DLink / D-Link resets going over bumps! Thoughts?

[tdressel]

Our programming team has been using the DLink (D-Link) DAP-1522. Life is great if robot is on blocks or on a test board. However, while running over carpet, it will set a bump in the carpet and viola - communications are lost and the unit can be seen rebooting itself.

Question 1: Anyone having issues with either the power momentarily being lost on their DLink or accidentally being reset?

More Info:
-DLink is securely on the robot
-all electrical connections from PDB (pwr board) to converter to barrel-plug and into DLink all look fine and are secure

-We do have additional DLink and will try those as well and see if the issue continues.
-If it continues, we may crack-open the DLink and possibly solder the power connector to the board - bypassing the code-barrel-plug/power-jack.

Question 2: Thoughts? anything else to try.

[MrForbes]

is power to the cRio being lost? or are you sure it's only the bridge?

[tdressel]

Good question. My focus has been on the DLink itself. I suspect not.

The reason for my comment is that I ran after the robot and was constantly watching the DLink. The LAN connection that the cRIO is plugged into will flicker wildly while communications are going on. When the DLink 'resets', all the LED's seem to turn off and then on. About a minute later, the DLink port for the cRIO flickers wildly again and we are off to the races.

When the reset occurs, the four LED's for the four ports blink off then on (or vice versa). If the cRIO was losing commuications with the DLink or losing power, the other LED's on the DLink shouldn't turn on and then off.

[MagiChau]

My team has a similar problem. We would randomly lose robot communication for the driver station, but I at the time did not think much of it. I need to inspect it further on Monday.

[Mike Betts]

Realize that the radio was designed to sit on an office shelf and not to get bounced around.

I would focus my attention on the power connector at the radio. This was a known problem with the previous device. Put a service loop with strain relief on the power cable and let us know how it works.

[Mike Betts]

Here is a slide I made last year for training my robot inspectors...

[tdressel]

In regards to the power cord (barrel plug), it's connected in such a way that the barrel plug end doesn't have any tension on it. No strain on the power cord.

I suppose what we could do is rig it such that the plug has no way of moving in the DLink power jack connector.

I know I'm 'reaching' but any thought that whatever mechanisim is used to reset the unit (you know, that little pin hole) might be connecting and causing the unit to reset?

Two thoughts on my part have been to (a) solder the power cable directly to the motherboard of the DLink and (b) ensuring that there is nothing in the area of the reset button mechanism that could be causing a momentary closing of a circuit.

[Mike Betts]

As I said, this was an ongoing issue with the old radio. I would see this four or five times in any given competition and the illustrated strain relief has never failed.

[Tom Line]

Just like PWM wires in victors, it is smart to hot glue ANY connection on the robot that does not have a positive locking feature that prevents it from coming loose.

It doesn't take much, it's easy to peel off, but it will keep you safe. It's smart to glue the reset button too on the older one since it had enough mass to trigger going over the bump.

We would shake our heads every time we saw a robot go to the fast-flashing (no connection) light during breakaway. 9 times out of 10, it was unsecured connections. Why beat yourself? It's hard enough to do well as it is.

[tdressel]

Tom, Mike, all,
Thanks for the suggestions. Hot glue it is. I'll have our electrical team review the diagram in the earlier post and the 'hot glue' ideas. I'm putting the glue gun in my backpack now to make sure we have one tomorrow night.

We're not in panic mode on this yet. Just don't want to head that way.

Again, thanks for the suggestions.
-Ted

[Mike Betts]

Tom, Mike, all,
Thanks for the suggestions. Hot glue it is. I'll have our electrical team review the diagram in the earlier post and the 'hot glue' ideas. I'm putting the glue gun in my backpack now to make sure we have one tomorrow night.

We're not in panic mode on this yet. Just don't want to head that way.

Again, thanks for the suggestions.
-Ted

Ted,

After a few days, post an update if this worked. I'd be interested to know if this is an issue with the new radios...

Regards,

Mike

[BitTwiddler]

While on the subject of the DLINK radio...

<R38> B. The radio power feed must be connected via the 5V converter (model # TBJ12DK025Z) to the marked 12 Vdc supply terminals located at the end of the PD Board (i.e. the terminals located between the indicator LEDs, and not the main WAGO connectors along the sides of the PD Board). No other electrical load can be connected to these terminals (please see the 2011 Robot Power Distribution Diagram posted online at www.usfirst.org/frc/kitofparts for wiring information).

