Chief Delphi - Flawed IR Design Confirmed

Page 2 of 2

Show 100 post(s) from this thread on one page

Chief Delphi (http://www.chiefdelphi.com/forums/index.php)

- Electrical (http://www.chiefdelphi.com/forums/forumdisplay.php?f=53)

- - Flawed IR Design Confirmed (http://www.chiefdelphi.com/forums/showthread.php?t=61250)

SWATRocks

10-01-2008 22:51

Re: Flawed IR Design Confirmed

So does anyone know where we would be able to get another IR sensor board? It worked perfectly one moment, got everything programmed, then it just stopped working. Oh well, as long as we can get a new one.
Thanks

EricVanWyk

10-01-2008 22:57

Re: Flawed IR Design Confirmed

Quote:

Originally Posted by RyanN (Post 674827)

Team Fusion had their first try with the IR board mounted on the robot this evening with bad results. Here are the problems we encountered:

With the programming, the digital input must be 1 in order to count it as pressed (this is bad because if one of the cables comes loose, it will result in a robot without any sense of where it is going)
Very gittery unless you run the motors for about 10 loops, then the robot drives for a few hundredths of a second before it stops.
The board will fry itself after a few minutes.

Yes, that's right, we fried our board, and I can't figure out what is wrong with it. The 5V regulator is putting out exactly 5 volts, the wires feeding the IR board are wired properly and working, and it WAS working perfectly.

We're not sure what happened, but while driving it, it all of a sudden took off on its own, hit our unwelded frame against a concrete wall. My first thought was it was a dead backup battery, and this was indeed true, so I replaced the battery with one reading about 7.8 volts, and it did the same thing, smacked into the wall. So I declared that we had problems and started to investigate, and yet, I found nothing wrong. I think we have it on video too of it going spastic on its own.

I definitely think there is a flaw in the design, a very dangerous flaw as the robot will all of a sudden start running on its own. Another flaw, if one of your wires comes loose, it will drive on its own. I'm going to contact FIRST and see what their conclusion is.

Ryan, remember that you will receive a series of pulses when the board is getting a signal. If you tie that input directly to a victor, you will see a series of pulses on your motor.

The IR board is to send commands, not state. Put a software latch or low pass filter in.

RyanN

10-01-2008 23:07

Re: Flawed IR Design Confirmed

Quote:

Originally Posted by Betty_Krocker (Post 674832)

Ok i may be over stepping my boundaries here, but I noticed the other day that when we tried to use our IR board that it too was fried, when i say this, i mean it turns on but it goes right to the first input and stays there, well i dont know that actaully because we havent hooked it up to anything but the light stays on and wont change, so is this what happens when its fried?

This could be possible in your situation.

Gary Dillard

10-01-2008 23:12

Re: Flawed IR Design Confirmed

Quote:

Originally Posted by EricVanWyk (Post 674846)

The IR board is to send commands, not state. Put a software latch or low pass filter in.

Can't you use it to send state, just not to toggle state? Read the input to set state, and remain in that state until it receives another, different state input.

Quote:

Originally Posted by SWATRocks

So does anyone know where we would be able to get another IR sensor board?

Still waiting for that info (I have a thread in the KOP forum looking for IR board and Trackball supplier info).

RyanN

10-01-2008 23:19

Re: Flawed IR Design Confirmed

Quote:

Originally Posted by EricVanWyk (Post 674846)

It is not directly tied to a victor through software. We have it setup to where if it detects that a button is pressed, then it runs a loop 10 times that runs the motors. This is not a coding issues at all. It worked perfectly for about 10 minutes. The IR board will not power on.

PhilBot

10-01-2008 23:57

Re: Flawed IR Design Confirmed

Quote:

Originally Posted by keen101 (Post 674123)

...And yes, I think it is a flawed design.
..and yes, it is my fault for connecting it wrong, but a better design would help.

Seriously but.... When you accidentally connect the 12V to one of the logic output pins, where will you put the diode to correct that design flaw :)

Phil.:rolleyes:

keen101

11-01-2008 01:13

Re: Flawed IR Design Confirmed

I have finally wired it up to a PWM cable. There should be no need for a diode now, since the PWM cable cannot provide reverse voltage. But, i suppose if i was really paranoid I could splice the red wire on it and add a diode there.

