Fan speed influences reading temperature hot-end

[maartenw 6]

A long time after working (not solving) on the reported banding problem (https://ultimaker.com/en/troubleshooting/view/5901-undesired-horizontal-ridges-on-regular-distance?page=1#reply-44932) I took my printer (Ultimaker-1) again to produce some prints. The prints using flexible polyester looked fine enough, but - same as for former prints using PLA - the layer bonding had to be improved. So I thought of decreasing the fan speed, which was till now kept at 100%.

For this decreased fan speed I found the print was of very low quality, worse than before and made me think of hot-end temperatures lower than the one set.

So I started to investigate. And yes, the temperature reading is strongly influence by the fan speed.

I set the fan speed on the Ultimaker controller and quickly changed back to the main screen to read the reported temperature. Most of the times the reading was more than 10 'C higher than the set value.

In a former discussions (https://ultimaker.com/en/community/view/6307-um-1-temperature-jumps-with-fan) and (https://ultimaker.com/en/community/view/6254-ultimaker-1-maxtemp-triggered-by-cooling-fan) it was mentioned that the wire connected to the fan had to be kept at a distance of minimal 10 mm from the wires coming directly from the thermocouples. This is impossible to realise as long as these wires all are through the upper part of the printhead. So the fan wire had to be taken out of the printhead and be given a route outside of this head.

To accomplish this I had to dismantle the head, and I decided to finally install the second hot-end, which I had on the shelf for a long time already.

After a lot of work, the print head with the dual hot-end was assembled. Now I could find whether the distance between the fan wire and the wires from the thermocouples was an improvement for the temperature reading.

For the measuring, the fan was disconnected to start with, and connected after setting the fan speed with the Ultimaker controller.

I noted the reported stabilised temperature at 0% fanspeed, then set the fan speed to a number (0..255) and then - after switching to the main screen - connected the fan shortly.

The reported temperature did not depend on the set fan speed, but changed immediately after connecting the fan.

In table form (lines reported in the sequence as tested):

Fan temperature

speed set read

0 250 250

2 250 246

4 250 249

8 250 249

20 250 249

30 250 263

40 250 267

50 250 267

70 250 264

99 250 258 (MAXTEMP)

after restart:

100 250 267 (MAXTEMP)

after restart:

255 250 250

125 250 264

151 250 264

175 250 262

200 250 263

225 250 261

240 250 250

250 250 250

235 250 250

230 250 250

225 250 262

227 250 260/252 (quick variation)

225 250 264/259 (quick variation)

(Sorry for the bad lay-out, I could not create a decent table. A list is showing bad as well....)

It is clear that between 25 and 230 a great difference exists between the 'real' temperature and the reported one.

The conclusion thus can be that a distance of more than 10 mm between the fan wires and the wires from the thermocouples is NOT solving the problem. It is also not worsening...

So I twisted the fan wires to prevent signals to be sent from the wires. This had the same results as reported above.

Next I layed the printer on its side and after disconnecting the fan from the printhead connected it directly to the electronics in the bottom of the printer. And YES! The temperature was stable for all fanspeeds when the fan was connected directly to the electronics. To check, the connection was made as original, where the fan was connected through the long original wiring at the print head, and the reported temperature was deviating as in the table again.

So the conclusion can be that the long wiring from the electronics in the bottom of the printer to the printhead influences in some way the reported temperature.

How can I get a long wire (1.50 m) with the right connectors to connect the fan directly to the electronics and give this wire a separated route?

[gr5 2,005]

Nice work!  My first comment has nothing to do with your question - just that you should be aware that leaving PLA or ABS in the nozzle at 250C for 30 minutes can easily cook it into a solid block that is difficult to get out without disassembly.  So I would have done this test at 180C.  But anyway getting to your problem.

There is a possible software solution and of course the obvious hardware solution is to use a shielded cable.  I prefer the hardware solution.  You can just cut the existing connectors off the existing fan cable and leave an inch of cable at each end and strip that and solder to a shielded cable - of course you need to use some kind of insulation such as heat shrink tubing to avoid shorting out the fan signal.

software solution:

http://marlinbuilder.robotfuzz.com/

If you go here - there is a checkbox (5th from the bottom) called "increase pwm frequency".  This supposedly can help your issue.  I believe this is "checked" by default in the latest UMO builds but I'm not sure.  It can make it more likely that your darlington transistor that controls the fan overheats and blows out but I have been using it like that (with fan running at 30% often) and had no problems so far.  However that transistor (Q1 I think?) does blow up often for many people.  It's very cheap and very easy to replace if you have even mild soldering skills (cut the leads at the device so you have something to pull on with pliers when you melt the solder).

Anyway that Marlin builder is an excellent one - extremely easy to create the hex file and then you can upload it into your printer using Cura "expert"  "load custom firmware...".

Edited June 29, 2015 by Guest

[maartenw 6]

Gr5, thanks for your comment.

I apologize for this late reaction.

I looked at your advise to create my own formware, but I was reluctant doing this not knowing how this firmware would work together with each Cura version.

On the other hand I made a separate wire from the fan connection at the electronic board beneath the bottom of the printer all the way to the fan at the printhead, which is lead at the outside of the printer and spiralized at the outside around the bowden tube. This works perfect! At any fan speed the temperature reading is constant AND when I bring the wire close (I mean close!) to the wire from the temperature sensor to the small pcb om top of the printhead, also the temperature reading is NOT influenced.