PID Tuning help needed

jay_s_uk

@Reefwarrior swap it. it will be unsafe until you do so

Reefwarrior

Cool thank you. I will swap it out this afternoon then, and see if it is better.

With regards to the temperature measurement. Is it normal for it to jump up and down a few degrees? The graphs show spikes, instead of a normal flat line....

deckingman

@Reefwarrior said in PID Tuning help needed:

Cool thank you. I will swap it out this afternoon then, and see if it is better.

With regards to the temperature measurement. Is it normal for it to jump up and down a few degrees? The graphs show spikes, instead of a normal flat line....

I think you'll find it's a lot more stable with a cartridge that is a better match to the hot end. The PID control can only do so much and when you have a heater that is over powered by a factor of 4 then you are likely to get overshoots when it turns on.

droftarts

@Reefwarrior said in PID Tuning help needed:

Is it normal for it to jump up and down a few degrees?

No, check your wiring. Either it's picking up interference (usually from stepper motor or endstop wiring), or there's a poor joint. It will also cause the PID tuning to fail, as it can't get a consistent reading, and it's probably why it's taking so long to tune.

Are you using a thermistor PT1000, PT100 or thermocouple? There are ways to make the last two more reliable, by using four wires rather than just two.

Ian

Reefwarrior

@droftarts Hi Ian,

It looks like one of those little glass bead thingies that measures the temperature....

droftarts

@Reefwarrior Sounds like a thermistor. Check your wiring, and make sure the leads of the thermistor aren't grounding on the heater block, and that the thermistor bead is making good contact with the heater block (ie not flapping about in a big hole!). Cartridge thermistors, or PT1000, are generally easier to use and get good contact with. What kind of hot end is this?

Ian

Reefwarrior

@droftarts E3d v6 clone. I will take a look at the hotend too....

deckingman

@droftarts Did you check the earlier posts? It transpires the OP has a 12V cartridge and a 24V system. So around 140 Watt instead of 35. Methinks, that's more likely to cause the oscillations.

droftarts

@deckingman Yes, read the whole thread. @Reefwarrior said "The graphs show spikes". While I'm sure the heater can create a spike up in temperature, it probably can't create a drop as quickly! But a dodgy thermistor can, though I guess it could be a power drop across the board when the heater switches on, if the PSU isn't up to it.

@Reefwarrior A screenshot of the temperature graph misbehaving would go some way to disambiguate the cause of the "spikes". Try wiggling the thermistor wires, see if that causes the spikes, too.

Ian

deckingman

@droftarts I guess it depends on the definition of "spike". I read this (quote)

Other strange thing is, there are quite a lot of oscillations with regards to the temperature being sensed. (Up to 3 to 4 degrees up and down)

As you say, we need a picture of the graph......

Phaedrux

A picture of the graph would be nice, but don't don't run that 12v heater at 24v just to make a graph for us. You're liable to melt something. Just replace the heater with a 24v version before you do anything else.

Reefwarrior

Hi all,

OK so I swapped the cartridge for a 24V one. Did the PID tuning... results:

Warning: heater behaviour was not consistent during tuning
Auto tuning heater 1 completed after 4 idle and 25 tuning cycles in 1494 seconds. This heater needs the following M307 command:
M307 H1 R3.064 C165.3 D10.06 S1.00 V24.3
Send M500 to save this command in config-override.g

Bothers me a bit with the 'warning'. Also it seems the temperature sensing thing is a PT1000 I think? It's a little silver cylinder that connects directly to the board, no daughterboard. Chinese brand.

https://imgur.com/p5oTiy3

there is the picture that I am referring to. Look at both the blue and the red - the curves are not smooth - it's like jagged. Even the 'MCU Temperature' seems to oscillate by about 1 degree up and down (including the decimals....)

Also got this after PID tuning my bed:

Warning: heater behaviour was not consistent during tuning
Auto tuning heater 0 completed after 4 idle and 25 tuning cycles in 1536 seconds. This heater needs the following M307 command:
M307 H0 R1.369 C277.3 D5.29 S1.00 V24.2
Send M500 to save this command in config-override.g

Phaedrux

@Reefwarrior said in PID Tuning help needed:

Also it seems the temperature sensing thing is a PT1000 I think? It's a little silver cylinder that connects directly to the board, no daughterboard. Chinese brand.

Why do you think Pt1000? It could be a plain thermistor.

Photos?

Yeah that graph looks a bit fuzzy which would explain the warning.

The overall curve of the graph looks ok though, so I think the values it provides should be usable.

Perhaps a good time to replace the temp sensor with something better quality.

Reefwarrior

@Phaedrux Unfortunately I am not in a position to be able to do that. I'm one of 70 pilots that got retrenched when our government decided to liquidate the national airline.

Phaedrux

I'm not sure that that means, but if you want to continue to roll with the equipment on hand, you can verify that the tuning results are accurate by setting a printing temp and seeing if the graph and reported temps are stable.

The fuzziness in the graph may also be a result of noise from running near other cables. You could verify that by temporarily isolating the thermistor wiring and seeing if that makes a difference.

Reefwarrior

@Phaedrux hi, yeah it's a plain thermistor NTC 100k; 3950B.

droftarts

@Reefwarrior said in PID Tuning help needed:

NTC 100k; 3950B

These are known for being pretty cheap, poor thermistors, which might explain your fuzzy readings.

Ian

Reefwarrior

Ok, seems I got it fixed, somehow. Next issue is now the heatbed....

Reefwarrior

Ok, bed issue solved too. Seems the algorithm is too sensitive in my case.

How do I mark this as 'SOLVED' ?