734 replies to this topic

[mjr]

http://www.vpforums....513#entry297158  thx mjr

 

Okay, replied there.

[Zablon]

I've given up on this    I find it really odd I have the same issue on 2 boards and no one else is having it. I would assume that means I'm doing something wrong, but....what ? *sigh* lol. 

[mjr]

I've given up on this    I find it really odd I have the same issue on 2 boards and no one else is having it. I would assume that means I'm doing something wrong, but....what ? *sigh* lol. 

 

I find it hugely mystifying too.  Maybe the common element is your PC?  Have you tried it on a different machine?

[boiydiego]

iam not giving up easyly just gone get the flipperbuttons everybody uses almost then iam going to play with the nudge setting,

only thing having a high deadzone would also mean that i must nudge to hard to get ball moving thats more a problem ,also 

i see lots of people mount the board before the shaker in front i just cant imagine that the shaker would not affect the nudge/ball , if just hitting the flipper (in zeroslope table ) so no contacter involved triggers it from 15cm away from board ..

Edited by boiydiego, 12 February 2015 - 04:10 PM.

[mjr]

Boydiego, are you using the Pinscape software or do you have a VirtuaPin plunger kit?  If you're using Pinscape, you shouldn't need a deadzone in VP at all, and a shaker motor shouldn't be causing significant motion on the ball.  What version of VP are you using, and what are your gain settings?  Did you see my post in your other thread asking about your electrical configuration?

Edited by mjr, 12 February 2015 - 06:54 PM.

[Zablon]

 

I've given up on this    I find it really odd I have the same issue on 2 boards and no one else is having it. I would assume that means I'm doing something wrong, but....what ? *sigh* lol. 

 

I find it hugely mystifying too.  Maybe the common element is your PC?  Have you tried it on a different machine?

 

 

yes 3 different machines...

[mjr]

 

I find it hugely mystifying too.  Maybe the common element is your PC?  Have you tried it on a different machine?

 

yes 3 different machines...

 

And are they different types of hardware and different OS versions?  I'm grasping at straws, but conceivable problems could be the type of USB controller hardware on the PC (which is usually built in to the motherboard, so if the machines all have the same type of motherboard you might need to find a 4th machine with different hardware) or the USB drivers in the version of Windows you're using (so again, if they all have the same Windows version, testing on a 4th machine with another Windows version could be worth a shot).

 

The problem you're having is so low-level that it seems like it's got to be something at the boot loader level, which is why I'm thinking the host hardware might be a factor.  If we hadn't already ruled it out, the symptoms would point directly at the version of the boot loader software you're using or the process you used to set it up, but no one can find anything you're doing wrong there.

Edited by mjr, 13 February 2015 - 12:58 AM.

[frodus]

Can you link to the files you're using? I recently had issues too and had to use different files.

[Zablon]

The systems I tried were all Win 7 64-bit, but different boards, etc. I do have an XP system I can try it on.

 

As for the files, they are the same files everyone else has, and someone already gave me theirs to try with no difference. I think it is telling that even many of the premade demos don't work, but some do just fine. Makes me think there is some hardware/code change on the boards..

Edited by Zablon, 14 February 2015 - 08:56 AM.

[boiydiego]

mjr i mounted the freescale on other location it disconnected it from usb but i pressed the reset button will that do anything to setup allready did ? also now deadzone 0  gain l/r 1300    u/d 2000  , the ball seems to allways wants to come back to its start position , there is also some aftershock from nudge viewable ..

Edited by boiydiego, 14 February 2015 - 01:34 PM.

[hauntfreaks]

Mike (MJR), I have spoken to you in the past... I had what i thought was a bad CCD... had it replaced and worked... I'm completing a second cab and have not had 3 ccd units replaced, so i believe its me (I also changed out the freescale board)... i've checked the wiring over 100 times... seems fine... in the past you asked me if  had an oscilloscope and didnt, but now have access to one... could you please send me the reading i should see...using the oscilloscope to test the CCD.... thanks

[mjr]

Mike (MJR), I have spoken to you in the past... I had what i thought was a bad CCD... had it replaced and worked... I'm completing a second cab and have not had 3 ccd units replaced, so i believe its me (I also changed out the freescale board)... i've checked the wiring over 100 times... seems fine... in the past you asked me if  had an oscilloscope and didnt, but now have access to one... could you please send me the reading i should see...using the oscilloscope to test the CCD.... thanks

 

Before you do that, I assume you've looked at the frame viewer in the test/setup software.  If not, you should give that a try first, since it's a lot easier.  That will let you see if the CCD is registering anything.  It could just be a matter of the light level being too high or too low, which the viewer can help you determine.

