Service Bulletin: Thermistor Issue in Aerostruder and Dual Extruder v3 Tool Heads

Thank you for your patience while we investigated and confirmed our findings.

There have been some reports that some printers equipped with the Aerostruder and Dual Extruder v3 tool heads were encountering temperature related error messages, and had difficulty maintaining and/or displaying the correct temperature.

After investigation we found that connectors for the thermistors in some Aerostruder and Dual Extruder v3 Tool Heads were insufficiently crimped by the manufacturer, a defect that was missed in our quality control process. We apologize for any inconvenience this may have caused and appreciate your business.

Based on our recent samples, we anticipate the number off affected units to be 5% of fulfilled orders, or less. Product quality and customer support have always been and remain among our highest priorities. View the full service bulletin for the potentially affected serial number range, steps to diagnose, and a no-cost resolution here.

Visit LulzBot.com/Support for a list of all past service bulletins.

Where is the serial number located? Mine was purchased on 11/30/17, and I can’t find any markings on either the toolhead or packaging.

There should be a sticker placed somewhere on the back of the toolhead where it rests against your extruder mount that will have the complete serial number for the toolhead. I attached an example picture for placement on the Dual Extruder v3.
V3 dual serial number sticker.jpg

My V3 does not have a sticker. Not even an e-steps sticker.

Reach out to the support team at LulzBot.com/Support and they may be able to find your probable serial range based on your purchase method. Include your order number. How is your tool head performing?

The temperature readings are a little noisy, but the toolhead works okay. Printing polycarbonate at 280C shut down once with an over-temperature spike. I wiggled the harness everywhere (sensor to rambo board) but couldn’t get the indicated temperature to also wiggle. Of note in my case - the noisy temp readings only occur when the extruder temp is in the range where the PID controller is active (set point +/- a couple of degrees). This leads me to believe that my toolhead doesn’t have a crimp problem, more of a system noise problem. I also had similar behavior with my dual V2.

I ended up putting a firmware band-aid on the problem. I added a slew rate filter to the extruder temperature measurement to limit the dynamics to what the system can normally produce (about 3 degrees C per second). So far so good.

We’d love to see those changes, where did you make them? Also, what firmware version are you running with that modification?

I PM’d the changes to Marcio about 2 weeks ago (in response to the similar troubles that others are having), but he seems pretty busy at the moment. I am currently using a modded (bltouch and a couple of custom toolheads) 1.1.8.8. The changes are all in temperature.cpp. Here’s a trimmed down version of the PM:

If you look at the extruder temperature dynamics, the system is only capable of producing about 1 degC/second of temperature slew. Anything faster than this almost certainly came from measurement noise. The noise itself usually looks spiky (i.e. the occasional noisy reading on top of small PID wander), but the large errors cause the control loop to try really hard. When I was printing polycarbonate I couldn’t go above 275C because the glitches would invariably trigger a shutdown.

I was lazy and didn’t characterize the temperature update rate with diagnostic messages. I just poked around in the code (I am currently fiddling with 1.1.8.4) and gleaned/guessed that the cycle time was 164 mS. So I made a slew limiter based on that value. I tried limiting the rate to 1 deg/second which kills off everything including ringing in the PID controller (too filtered) and 5 deg/second which lets through more noise than I like. I verified that autotune produces very similar gain numbers with the filter disabled and with it set to 3 deg/second, so the PID controller shouldn’t be too irritated. The filter will probably allow a bit higher D gain since it is no longer chasing shadows.

Anyway, here’s the code (in temperature.cpp), do whatever you want with it.

