint breaks cleanly on the metal to HXTAL interface. See images 2 and 3 for closeups of the failure.


The joint broke cleanly where I had scored the HXTAL. The only thing that is not 100% is the metal to HXTAL interface. The HXTAL bond appears to be clean, unlike last time where there was a lot of oxidized metal on the glass side of the breakage.

This leads to some questions:


1) It should be noted that there is no compressive force on the HXTAL, I just rammed it into the ground with no load. Was my test unfair in consideration of what the table would experience because of no load, where I am effectively "PEELING" the epoxy from the scored HXTAL (weakest) edge and proceeding to let it "PEEL" all the way until the glass is come undone? After all, it withstood all of the tests in TEST METHOD parts #2 and 3.
2) In light of passing method parts #2 and #3, should I be concerned with this kind of failure or would this be expected because with test part #5, I am expected too much force to be absorbed into the adhesive, where the force is concentrated on primarily where the glass edge meets the metal bar - since the metal bar will flex slightly under shock?

In regards to trying out other adhesives, I am still investigating and considering. Perhaps I should have been less hasty and anchored the glass to the floor and rocked the fence post attachment to see if it would break like that, but I did not so that would have to be another test. At the moment, it would be beneficial to have opinions from a structural standpoint if test method #5 is fair/unfair comparison, as a test for the table joints, especially with consideration that the glass (200 lbs.) will be compressing the joint and the width (6" each) of the frames will provide some rocking stability.

Many thanks,

Tom Fuhrman
06-24-2012, 02:33 PM
staying at a hotel in Amsterdam and all the desk tops are 3/4" glass and are held on to legs exactly as you are trying to do with an adhesive and it appears these have been installed for quite afew years with no severe problems. I would suggest you contact Bohle as they may be the ones who supplied the adhesives. Seems this type of joint from glass to stainless/or aluminum is very common in Europe.

Pete VanderLaan
06-25-2012, 05:36 PM
and just look at the mess Europe is in!

Matthew Leung
06-26-2012, 05:28 PM
Har har har, I'm guessing they'll need more than some UV glue to fix everything together. I guess I should consider myself lucky if that is all I need to fix my table then :D

I looked up Bohle, and it seems to point towards a variety of UV curing type adhesives. Invariably I will end up testing some, hopefully with a thicker viscosity to reduce run-off from glue joint.

The alternate glue suggestions are very helpful, but I am looking for an absolute sense of what is a fair test for my glue joints, as ultimately the UV glue joints will have to undergo the same tests to evaluate stability.

As I mentioned, I was able to apply static load of 420 lbs on the fence beam and a dynamic load of which part of it (my weight of 150 lbs) is dynamic. However this is where the force vector is mostly parallel to the glue joint. All it took to break the joint is to drive the fence post at an angle, with the HXTAL'd metal-glass assembly at the bottom striking the ground, causing it to 'pop' off.

That is a little unsettling, especially given that the force being driven was not very high (hard to estimate but not like 200 lbs going into the ground at an angle), but at the same time, I won't be having a giant pick up the table and smash it into the floor at an angle either!

Is this a viable amount of energy that a table could experience in this way? I don't know that much about adhesive engineering, but I guess considerations are peel strength and shear fracture strength?


Matthew Leung
07-04-2012, 04:01 PM
Does anyone know anything about forces on adhesives on this forum? Pete, you mentioned Quartering Shearing, I don't know what that means, I even googled it and it came up with squat.

An update: I was able to get in touch with technical sales at CR Laurence, and they were able to answer many of my questions I had about UV glue. They are helpful once you get past the sales rep and talk with someone that has working knowledge. I'll be placing an order with UV703 from CRL or with another vendor for Loctite 349.

They say also that the glass to UV glue bond is on a molecular level, but similar to HXTAL, UV glue cannot provide a bond on the molecular level on the metal to glue interface, apparently welding or similar can.

