smoothieware / smoothieboard2 Goto Github PK

The XT30 connectors are rated for 15A continuous (30A surge). Some users may exceed this rating between the FETs including bed.

Adding a second XT30 connector in parallel allows using the same connectors as other sections while leaving the option of only using a single connector if not running over 15A through the FETs.
This modification will allow 30A continuous to be used between all Mosfets on the circuit which is closer to the rating of the Fets and likely limited in some cases by the output connectors.

This mod requires moving status LED resistors as well as board thermistor but improves routing quite a bit

Tested a voltage divider on UVLO which takes Vmot and divides it down to a usable voltage for the UVLO.

Currently, the pullup is tied to 3.3v and the endstop pin 1 is tied to 5v by default but is switchable to 3.3v via jumper.

If we tie the pullup resistor to pin 1 on each connector instead of 3.3v this will allow the voltages to match based on what is selected via jumper.

T1 seems to be noisier than the other ADC pins. Testing seems to show that the pin location on the chip and it's proximity to the PHY clock could be the issue. Swapping T1 for T4 is currently an acceptable fix (output of buffer swap) since T4 is used for the onboard thermistor and does not operate in a range which requires the same level of precision as hotends and such which operate in a higher range of temp.

Further testing on all gadgeteer ADC pins should also be done for comparison on noise.

When board is resetting an MCU is in it's boot state it causes the SSR inputs (and most others) to pulse on momentarily. This was partially mitigated by adding 10k pullups to the SSR outputs.
Since this is a condition where the enable pin is racing with the input pins, and by design the enable on the IC is slower than the inputs I found that we can eliminate more of the pulse by running the enable pin from 5v with a 4.7k resistor instead of a 10k. Then by running the inputs on 3.3v with 10k resistors they are slightly slower than normal.

In testing this eliminated most of the pulse. Currently, all the mods are done except that R80 needs to be swapped for 4.7k.

High pitched whine regardless of driver settings and current. Lower current helps but does not eliminate it.
We also tested various ranges of current based on sense resistor value from .05ohm-.1ohm
Solution: Do not use metal element resistors. Thick film is designated in the datasheet. All noise eliminated.

Ever Ohms Tech MA251220FR075FZ

Bournes CRM2512-FX-R100ELF
This issue is solved but put here for educational reference.

Add fiducials to the PCB to ease PnP setup and alignment.

I placed the upper fiducials with 10mm offset from corners and lower as 7.5mm. This will help differentiate visually if the board is placed in the machine rotated 180deg. Backside fiducials are mirrored from the topside. So they have different reference points entirely to error the machine if placed upside down.

Board thermistor trace routing takes quite a path around the board. By routing up and left instead of down and left it should shorten trace width while also moving components into an area where there are less components.

When adding a second FET connector it will be necessary to move the thermistor location, routing can also benefit from moving the thermistor closer to J40 in the process.

Example alternate placement/routing suggestion.

The template for extensions refers to "Gadgeteer.lib", but I could not find this file anywhere.

Where are the Smoothieboard2 files?

If possible rotate C111 by 90deg. Current orientation makes it difficult to rework/inspect solder joints without damaging or removing the Gadgeteer headers.

MT9700 and other backfeed protection/ current limiters only protect when disabled. Currently, the USB 5v acts as a one shot. It will stick on and backfeed into the host computer if 5vreg is enabled.

Solution was to add an ideal diode (per Kliment) which auto negotiates the issue.
Tested a deadbug build which uses an HBT3906 and a SQ3493 Mosfet.

First off, that is a great looking board, thanks for all the work.

My question is, as there appears to be room in the 2660 version of the board, is it possible to add step/dir/en/gnd headers/JST PH or similar for some or all of the driver outputs?

Probe input is setup currently with a comparator which requires an active signal between 3 and 25v to operate.

Currently, the most common use case (Microswitch, BLtouch) require a pullup to 5v in order to operate. The circuit right now has a NO jumper to 3.3v with a pulllup resistor. A couple things need to be done.
1.) Pullup swapped to 5v instead of 3.3v
2) Jumper changed to NC instead of NO. Those with HV probes need to cut this jumper.
3) since this jumper being NC causes a possibility of accidental HV connection without disabling the jumper we should add a diode (such as the WS5817 currently onboard) to block in case of accidental connection. This should still be cut for HV use even with the diode...but it should prevent board damage.

I want to plan upgrade for v2.
How can i look at schematic and pcb?

Having Vmot right next to the logic pins and probe connections is risky since a backwards connector or misalignment can cause issues.

There is currently a small Vmot header next to the Vmot input which is less than 75mm away. Users who require a HV input will already be required to cut the pullup jumper (JP7) so running an extra wire is worth the protection for the limited number of people who use those probes.

Changing the probe input to a 3 pin header and deleting the V++ (Vmot) will add some user error protections.