I just discovered rule this yesterday afternoon while looking for something else. Ay-yi-yi! We've been running ours directly off the 12Vdc terminals. It still works but I hope we haven't somehow damaged it. Maybe everybody else but us knew this but you might want to check your robot to be sure that the converter is being used to provide the correct voltage to the radio.

[Mike Betts]

... Ay-yi-yi! We've been running ours directly off the 12Vdc terminals. It still works but I hope we haven't somehow damaged it...

I doubt if it was damaged. I think the main reason for powering from the dedicated terminals is to provide a kind of UPS function. The battery voltage can crash to very low levels when you put large demands on it (like four cim motors at near stall...).

FIRST is trying to make the communications as robust as possible.

Change the wiring to their recommendations and move on...

[Mark McLeod]

Ay-yi-yi! We've been running ours directly off the 12Vdc terminals. It still works but I hope we haven't somehow damaged it.
As long as you used the 12v-to-5v converter it'll survive.

The real problem with using the red/black wago connections for the radio is that a significant drop in battery voltage, e.g., full-power robot pushing contest, will risk resetting your radio.
The special 12v connector (and the 24v cRIO power) are protected against voltage drops to avoid resetting either one of those devices during a competition match.

[JCharlton]

This probably isn't your problem, but just for the record we had hellish problems last year losing connection. This persisted through two regionals. Our drivers spent more time with their hands in their pockets than on the controls.

After getting back and experiencing the same problems on a completely different robot we investigated further.

We found vigorously shaking the radio had NO effect, but gripping it lightly on the top and bottom killed it instantly! We took it apart and looked at it under a magnifying glass for aluminum shards, cracked traces, anything. Finding nothing we put it back together and it worked just fine. Go figure.

That's why I'm a mechanical guy.

[pilum40]

While on the subject of the DLINK radio...

I just discovered rule this yesterday afternoon while looking for something else. Ay-yi-yi! We've been running ours directly off the 12Vdc terminals. It still works but I hope we haven't somehow damaged it. Maybe everybody else but us knew this but you might want to check your robot to be sure that the converter is being used to provide the correct voltage to the radio.


I just got through purchasing another DLink because our team wired it directly and fried it. $107.00 from Office Depot because we didn't RTM! Please learn from our eagerness to get the robot rollin'!

[Al Skierkiewicz]

If you are loosing something when you run over bumps check these things...
1. lightly tap the red button on the main breaker. Do lights on the robot flicker? Replace main breaker it is defective.
2. Does your cRIO move around? It may be shorting to the frame of the robot when it moves. Some RS770 motors are being delivered with a short to case. If you are using these motors check continuity from each motor terminal to motor case. You may be sending power hits to the Crio.
3. Triple check the power wiring coming from the radio connector on the PD. These are notorious for being improperly terminated. The wires should not pull out when tugged and there should be no visible whiskers that can short the two wires together.
4. Check that the battery terminals on the PD are snug and have a lockwasher under the nut that hold the terminals in place.
5. Check the battery connections by moving all #6 wire with the robot turned on. Flickering indicators are a sign that something is loose.

[Danny Diaz]

Any body ever heard of a converter that is bad? On our converter, when we power 12v in we get 7.3v out. That doesn't necessarily concern me, what does concern me is that when we plug in the power to the D-Link we get nothing; we tested voltage, and we exactly get 7.3v out of the plug (as expected - the plug that came with the D-Link gives 5.2v in the same orientation) yet the D-Link is only pulling 2mA from the converter. Something seems bad here...

We do know:

1. If we bypass the converter, everything "works" (we accidentally did that before adding in the converter). This tells me the extra plug we bought (same size as the one that came with it) fits and works.

2. If we plug in the power cable that CAME with the D-Link, everything works fine (even after the fact).

3. With the converter in-between, voltages look fine but we're pulling almost no current.

Does this sound fair to say the converter is possibly bad?