I will be running the same test procedure on UV glue as I did on HXTAL for yield strength comparisons. I hope it gives me the resilience I am looking for.

Happy 4th!

Pete VanderLaan
07-04-2012, 04:50 PM
imagine a square. then imagine the force coming at the square from a diagonal across the corner to the opposite corner. In architectural terms it's called quartering shear or a quartering wind. The force gets applied very differently than it would hitting one of the flat faces.

Lawrence Ruskin
07-05-2012, 10:04 AM
I've used an epoxy called 330 that you find in lapidary shops it's supposed to stay water clear and it does unless you expose it to the elements.

I store my glass up on the flat roof of my house 'cause I can see what stock I have at a glance.

I put some glued pieces up there a bunch of years ago and when I went to look at them after several years, the glue was flaking and it had turned yellow.It had lost it strength.

So when you glue up a piece of furniture you have to worry about heat and cold, sunlight, and the fact that the glass and the metal will expand at different rates.

Yes, I know UV glue sets up with sunlight. Yes I mix the heck out of the glue.
I don't trust glue when it comes to people's safety.

Pete VanderLaan
07-05-2012, 11:12 AM
Resistance to solarization is one of Hxtal's claims to fame. It has been stress tested out to 25 years with no discernable changes. Staying clear until it turns yellow is not acceptable.

Doug Keller
07-05-2012, 12:58 PM
imagine a square. then imagine the force coming at the square from a diagonal across the corner to the opposite corner. In architectural terms it's called quartering shear or a quartering wind. The force gets applied very differently than it would hitting one of the flat faces.

Try landing a light airplane with a quartering tailwind.

Matthew Leung
07-05-2012, 10:28 PM
Lawrence: The table is for personal use, for indoors, low humidity, air-conditioned.

What is the most cost effective way to fashion a UV lamp? I know they have industrial lights from several hundred to several K, but that is simply not an option, I will likely be doing this just once or maybe a few other times.

I was thinking of buying a nail salon gel nail curer that has 6 x 9W (54W) UV light fluorescent tubes. I think they emit a broad UV spectrum centered around 360nm and offer a coverage similar to my puck sizes I need to zap. Question: In terms of power, is this sufficient for gluing to spec?

Also, anyone got any opinions on Loctite 349 vs. CRL UV703?
Other than that it will fall apart one day, spontaneously combust the table, and make an uncontrollably expanding black hole in my living room! It's gonna be glued because that is the design concept, I just need to find my best bet.

HXTAL seemed great for glass, but without sufficient explanation from anyone about why an angled blow releases the bond from the metal, I have to move on to alternatives such as UV glue.

Thank you,

Pete VanderLaan
07-06-2012, 06:12 AM
Matthew, I can lend you the proper lamp for UV Bonding. I can ship it to you. Don't buy a light for just a few uses. They're expensive and it's really easy to get one that is not in the proper wavelength.

Matthew Leung
07-07-2012, 07:26 PM
Thank you Pete, that is incredibly generous and helpful of you. I will PM you.

Tom Fuhrman
07-09-2012, 08:01 AM
Just back from European trip and saw the joint connection you are trying to do everywhere. Even the coffee table in our room on the ship. From what I could see by looking at tyem , they might be using an adhesive double sided clear film to hold them together. A lot of the new shower enclosures are using this type of adhesive for all sorts of bonds. I think you need to check with some of the larger flat glass fabrication companies that do this type of thing all the time. I think CRL handles the adhesive films. I could pick up the entire table with 1" thick top that was bonded by a 2" circle in the middle and it didn't budges and these have taken a major beating in their environment for 5+ years. Table was about 3" X 1.5' in size. It didn't appear they were using an epoxy. The europeans are way ahead of us when it comes to flat glass fabrication and usage in architectural and furniture applications. All the drink tables in the theatre,i.e. 100+ were joined this same way.

