Brownout reset general cause?

Hi,

I recently fried my atmega1284p on my sanguinololu 1.3a and just managed to burn a new bootloader and upload marlin.
Now I'm running into a brownout resets as soon as I connect to the board in pronterface. I read a discussion on marlin about this error occurring in GEN7 boards because either and LCD is connected or some resistor is valued wrongly, but find no indication that the same issues could happen on the Sanguinololu. (https://github.com/ErikZalm/Marlin/issues/174)

I measured the the 5V on the underside of the USB while connected to the main power supply and I get stable 4.68V which wouldn't be enough to trigger the brownout detection set to 4.3V.

So what could cause a brown out?
1) Is it possible that the current fluctuations are so fast that I cant detect them with my normal Voltmeter?
2) Could it be that I've blown some of the smaller capacitors and they are causing current fluctuations?
3) Could I just set the brownout detection lower? (the reprap article about the Sanguinololu list the 2.7V for the 644p) and hope the chip operates normally.

I'll appreciate every pointer

There is a 5v coming from usb, that can be measured on the usb connector itself (search "usb pinout"). And there is a 5v made by the regulator on board, meaning the board should have 5v also when the usb is not connected. Measure them separately perhaps that helps pointing out an issue.
Either way i think the 4.68v look suspicious. The 5v regulators have some error margin, but in few % range, perhaps it could be 4.92v ok like that, but as low as 4.68v that it should not be that much of a difference. Maybe you have a short somewhere, or maybe the regulator is defective. You could try measure the resistance (offline) between the 5v line and gnd of the board see if that is unusual.

Quote
sungod3k
1) Is it possible that the current fluctuations are so fast that I cant detect them with my normal Voltmeter?

Yes.

---

Generation 7 Electronics	Teacup Firmware	RepRap DIY

@Noobman: I noticed the LM7805 getting superhot fast but I read that this happens as a result of the difference between regulator and usb output and accepted it.

I actually measure 4.61V at the legs of the regulator, so Im not sure if it fried itself over time in normal use of if it took an extra hit when I fried the chip.
I get 5.03V from the USB cable and when I disconnect the USB I only get 2.5V (that seems totally of) on the regulator pins, so looks like I need to replace the LM7805.
Is there a stronger alternative to it which wont get is hot? Or is it more likely that some blown capacitors are causing this low voltage?

@Traumflug: time to get an oscilloscope

Edited 4 time(s). Last edit at 06/20/2014 02:47PM by sungod3k.

Some little progress

I removed the voltage regulator and get now a clean 4.98V when the board is completely empty. I Also removed the pololus which I hadnt done before and found that one of them was causing the drop in voltage to 4.3V and got hot while the 3 other only the it drop to 4.6V.

So I left only the three good ones on the board and plugged in a motor and actually connected to pronterface without brownouts. I also tested with the bad one and got my brownout back. So it seems there is one bad driver is influencing the whole board.

However marlin is still reporting an "external reset" the motor doesnt move and the extruder thermistor shows very weird values. It gives me a permanent 122C. I checked the config but they are still set to 1 (100k). The bed thermistor works normal

Quote

Connecting...
start
Printer is now online.
echo: External Reset
Marlin1.0.0
echo: Last Updated: Jun 21 2014 07:28:56 | Author: (none, default config)
Compiled: Jun 21 2014
echo: Free Memory: 13445 PlannerBufferBytes: 1232
echo:Hardcoded Default Settings Loaded
echoteps per unit: etc..

So it looks like the printer is in some sort of purgatory where it connects and is then resetted at once with only some function working properly.
I also checked the capacitors but they are generally rated for 50V and I see no damage.
With Sprinter as firmware I dont get resets, but except for the bed thermistor nothing else works.
Playing around with the different pololus, thermistor motor and endstop combiations still gives me below 4.8V in general.

Edited 4 time(s). Last edit at 06/21/2014 04:35AM by sungod3k.

The 7805 gets hot when it has to deliver a big current. The linear voltage regulator works by burning as heat the difference in voltage, so the watts dissipated depend on how much current you draw at 5v. If you draw 1A, it has to dissipate (12v-5v)*1a so 7watts. But the 5v consumption on your board should be very small, so normally that regulator should not get hot. If it gets really hot, that points to a probable short on 5v line, and that short draws big current, which makes the regulator trying to dissipate too much and that process could destroy the regulator.

Perhaps there was a short on that stepper driver. You can measure resistance between 5v and gnd on that pololu driver and compare to the others.

Or perhaps you also have some short on your board. You can measure resistance between 5v and gnd of your board.

I removed the 7805 and that results in 4.98V when the board is completely empty.

I took a closer look at the drivers

• 5V - GND driver 1 alone = 13.7 ohm
• 5V - GND driver 2 alone = 9.7 k ohm
• 5V - GND driver 3 alone = 0.47 k ohm
• 5V - GND driver 4 alone = 1.5 k ohm

I have no reference knowledge but that seems high

Also when I measure the 5V - GND empty board get 24.2 k ohm,
and contrary to my expectation that high resistance driver would increase overall resistance on the board I got these results.

• 5V - GND board with driver 3 = 40 ohm
• 5V - GND board with driver 1 = 10 ohm

Just interpreting the difference in Voltage when the empty board gives 4.98V and the lowest resistance driver (#3) on the board gives mit 4.85V, and the highest (#1) 4.3V, does that mean the drivers are the bad guys here?

When you measure the resistance between 5v and gnd, on a place where is an IC, basically the ic may see a sort of supply and some parts of the ic can get more or less activated. So the results are probably not going to be stable and not very relevant, but at least should be in some reasonable interval.

The 5v is logic supply, so generally this consumption is very low in few mA range. A driver could draw like 5-10-20mA, and atmega could draw like 100mA, more or less around that. But the 10-40 ohms range is not reasonable, because if a driver tries to draw 5v/10ohm = 0.5A at 5v, that certainly is a problem. While their 5v-gnd measurement seems ok, it means there is some other point they draw 5v from.

Try measure between ground and each of the points MS1, MS2, MS3, Sleep, Reset. These are also tied to 5v, so perhaps the short goes through one of them. Perhaps that should help identify the problem driver.

At least one of the above mentioned pins on every driver has resistance of 50k ohm or higher, only few of them have a "normal" rage of 40-100 ohm.

I would need to order new steppers for a possible new board anyway, so I guess I test with new ones next.

Those pins are logic input pins, so they consume in miliamps range. If the resistance is high, they consume almost nothing, and that is how it should be. The MS pins have pulldowns like 100k. So the range of 100k-50k is ok. For sleep and reset also at least in kohms range. The range like 40-100ohms thats probably not ok. For 5v, that points at a consumption like 50-125mA which is well outside of what a logic pin should be like.

Pin measuring is not really an method of getting good conclusion, just sometimes perhaps could point to a problem if its an evident short or something like that.

You could try taking the driver which measures the higher resistances of all, hopefully that should be ok, and try some tests with that single driver installed. Also if the driver are susceptible to draw too much, do the upload without them so they dont deplete the 5v like it was in your measurements. You will also have to try replace the 7805 with a new good one.

Well now the board is completely gone I had a friend over and he sat the board on a metal surface and now I barley get 4.7V on an empty board and there its not connecting anymore.
Thanks for all your explanations and patience.

Im gonna go RAMPS now, do you know a good source? I weary of any clones