Fired up heat bed

Just completed the first test power up of an aluminum heat bed I fabricated for my new printer. I'm testing this through a Variac 120v AC transformer so not to over heat the resistance wire until I get the power load sorted out. I strung 16 ga. nichrome wire through 7 ea. 20 X 80 V-Slot extrusions. I insulated the wire from the aluminum bed by encasing the nichrome wire in PTFE tubing. Worked very well. The V-Slots are internally pinned together down the length of mating joints and then capped on each end with overlapping aluminum angle screwed to the ends of the extrusions. Eventually I will position the bed and run the wires and thermo-coupling through the stainless table top so everything is mounted out of sight.

I used an oven door glass I scrounged for the build plate, but its not full width of the bed. I'll get a correct sized piece after I get everything running properly. It took 30 minutes to heat the bed and glass to 100°C at 120v. This is a 20lb chunk of aluminum so I'm not surprised about the time. The wire measures 475" @ 10.9 ohms. That's approx. 1309 watts @ 120v...drawing 10.9 amps if I applied the web formulas correctly. The wire itself reached a maximum temp of 175°C which is about 100° below the max work rating of the PTFE tubing....the melting point is much higher so I think I'm safe with the design. For faster heating, I'd like to increase the watts to about 1800 @ 12.75 amps using 140v to 150v. Does anyone know if that's possible using a 2000 watt continuous duty step up transformer? I don't want to go to 220v mains and step down.

The printer itself is incomplete as you can see....waiting on printed gears for the XYZ synchromesh cable drives to arrive before adding the gantry assembly, steppers, and wiring. Still haven't decide what micro-controller and drivers to use....any thoughts? The_Digital_Dentist suggested these heavy duty drivers Stainsmart.com that I like for parallel connecting the dual Z and dual Y axis motors.

Thanks....I appreciate any input.




Edited 1 time(s). Last edit at 05/11/2015 11:10PM by simspeed.

Wow, that´s a huge table!
Do you need to heat up the whole table all the time?
I´m pretty sure the resistive wire hung loose in the middle, while it was hot? You might need a support structure there to have constant distance to the glass.
Or you could split the heating element in two pieces (or even four ) and heat up only a small part of the table for smaller parts.

The ovendoor glass might not be the right choice. I´m sure they are built to keep the heat inside the oven. ( e.g. infrared filtering/ semi reflecting mirror )
Wait until you get the real glass, until you do some power calibration. I have seen false data readings from an IR-temp probe, because of the reflecting glass .

A 230V triac ( dimmer) can usually handle 2000W resistive load or more.
Remember to put a controllamp parallel to the heat element, so you can´t forget to switch it off.
-Olaf

Hi Olaf... The wire hanging down in the middle is the thermocouple screwed into the bottom of the bed. I don't have it hooked up to anything at the moment. Eventually it will drop down through the table top. The nichrome wire is securely encased inside the aluminum extrusions looped back and forth at each end to make the circuit where the two ends of the ptfe tubing make the orange extension cord connection.

The glass sits flush on top of the aluminum extrusions. You are right about breaking up the heating element into multiple loops to heat portions of the bed. I may eventually build another bed that way. I'd have to turn the extrusions 90 degrees to make that practical. For now I just wanted to test the loop and ptfe insulation design to see how well it worked. I'm pretty happy with it. You may be right about the glass too....although this doesn't appear to be reflective glass. Actually I'm not sure its even hi-temp glass. I heated the bed up with out the glass and then with glass and the temps above and below were essentially the same. I'm using a K type thermocouple that's threaded into the aluminum from below so the glass top is not a factor there.

Yes, I will wire everything up neatly. I have some high amp illuminated rocker switches on order for the bed and printer power supplies to show when power is on or off. The stainless table top has about 6" clear at one end of the bed...that's where I'll cutout for switches and controls. I may add an end plate on the upper frame that hinges up for the controls. I'm not familiar with the dimmer....I'm hoping someone can advise if a step up transformer is feasible or not. Thanks for your comments.

Edited 1 time(s). Last edit at 05/12/2015 02:41AM by simspeed.