Matthew Leung
07-09-2012, 03:03 PM
Hi Pete,

We were discussing thickness of glue line for UV glue bonding. (I wanted to discuss it here so everyone can have access to it.)

I was wondering what your feedback is on 5-15 mils MAX glue line. All bond surfaces themselves are flat, but I do not believe the glue line would be constant but varies slightly over the surface of the puck areas because of slight non-levelness between the 4 pucks. This means that at some point, there is zero clearance and the glass is probably touching the metal because it has to hold the table up:

Using a high viscosity UV glue designed for "thicker" bonds, is there risk of creating the ugly fern lines? What are some recommendations for safeguard against these artifacts?

By the way, would we be running into some potential issues with thermal expansion - not that I am expecting it in an indoors environment - since now the glue line is so thin with UV gap sensitivities?


Pete VanderLaan
07-09-2012, 03:43 PM
15 mils sounds like an awful lot. As to the high spot, everything has it. In the schott stuff, you can see the fresnel patterns at the high spots where there is virtually no adhesive. Not many surfaces in the world are perfectly flat. I think that Tom has had the good suggestion. I would look in to that tape.

Matthew Leung
07-09-2012, 05:15 PM
I did some reading on structural film adhesives (3M), they seem good for some things, but it seems to introduce a host of some other problems like applying 50 psi across all bond surface and then heating to 350F, but tensile and shear strength appear comparable, but cleavage is much lower. That being said, cleavage/peel strength isn't something listed on UV glue datasheets I've seen.

For now, I'd like to run strength and aesthetics sample testing on UV glue with induced gap imperfections to see if I can gain some confidence with it.

Practical advice about UV glue technique (artifacts, return of thermal expansion discussion since no spacers, etc.) instead of divergent solutions would be most useful to me.

I was visually estimating 15 mil gap earlier because I couldn't get my caliper between the puck and table. I was able to use sheets of paper to better evaluate clearance:

On 2 pucks, the clearance is <3 mils, I cannot fit any paper inside, they look nearly flush.
On 1 puck, the clearance is ~3 mils about 0.5" in, after which the paper gets stuck.
On the last (worst) puck, the clearance is about 6 mils (2 sheets) about half way in, after which the paper gets stuck.


Rick Wilton
07-10-2012, 10:25 AM
from what I've seen about the 3m tape is you need 15 psi and that is easy to achieve with their roller. You don't need 15 psi all at once so a roller or squeegee works fine.


watch this video about VHB


Tom Fuhrman
07-10-2012, 06:11 PM
Matthew: jump on a plane and visit some of the glass shows in Frankfurt, Hanover, or Milan and I'm sure you'll find the proper vendor that can address all your issues and has a lot of experience in making these type of metal/glass laminations. Check out the dates and locations in US Glass magazine. It will save you a lot of time and energy in the long run. Italians have been doing this successfully for many years.

Matthew Leung
07-10-2012, 06:36 PM
Interesting link, I didn't know the applications were so diverse. Maybe one day it will come down to experimenting with the adhesive films, but for now I will try my luck with UV adhesives and see if the samples past strength testing.

Pete VanderLaan
07-11-2012, 05:14 AM
50 PSI is easy to apply with a conventional roller for formica

Matthew Leung
07-25-2012, 01:31 AM
Hey guys, I just wanted to follow up with my conclusions on UV Glue tests:

I ended up testing with Loctite 349 and CRL UV703. To simulate joint imperfection, I added a 2 layers small swatch of packing tape to an area close to one edge of the joint. It introduced about a 8-9 mil gap on one end. For curing, I used a 56 watt 365nm gel nail curing device and was pleased to find that this worked perfectly well for curing. I let it bake under the light for 20+ minutes for extra measure.

Using the same test methods as before (fence post with the metal + glass bond sticking out attached at the end like an "L"), I found that the UV glue provides more than adequate static load strength like HXTAL did. But also the increase UV glue flexibility made the joint structure more resilient to hard, brisant raps. It was so resilient I broke the fence post apart first (on Loctite 349 sample), before the joint failed!

