24V: The other way

[Anton]

Hi all.
I've got Max v3, and just like everyone, I have problems with a heatbed - it's definitely not enough powerful to print ABS, it hardly reaches 100, even after I tune up the voltage to 14.5.
I've read how people make RAMBo run on 24V. It's fine for everything except heatbed, it's fused to 15A, so the voltage is basically limited to 14V.
The common solution is to use a solid-state switch for that. The shitty ones heat a lot, and if they burn out - they leave a bed in ON position, which can easily cause fire. The good once are... EXPENSIVE (>$100). And you need two power supplies - 12V and 24, which is also terrible.

I want to try (theoretically) two other ways.

1) What if we use only 24V supply, and power up the entire board with it. BUT we can put a 24V->12V DC step-down converter near the heatbed. So the board outputs 24V, up to 15A, which is converted to 12V 27A (considering 90% efficiency), which is right what we need. It will require a good converter with a good fan, but even if it burns out, it will just stop the power supply.
Probably the PWM of the bed can be a problem for the DC-DC converter. Is it? Can we deal with it somehow?

2) What if I order another Onyx heatbed, connect it to the first one consecutively, doubling the resistance, and both of them together under the glass (I think they will fit in). So if I also double the voltage (to 24V), it will result in the same current and double heat generated! Well, it's not really efficient, the upper heater will shield some heat from the glass, but I think it can still make a big difference almost for free. At least it should easily reach 110!

[Xenocrates]

To be clear, there is a perfectly fine cheap SSR for this, the Auber Instruments 100A DC-DC SSR.

Method 1: The 24V will output enough current to the hotend to smoke the hotend whip connector in short order, as you are greatly increasing the power dissipated in the contacts, which are already amperage constrained. Not only that, but the heated bed will still pull the same amount of power, as you have the same voltage and resistance, which is what determines current. So it does nothing and sets the whip for the hotend on fire.

Method 2: While this will theoretically work, you would need to figure out a good way to connect them in series while they are stacked, which is somewhat difficult as all the contacts are on the bottom side, and the thermistor would need to be insulated from the other board, possibly by slightly enlarging the mounting hole and using an E3D cartridge thermistor as the sizing is close. Then the boards would need to be joined using thermal interface tape or thermal paste/epoxy (At your option). Considering the wiring in-between, there is a significant amount of space to bridge, and a custom machined copper or AL interface would be about the only satisfying way to connect them thermally. Also, a second heated bed is ~70$, and even if you had to buy a 100$ SSR instead of the 20$ one we pretty much all use, you would end up loosing the cost benefit in thermal interface materials.

I'm not going to say it's stupid to come up with outside the box ideas like this, but good on you for running them past us online before trying it, and check the documentation (And your math on electrical stuff), a bit more thoroughly before buying anything. Either of them would have given you nasty times trying to make them work.

[Anton]

Hm... Agree on method 2, it will be hard to join things together.
About method 1: probably agree on hot-end, it might be not 12v tolerant, but disagree on heatbed part - the board will run on 24V, it can give up to 15A to heatbed. Then the DC converter gets this 24*15=360W of power, will loose 10%-20%, and the remaining 300W will go to the heatbed. The idea is to increase a voltage on RAMBo, so the current can be lower, the board wires will not overheat, the fuse will be OK.
And the same can be done to hot-end.
But... OK, still it doesn't solve the problem of PWM to control the output. I'm pretty sure step DC-DC converters will not be able to work, if the switching frequency is close to the input frequency, generated by RAMBo.

[Anton]

Thanks for the link of useful SSR, by the way!

[Xenocrates]

Yes, more power will be available to the heated bed. However, the heated bed is ~1.1 ohms. At 12 volts, it will draw 10.9A no matter how many amps there are. It's why a lightbulb that is rated for 60W doesn't blow up when you put it in a 100W fixture, and why a computer can use a 1KW power supply when it only draws 200W. You need to increase the voltage or lower the resistance to increase power.

Given:
P=I*V and I=V/R, thus P=V*(V/R), thus P=V^2/R. Note that amperage of the supply can be factored out of the equation for power entirely, and that it is wholly dependent on the voltage and resistance.

[Anton]

Ah... sure. Sorry, I was drunk and sleepy yesterday
It will obviously not work as I described. Thanks for the link once again!

