24V PSU for printer

Hi,

I'm constructing a fairly large-scale printer with and trying to power it with 24V. The issue that I ran into is the custom-made heated bed I have draws about 450W of power, which means I need a higher power rating for everything to work. Can't find a 24V led-strip-type psu rated for 720W at 110V anywhere (only 240V and I'm in the USA). Can't easily use an ATX psu since those are typically 12V, and I've been told it's not a great idea to try to wire power supplies in parallel without special control circuitry.

I was wondering however that since there are 2 sets of power wires going into a RAMPS board from the psu, if I couldn't simply use 2 different power supplies, say 2x 360W 24V supplies, and have those connected to the different RAMPS inputs. I believe one input powers the electronics and hotend, while the second is responsible solely for heating the bed (had that one disconnect and printer continued working just fine, only the bed temperature dropped), so each psu will essentially be feeding a different circuit, one the bed and one the extruder, motors and the rest of the electronics. Are there any problems with having a setup like that?

And yes, I'm a beginner at electronics so please easy on the bashing if I am missing something obvious.

My advice is that you read as much as possible on the subject, and ask for help from people who are quite knowledgeable and experienced in the proper forums.

For example, if you need help for the RAMPS, ask in the RAMPS forum, but before you do that, read the pages about the RAMPS board in the wiki.

Also, I would honestly be worried about the risks involved in dealing with your custom heated bed, so I would ask for help in the Safety forum.

Finally, I would post pictures and links to the specifications for your heatbed and other components of your "fairly large-scale printer", right now your description is too vague for anybody to be able to provide any concrete help.

Jameco has 24V Mean Well 600W and 750W supplies that run from 120VAC. My SE-600-24 set me back $90, and you can also find used/pull bargains for the SP-750-24 on eBay from time to time.

In any case, I think that you may want to check your math about the maximum capacity of the RAMPS heatbed MOSFET. You should probably be using an external SSR/heat sink to control a large heatbed.....

vreihen is right about the ramps board, you are needing over 20 amps for the bed to operate and I doubt your ramps board can handle it.

Thanks for the feedback,

@AndrewBCN: A heated bed is essentially a big resistor, so I doubt that are any hidden dangers aside from that. Mine has a resistance of 1.3 ohms and is rated for 24V (thus about 450W), it's just a resistance wire encased in silicone. Other components are an arduino mega, ramps 1.4, and 5-6 fairly standard nema17 motors for the axes+extruders, as well as 2 standard hotends powered by 24V ceramic heaters. All that should be ok with about 360W of power if the bed is factored out. If something more specific is needed, I can provide the specs but I don't see how it would help at this point.
I think my question is essentially about the soundness of the electronic arrangement I suggested in general (2 power supplies to the 2 different ramps channels). I did read about the ramps, but couldn't find an answer to that and don't have the knolwedge yet to figure it out myself, so I asked. Thanks for the suggestion of asking on the ramps forums, I'll try that.

@vreihen & ggherbaz: from what I found the ramps MOSFETs are rated at 55A each (must be one of them that powers the heated bed), so that doesn't seem to be an issue unless I'm missing something else. I certainly could see a possible issue with the setup in general since if the resistance jumps at the screw terminals that much voltage could fry them pretty easily, but at least I couldn't find a current limitation as you mentioned for the heated bed on the ramps.
I'm definitely willing to look into an SSR for the heated bed if that turns out to be necessary, but it'd be nice if I could use the available circuitry in the ramps.

Edited 1 time(s). Last edit at 04/02/2015 03:02AM by anvoice.

@ anvoice

You are still missing quite a few points, that was obvious with your first post and the situation has not changed much even though vreihen, ggherbaz and I are trying to help here. What worries me most though is that you don't quite realize the danger involved when trying to deal with 24V 20A PSUs without proper knowledge, so I'll repeat my advice that you ask for help from people who are quite knowledgeable and experienced.

Good luck.

Every RAMPS board that I've ever seen has a 5A input for the heaters/motors and an 11A input for the heated bed. The voltage input terminals and yellow poly fuses standing proud next to them are all sized accordingly. It has been reported that several Chinese RAMPS boards have capacitors that are only rated for 17V, and running those at 24V will let the magic smoke out.

The input terminals are fragile even running them under normal conditions. This is a picture of a Sintron RAMPS board from an Amazon review:



Like I said, use an SSR with anything larger than a standard Mk2 200x200mm heatbed.....