eeeeweeezeee

11-01-2008 01:16

Re: Flawed IR Design Confirmed

when i connected mine up, i very carefully checked with clip leads the first time and had someone with their finger on the off switch for the robot when i hooked it up earlier today. always test things like this very very carefully, and make sure it does exactly what its supposed to immediately. if it doesn't, pull power and check everything before trying again. speaking of needing new IR boards, does anyone know where to get them, and how much they cost? apparently IFI is out of the loop. I didn't burn out our existing ones, but they are kinda neat and if they are cheap I'm going to want some for personal projects.

keen101

11-01-2008 01:34

Re: Flawed IR Design Confirmed

I don't know where you can get the ones from diversified, but here is a link to buy the original TinyIR2 modules, which have up to 20 commands!!!!

http://www.tauntek.com/tinyir2-learn...l-receiver.htm

The Lucas

11-01-2008 10:15

Re: Flawed IR Design Confirmed

Quote:

Originally Posted by keen101 (Post 674955)

If you are using the PWM cable to connect it to the RC's PWM Output (which draws power from the 7.2V backup battery), the 0.6V voltage drop of the diode would likely cause the voltage to drop below the required 7V (or only work for a short time since fully charged backups can be up to 8.4V). Compatibility with the backup battery might be the reason there is no diode already on the board.

P.S. you can get reverse voltage from a PWM cable if you flip the connector and plug in only 2 pins (red to black). I've seen them plugged in that way on a crowded RC.

Al Skierkiewicz

11-01-2008 10:56

Re: Flawed IR Design Confirmed

The 78XX series of regulators need at least two volts across the device to meet regulation spec. That is why the RC drops out below 8 volts, it also has a 7805 internal. (BTW that is why the current in the RC is limited to 1.5 amps, the max current spec on a 7805) When the input falls below the required 2 volt differential the output falls at the same amount. This is a problem that gets power supply designers into trouble all the time. If the input ripple from an AC simple linear power supply (wall wart) falls below the 7 volt minimum, then the output of the 7805 suddenly has ripple.
The 7.2 volts battery may measure at 8.4 volts when it is charging or slightly after. It is generally at 7.2 volts, or 1.2 volts x 6 cells for NiCad battery chemistry.

keen101

11-01-2008 10:58

Re: Flawed IR Design Confirmed

Quote:

Originally Posted by The Lucas (Post 675125)

P.S. you can get reverse voltage from a PWM cable if you flip the connector and plug in only 2 pins (red to black). I've seen them plugged in that way on a crowded RC.

Not really, because if the wire only has two pins, then it's not really a PWM cable it it?:D

esquared

16-01-2008 12:50

Re: Flawed IR Design Confirmed

Quote:

Originally Posted by Kat Kononov (Post 674730)

If I add a diode to the IR board, will it prevent it from breaking in case the polarity is accidentally reversed? If so, where should I add it?

IMHO, a small series fuse in line with your +12V input and a reverse-biased (cathode to the +12V, anode to GND/return) diode in parallel with your +12V/GND leads at the input.

How to pick a diode:
1.) Low forward voltage drop when conducting otherwise your IC still sees too large of a negative voltage across it's terminals
2.) Non-repetitive forward current rating - need to match this value (and the time it can withstand this value) to your fuse to ensure the fuse will blow before your diode does!
3.) V(br)r (reverse breakdown voltage) > the voltage you are sitting across when not conducting

How to pick a fuse:
1.) Fast-acting, since this board doesn't have a lot of inrush current (inrush current = initial current spike when powering on and charging up capacitors and associated IC's) or current spikes where you wouldn't want the fuse to blow.
2.) ~1.5x your expected maximum current draw to avoid nuisance blow, as in 1. Easy to spec, since the LM78L05 can only source 100mA, anything above that is an error condition.
3.) Small - no giant breaker or glass BUSS fuses, lets stick it right on the board

Possible solution:
Diode - SD103AW,BW, or CW
Fuse - 0603FA250-R

Alternatively:
Use a series diode (heck even the one mentioned above) with an LDO (Low dropout regulator) instead of the standard linear LM78L05. Something such as the MIC2950 which has a dropout voltage (as Al S. mentioned) of 0.3V@100mA would mean that your circuit would work down to 5V + 0.3V+0.6V (forward voltage of the diode above) or 5.9V! If your backup battery is down to 5.9V you're in trouble :)

At the end of the day, after typing this up someone will find a rule or a new Q&A stating we can't modify the boards anyway :rolleyes:

RyanN

20-01-2008 10:56

Re: Flawed IR Design Confirmed

Just to let anyone reading this post because your IR board died, IFI is now selling them: http://www.ifirobotics.com/first-store.shtml

All times are GMT -5. The time now is 07:56.

Page 2 of 2

Show 100 post(s) from this thread on one page