[DeltaCon]

I would not understand why anyone would go an untested route while a fully tested and much used alternative is available. There is really no need for a second PSU when you already want to go the stepdown route. Just stepdown your 24V PSU to 12V and run the board from there. Then you have 24V available (please turn that down to about 20V to save your Onyx) and you run that through the not so expensive Auber as Xenocrates suggested. Well tested and a certain way to get things going.

Another remark: If your bed can keep hold of 100C there is really no need to go 24V. Why would you need more than 100C temp? You are not boiling eggs on your printer are you? Ofcourse it is very luxurious to be able to reach 100C in 4 minutes rather than 45 minutes....

There is a great thread available about this subject, with a lot of input from Xenocrates and other Electric Fanatics like him (don't count me into this!). Suggestions for the to-buy-stuff are in there: http://forum.seemecnc.com/viewtopic.php?f=43&t=7766

[gchristopher]

I would not understand why anyone would go an untested route while a fully tested and much used alternative is available.

Fun! We're all using DIY/amateur machines and trying out new things here. True, 99% of the time my ideas aren't as good as what someone else has done, but each time I learn something interesting that I wouldn't have if I'd just copied someone else's solution.

Another remark: If your bed can keep hold of 100C there is really no need to go 24V. Why would you need more than 100C temp?
.... to be able to reach 100C in 4 minutes rather than 45 minutes

I see a difference in ABS adhesion on PEI between 100C and 110C, especially for prints where corner lifting is a concern. For the warmup time, don't forget the dish towel or piece of styrofoam on top of the bed to trap the heat in! That makes a big difference, also.

[Xenocrates]

I'm going to agree with Gchristopher. While Mike and I have put serious effort into making 24V conversions a well documented modification, there's still room for improvement. For example, Mike worked out which SSR would work best for our application, and looked into things like wire derating for temperature, but Jim (I think) added the novel bit of with an external 24V supply, having a little relay to switch between 24V and normal power, then I added a safety cutout relay where the bed current passes through a mechanical relay (On both the 12 and 24V circuits) rated for ~25A, with a circuit that goes through one or more thermal fuses embedded in the snowflake. At each step, someone thought to themselves "I could do this different", or "You know, this would be handy and easy", and in each case, things were made better for everyone. I bet there are a few more improvements that could be made to the bed assembly. For example, Mike has pointed out moving the FSR's outwards improves the sensitivity and bed stability, while others have been working on insulating the underside further. I myself will probably combine those with a PT100 based bed sensor and a copper heat spreader, if I ever get off my ass and build a printer of my own design.


Additionally, for crazy folks like me, the higher bed temperature for things like carbon filled Polycarbonates helps a lot.

[gchristopher]

Related to the 24V upgrade and making up your own solutions: the dinosaurs haven't posted pictures in a while, but they're making great progress on moving the 24V power supply to an external metal enclosure. Since the LED-power-supply-style wiring doesn't fully enclose the AC leads in their own grounded container, this brings the entire printer closer to best wiring practices. It also cleared up a lot of space in the interior, and gave them a chance to playing with making their own power cable with anderson powerpole connectors. That was a positive learning experience. Powerpole connectors are great.

[img]http://gchristopher.net/3dprinting/powe ... review.jpg[/img]

[DeltaCon]

I agree guys, I am sorry, my urge for safety overtook my curiosity for new thing!

[Anton]

Why I need a better heat-bed:
1) As gchristopher pointed out, ABS needs 110, sometimes it's essential.
2) Waiting for 30 minutes to reach the temperature, having to cover the bed with something every time - that's a shame.
3) Printing small detail in PLA, or printing a huge detail where you need to fight warping/curling - when you need fans on 100%, cools the bed down to 50 easily. Not being able to keep the temperature during print - disaster which can't be fixed by simple tricks, you just need more powerfull heating. But tuning the input voltage to 14.5 helped significantly.

[3D-Print]

I love my 24V upgrade (http://forum.seemecnc.com/viewtopic.php ... 125#p86387). Bed heats from room temp to 100 degrees in just under 3 min. Once at temp it is money stable with the SSR and calibration. Without a question a great upgrade.

I do worry about fires? Yes!! Thus as mentioned.... I'm currently adding a number of thermal breaks throughout the printer.