 

For the scope, you can hook it up to the different leads that connect between the KL25Z and the CCD and check that the waveforms look like what they should according to the CCD data sheet.  First just verify that you're getting power to the chip - you should see a 3.3V straight line between pins 1 and 13 on the CCD.  Next, I'd check the clock input (leads 4 and 10 on the CCD - should be wired together).  Connect the scope leads between here and ground.  When the KL25Z program is running, you should see a 3.3V square wave at a frequency of about 42 kHz.  That's the signal from the KL25Z to the CCD that tells the CCD to send data.  If that looks good, look at pin 2.  You should see a short pulse there about every 30 ms - so set the scope to look at about 30 Hz, and you should see a mostly flat line with one pulse (or set it to 10 Hz and look for about 3 pulses).  That's the signal that starts an exposure grab and read cycle.  If all is good so far, the KL25Z is working.  The remaining question is whether the CCD is holding up its end of the deal.  Look at CCD pin 12 - that's the analog data out pin.  Look at about 30 Hz here too - that will put approximately one read cycle across the width of the scope display.  The readout here is the voltage level coming out of the CCD output pin, which should correspond to the photo brightness level the CCD is reading on the sensor.  A high voltage (around 3.3V) is full saturation (i.e., the brightest level it can read), and a low voltage (about 0V) is full darkness.  If you shine a light on the sensor and put something opaque to cover part of the sensor, you should see a pattern of high and low voltages on the scope that matches the bright and dark areas on the sensor.

 

If you don't see anything on pin 12, you might try repeating that last test on pin 6 (rather than pin 12).  The TSL1410R is physically built out of two separate sensors that we daisy chain together.  Pin 12 has the daisy-chained output from the second sensor.  If the wiring between the two halves isn't set up properly, or if the second half-sensor is bad, you might still be able to see the output from the first half-sensor, which is on pin 6.  If you wired it according to my diagram, pin 6 and 12 should be wired together so you should see identical results, but it could be worth checking anyway in case there's a bad solder connection or broken wire.

 

Hopefully that'll get you started. Let me know what you see and I'll do what I can to help you interpret it.

mjr i mounted the freescale on other location it disconnected it from usb but i pressed the reset button will that do anything to setup allready did ? also now deadzone 0  gain l/r 1300    u/d 2000  

 

No, pressing the reset button just restarts the software.  It doesn't lose any of the saved settings (those are stored in non-volatile flash memory on the KL25Z, so they survive resets and power cycles).

 

This *doesn't* apply if you do the boot loader reset procedure, where you unplug the board, hold down the reset button, and plug it back in.  That wipes everything and goes into boot loader update mode.  But if you just pressed the reset button while it was plugged in, that just does a soft reset.

 

Gain in the 1000-2000 range for VP 9.9 seems about right - that's in line with my settings.  Note that if you have the u/d and l/r settings that different, it's probably a matter of your taste rather than realism - the accelerometer itself is very linear on both axes, so it should be closest to reality when the gains are the same on both axes.

[hauntfreaks]

before you do that, I assume you've looked at the frame viewer in the test/setup software.  If not, you should give that a try first, since it's a lot easier.  That will let you see if the CCD is registering anything.  It could just be a matter of the light level being too high or too low, which the viewer can help you determine.

 

Mike here is a video of the CCD test (CCD was in total darkness).... and pics of the wiring of my setup.... this is also the same setup as my other cab which is working perfect... other tests were 2 other freescale boards and hooking the CCD to my working setup, I also did put all wires on my meter and there all good, no breaks... all gave the same test results.... and this is also the third CCD from mouser, I can't see 3 in a row being bad, its why i'm blaming myself.... 

 

 

 

Edited by hauntfreaks, 19 February 2015 - 06:06 AM.

[Spektre]

Has anyone released a fork for this code using an analog input and an A/D? Lemming had said he worked on it.

[mjr]

 

before you do that, I assume you've looked at the frame viewer in the test/setup software.  If not, you should give that a try first, since it's a lot easier.  That will let you see if the CCD is registering anything.  It could just be a matter of the light level being too high or too low, which the viewer can help you determine.

 

Mike here is a video of the CCD test (CCD was in total darkness).... and pics of the wiring of my setup.... this is also the same setup as my other cab which is working perfect... other tests were 2 other freescale boards and hooking the CCD to my working setup, I also did put all wires on my meter and there all good, no breaks... all gave the same test results.... and this is also the third CCD from mouser, I can't see 3 in a row being bad, its why i'm blaming myself.... 

