Jam detection and speed reduction

Hi there,

I have been using the jam detection feature in the firmware and this seems to work quite well. What is not working so well is the speed slowdown feature. This may work for some situations but is no good in others. In my case keeping the speed the same and slowing down the flow rate would be a much better option. I make large complex prints that can take up to 2 weeks to print(printing 24/7). There are numerous overhangs during the printing. If an overhang starts warping upwards and thus putting  pressure on the nozzle and this in turn causes  slipping, the system works and the speed slows down.

But slowing the speed down only reduces the filament flow proportionally and so does not really stop the possibility of the filament drive gear chewing out the filament. You have to slow the speed significantly to try to stop the filament from being chewed out. 
An overhang may only occur on a small part of a layer and only on a certain number of layers. But my printing speed is now reduced for layer after layer(unless I'm constantly checking), thus significantly increasing  printing time. 3D printing is such a slow process anyway, anything that slows down printing time is the last  thing you want. I use Repetier informer which is great, it tells me when there is a jam but unfortunately it doesn't tell me when the speed has been slowed.

So my question is, is it possible to have an option in the jam detection section of the firmware to slow down flow rate and not speed when there is certain amount of slippage? I realise this would also mean the firmware would have to detect when the slippage was reduced to say 5-10% and then increase the flow rate back to 100% otherwise flow would be reduced for normal parts of the layer. If the firmware is already monitoring slippage this should not be hard to implement should it? 



  • Interesting case. There are in deed several reasons why slip occurs. Also on first layer you can easily get slippage from over extrusion if you are too close. Other reason is just not enough power so you really need to slow down print to come up. So both cases can happen and in a perfect world it would only be speed reduction.

    Reducing flow instead of speed is not hard as you just call the flow adjust instead of speed adjust and continue. But if i remember right it will not go back to 100% so that is not what you want.

    What you might do is use the event system and set continuously flow rate to computed slippage. I'm quite sure that will result in a bad quality for 2 reasons:
    1. It only affects new moves so it always lags current slippage.
    2. Measurement of slippage is imprecise. If I test this it writes different values, which is why I need a threshold.
    3. Depends also on number of measurement points of jam detector.

    So if we assume a precise detector with precision better then 10% we could make 10% steps so in normal print we get 100% flow. That would improve quality a good way. Only if you come from a problematic part into a good part that beginning will be too thin and eventually start getting thick when you reach the problematic region again.

    I hope you see the problem of lag. Time to get new jam lag and then more time till changes take effect.

    I think this would require extruder handling like with advance so we can add or omit extruder steps based on current slippage. That would reduce the buffering moves delay which can be anything from 0.1s to 10s delay.

    Also I think we should then have a average slippage computed so it does not change dramatically from a single bad measurement. 

    Need to make some tests with my printer to see if that might work also I print with PLA so not have that problem especially not having such long prints where it gets noticeable.
  • Thanks for looking into this for me.

    Based on your comments about lag maybe a solution would be to continue slowing the speed when there is slippage. As I said this does work if the speed is reduced enough. Then all the firmware needs to do reset the speed back to 100% when it detects slippage has been reduced. It then doesn't matter if there is some lag because you will always have the correct amount of filament flow. As one layer can take up 20 - 30 minutes to complete a few seconds lag will not make any difference to the time. Sure overall print time will be increased but not nearly as bad as having the slow speed layer after layer.

    If this could be implemented it would make a huge difference to print times.

    Thanks again 
  • That is really complicated. At some point we also have to say filament is out, which is in my case the main reason for jam detection. It is what is happening in 95% of detected cases for me.

    I have to think about this as it is a non trivial change that can give me lots of trouble and support questions.

Sign In or Register to comment.