If anything goes wrong with the temperature control the heater wire will melt/burn the PTFE, possibly producing unhealthy gasses. You want something made of glass fibers like the insulation they put on thermocouple wires. Try something like this instead: [goo.gl]
It might be a lot less expensive to wrap the nichrome wire on pegs then insulate it between two sheets glass fiber cloth - no danger of anything burning/melting.

You're going to a lot of trouble/expense to make a big machine- I wouldn't use printed gears or anything else (except cable clips) in it.

What size extruder nozzle to you plan to use?

Hi Dentist....thanks for your comments. I've attached photos showing how the wire and insulating PTFE are looped through center holes in the extrusions. Your suggestion for the glass fiber sleeve would probably work using the larger hollow voids instead of the smaller tap bores.

I agree overheating the wire would melt the PTFE... but in this application, the total resistance of the wire in length and gauge is greater than 120v mains power can heat to melting temperature. So if I don't input power significantly greater than what's available through a standard 115v outlet it should be good. As it stands now the max temperature the wire reached at full 120v was 175°C. The question is what is the power limit where the wire heats beyond the PTFE melting point? Much greater it appears.... thankfully I don't need that kind of heat or power input for this application. I understand you are an electrical engineer so please correct me if I'm wrong.

Yes you are correct, this is not a cheap machine to build, but I have specific need for a print envelop this size for a separate project. Building this one, rather than buying a comparable machine or contracting out a big print job, is the most economical route for me.

I'm having the gears printed on a Stratasys machine because none are available that fit this configuration. My print guy emailed me last night that the synchromesh cable fit was perfect and the parts will be ready to pick up today. If they work out as planned, then I'll cast stronger ones in polyurethane for production work. What kind of "cable clips" are you referring to? I'm not sure what that means.

For the big prints I imagine I'll use .8 mm or even 1 mm tips on a Volcano hot end or something like that. The print speed for this design won't be great, so a larger output volume will be needed to compensate. Any thoughts on the step up transformer question I posed? Thanks...TP.

A 2 kW transformer will be very heavy- maybe 15-20 kg. Then there's the issue of supplying power to it. Most domestic wiring circuits may not have 2kW without heavier gauge wiring and a bigger breaker in the box. You may have to go to one of the higher current outlets/plugs, too. something like this: [www.ebay.com]

You would connect the secondary to add to the line voltage- see [www.emersonindustrial.com]

Make sure you ground things properly!

Thanks for the feedback Dentist....pardon my ignorance; I was thinking of something more like this for the step up transformer. 2000w step up transformer

I don't see why that wouldn't work. I'm amazed they can do it so cheaply.

Would the use of a ground fault receptacle mounted on the printer chassis work as a proper ground and circuit breaker in the event of a short at the heat bed? I see the need to add a ground wire from the bed chassis back to the GFCI to connect up to the ground wire in the power cord back to the wall receptacle? (not shown) Thanks...TP.

Hi Simspeed ... I think this is a really interesting way of building a large heated bed. Thank you for sharing it. I need to build a heated bed of a similar size. I'm thinking of putting a thin PEI (or glass) sheet directly on top of the 20/80. So I'm just checking if you have any further advice from your experience with this design.

How accurately flat have you managed to get your bed, and does it maintain its flatness when heated? By my calculations, building a bed out of 20/80 V-slot like this weighs about the same as 1/4 inch aluminium, but is about 14 times as stiff and presumably less prone to heat warping too.

Does the nichrome heating wire give a consistent temperature across the bed surface? How do you think the nichrome wire heating compares with other options such as a silicone rubber heater on the underside?

Do you think the bed would still heat up well with the nichrome wire running through the larger holes through the V-slot? (.... as the nichrome would basically be heating the air in the holes, and then the hot air would be heating the aluminium 20/80.)

Also, you mentioned that "The V-Slots are internally pinned together down the length of mating joints" - how did you do this?

Edited 1 time(s). Last edit at 07/21/2015 11:29PM by SydneyDesigner.