@AndrewBCN: as I mentioned, I'm a beginner so I'm sure there's much I don't realize. I tried providing extra details as you suggested. If more info is needed, please specifically mention what I'm missing. If you just wish to bash me for not "improving the situation" without saying what I should improve, I would suggest stopping here. Other than that, thanks for your suggestions.

@vreihen: seems to make sense, my ramps 1.4 terminals for the heated bed got somewhat charred at some point due to poor electrical connection = resistance jump I believe, and that's a Prusa i3 bed. Would this particular relay work? Looks like it's close to the specs I need, 25A (more than bed should draw at 24V). If I'm understanding this correctly, I'd connect the ramps heated bed wires to one side, and connect a second power supply capable of giving the required 450W to the other, while connecting the thermistor wires to the ramps. Then the lower current ramps signal would switch the SSR and allow a larger current go flow to the bed. Is that the gist of it?
The Jameco 750W power supply you mentioned looks good, but that wouldn't solve the ramps board limit problem, and I don't really want to pay $240 for a psu... The 600W would leave me with only 150W to run the motors/hotends and everything else, which is pushing it.

Edited 2 time(s). Last edit at 04/02/2015 05:57PM by anvoice.

That's the same SSR that I'm using. Just make sure to get the heat sink for it! Also, beware the counterfeit Fotek parts eBay:

http://forum.seemecnc.com/viewtopic.php?f=43&t=7161

Here's a good thread on SSR's and wiring them:

http://forum.seemecnc.com/viewtopic.php?f=43&t=7766&p=66815&hilit=ssr#p66815

That's great info, thanks! Won't forget the heatsink. Also found another good article on the topic here:
http://3dprintboard.com/showthread.php?7167-Clarifying-i3-i3v-heat-bed-and-heat-bed-relay-wiring

About the fake SSRs: I actually can't find any of the "real" ones anywhere, assuming all the ones without the beveled bottom right corner are fake... The ones on amazon, newegg, etc., and all the ones on ebay all have the same square corner
If you have a good reliable source for these, I'd be glad to hear it. Otherwise I might have to go with one from a seller with good feedback on ebay for lack of alternatives, and monitor it closely for a while to make sure it doesn't start a fire.

Would there be any downside to using a relay rated for a higher current (e.g. 40A) instead? I don't think the power FETs should care much about providing a smaller current than maximum, but would it heat more or something of the sort? I feel it could provide a safety cushion in case the SSR turns out to be fake and has subpar specs. Doesn't cost that much more for a 40A than a 25A one.

Edited 1 time(s). Last edit at 04/03/2015 03:51AM by anvoice.

Suppose the only difference between the 25A and 40A parts are the stickers on them? Not saying that the pirates have done this, just that I have a 100A SSR with screw terminals small enough that even a blind UL inspector on his first day at work would not approve for that current.

I read somewhere (on the SeeMeCNC forum?) that someone had found a different brand SSR that worked well and had a traceable pedigree back to the manufacturer. My Fotek SSR was sold directly by Amazon (not a third-party marketplace seller), and I'm still 50/50 on whether it is genuine. I intend to replace it with a non-Fotek part one of these days for peace of mind, but for now I have it mounted out in the open with nothing around it in case it burns up.....

Think I'll go with the 40A one just because the ship date's quicker and the price is half a dollar more. I'm willing to bet it's a clone due to the low price, but I just can't find a dc-dc relay anywhere else. Will take a chance and see how it goes. Judging by reviews on amazon of this SSR it's fake and poor quality there.
Time to make a 3d printer for electrical components perhaps?

Thanks for your help.

I have a few comments on this thread:

1.IMO, mains power is the only practical solution for high power heated beds. Even with a 24C supply, above about 400W the current gets too high to manage easily. Of course, use of mains power raises additional safety concerns, such as protective grounding and adequate insulation. If in any doubt whatsoever, consult an qualified electrician.

2. All SSRs need a heatsink to run them at anything above a small fraction of their rated current. Both the DC and the AC ones typically drop about 2V at maximum current (check the datasheet for actual maximum values). So if you use a DC SSR to power a 24V 20A heated bed, that's about 40W of heat generated in the SSR that you have to get rid of. OTOH a 240V heater of the same power will draw only 2A, so that's only 4W of heat to get rid of.

3. Any relay (whether SSR or mechanical) can fail short circuit. So can the mosfet on a RAMPS. So if you are going to leave a printer with a high-power heater unattended, at least fit a smoke alarm in the room containing the 3D printer and have a fire extinguisher at hand. Preferably, fit an over temperature cutout - these are obtainable very cheaply as microwave oven spare parts.

