Skip to content

June 26, 2017

3D Printer Noise – Not a Fan

One of the things that 3D printer reviews tend to gloss over is just how loud the popular Cartesian or Delta 3D printers can be. It doesn’t matter if the printer is a typical open frame type or closed up in a box, quiet is not a word that comes to my mind.

Given that the speed of the printer will affect the steppers, thereby transmitting said vibration through the printer frame itself, there’s not a whole lot you can do about some of it. Granted there are stepper dampers that are said to “lessen” the vibration transmission, or isolation feet so the printer doesn’t use the table/bench it’s sitting on as a “speaker”…

I tried the stepper dampers, ah, yeah. Didn’t seem to do as much as I’d thought. The budget dollar store rubber feet had a far greater impact though.

Fan Noise

However, what about when the printer is simply sitting idle? On a delta printer there is a fan on the hot end that needs to run continuously. Replacing that fan with a ball bearing low db fan is a good move. The original one that came with my printer sounded like a cement mixer.

Typical specs for the “cement” mixer would be as shown:

Noisy Fan

Considering the specs were from a ball bearing fan, it’s a noisy little brute. Basically, look for a 6,000 RPM, 0.15A max, and dB-A of 21 (lower is better). If you order these from eBay or some other equally sketchy source be aware that any specifications you see are going to be questionable. At best.

I picked up some Coolertec case fans OW4010DFS-2P hydro bearing, 15dB-A. You don’t more than 5 or 6 CFM. NASA wind tunnel fans not needed. The Coolertec fans? Wonderfully quiet, darned hard to get.

IMG 1175

Power Supply Fan

The second culprit, major source of annoying white noise is the power supply fan. On a typical 360 watt (12V 30A) switching supply I found that the fans on most of the ones I’ve worked on or bought, run all the time. Loud!

IMG 1151

I’d removed all the grating from one supply, replaced the fan with a quieter one and it’s still noisy.

By some sheer freak of nature I ordered a power supply from Amazon and when I installed it, the fan wasn’t running. I thought it was fried from the get go. I checked the voltage and it was fine. Once running for a while during a print job the fan would come on, run for a minute or two, shut off again.

I started to investigate why this happened.

The majority of these 12V 30A supplies “specify” a thermal controlled fan, so as the power supply heats up from current draw, the fan will cut in and cool it off. Oddly enough even though thermal control is stated, my experience is that the fans run all the time.

So….why? That’s the question.

Thermal Control

I took apart a power supply where the fan cycled and one where it didn’t. Started to look at what the differences were.

IMG 1153

It was surprising to find that the majority of the ones I have, from 4 different sources/companies, all use the same PCB. Or a minor variation thereof.

I tracked the difference down to a jumper of the “fan runs all the time”…see that jumper RT2? Yeah, there is supposed to be a thermal relay component there. But hey, that cost .97 so…jumper. Or as I found with a lot of offshore stuff, cost, not quality.

IMG 1154

The part that’s missing is a simple thermal relay, N.O. (normally open). It closes at 45C and opens at 33C.


To get the PCB out of the case is a little tricky. Remove the two screws that hold the heatsink bar on the MOSFET’s at the end and edge of the case.

IMG 1155

I found on my power supplies the screws were two different lengths. Thus, whatever comes out of the spare parts bins when they were assembled. There is an aluminum plate with some heat sink material on it, that needs to come off the side MOSFETs (where the thermal jumper wire is). You need the wiggle room to get the PCB out.

Next are the PCB screws. There should be five in total. One in each corner, one in the middle. Mine was built on a Friday at 3 minutes before quitting time, I only got four screws.

IMG 1156

To get the PCB out, slide it slightly forward (so the back edge of the PCB will clear the rear inset aluminum; about 1/8″). Notice on the side where there are no MOSFETS the PCB is notched so I just lifted that side up, but left the opposite edge with the MOSFETs where it was. Sort of rocked it back towards me. It’s a tight fit so don’t get heavy handed with it. It goes back in the same way.

Power Spply PCB

Once out, it’s easy to work on.

Thermal Relay

If you’re thinking the relay would attach to the case or the MOSFETS, ah, nope. It doesn’t. It actually goes inside the toroid coil on the PCB. And you glue it in there.

Start by bending the leads on the relay.

IMG 1159

Now you need some “glue”… but you want thermally conductive glue.

I found some on Banggood and eBay called “silicon thermal plaster”. It’s not inexpensive either and I decided that since I didn’t know what it was, couldn’t find any specs on it, and lastly didn’t feel like waiting for months to get it, I’d source somewhere else.

Which lead me to M.G. Chemicals. Who, oddly enough, seem to have a wonderful assortment of off beat items that fit exactly what I need.

One of which happens to be called, “Thermally Conductive Adhesive”.

IMG 1161

M.G. sells two kinds of this epoxy (it is a two part adhesive). One is called paste the other is flowable. I thought I would need the paste, but that was not available. I got the flowable instead.

I was hoping it was not as runny as five minute epoxy and I wasn’t disappointed! It’s about as “flowable” as silicon. I.e. you goop it, it stays gooped where you goop it. How’s that for technical terms…

As with the “silicon plaster”, 25ml isn’t inexpensive either. As in $31 CDN.

BUT, if you’re doing a couple of your own power supplies, a couple of friends as well, it cuts that cost down.

Seriously, and I do stress this point, when you buy something from a reputable company who supplies MSDS sheets, you know using it won’t cause a third eye to grow on your thorax three weeks later.

I got two mixing tubes with the adhesive and let me tell you, not only do they mix the epoxy PERFECTLY, you’re going to get a bonus hand work out shoving the mix through the tube…:-) The tubes are obviously a one time use but if you were doing a small restricted access and several of them, arrgghh matey, they be worth their weight in gold.

Otherwise, just put in a piece of paper and mix it like you do five minute epoxy.

Put the thermal glue on the edge of the toroid coil and don’t over use it. You don’t want to plug the air off from getting through the coil. Mine actually looks thicker than it really is. Maybe because I used a popsicle stick to put it in. I suggest a Q-Tip…

IMG 1162

The epoxy said it had a 45 minute working time and a 24 cure time. I left mine for about 6 hours to setup, then proceeded with the wiring. When I checked the joint 24 hours later, it was just like epoxy, hard as a rock.

Relay Wiring

Cut the jumper on the PCB and remove the two short wires.

IMG 1163

I used a solder sucker to clean out the holes…

IMG 1164

The relay is not polarized so there is no plus or minus to watch for. Simply solder in the two leads from the relay.

IMG 1165

Reassemble the PCB in the case, put the aluminum edge plates back in, PCB screws, case screws, etc.

IMG 1166

Make sure the relay wires are not chaffing on the heat sink clamp (yes, it’s a typical Mickey Mouse way the offshore guys do it) and button the case back up.

Rather than pop mine into a 3D printer I grabbed a bunch of LED strings and tested it.

IMG 1167

Kind of seasonal I must say, bright was the other thought…but sure enough, 10 minutes later I tested the heat of the edge MOSFETs, 45.6C, the fan came on, dropped the temp down to 33.2C and shut off. It works perfectly.

Read more from 3D, Electronics

Comments are closed.