-Danny

[Matt Krass]

Any body ever heard of a converter that is bad? On our converter, when we power 12v in we get 7.3v out. That doesn't necessarily concern me, what does concern me is that when we plug in the power to the D-Link we get nothing; we tested voltage, and we exactly get 7.3v out of the plug (as expected - the plug that came with the D-Link gives 5.2v in the same orientation) yet the D-Link is only pulling 2mA from the converter. Something seems bad here...

We do know:

1. If we bypass the converter, everything "works" (we accidentally did that before adding in the converter). This tells me the extra plug we bought (same size as the one that came with it) fits and works.

2. If we plug in the power cable that CAME with the D-Link, everything works fine (even after the fact).

3. With the converter in-between, voltages look fine but we're pulling almost no current.

Does this sound fair to say the converter is possibly bad?

-Danny

That does sound suspect, I remember when we wired ours up I probed it with the multimeter before connecting the radio to confirm polarity and it was reading around the expected 5V under no load. Some SMPS don't work well without load, but if you're connecting it to the router and not getting anywhere, that's weird.

Are you sure your router is good? It seems strange it wouldn't try to pull more current. If the problem was in the converter not delivering more current the voltage should take a dive.

Matt

[kamocat]

Our programming team has been using the DLink (D-Link) DAP-1522. Life is great if robot is on blocks or on a test board. However, while running over carpet, it will set a bump in the carpet and viola - communications are lost and the unit can be seen rebooting itself.

Question 1: Anyone having issues with either the power momentarily being lost on their DLink or accidentally being reset?

More Info:
-DLink is securely on the robot
-all electrical connections from PDB (pwr board) to converter to barrel-plug and into DLink all look fine and are secure

-We do have additional DLink and will try those as well and see if the issue continues.
-If it continues, we may crack-open the DLink and possibly solder the power connector to the board - bypassing the code-barrel-plug/power-jack.

Question 2: Thoughts? anything else to try.
What about that button on the side of the D-link? (I believe it has circling arrows) Is it getting pressed when the robot hits a bump in the carpet? Could that be resetting the D-link in some way?

[NetPlanet]

Last year we had a similar problem with communications. We lost it every time we went over a bump! It is still a mystery for the most part, but I am sure that it was the terminals to the battery. The few times we didn't lose communications were matches where we cleanly redid the leads to the battery beforehand. I'm sure the battery box was also to blame. It was solid and kept a strong hold on the battery, but the whole thing swung around as the robot moved, which probably loosened the leads.

[MagiChau]

Last year we velcroed our black bridge on, kept the battery in place with surgical tubing, screwed in a bolt for the battery connector end attached to the Power Distribution Board. I don't think we had any problems last year during a match related to communication.

[Al Skierkiewicz]

Danny,
There are two things that concern me. The first is that the DAP works on 12 volts when you bypass the convertor. This should not happen, it should blow up the DAP at 12 volts. The second is that you are measuring 12 volts into the convertor but only getting 7.3 volts out.
I have to ask (even though I know the answer) is the convertor wired exactly according to the drawing. The +12 volt radio connector on the PD wired to the convertor input leads and the DAP connected to the output leads?

[Danny Diaz]

The first is that the DAP works on 12 volts when you bypass the convertor. This should not happen, it should blow up the DAP at 12 volts.
That's what I thought, too, and nearly had a heart attack when I was reviewing the schematic and saw the need for the converter. Eric VanWyk looked over the components inside the DAP, and he says all of the power components should be capable of handling 12v but the manufacturer of course will not sign off on that (for good reason). After plugging in the DAP into 12v, I was concerned I may have fried something, but when we plug the DAP into the wall using the supplied transformer/cable it works flawlessly (at least as flawlessly as we can tell). I think we're going to get a new DAP and have it stand-by just in case something turns up, but we cannot find anything wrong with the one we have.

The second is that you are measuring 12 volts into the convertor but only getting 7.3 volts out. I have to ask (even though I know the answer) is the convertor wired exactly according to the drawing. The +12 volt radio connector on the PD wired to the convertor input leads and the DAP connected to the output leads?
That's a fair question, but I wonder why you said, "only getting 7.3 volts out" - when I saw 7.3v when we first started playing with it I thought that was pretty suspect since I figured we should be seeing something around 5v since the converter is a 12v/5v converter. We absolutely have the converter's 12v input line connected correctly to the PDB's 12v output just to the left of the 24v line for the cRIO (and we're correctly measuring 12.3 volts from that line), and then the supposed 5v output (from which we're measuring 7.3 volts) from the converter is wired correctly to a connector that plugs into the DAP's power plug (EDIT: Yes, and we have a fully-charged battery).