Hi SydneyD... Thanks for your comments and questions. I'm planning to lay glass from an oven door on top of the aluminum bed for starters. You can see it in the photos above. Since building and testing the bed shown, I've put it aside while I work to complete the rest of the printer. This is my first build so I'm taking it pretty slow to be as sure as I can that the motion drives are designed and built well enough for the stress loads a Cartesian printer this size will likely generate. So to your questions here's where I am:

• I plan to check the bed flatness with a dial indicator once the XYZ gantry is installed and powered up. I added 4 adjustment screws to get a an initial bubble level. 3 screws would be easier and more accurate most likely but I’m unable to center a screw on either end of the bed because of the tight nichrome wire loops underneath the header covers. If I build another bed using this design, I’ll probably run the extrusions crosswise rather than end to end. That way I’d be able to zone the bed to heat smaller areas rather than heat the whole bed every time I want to print. That will be considerably more difficult given the additional end loops and extra extrusion pinning. The bed actually bubble levels pretty easy but we know that’s the easy part. The bed feels really solid, but I wouldn’t want to bang it around for fear of movement along the extrusion joints.

• The bed takes about 20 minutes to come up to temp, but once there the heat reads pretty uniform everywhere. I tested it 4 times cold to hot and back again, and it repeated readings consistently each time. I tested with a temp gun and a thermocouple mounted to my Amprobe meter. I’ve never made or used another type of heating element; so I have no reference to compare to. I originally planned to lay the nichrome wire in the open v-slots and pour a silicone surface pad, but once I looked as the internal structure of the 8020 I recognized it would be pretty simple to route the nichrome through the channels to get an even distribution and keep everything clean and insulated inside the PTFE tubing. Turned out about like I expected thankfully.

• I considered that option first, but decided the use the smaller core holes because the PTFE and the 16 ga wire fit perfectly. I get direct contact to the aluminum through the PTFE so I think that generates quicker heating than air contact. It also allows easier routing of the end loops in my opinion.

• Pinning the extrusions together along their length occurred to me while building the aluminum frame onto the stainless table. I used the same technique in several locations to tie everything together. T-slot type extrusions us threaded nutserts down in the channels to clamp fixtures and such to the surface of the extrusions. I simply reversed the nutserts and doubled nutted them back to back using 5mm hex setscrews. You can adjust the space between the nutserts tight enough so that they clamp the two extrusions under tension when installed at either end of an extrusion pair. I then took a long ¼” metal rod and tapped the double nut down the joint spaced equally along the length. It’s a tedious process…but it works well. Easy to get the clamp too tight or too lose so takes a little patience.

So far I’ve completed the build of the Y axis, I followed Digital Dentist’s recommendation and mechanically tied both sides of the gantry legs together using a closed loop synchromesh cable routed under the stainless table top. I’m using two 2.4Nm Nema 23s, one at each end of the loop on either side of the frame.

I fabbed up the gantry itself for the X and Z axis and finished rigging the cable and motor for the X. Again I followed DD’s recommendations and decided to tie the left and right Z motors together using a cross shaft throught the gantry header. I’m using synchromesh cable there again for the linkage to a pair of 0.9Nm Nema 23 motors. The single X motor is a 59Ncm Nema 17 again using synchromesh cable for the hotend movement.

Bearing Blocks for the cross shaft are being printed locally by a guy I found who has a Staysts printer and CNC machine equipment in his garage. I am envious of his setup. Those pieces should be ready later this week. He also printed synchromesh drive pulleys for me for each of the 3 different motor shaft sizes I’m using. We’ll see if they can handle the cable stress once the build is finished. If not, I’ll get him to machine adapters so I can us the factory 5mm ID pulleys from SDP/SI.

I mounted the PSU, external drivers, and smoothie board to the frame this weekend. I plan to wait till I have the axis working before I bring the heat bed back and start wiring up the SSR and mains power for the bed. I’ll post updates for the build once the gantry is mounted. Best of luck to you on your machine SydneyD…. 

Edited 1 time(s). Last edit at 07/22/2015 12:35AM by simspeed.

Thanks for your reply Simspeed - that gives a better picture. What cartesian design is your printer based on? Sounds like you are close to getting it printing? (I'm building a CoreXY and still a way off finishing it.)

I'd be wary of heating partial bed zones as the differential expansion of the heated part would be constrained by the cooler parts, and that might cause the bed to warp - perhaps permanently. What do you think?

Another option for bolting the bed together is to bolt the seven 20/80s onto a couple of 20/20s. Then the precision faces of the 20/20s mating surfaces should keep the 20/80s flat.