---

Large delta printer [miscsolutions.wordpress.com], E3D tool changer, Robotdigg SCARA printer, Crane Quad and Ormerod

Disclosure: I design Duet electronics and work on RepRapFirmware, [duet3d.com].

Others have identified a big pitfall of an AC-powered heated bed being the 50/60 hertz crosstalk it injects into everything. Do you think that this is a problem, or are they over-reacting?????

I have never heard amyome saying that crosstalk is a problem with using a mains powered bed. I can see that there will be some capacitive coupling between the heating element and the thermistor, but this should be less of an issue than capacitive coupling between the hot end thermistor and hot end heater wires. Both are easily countered by a filter capacitor on the thermistor input - which is a standard feature of most printer electronics.

---

Large delta printer [miscsolutions.wordpress.com], E3D tool changer, Robotdigg SCARA printer, Crane Quad and Ormerod

Disclosure: I design Duet electronics and work on RepRapFirmware, [duet3d.com].

While AC sounds interesting, I'm new enough to electronics to not want to risk messing with mains voltages...

I do have an update: the SSRs I got from ebay weren't so great: packaged so poorly one was visibly damaged in shipment, other burned out in seconds switching 15A (rated at 40A). Inside was a BUX98AP power transistor, rated at 25A max, so while I'm not sure why it failed at 15A (damage during transport perhaps) it was not rated correctly. So I would use caution when going ebay on SSRs, although it's the only place I could find DC ones.

I did however work out what seems to be a pretty nice solution: using a circuit with a IRFP4368 power FET to handle the switching. It's rated for far over 20A, and at that current will only produce about 4W of heat, which is far better than a typical SSR. It also doesn't cost too much more than one of the Chinese SSRs, and at least you know what you're getting.

Of course, I'll try taking precautions against short-circuits and such, will most likely include a fuse and look into temperature-sensitive safety elements. Just need to make sure they don't have too high a resistance, or they'll be dissipating too much power at 20A.

Conventional DC SSRs drop around 2 volts at rated current. This voltage is needed drive the output mosfet or transistor. They invariably require heatsinking if you use them at more than a fraction of their rated current - which is probably why yours burned out.

Given that in a typical 3D printer setup you do not need isolation between the control signal and the output, a mosfet is generally a much better solution for controlling a DC-powered heated bed. The important thing is to get one specified with a low Rds(on) at the gate voltage you will be driving it at or lower and the current you will be switching or higher. From the Rds(on) you can work out the power dissipation.

The IRFP4368 is not a good choice, because it is designed for higher gate voltages than 5V printer electronics provides. Try IRLB8743 or (better) IRLB3034.

Edited 2 time(s). Last edit at 04/24/2015 05:15AM by dc42.

---

Large delta printer [miscsolutions.wordpress.com], E3D tool changer, Robotdigg SCARA printer, Crane Quad and Ormerod

Disclosure: I design Duet electronics and work on RepRapFirmware, [duet3d.com].

It was heatsinked properly. Given that the (underrated) transistor was still rated at 25A, it must have burned out because either it was damaged during shipping (virtually no packaging), or the power transistor was an inferior quality copy.

I wasn't aware on-resistance varied considerably with voltage, good to know. However, comaring the data sheets the max resistances seem to be about the same for both the IRFP4368 and the IRLB3034, so I'm not sure it'd make that big a difference. The only big difference I can find between the two in specs is the Vdss, which is much larger for the IRFP436, I thought that mainly meant it can deal with higher voltages. Their Vgs seems to be the same, so I doubt there's a difference about how they'd be driven. In addition, I'd guess the TO-247 package is a bit more reliable overall due to its intrinsically better power dissipation.

Quote
anvoice
It was heatsinked properly. Given that the (underrated) transistor was still rated at 25A, it must have burned out because either it was damaged during shipping (virtually no packaging), or the power transistor was an inferior quality copy.

I wasn't aware on-resistance varied considerably with voltage, good to know. However, comaring the data sheets the max resistances seem to be about the same for both the IRFP4368 and the IRLB3034, so I'm not sure it'd make that big a difference. The only big difference I can find between the two in specs is the Vdss, which is much larger for the IRFP436, I thought that mainly meant it can deal with higher voltages. Their Vgs seems to be the same, so I doubt there's a difference about how they'd be driven. In addition, I'd guess the TO-247 package is a bit more reliable overall due to its intrinsically better power dissipation.