It's just that when we try to power the DAP through the converter, nothing happens - no lights, no nothing, and while plugged in the multimeter says 2mA when I cut one of the wires and put the multimeter in series between the cut. But, when we unplug from the converter and use the supplied walwart (plugged into the wall, of course) everything works great on the DAP.

-Danny

[Matt Krass]

It's just that when we try to power the DAP through the converter, nothing happens - no lights, no nothing, and while plugged in the multimeter says 2mA when I cut one of the wires and put the multimeter in series between the cut. But, when we unplug from the converter and use the supplied walwart (plugged into the wall, of course) everything works great on the DAP.
-Danny

Just curious, are you able to measure the current draw and the voltage at the same time? it seems odd that you'd be pulling so little current at 7.3V. Like I said before, I'd expect the voltage to dive if the converter was damaged, other than that I would agree it sounds like a faulty converter.

Worst case, you could design an elaborate mechanism to trail an extension cord to your robot ;)

[Danny Diaz]

Just curious, are you able to measure the current draw and the voltage at the same time?
While one multimeter was reading the 2mA with the plug trying to power the DAP, we used another multimeter to verify the voltage going INTO the converter, and saw we were still getting 12v INTO the converter - I remember saying out loud, "now we need to measure the voltage coming OUT of the converter," but because the "other" solder joint was still heat-shrinked for the output side I think it was quickly forgotten. Thanks for the reminder, I'll do that today.

-Danny

[Matt Krass]

While one multimeter was reading the 2mA with the plug trying to power the DAP, we used another multimeter to verify the voltage going INTO the converter, and saw we were still getting 12v INTO the converter - I remember saying out loud, "now we need to measure the voltage coming OUT of the converter," but because the "other" solder joint was still heat-shrinked for the output side I think it was quickly forgotten. Thanks for the reminder, I'll do that today.

-Danny

If the voltage is low as I suspect you may have a bad converter, or when you attached your ends and did the heatshrink you got a short, and the converter is shutting down to prevent damage. (Do we know if it has this feature?)

If it doesn't have such a feature it might be that you have a bad converter because of a short damaging it. I'd definitely double-check all your connections in this case and make sure there are no shorts and see if that resolves your problem.

If you measure a high voltage (like 7.3V or so) at that current.. I'm flabbergasted?

I wouldn't personally be too concerned about the 7.3V, some SMPS and even linear regulators float high without load, just because it didn't occur when I checked it doesn't mean it won't.

Do report back please, I'm curious how this works out,
Matt

[Al Skierkiewicz]

Tis a puzzlement.
I would have guessed that the failure mode on this device would a zero volts or 12 volts. Since it is not zero, 7.3 seemed to be low. I have not found any real data on this device on the manufacturer's website so there is no way to point at a particular type of convertor. So I am going to say it is defective but this is what I think might be the failure. If it is a switching type of regulator, the frequency might be so far from design that the output is not able to generate any current. Have you looked at the output with a scope? I am guessing there is a fair amount of AC voltage on the output and your meter is measuring that and interpreting it as 7.3 volts DC.
Of the teams that have reported that they hooked up their DAP to 12 volts and killed it, some may have reversed the voltage prior to the connection. Maybe you are just super lucky. With that kind of luck, I hope we don't play against you this year.

[Matt Krass]

Tis a puzzlement.
I would have guessed that the failure mode on this device would a zero volts or 12 volts. Since it is not zero, 7.3 seemed to be low. I have not found any real data on this device on the manufacturer's website so there is no way to point at a particular type of convertor. So I am going to say it is defective but this is what I think might be the failure. If it is a switching type of regulator, the frequency might be so far from design that the output is not able to generate any current. Have you looked at the output with a scope? I am guessing there is a fair amount of AC voltage on the output and your meter is measuring that and interpreting it as 7.3 volts DC.
Of the teams that have reported that they hooked up their DAP to 12 volts and killed it, some may have reversed the voltage prior to the connection. Maybe you are just super lucky. With that kind of luck, I hope we don't play against you this year.