 

I agree, three bad ones in a row does seem impossibly unlikely.  So it looks like the CCD thinks it's almost fully lit even though you have it in total darkness.  I'm assuming that shining a light on it has no effect - i.e., it still just shows up as all white with the min and max around 64000?  I think what this means at the hardware level is that the analog output pin on the CCD (or, equivalently, the analog input pin on the KL25Z) is basically pegged at 3.3V.  It might also just mean that it's disconnected (internally or externally) and that the KL25Z is pulling up the disconnected pin to 3.3V, but I don't think it would make sense for it to use a pull-up on a pin allocated as an ADC (analog in) port.  One thing to try might be to disconnect just that wire - the one between KL25Z port PTB0 (pin 2 on jumper J10) and CCD pins 6 & 12 - and run the exposure test again.  If you see the same pattern (all white), try connecting KL25Z PTB0 to KL25Z ground (J9 pin 14) - that should make the exposure meter show solid black.  If *that* doesn't happen, it's almost certainly something on the KL25Z side rather than the CCD.

Edited by mjr, 19 February 2015 - 11:16 PM.

[Zablon]

hauntfreaks: just a question since you are using push pins on the KL25Z, did you try swapping out the existing one that is working with the one that isn't to see if it makes the other one stop working?

[hauntfreaks]

 

 

here is the testing pulling the PTBO wire.... it might be hard to see but its a dark grey and white pattern that has no change from darkness or bright light.... I also tried the grounding as you said... still the same, no change.... I have also hooked it to 2 other KL25Z boards and to my working cabs KL25Z 

 

 

Zablon.... I have hooked this CCD the working cab, but not vice versa....  

[mjr]

here is the testing pulling the PTBO wire.... it might be hard to see but its a dark grey and white pattern that has no change from darkness or bright light.... I also tried the grounding as you said... still the same, no change.... 

 

Just to clarify, the video is with the PTB0 port on the KL25Z completely disconnected from anything?  Looks like you're getting random voltages in the video - if you look at the min/max you can see they're randomly jumping all over the place.  A mid gray on the display is about what you'd expect with random inputs, as it's an average of random values.

 

Random voltages with the wire disconnected strikes me as what you'd expect, so I think this means the KL25Z is okay, at least as far as the ADC port.

 

However, you *should* see all black when you connect PTB0 to ground.  That should peg the voltage on the ADC input to 0V, which should register on the digitizer as values all very close to 0, which should show on the exposure meter as all black.  When you say you grounded it, you are talking about connecting the wire to the *KL25Z ground*, on J9 pin 14, right?  Something seems wrong on the KL25Z side if you're seeing random voltages (gray exposure meter) in this case.  Maybe you need to try grounding it through a resistor (just about any size should do - I'd try something around 1K or 10K).

 

I have also hooked it to 2 other KL25Z boards and to my working cabs KL25Z 

 

Again, to clarify, the results of that were...?  I assume you mean you tried each of these combinations and saw each of these results:

 

Cabinet KL25Z + Cabinet CCD  ---> working properly

Cabinet KL25Z + new (broken) CCD  --->  all black pixels

New KL25Z #1 + Cabinet CCD  ---> working properly

New KL25Z #2 + Cabinet CCD  ---> working properly

 

If it does turn out that your CCDs really have been breaking over and over, and you can't think of any obvious trauma you subjected them to, one possible culprit is static electricity.  These devices seem very solid mechanically, so I don't think you could easily break them with even slightly rough physical handling, and the solder pads are big enough that you're probably not going to kill it by ordinary soldering (assuming you're using a good soldering iron for electronics and not a welding torch, and that you're taking the ordinary amount of care that you'd use with any solid state device).  But I wouldn't be surprised if they're susceptible to damage by static discharge, given that their whole function is to manipulate very tiny amounts of electric charge.  For anyone else ordering these, please be attentive to good practices for handling static sensitive devices - ground yourself before handling it by touching some bare metal on a plugged-in PC or other electrically grounded appliance, and don't walk around on carpeted floors with it.

[hauntfreaks]

did another fast test... I grounded the PTBO to #14 (this time correctly) and the test CCD screen went black, as you said it should....

 

I also did a voltage test on #1 and #13  

results:

bright light = 3.3v

total darkness = 3.04v - 3.02v

[mjr]

did another fast test... I grounded the PTBO to #14 (this time correctly) and the test CCD screen went black, as you said it should....

 

I also did a voltage test on #1 and #13  

results:

bright light = 3.3v

total darkness = 3.04v - 3.02v

 

That's a little strange. Those pins are just the power supply (CCD pin 1 is ground, 13 is positive), so I'm a little surprised that they'd vary at all from the 3.3V supply voltage.  You might try measuring the current through those same pins - the drop in voltage makes me wonder if there's a short somewhere that's overloading the KL25Z power supply.  

 

The analog outs are pins 6 and 12, so the voltage between pins 1 and 6/12 *should* vary according to the light level when it's working properly - close to 0V in complete darkness and close to 3.3V in bright light.