Gordon

#define TEMP_UPDATE_PERIOD 0.164
#define TEMP_SLEW_RATE 3.0
float tempslew(float last, float next)
{
static int init = 100;

if(init)
init–;
// updates at 164mS or ??
if(init)
return next; // default nop behavior

if(next > last) // upslew
{
last += TEMP_UPDATE_PERIOD * TEMP_SLEW_RATE; // x deg/sec slew
if(last > next)
last = next;
}
else if(next < last) // downslew
{
last -= TEMP_UPDATE_PERIOD * TEMP_SLEW_RATE;
if(last < next)
last = next;
}
return last;
}

void Temperature::updateTemperaturesFromRawValues() {
#if ENABLED(HEATER_0_USES_MAX6675)
current_temperature_raw[0] = read_max6675();
#endif
HOTEND_LOOP()
current_temperature[e] = tempslew(current_temperature[e], Temperature::analog2temp(current_temperature_raw[e], e));
// current_temperature[e] = Temperature::analog2temp(current_temperature_raw[e], e);

Sweet, thanks for sharing!

My aerostruder seems to have this exact problem. See the attached image. Would it be possible to solve this with the firmware mentioned here?

Or. If my problem actually IS a thermistor issue. Can I get to the thermistor and fix this myself. What is the issue? Bad connections?

@eirikso Yes it sounds like it is a thermistor issue. Check Lulzbot’s service bulletin here to see if you might be affected - https://www.lulzbot.com/sites/default/files/Thermistor_Service_Bulletin_2.pdf

If so, call them (phone support > online tickets) and provide them the info they request and you should get a free replacement setup. I am unfortunately also in the bad batch by SN however I crimped my thermistor connections and put a silicon sock on the extruder and things have almost reached my OCD levels of perfection. Good luck & happy printing! -Nate

Has this issue been fixed with all newly shipped units. Mine should be here tomorrow, ordered directly from you.

All the new toolheads that are above the serial range listed in the service bulletin will have already been checked to make sure they aren’t affected by this issue.

I just started having this problem when printing above 240C. I have Mini Aerostruder sn: KT-CP0129-0276 so I’m definitely in the problem range. However, I checked the crimp connectors and they seem fine - When I just heat the extruder and wiggle the wires I don’t see any noticeable noise. This makes me wonder if this is a firmware issue (e.g. PID/slew rate) per GJennings.
Does Lulzbot have a firmware revision forthcoming?

P.S. I ordered this Aerostruder direct from Lulzbot on 4/21/18 so either it was checked/repaired or one of the old ones slipped through…

There’s definitely no problem with the thermistor in my Mini Aerostruder.
After a number of tests and reading several related posts on this topic, here’s the conclusion I’ve come to:

  1. Printing at temperatures above 235C, you can’t run the cooling fan without compromising the hot end temperature. This is partly due to the design of the print cooling duct which blows a substantial amount of air directly onto the hot end block. However, it is also due in part to the PID settings which are not optimized for the Aerostruder printing at these elevated temperatures. You can increase printing temperatures up to around 260C with very low fan speeds but the hot end will run around 10-12C below the set temperature. Here you start running into sketchy territory where the controller may decide that you have runaway and will shut down. The block heater can deliver enough energy to drive the hot end even with the cooling fan on. However, the PID settings won’t drive it hard enough to get there.
  2. With no print cooling fan, you can easily maintain any temperature up to and beyond 260C
  3. Users who have installed the standard E3D silicone sock have pretty much solved this problem.

I suspect that Lulzbot only tested the Aerostruder with high temperature filaments for which cooling is undesirable (e.g. ABS). Unfortunately we print a lot of CPE and CPE + for which cooling is necessary in order to get the best resolution. I strongly recommend that this toolhead be provided with the E3D sock. Further, I would really like to see Lulzbot modify the cooling duct design to direct the majority of the airflow downwards rather than directly at the hotend. In addition, I think Lulzbot should optimize the PID and related settings to work reliably with higher temperature filament.
With the roll out of the new Mini, this issue is going to have to be taken very seriously.

Providing the E3D silicone sock should be a no brainer. Socks for the Dual Extruder should also be designed and available.

We have actually optimized the PID settings for the Aerostruders more powerful fans. To have those new settings be enacted you have to first flash the firmware then through the console enter first M502 then M500 this will restore settings to default and save those settings allowing the new PID settings to be enabled

Would you be so kind as to give us the minimum version number of the firmware that contains these changes?