Hmm, interesting theory, I hadn't thought of that. I'm quite curious now, I wonder if they have a scope.

Matt

[tdressel]

Ted,

After a few days, post an update if this worked. I'd be interested to know if this is an issue with the new radios...

Regards,

Mike
I started this two days ago. Yesterday, we had a short round of testing where the original DLink (DLink-1) was used and though the power cord was tethered such that the barrel plug would not move - we still had our issue of it resetting. We also did a good job of using a glue gun on the barrel plug (power cord) and still the issue persisted.

Tonight, we tested this again but instead of the original DLink-1, we used a 2nd one (DLink-2). Worked just fine till the our battery casing and battery fell off! We have a piece of carpet rolled out in the hall and we placed chunks of 2x4 underneath the carpet. The robot hopped around all over the place. The 'reset issue' didn't occur. Once we fixed the battery, we proceeded to test.

DLink-2 seemed to work fine. Though, we might have had some lag issues.

Replaced the DLink-2 with the DLink-1 and though the reset issues didn't occur again, the robot exhibited some lag where the driver tried driving but the robot would not respond for a up to five seconds at a time. Not sure if we have a different problem or the same one.

Tomorrow, we plan to do the same bumpy-carpet testing with both DLinks mentioned as well as a 3rd one (DLink-3). Hope to report back some findings.

Suggestions for any type of specific tests to run? I figure we had the symptoms with lumpy carpet. I'm not sure what to do about the 'lag issues' since we don't have a camera mounted and I'm not sure if it's interference with the battery and electronics. I saw another thread that suggested placing the DLink about two feet above the electronics.

[Danny Diaz]

If I haven't said this yet (and I don't think I have), sorry Ted, I didn't mean to hijack your thread. ;)

If the voltage is low as I suspect you may have a bad converter

We did the test today, and with the DAP power plugged in we again read 2mA being sourced, but when we're sourcing current from the 5v output the voltage on the converter drops from 7.3V to 4.0V. We're satisfied with calling it "blown", and we've ordered a replacement (2 actually). What's really cool is that the 5v output on the PDB works wonderfully to power the DAP, and its maximum current draw is only .5A higher than the camera - we're not even using the camera this year (programmers decided they had enough to chew on given the control challenge they were given for the rest of the bot), so we'll just keep using that until the new converter comes in (too bad that's not competition legal).

The only thing that interests us is that before we realized the converter was required for powering the DAP, we were using the converter as a benchtop 5v power supply for testing 5v sensors using competition batteries as power sources, and the tests all seemed successful (and the sensors draw about 60mA) - the tests are no longer successful, we cannot power squat. Perhaps the 12v input got its polarity swapped incorrectly once or twice, causing the converter to blow? That would be a crappy converter for sure without polarity protection. Anyone know if the device has polarity protection, I couldn't find any real specs on it to determine this. I do know that in its current configuration if you reverse its input polarity the 5v output drops from 7.3V to 0.2V until you swap the polarity back again (we just HAD to test it).

-Danny

[DtD]

Danny,
There are two things that concern me. The first is that the DAP works on 12 volts when you bypass the convertor. This should not happen, it should blow up the DAP at 12 volts. The second is that you are measuring 12 volts into the convertor but only getting 7.3 volts out.
I have to ask (even though I know the answer) is the convertor wired exactly according to the drawing. The +12 volt radio connector on the PD wired to the convertor input leads and the DAP connected to the output leads?

The DAP seems to have a switching voltage regulator inside it immediately following the power input, so I'm not surprised it worked.

~David

[Al Skierkiewicz]

Ted,
It is best to mount the DAP-1522 away from motors and large metallic objects but two feet from electronics is kind of extreme. There should be very little lag in controls with this device. Are you getting any indications of loss of signal, dropped communication, etc.? Does the robot run correctly if you tether through the DAP? (That is an extra added feature this year. You do not have to disconnect the wireless to tether.)