The difference is that the on resistance of the IRFP4368 is specified at 10V gate drive, whereas for the IRLB3034 it is specified at 4.5V. You most definitely cannot rely on the on resistance spec of the IRFP4368 if you give it only about 5V gate drive. I repeat, the IRFP4368 is not a good choice for this application, unless you add a gate driver IC to increase the gate drive to 10V or more..

Edited 1 time(s). Last edit at 04/24/2015 07:51AM by dc42.

---

Large delta printer [miscsolutions.wordpress.com], E3D tool changer, Robotdigg SCARA printer, Crane Quad and Ormerod

Disclosure: I design Duet electronics and work on RepRapFirmware, [duet3d.com].

You can run the heater on 24VAC from a transformer. Transformers are cheaper (I think I paid $50 for mine, shipped.) and more reliable than switching power supplies, but a hot heavier (0.5 KVA transformers weigh about 20 lbs).
Here's one: [www.ebay.com]

You would use an SSR to switch 117VAC into the primary side of the transformer and wire the 24VAC out directly to the heater. You don't need to run 10 gauge wires between the heater and the transformer. A couple twisted pairs of 16ga works fine.

I've been using this arrangement for about 2 years in my printer and it works perfectly. The SSR doesn't even get warm.

@ dc42

Although I am not 100% certain about this, it would seem the OP is using the RAMPS heatbed output to drive an external circuit with the IRFP4368, and the heatbed output from the RAMPS is likely to be 12V. In other words, he should be fine with the IRFP4368.

@ the_digital_dentist

Replacing the 24V DC switching power supply with a 0.5KVA transformer is an extremely bad idea, for many reasons.

Quote
AndrewBCN
Replacing the 24V DC switching power supply with a 0.5KVA transformer is an extremely bad idea, for many reasons.

Mostly for the sake of your back if you move the printer around!

The extra weight can help stabilize your printer on the work surface, and from a lot of printers I've seen, a little extra stability wouldn't be a bad thing.
A transformer is MUCH more reliable than a cheapo switching power supply.
A transformer is MUCH safer than wiring the heater directly to 117VAC.
A transformer costs about as much as an equal power capable switching supply.

Other than the fact that dropping a transformer on your foot is MUCH more painful than dropping a switching power supply on your foot, what are a few of the many reasons it's an extremely bad idea?
Maybe I failed to consider something...

My printer has become quite heavy, and is a bit of a burden to transport due to the weight (about 60 kg). I'm considering the possibility of reworking my bed heater arrangement to eliminate the transformer so I can shave some of the weight from the machine. I may ditch the heated bed completely which will lower both the machine's weight and the moving mass...

off the top of my head: bulk, weight, cost, no regulation, low efficiency, magnetic losses and consequent heat, hum, 24V AC @ 20A is much more dangerous than 24V DC @ 20A which is already dangerous enough as it is...

Quote
AndrewBCN
@ dc42

Although I am not 100% certain about this, it would seem the OP is using the RAMPS heatbed output to drive an external circuit with the IRFP4368, and the heatbed output from the RAMPS is likely to be 12V. In other words, he should be fine with the IRFP4368.

If he's driving the mosfet gate directly from the heated bed output, then the PWM will be inverted, i.e. heater at full power when it is supposed to be off.

---

Large delta printer [miscsolutions.wordpress.com], E3D tool changer, Robotdigg SCARA printer, Crane Quad and Ormerod

Disclosure: I design Duet electronics and work on RepRapFirmware, [duet3d.com].

Quote
dc42

Quote
AndrewBCN
@ dc42

Although I am not 100% certain about this, it would seem the OP is using the RAMPS heatbed output to drive an external circuit with the IRFP4368, and the heatbed output from the RAMPS is likely to be 12V. In other words, he should be fine with the IRFP4368.

If he's driving the mosfet gate directly from the heated bed output, then the PWM will be inverted, i.e. heater at full power when it is supposed to be off.

The OP has not posted a drawing of the circuit so for now we can't assume anything, right?

Quote
AndrewBCN
off the top of my head: bulk, weight, cost, no regulation, low efficiency, magnetic losses and consequent heat, hum, 24V AC @ 20A is much more dangerous than 24V DC @ 20A which is already dangerous enough as it is...

