Skipped steps

Ever since I swapped out the old desktop that used to run my Sells Mendel with Gen6 electronics (desktop went PSU failed it was very old) and now run of a Win 7 netbook, I get skipped steps on some prints, not all prints though and sometimes I can print eh same part again and it doesn't happen.

I am wondering if my electronics might be on the way out, sometimes its the Y axis other times the X axis. I have some sanguino boards that I sell with my Prusa kits I could swap in with a bit of fiddling, or do you think it might be power spikes. I'm the only one home when I print and nothing else is running in my workshop either.

This is starting to get annoying having to throw away parts. It usually steps sideways maybe 2-3 steps and after a few layers steps back again. Really not sure what's going on so any suggestions gratefully received

---

__________________________________________________________________________
Experimenting in 3D in New Zealand

It is very unusual for it to correct itself, other than pure luck of missing the steps in both directions equally. Usually skipped steps are in the same direction due to some slight asymmetry. If it self corrects it is usually a loose pulley or a loose nozzle.

---

[www.hydraraptor.blogspot.com]

Yes I can't find anyhting loose at this point still looking though, it does skip for several layers. What it does is mover over a step then other sometimes only 1-2 other times 2-3 and it aways comes back again in exactly the same pattern.

It doesn't do it every time which is the frustrating thing, and as I said sometimes on the y mostly on the x though and this is a new thing as it's been printing fine for over 6 months every day. It can't be a temp thing either as it's winter so not hot in my workshop

---

__________________________________________________________________________
Experimenting in 3D in New Zealand

Is the amount it moves 4 motor steps? I.e. for a 200 stop motor and 8 tooth pulleys it would be 0.8mm.

Normally when a motor skips steps it stalls so it skips more than 0.8mm, but it would have to always be a multiple of that. It then makes it very unlikely that it would skip exactly the same amount in the other direction.

A noisy end stop can cause moves to stop short, but again it is hard to see how it would correct itself.

---

[www.hydraraptor.blogspot.com]

It could be without measuring I'm only guessing. I have optos on Gen6 not mechanical switches not sure if that makes a difference

---

__________________________________________________________________________
Experimenting in 3D in New Zealand

The only difference is optical ones are susceptible to stray light as well as electrical noise.

---

[www.hydraraptor.blogspot.com]

And the optical flags getting bent making the bed, extruder, or nozzle grind into a hard crash when the printer tries to home.

If you use your printer a bunch, this will happen. Because of this, I believe in mechanical enstops all the way.

I doubt it is mechanical, if it is happening on both the X and Y axis, but I would check to make sure both smooth rods are oiled and there is no bald or rough spots being formed through wear and tear. Check to make sure linear bearings aren't wearing out by listening to them glide when the power is off and you are moving them by hand.

It does sound like electrical interference. Is the missed steps random along each axis, or always in the same place(s)?

Random, the only things in my workshop are a freezer and a file server, the freezer is on a different plug. I'm thinking og getting a UPS for the printer.

It is odd though that this only started when I switched to using my netbook from a desktop

---

__________________________________________________________________________
Experimenting in 3D in New Zealand

Laptops usually have a lot of noise on their ground when driven from the mains. The ground lead of the PSU is linked to mains earth but the cable is a couple of metres long and has 2 - 5A flowing in it. Not surprisingly this gives noise at the laptop end of the cable and its ground is a few hundred mV above mains ground. You really notice this if you are trying to use one as DSO.

If I was designing a laptop I would use a two core shielded cable from the PSU. I would ground the shield to mains earth at the PSU end and connect the OV to the shield at the laptop end. That way the 0V current is not flowing through the shield so the ground of the laptop would be at earth potential and not noisy.

Also machines have a lot of ground noise at the controller due to large currents flowing in the ground to the PSU. You then connect the two grounds with the USB cable. Noisy currents then flow up and down the USB lead. This is one of the reasons I prefer Ethernet, because it is isolated by transformers at each end of the cable. USB where both ends are mains powered is problematical due to the ground loop created.

---

[www.hydraraptor.blogspot.com]

I had the exact same problem, and realised it was when I switched to a new laptop power supply that I got cheaply on ebay, it interrupts prints when plugged in to /the other side of the house/!!! See if the problem happens when you're running from battery. If it does, probably some funky power management thing, otherwise, bad psu.

In the past, I had this kind of problem whenever I brought my RAMPS Prusa into the office. Turning on the lights in the room would sometimes stop the prints. The one thing different between home and office was the laptop I was using away from home. Windows 7 Pro, HP Laptop.

There may be something to this.

I have a new theory which I am going to test later, I'm wondering if it happens when the netbook reaches full charge and then the PSU switches from charge to trickle

---

__________________________________________________________________________
Experimenting in 3D in New Zealand

NelsonRap Wrote:
-------------------------------------------------------
> I have a new theory which I am going to test
> later, I'm wondering if it happens when the
> netbook reaches full charge and then the PSU
> switches from charge to trickle

It was better only moved 2 steps so the part is usable, guess I'm going to either look for a cheap desktop to use again or convert to a sanguino board with SD card

---

__________________________________________________________________________
Experimenting in 3D in New Zealand

Could it be that the pulley is sliding (back and forth) over the motor shaft to create a shift that corrects itself later?

I had some issues with the usb -> serial -> extruder controller. Some bytes are random corrupted every few minutes when the printer is running, I added a checksum in the protocol which help a bit, but still could cause random de-sync at random time. (with a timely hotend issue I gave eventually)

I guess ground loop and noises are the root of the problem. Just wonder if anyone has seen any USB-optocoupler product?

Do you have ferrite cores on your USB cable? I was getting random microcontroller reboots on one of my machines until I switched to a USB cable with ferrite cores.

billyzelsnack Wrote:
-------------------------------------------------------
> Do you have ferrite cores on your USB cable? I was
> getting random microcontroller reboots on one of
> my machines until I switched to a USB cable with
> ferrite cores.

So did I.

I have found that my machine run perfectly until a very certain type of structure is printed. It is when a solid base is printed for an internal structure. I find that Slic3r does a very short move, very fast zig-zag at the edge of the solid (usualy disk, base of a hole) insert. My X has a rather high mass and the extreme reversal/acceleration causes the loss of a few steps and it usualy corrects itself, just happens to lose the same amount (or real close) but the other way. Ended up dropping my Jerk to less than 2 and max accl to 900mm/s/s but will still do it just not as much. Using much less than these settings gives a perfect print but it is just embarrassing to watch, almost as nervous and jerky as I am. Took a while to figure out that it wasn't just random noise but a very paticular type of structure being printed.
Todd

Just to leave a note of what I have found just now.
In case ferrite core doesn't solve the problem, I would be looking for product based on ADUM4160 - a USB isolation chip made by Analog Devices.
Isolation of signals and powers (requires powers from both host/device side to power the chip and provide reference)
There are products built upon that selling at around USD 20 in China.

A cheaper alternative would be USB-FTDI-RS232/UART-Optocoupler product, selling at around USD 10 in China.