[Matt Krass]

If I haven't said this yet (and I don't think I have), sorry Ted, I didn't mean to hijack your thread. ;)



We did the test today, and with the DAP power plugged in we again read 2mA being sourced, but when we're sourcing current from the 5v output the voltage on the converter drops from 7.3V to 4.0V. We're satisfied with calling it "blown", and we've ordered a replacement (2 actually). What's really cool is that the 5v output on the PDB works wonderfully to power the DAP, and its maximum current draw is only .5A higher than the camera - we're not even using the camera this year (programmers decided they had enough to chew on given the control challenge they were given for the rest of the bot), so we'll just keep using that until the new converter comes in (too bad that's not competition legal).

The only thing that interests us is that before we realized the converter was required for powering the DAP, we were using the converter as a benchtop 5v power supply for testing 5v sensors using competition batteries as power sources, and the tests all seemed successful (and the sensors draw about 60mA) - the tests are no longer successful, we cannot power squat. Perhaps the 12v input got its polarity swapped incorrectly once or twice, causing the converter to blow? That would be a crappy converter for sure without polarity protection. Anyone know if the device has polarity protection, I couldn't find any real specs on it to determine this. I do know that in its current configuration if you reverse its input polarity the 5v output drops from 7.3V to 0.2V until you swap the polarity back again (we just HAD to test it).

-Danny

Oh goody, we're back in the land of sensible result! :)

You had me really wondering for a while, sorry your converter is toast, but it sounds like everything's gonna work out for you. Maybe when testing sensors your converter got shorted by accident? Since it's a vital part of the robot and we're down to under a week, I'm sure not going to try shorting ours out now to see what happens!

[tdressel]

Ted,
It is best to mount the DAP-1522 away from motors and large metallic objects but two feet from electronics is kind of extreme. There should be very little lag in controls with this device. Are you getting any indications of loss of signal, dropped communication, etc.? Does the robot run correctly if you tether through the DAP? (That is an extra added feature this year. You do not have to disconnect the wireless to tether.)

As for indicators of lose of signal - we did have that two days ago when I first created this thread - where the various LEDs would go dark as if in a reboot sequence and about a minute later, the Driver Station on the laptop would report that communications were back up. During this time, the cRIO had power and looked fine. The second DLink is more of a momentary lapse. During the time where students are reporting they are pushing controls and the robot is not moving - I viewed the DLink LED's and they were flashing away. The LAN LED that the cRIO was plugged into was flickering the most. To me, that indicates communications are there - just that the robot is not responding.

Not sure what you mean by 'tether'. Using the "How to configure your radio" doc from FIRST, we have set up our DLink to be in AP mode. Therefore, wifi comms wise, we have just the DLink in AP mode (10.11.11.1) and the cRIO (10.11.11.2) communicating with the laptop (10.11.11.9). There is no separate WAP (wireless access point) or Router such as we have had in the past two years.

The two feet comment came from this thread (see 'billbo911'):
http://www.chiefdelphi.com/forums/showthread.php?t=91779

Honestly, it's currently attached to the same board as the 'big battery', the PDB (Pwr Dist Brd) and the various Jaguars. So, not sure how to rule out 'noise' and interference of the wifi signal except to go up and away from the rest of the electronics. 6 inches, 1 foot, 2 feet - I'm game to try them. I figure at two feet, I've completely reduced that 'wonderment' as to why we see a lag every once in a while.

Thoughts?

[Al Skierkiewicz]

Ted,
A minute of lost communications is what you will get when the DAP reboots. It takes around 55 sec. Just to be sure, you are using the power convertor connected to the +12 volt radio output of the PD. This is a protected and regulated power supply that is independent of the battery voltage swings.
Tether is a method of controlling the robot without the radio using a cable between the driver's station and the robot.

[tdressel]

Right! Tethered. No tethering here - in all these interactions, we are not tethered. All intereactions are over wifi. We are using the converters. Others on this thread have mentioned power issues with their converters but we are not seeing thier symptoms.

[Matt Krass]

Not to possibly make this seem like more problems, but are your sure that the problem is still with the router?

Momentary lapses sound like it could be Jaguar faults, did you happen to see what the lights on your Jags were doing when these lapses occurred?

Matt

[LeadU2Fun]

We had a problem with momentary lapse of not being able to drive. I was able to actually hear the click of the auto reset fuses when I was troubleshooting. Found out my students were powering the cims from the 20A auto reset breakers instead of the 40A ones. We reconnected to the 40A circuits and haven't had a problem since.