Aluminium Heat Bed

I have a 1/4" x 12" x 12.5" cast aluminum tooling plate bed with a 450 W kapton heater powered by a 24V transformer with AC power switched by an SSR. I turned the max PID frequency in smothieware down from 20Hz to 8 Hz. The thermistor is screwed to the bed. There is currently no insulation on the bottom side of the bed, except for a small piece of silicone foam that is covering the thermistor, though I may try adding full coverage soon. It heats to 110C in under 5 minutes.

I have been struggling with a noise problem that seems to be due to the bed plate resonating, so I may ditch the heated bed completely and switch to printing on PIR foam instead. That will eliminate all the wiring to the bed, the transformer and SSR, and the need to level the bed (not that leveling is an issue for me), and greatly reduce moving mass in the Y axis.

That sounds a good construction to me, albeit with a very large thermal mass (because of the thick aluminium) and a lot of heating power (about 5Kw/sq.m). Having the thermistor screwed to the bed should mean that it follows the bed temperature quickly. Can you confirm that it is screwed to the centre of the bed, and not at one of the edges?

What sort of noise problem are you referring to - do you mean that the bed is vibrating at the PWM frequency, or tjhe AC mains frequency, or something else?

---

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 don't know how you could get 10C swings with that arrangement. I have a 200m x 200m x 6mm aluminium tooling plate bed with a 290W heater. That only gives 1C swings with bang-bang.

---

[www.hydraraptor.blogspot.com]

Thank you to all the smart contributors to this thread. Bookmarked for ref.

I hope I am not 'hijacking' but rather creating a bigger pool of information for this page.

I am currently assembling a 3thk Alu plate (305x250) with 6 off 50W 5ohm power resistors in parallel to give 0.8333 ohm.

About 14.4A or 172W at 12v, by my sums. fed via a 25A DC SSR

Can I ask-

Do I need a flyback diode across the SSR output?

As dc42 stated; The wiring diagrams for heating/ resistive loads generally do not show a resistor, however, as these are surely wire-wound resistors, won't that be a partly inductive load?
Will it matter?
What sort of spec. diode should I be looking for?

Also
The notes in Marlin would imply to me that PID control is recommended for this set up

Can/ should the PID switching speed be altered?

Should MAX BED POWER be reduced?

Sorry to drift away from dave3d's original question but with all the above comments, this thread is becoming a bigger resource. It would be useful for a wiki page.

Thanks

Alan

Wirewound resistors are generally wound non-inductively, so the wiring will probably have more inductance than the resistors.

Have you considered using a Kapton or silicone heater instead of resistors?

Your DC SSR will almost certainly have a higher voltage drop than a mosfet, possibly as high as 2V. Check its datasheet to see what voltage drop it has, and allow for it in your calculations.

---

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 just read this on reddit in a discussion over heat beds:

"running an SSR like this in PID can also contribute to thermal runaway so be sure to switch your heated bed to bang bang control and check the frequency of the temp checks."

"As someone mentioned here, you'd want to set your bed PID to bang bang mode otherwise you'll melt the SSR. Also disable PWM.
For safety, I'm putting a 10A fuse into the mains circuit, so I can hopefully salvage my SSR if the bed decides to short out."

Thanks dc42.
Resistors weren't my ideal choice. I just happened to have some that add up to that number. I couldn't find an ideal 12v heater pad of my size and power requirement.The alu would be a heat spreader for a glass sheet on top.
I hadn't considered the volts drop. Fotek SSR 25DD data sheet only mentions leakage current 1mA nothing about V drop.
Being on Reddit doesn't make it more true Kr. SSRs are designed for fast PID controlled switching - no one mentioned melting
The notes in Marlins config specifically reference PID control of SSRs. People do use it.
Trusted opinion seems to vary though.

A

Thanks for all the comments, some of which were over my head, but thanks anyway. After careful consideration I have decided to change to 24v for my aluminium bed. I have read that putting 24v through the 12v connections will heat the bed in 2 mins, which would be fantastic. Printing directly on the aluminium bed and doing away with the glass would also be an advantage.

I found my existing 12v PSU is voltage adjustable after all; my mistake, I am new to this. It can be screwed up to 14.5v but as I mentioned I do not wish to supply my Arduino/Ramps/LCD panel at a higher voltage.