I wouldn't characterize any of those reasons as extreme and some are dead wrong:

Size isn't really much different, though transformers tend to be more cubic while switchers are usually sort of brick shaped.
You've got me on weight, which I already stated.
Cost is about the same.
Regulation is 100% unnecessary. You're going to be using it to power a heater which you'll be switching on and off based on the temperature in the bed. Why would you need regulation to power a heater?
Transformers are MORE efficient than switchers, but I don't know why efficiency would matter (other than heating the transformer/power supply). Do you think you'd notice a difference in your electric bill comparing a transformer driven heater to a switcher driven heater? If so, maybe the best solution for you is to do something like fly kites instead of 3D printing.
Heat and hum- not an issue. Even after a multihour prints, my transformer doesn't get more than a few degrees above ambient and makes no audible noise. Once the bed reaches print temperature the power used to maintain that temperature is much lower as you can see when the duty cycle of the heater drops (watch the LED on the MOSFET switch on the controller board). A switcher DOES put more noise back onto the power line than a transformer.
As for the old Edison vs Tesla debate about AC vs DC safety, at 24V there isn't much difference in safety between AC and DC. Hardly extreme...

A transformer is definitely more reliable than a switching supply.

The best argument against using a transformer is just the weight. If you want as close to absolute reliability as possible, use a transformer.

Edited 1 time(s). Last edit at 04/24/2015 09:23AM by the_digital_dentist.

Couple of points:

1. Can you really get a 24V 20A transformer for the same price as a 24VDC 20A power supply? I'd be very surprised if you can - but please surprise me!

2. Although DC is not safer than AC, the LED PSUs are short-circuit protected, so perhaps safer than a transformer in respect of a short circuit. But you can always add a fuse to the transformer.

---

Large delta printer [miscsolutions.wordpress.com], E3D tool changer, Robotdigg SCARA printer, Crane Quad and Ormerod

Disclosure: I design Duet electronics and work on RepRapFirmware, [duet3d.com].

A fuse is always essential when powering anything off the AC mains. I have a 10A circuit breaker that I use as the main power switch in my printer.

I paid $50 for my transformer, shipped, via ebay. It's old (I think the date code says it was made in the 70s- which attests to the reliability of transformers) and used, but works fine. Here's a similar deal: [www.ebay.com]

A quick look at 24V 25A switchers finds about $90 to be the going rate. Here's an example: [www.ebay.com] A lot of the 0.5kVA transformers will cost about that much with shipping.

These are just two examples, and of course, you can pay a lot more, and sometimes less for either. So I'd say that cost wise, they are about the same. However, if you factor in reliability, the transformer will probably cost you less in the long run, as you won't be replacing it (unless you decide it's just too damned heavy!).

I used to restore a lot of antique radios from the 20s and 30s and the one (and only) part you could count on to be still working fine (unless it was struck by lightning) was the power transformer.

Edited 1 time(s). Last edit at 04/24/2015 11:14AM by the_digital_dentist.

So you are talking about a limited availability used transformer. The one you linked to is rated at 0.5KVA. A 24V 20A (480W) switching supply from the site you linked to is also $49. So about the same price if the transformer has free shipping (I can't tell because I am in the UK).

btw a 10A fuse in the primary circuit might not trip if there is a short on the secondary side, it depends on the transformer. More years ago than I care to remember, I was tasked with checking out whether the circuit breaker on the primary side of a transformer would trip if the secondaries were shorted. Even with the circuit breaker in the primary correctly rated, it would not necessarily trip. Winding resistance and leakage inductance limit the primary current - and therefore the secondary current as well - although not enough to avoid burning out the transformer.

---

Large delta printer [miscsolutions.wordpress.com], E3D tool changer, Robotdigg SCARA printer, Crane Quad and Ormerod

Disclosure: I design Duet electronics and work on RepRapFirmware, [duet3d.com].

I will indeed be driving the IRFP4368 with the RAMPS heater output (D8), which will be at 24V because that is my PSU voltage: in order to get an appropriate signal, I will use a 15V zener diode across the gate and source. The IRFP4368 will thus be getting about 10V and, being an N-channel mosfet, will be on when the heating signal (now the triggering signal) from the RAMPS is on. So yes, it will be driven with 10V, not 5V. Thus the PWM signal should not be inverted. The idea came from an electronics forum member who seemed to know what he was talking about.

I have 2 PSUs because I can't find 1 with the required power for a reasonable price. One will power the RAMPS and related electronics, the second (400W, so slightly underpowered for the 450W bed) will be solely used to heat the bed.

A transformer would work, but I can't keep buying more stuff for this project when I already have the parts to make it work with DC... I think the 24V DC circuit should work fine in this case.

The schematic is below:


Edited 4 time(s). Last edit at 04/24/2015 09:06PM by anvoice.