Extruder controller issues

Hello,

I am working on building my first Mendel and I am trying to diagnose a problem I have been having when attempting to drive a stepper motor with a Makerbot v2.2 extruder controller. I programmed the extruder controller with the suggested test routine (located here) and observed the following behavior:

1) Immediately upon reseting the extruder controller board, the motor emits a relatively high pitched whine that decreases in frequency to a steady level and persists for approximately 10 seconds. During this time I can feel the stepper motor clicking, none of the 4 LEDs located beside the A3949 chips are blinking, and the upper A3949 chip gets very hot to the touch while lower one remains cool. This occurs no matter what value the onboard trimpot is set to.

2) At this point the whining stops and two of the LEDs beside the A3949 chips begin to blink. As I gradually turn the trimpot clockwise the stepper motor shaft eventually begins vibrate back and forth but not rotate. The violence of the vibration increases if I continue to turn the trimpot clockwise, but the the shaft will not rotate.

I don't believe the problem is with the stepper motor, a Lin Engineering 4218L-01-10, since I tried an identical stepper from the y axis and it behaved in the same manner. When I turned the trim pot a quarter turn clockwise and measured AC voltage across the 1A and 1B pins without the motor connected I got 0 VAC. Under the same conditions I measured ~7.8 VAC across the 2A and 2B pins.

Based on these observations I am inclined to believe that one of the A3949 chips is fried and will need to be replaced. Any advice/insights would be greatly appreciated. Thanks.

Hello,

I did some additional poking around and discovered that when the lower A3949 chip is supposed to be driving the the stepper motor, there is ~12.5 V between the VBB and GND pins as expected. The voltage between the VREG and GND and VCP and GND pins, however, is 0 V. According to the data sheet for the A3949 chip (located here), the chip will disable both outputs if either the VCP or VREG is low voltage.

As result, I am no longer sure if the A3949 itself is damaged or if it has simply disabled both outputs do to low voltage on VCP and VREG. Does anyone know why both the VCP and VREG would be low for only one of the A3949 chips and/or if this could could result from damage to the chip itself? How can I begin to diagnose what is going wrong?

By way of background, I have only modest knowledge/ability when it comes to electronics and limited familiarity with the extruder controller circuit design. As for diagnostic tools, I have access to a fairly good multimeter and possibly a scope if required.

Again, any help and/or ideas would be greatly appreciated!

Thanks,
Ben

Hi,

Did you ever get anywhere with this?

I've got similar sounding problems with the same board I bought assembled from Makerbot a couple of months ago. At first I thought it was firmware but after spending quite a lot of time looking through, I'm pretty sure it isn't.

At power up the outputs from the atmega seem to randomly decide which of the four H-bridge pins (D5-D8) to drive high, if a pair is high (D5+D7 or D6+D8) then I get a whine from the motor if connected (nema 17 stepper). D5 and D6 are PWM and when driven, the waveform looks good, but they don't respond to turning the pot. No change or movement when I activate Extrude on the reprap host. All other functions of the board, temperature, heater switching etc are working correctly.

While poking around I noticed that sometimes when I touched the board the LED's on the H-bridge lines would flicker, then stop when I let go, if the motor was attached, it would turn. I checked solder joints, reflowed many, ultrasonic cleaned the board and metered the tracks checking for continuity and shorts, all seemed good.

After more checking I found that if I touch D10 on the header near the voltage regulator then the LED's flicker at what looks like mains related frequency in a pattern. If I keep touching D10 while adjusting the pot, the PWM values change in line with what is expected. Again, the motor will rotate if it's connected but the LED activity is the same if it is not. I've tried pulling D10 high and low with a 10k resistor, both of these stop anything from happening. Checking with a scope, when I touch D10, I'm applying a 50Hz sine wave through general pickup.

I also checked the effect of the 16MHz crystal - With 8Mhz all the data to the reprap host goes wrong but the cycling of the LED's is still the same.

I'm presently concluding that the H-bridges are fine but the control lines from the atmega, D5-D8, are not being driven correctly.

Is anyone else finding similar problems ? Any help would be much appreciated.

Cheers

Kirtlesog

D9 and D2 should be connected to SCL/SDA on the motherboard. They are the step and direction signals, which is why the motor moves when you touch them. See [reprap.org]

---

[www.hydraraptor.blogspot.com]

Thanks bigtime Nophead.

I've used I2C quite a lot but always chip to chip so that's how I connected it. I'd looked at the wiring diagram various times but just got blind to the mistake. Shows how wrong assumptions can be!

A quick test seems to work well, I'll wire it up properly later and let you know how I get on.

Still, it seems a somewhat strange way of doing it.

Cheers

Kirtlesog

Hi,

Yes, I was able to resolve this issue but forgot to post an update. I went back and rechecked the H-bridge in question and found that one of the output pins was internally shorted to ground. I replaced the chip and was able to get my stepper extruder motor working. The upper H-bridge still runs hot when compared to the newly replaced one, however I am planning to hold of on replacing it since everything seems to be working now.

Thanks,
Ben

Ben,

Glad to hear you resolved things. Nophead was right on the money for my problem and things started working properly after that.

The 3949 bridges do tend to get hot very quickly. For the record I noticed that they get hottest when the system sits there with one coil under PWM and nothing moving, which seems to be a common resting point in the software after movement. This is probably down to switching losses in the bridge rather than pure forward current.

I've cut up a heatsink and clamped it to the board using clothes pegs and plenty of thermal grease, they seem quite happy with this. If they start showing problems I have an L298 chip (dual bridge) which is in a much bigger package and easier to heatsink, it should wire up without too many issues.

While messing about I broke my thermistor, it's a real pain waiting for a new one!

Cheers

Kirtlesog