I have decided the easiest way is to mount two power supply units, one 12v and one 24v, side by side under the printer (an OrdBot). Physically they are the same size. The one I have ordered is rated at 15 amps.

I have an aluminium Mk3 bed that I often use at 100C for ABS - my warm-up time is less than 10 minutes from a 12v supply.

However, I have 2 layers of cork sheet (each about 2.5mm thick) held underneath the bed by a thin aluminium sheet. During warm-up I also put a piece of corrugated cardboard on top to help things along, until the print head is about to move.

A final follow up. I have now fitted the second 24v power supply unit to power just the bed. It is rated at 17 amps.
I still use on/off control via a SSR. I adjusted the output voltage of the 24v PSU so it gives around 15 amps, just to allow a margin.
It now heats up to 110 deg C in about 3 mins!

Don't trust an IR gun to give you an accurate reading on aluminum. This is second hand knowledge and I can't authoritatively say more.

Thats the direction I'm heading. I have the Power supply and SSR. Are you staying in bang bang mode? Ive read that using PID can burn up an SSR.

Quote
dave3d
A final follow up. I have now fitted the second 24v power supply unit to power just the bed. It is rated at 17 amps.
I still use on/off control via a SSR. I adjusted the output voltage of the 24v PSU so it gives around 15 amps, just to allow a margin.
It now heats up to 110 deg C in about 3 mins!

Ecky: I measured the temperature using the printer bed thermistor.
I have previously used an IR gun to check the temp but it gave a low reading. It is due to the emmissivity of the aluminium surface. A highly polished surface will give a low reading. A matt black surface, a higher reading.

krwynn: on/off control works fine. It overshot to 114 deg C initially then quickly settles down and cycles within a 2 deg C band or +/- 1 deg C. This is adequate for bed control I think. I cannot see how an SSR can work with PID control. The SSR is itself an just an on/off switch. How can it work proportionally and give a variable output? Unless it switches very very rapidly.

I am very happy with the setup. The bed now reaches temp just before the nozzle! It means I can be printing within 5 mins of turning the printer on. I only do ABS and it is much better than I had before. I have also done away with the glass sheet and clips which reduces weight. I have kapton tape directly on the aluminium. So far so good.

Edited 1 time(s). Last edit at 03/29/2015 02:01PM by dave3d.

The proportional control of pid is just a matter off switching on/off/on/off quickly as the target temperature approaches to avoid overshoot. As it gets very close the off portion becomes longer to maintain temperature but not overshoot. These times and values are as defined Kp Ki Kd

The hotend uses pid control, anyway, hence why the flashing indicator light (On my RAMPS board) flashes at different rates as it gets close to temperature. It's not putting out a variable voltage its just switching on and off quickly.
The current consumption of the hotend is low enough that a small onboard MOSFET transistor can handle switching the power through it, unlike the power of a heated bed.

As for 'burned out SSR's' - The simple answer is to check the cycle rate of your SSR against the cycle rate of your firmware.

In my case a Fotek SSR 25 DD has a min switching time of 10ms which gives 100Hz.
Marlin firmware switches at just under 8Hz

A 'safe margin' I would say.

For on/off (bang bang) control, which is just like a regular thermostat, a mechanical (high current) relay is better, an SSR is a pointless expense that just drops voltage passing through it, if dc42 is to be believed and I think he probably is.

Much as I like the idea of pid control, I've just gone for on/off myself. It is 'near enough' and the thermal inertia of my print bed is such that it's not going to change temperature very quickly.

a

Quote
alan richard
For on/off (bang bang) control, which is just like a regular thermostat, a mechanical (high current) relay is better, an SSR is a pointless expense that just drops voltage passing through it, if dc42 is to be believed and I think he probably is.

Just to clarify:

DC/AC SSRs are great for controlling mains-voltage heated beds. They drop about 2V, but that is of little consequence when switching mains voltage at relatively low current. SSRs need heatsinking unless they are run well below their rated current, but even so, an SSR-10DA (rated at 10A) should be fine for driving a 230V 500W (2.1A) heat bed without a heatsink, and an SSR-25DA can be used for more powerful heat beds.