My design turned out very similar to the 3DP Unlimited large format printer. The XYZ movement is all in the gantry carriage with the bed fixed to the stainless cart I bought for the project. I dropped my extrusion parts off for machining today and looked at the finished bearing blocks as they came off the Stratasys printer. Man the detail from that printer is perfect and the parts fit perfectly.

I'm not sure the differential expansion would be so great in that context. The extrusions are clamped together at contact surfaces, so the heat would spread from the heated part through to the non-heated parts uniformly, but dissipate the farther the heat travels from the source. The end effect may be that it takes longer to reach temp or may never reach temp at the desired zone. I'm just guessing of course. Yes indeed using a couple of transverse 2020s on the bottom to flatten the 8020 bed would be a wise move. The end headers are made of overlapped aluminum angle but they don't provide the rigidity the 2020 would. I may just take your advice and add those to the bottom.

I bought some insulation board to cut to fit to fill the air gap between the heat bed and the stainless table top. I'll wait until I remount the bed before I install that. I bought heavy duty aluminum foil and spray contact glue to wrap and seal the insulation board after sizing for the installation. The air gap below the bed was non-insulated during the heat test so I'll likely get higher temps and quicker time to temp when finished.

Here are a few photos of the build as it stands today.

Ah, thanks for the pics - now I understand what you are building (and why you need the trolley!). Exciting project. Are you planning to add any form of heated enclosure?

The design doesn't lent itself well to a full enclosure. I do have a way in mind, but doubt I'll try to implement it for this machine. I've drafted plans for an altogether different type of machine that will get a full case enclosure with a lifting lid and double swing open doors. I really love working with the v-slot aluminum extrusions. Doors with magnetic latching, like I've used on the end flap doors with this machine, are easy to fab and work flawlessly. I plan to use this machine to make component parts for the new design that requires a large bed like this one to build. I don't want to show that machine design yet....but here's the enclosure design with clear panels rather than insulating panels that will be used in the actual model. This accommodates a 609 mm X 2134 mm bed with 1 meter Z travel. It won't be Cartesian...

Hi again Simspeed - I was just wondering how things are going with your 80/20 bed design as I'm close to building mine. Any latest suggestions/modifications?

Hi,

I just come to this old topic.
I'm amazed by the huge energy consumption of the heatbed : 1800W !
The bed is big but I thought an AC heatbed should heat that up with less power.
How do the bed works now ?

++JM

Quote
SydneyDesigner
Hi again Simspeed - I was just wondering how things are going with your 80/20 bed design as I'm close to building mine. Any latest suggestions/modifications?

Hi Sydney...the bed design was the best part of the build unfortunately. After putting the motion components in place, I realized my worst fears about the mass and weight of the gantry using cable/belts rather than acme or ball screws. Just didn't work well enough to continue. Rather than convert the YZ axis to screw drives, I redesigned the rig to a Core XY layout using the c-channel v-slots for each corner. The bed will remain stationary as planned but the Core XY motion platform will move along the Z axis suspended at each corner. That design uses a lot of printed parts for the corner connections, rail sliders, and such, which proved too costly to contract out. So I'm currently collecting parts to build a delta printer that I'll use to make the many printed parts for the Core XY machine. Convoluted isn't it... . Doesn't matter, I'm having fun designing and building to suit my goals. Eventually one day I may actually produce a working machine....LOL.

Your suggestion to under brace the 8020 bed across the section lengths for flatness was a good one. I'd recommend just that to anyone considering a similar design. You will need a flat surface such as tempered glass on top of the bed to print on. Best of luck with your build.

Quote
J-Max
Hi,

I just come to this old topic.
I'm amazed by the huge energy consumption of the heatbed : 1800W !
The bed is big but I thought an AC heatbed should heat that up with less power.
How do the bed works now ?

++JM

Hi J-Max....I have yet to increase wattage to the bed from the initial build so can't comment past that. My only experience with wiring and electrical loads comes from this build, so I'm not knowledgeable enough to offer any real understanding beyond what the tests above produced. I guessing obviously, but the resistive load over 475" of 16ga nichrome wire was absorbed within the mass of the aluminum bed such that high wattage is required to reach the temps shown. What kind of wattage does it take to heat a cooking griddle to plus 100c? Thanks...

Thanks for the update Simspeed. All the best for your new design. Please post when you get it going.

The end goal is supposed to be a working machine? Now I know what I've been doing wrong