I tested the CRL UV703 first, and it took a good 8 strikes into the ground to delaminate the glue. I think it is good that the failure was not sudden like with hxtal and compromise of the joint integrity was both tangible and visible over the last few strikes before complete separation. Glue separation occurred on both the glass and metal upon joint failure.

Loctite 349
Similar results were obtained with Loctite 349 except that it took more energy to break the glue joint. Looking at the joint remnants, they both appeared as semi-flexible films. It was noteworthy that Loctite 349 has much better peel adhesion on my samples than CRL UV703 (meaning harder to pull the partially failed glue film off both glue substrates metal and glass).

Also, I think taking advantage of higher viscosity property by creating a fillet surrounding the chamfered edges of the puck increases peel resistance, which seems to be this type of joint's greatest weakness in my tests.

I will be going with Loctite 349 for finishing my table. Thank you all for your helpful input and advice.


Eben Horton
07-25-2012, 09:17 AM
im not an expert on adhesives, but i think your testing is a little flawed. In my limited experience, glue joints fail over time and not right away. UV light or other variables can break down a glue joint and it will simply just give way with no explanation.. even if you tested the strength of the joint a day or 2 after you glued it.

I dont think you will truly know what glue is best unless you are willing to spend a few years watching how the glue ages.
Again, i could be wring here.. but consider what i am saying.

Matthew Leung
07-25-2012, 11:54 AM
Well, I am testing for initial bond strength for certain characteristics that I am concerned about such as peel resistance from rocking and occasional bumps - at much lower forces than I am introducing with the destructive testing I was conducting. This explains my tests. As I have mentioned before, there is nothing special about where this table is going: it's air-conditioned and low humidity, no direct sunlight and in general pretty little light exposure.

As for longevity testing, Loctite 349 specifically suggests bonding metals to glass and show some joint strength over time and environmental aging. I rely on these figures because in reality, one has to stop the R&D at some point and finish the project.

Do you have evidence to suggest that I should be concerned specifically with metal to glass bonding with Loctite 349?


Allan Gott
07-25-2012, 09:34 PM
Don't let a horse sit on it

Greg Vriethoff
07-26-2012, 12:54 AM
While I don't have any "sage" advice to offer either way, the only thing I'd like to add is please continue to share whatever results are bore out from this endeavor. I appreciate the time and trouble you have taken to present this issue, your methodology, and the results you have presented here.

The course you have chosen may not be the "solution", but you have come by it through hard work, perseverance and humility.

I applaud you and your efforts.

This is what (I feel) this forum is about. Sharing our experiences whether they be successes or failures.

Even if five years from now a horse sits on your table, and breaks the joints, you have done your due diligence to resolve this issue to the best of your abilities.

Thank you.

Matthew Leung
07-26-2012, 03:48 PM
Haha Allan good advice! There's no telling, but I'll try to keep the horses out.

Greg, I'm glad to help build a resource that others can collaborate and benefit from as well. Thanks all, for giving me good leads and technical advice to come up with an reasonable solution.

I'm probably doing the glue up on the actual table this weekend, so I'll let you guys know how it goes.

Matthew Leung
08-09-2012, 05:05 AM
Some things with work came up so I didn't get a chance to prep and glue the table up until today. The UV glue cured as expected and behaved like during my tests so there were no surprises. I was surprised though at how fast the moisture reabsorbed into the glass and metal so the race to getting the adhesive on all 4 pucks and lowering the hydraulic/glass assembly was an adrenaline pumping activity.

I decided to sacrifice some aesthetics for joint strength by allowing a fillet to bond on the metal puck's chamfer and the glass surface (instead of cleaning it off before the full cure).

I am happy that the joints seem very robust and optically free of artifacts. I'll try to post some pictures of the finished product when things calm down.

Thanks again for all of yall's support and direction.