The problem with DC/DC SSRs is that they also tend to have around 2V voltage drop at full current, despite the fact that the mosfets they use for the output stage can drop much less voltage. I believe this is because they derive the gate voltage needed to turn on the mosfet from the secondary side. The voltage drop has two consequences. First, it reduces the voltage at the heated bed, resulting in a reduction of heating power. For example, a 12V heated bed run from 10V will only produce 70% of its rated power Second, it generates heat, e.g. 2V drop @ 20A = 40W of heat to get rid of.

It would be possible to make a DC/DC SSR that derives the gate voltage from the primary side, but I don't know of any SSRs that do this. I'm tempted to manufacture some.

An automotive relay is an alternative to a DC/DC SSR, but there will be quite a lot of arcing on the contacts, so it may not last very long. But it is cheap and drops negligible voltage. Remember to connect a flyback diode in parallel with the coil.

Unless you need isolation between the heated bed and the printer electronics, then the best solution is usually a mosfet, or two or three mosfets connected in parallel.

Edited 2 time(s). Last edit at 03/30/2015 10:58AM 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].

i have the same issue where my Mk3 bed wont go above 70c the thought i have is what about buying a second 24V supply and a relay (12v/24v) then running the 12v bed power cable from the ramps to a relay thats 12v on the coil then run the 24v power supply thro the switching side direct to the bed wired in 24v configuration that way whenever the ramps tells the bed to warm the relay will aloow for 24v direct to the bed with no other load however unless you got a solid state one you would get the click each time it turned on and off

anyone else see an issue with this?

I insulate the bottom of my bed with Frost King Pipe Insulation tape from local hardware store. Here is a link for a pic. Rated up to 250 F 120C so safe for the temps I use; cut the time to heat my 350mm x 450mm to 5 minutes. and no bounce in temps. If I want it to heat up super quick I will toss a piece of cardboard on top of bed while it is heating.

Quote
mmutch
i have the same issue where my Mk3 bed wont go above 70c the thought i have is what about buying a second 24V supply and a relay (12v/24v) then running the 12v bed power cable from the ramps to a relay thats 12v on the coil then run the 24v power supply thro the switching side direct to the bed wired in 24v configuration that way whenever the ramps tells the bed to warm the relay will aloow for 24v direct to the bed with no other load however unless you got a solid state one you would get the click each time it turned on and off

anyone else see an issue with this?

No problem as long as both PSUs share the same wall socket or the 24V-GND is completely separate from the rest of the printer. ( fed directly to the heatbed and +24V is switched by the relais.)
-Olaf

Is there a source for the wiring diagram for a 24V Power Supply and an SSR that can be referenced? Perhaps a complete one?

Yes. It's called Google.

1.5 seconds of google searching later

-a

Make sure the Solid state relay is a DC/DC one and a name brand one. The cheap ones on Amazon and ebay tend to be DC/AC with a new label. You can tell they are fake when they get hot, allot of them will be sold with a big heat sink. I have a Crydom D06D60 LINK just gets slightly warm with a my 360 Watt silicone heater. I tried the cheap FOTEK SSR ESSR-DD100A and it got super hot even with a heat sink and failed after a week in the closed position so be careful.

Quote
AquaticsLive
Make sure the Solid state relay is a DC/DC one and a name brand one. The cheap ones on Amazon and ebay tend to be DC/AC with a new label. You can tell they are fake when they get hot, allot of them will be sold with a big heat sink. I have a Crydom D06D60 LINK just gets slightly warm with a my 360 Watt silicone heater. I tried the cheap FOTEK SSR ESSR-DD100A and it got super hot even with a heat sink and failed after a week in the closed position so be careful.

Which is why I was wondering if there were some 'solid' wiring diagrams, BOM's etc, regarding this modification. I'm certain with all the experience out there a quick lessons learned, here's the schematic, parts list and misc. other things would have been generated.

I watched Thomas's video LINK and went for it, but yeah looked around too. Bang Bang is what I used, but going to switch to PID once I get a Nice Big Capacitor to sit in between the SSR and the power supply the fast switching of that big of a load is not safe for the power supply.

Edited 1 time(s). Last edit at 06/16/2015 01:52PM by AquaticsLive.

well If you buy a china Product, the Data sheet is in Mandarin or Non existent mine was a Kit with 4 Pads it did say 12/24VDC, but they had it hooked Directly to..