Stec Records Forum - Discussions about software and hardware used to make music.
gfxgfx
 
Welcome, Guest.

Login with username, password and session length

November 22, 2017, 11:44:58 PM
logo
gfxgfx
gfxgfx
 
Notice - We are experiencing problems with our online registration system.  For now, you can register by sending us a message using our contact form (be sure to include a preferred user name).
gfxgfx

gfx* Home Help Search Calendar Login Registergfx
The Stec Records Forum  |  Audio Gear  |  MPX G2  |  Lex-2186 Error During Start Up « previous next »
Pages: [1] Go Down Print
Author Topic: Lex-2186 Error During Start Up  (Read 9782 times)
Keith
Lexicon Repairs
Administrator
Full Member
*****
Offline Offline

Posts: 120


Stec Records Forum PD


WWW Email
« Reply #13 on: February 03, 2013, 03:46:46 PM »

This is a perfect example on an article I wrote on this forum bout error codes or error messages Error translation This also applies to any other unit with built in self diagnostic tests.
« Last Edit: February 03, 2013, 04:08:05 PM by Keith » Logged

Mapex Pro M Maple standard, Pearl 800 series cymbal stands, Vex hi hat stand, Pearl P201P double bass pedals, Tama mid level throne, All Zildjian A cymbals, 21" Rock Ride, 18" med thin crash, 16" med thin crash, 10" splash, 14" new beat hi hats.
bsellon
Chief Administrator
Administrator
Hero Member
*****
Offline Offline

Posts: 770


Time reveals truth.


WWW Email
« Reply #12 on: January 27, 2013, 12:07:58 PM »

Excellent! Must be a pad adjacent to that pin that's a little too close.  Solder has a low melting point so it may change shape, slightly, over time even at normal operating temperatures.  I've had gear get quite hot out doors in the heat and work with no problems but it's probably not a great idea to have it vibrating when it's very hot; combine loosening of the binder with inertia. Good argument to include some kind of temperature monitor on the board, if only to log it.   It could easily have been an incorrect or bad batch of solder used for a batch of boards so you may have future problems along these lines.  If that's the case, please report the spot here so we all have a log of potential problem spots (expand the log I should say). Nice work.
Logged
Frank G.
Newbie
*
Offline Offline

Posts: 7

unregistered visitor


Email
« Reply #11 on: January 27, 2013, 12:42:59 AM »

Hi Bob. Problem solved! I desoldered and lifted pin 44 (8FS) of  XC3042  off its pad on the pcb and found that I no longer had a short. I turned on the unit and found that I now had a clock pulse on pin 44. I jumpered pin 44 to the 8FS test pad which is almost adjacent to pin 44. I turned on the unit and it initialized and ran without any errors. I now have reverb and delay and it appears that all works well. It appears that pin 44's pad was somehow shorting to ground. Thank you very much for all your assistance and for hosting this site. I couldn't have repaired my unit without your guidance and the information I found here..

Best regards,

Frank


Logged
bsellon
Chief Administrator
Administrator
Hero Member
*****
Offline Offline

Posts: 770


Time reveals truth.


WWW Email
« Reply #10 on: January 25, 2013, 01:49:11 AM »

Hmm. It still could a bad FPGA but that does suggest a short on the board.  Certain types of flux can oxidize in weird ways causing continuity between traces.  Scary shape but it could also just be a metal fragment floating around inside.

The drill is to follow the trace between the chips looking for anything unusual but also looking for possible cut/repair points where you can cut the etch to separate the two pins but safely and easily repair the etch.  Once the etch is cut, you check to see what side is still shorted to ground and work your way back to the chip.  With any luck, the short is somewhere between the two pins and you can cut the etches out of the circuit and add a little eco wire between the pins on the chips or a jump off point.
Logged
Frank G.
Newbie
*
Offline Offline

Posts: 7

unregistered visitor


Email
« Reply #9 on: January 25, 2013, 01:19:46 AM »

Thank you Bob. I  verified that there is no 8FS signal out of  the XC3042. Furthermore I discovered this evening that 8FS is in fact shorted to ground.  The short can be in the pcb, the 2186, the lexichip or the XC3042. Any suggestions on what to check first? I was suspecting that 8FS input pin 65 on the Lexichip might be shorting with ground pin 64 but after looking at these pins closely with a loop I see no visible evidence of a short.
Logged
bsellon
Chief Administrator
Administrator
Hero Member
*****
Offline Offline

Posts: 770


Time reveals truth.


WWW Email
« Reply #8 on: January 24, 2013, 07:46:12 PM »

Yes, the 8FS should be present at all times. Sounds like that's the problem. You want to check the signal at the output of the XC3042 chip itself. With any luck, its a bad etch or solder joint between the chips (just add a wire).

If there no signal out of the XC3042, the XC3042 FPGA (Field Programmable Gate Array (Bob - I think))  is probably bad. As it turns out, that's not as serious a problem as it could be. I'm not sure of the exact phase alignment between the clocks (I'm thinking they are very close) but you might be able to just divide the next higher clock externally.  I don't think there is any rocket science  here; the FPGA just implements a clock divider.  I found an example here: http://www.electronics-tutorials.ws/counter/count_1.html

I don't think this is related but another thing to keep in mind is that the FPGA is programmed automatically on power up from the ROM. If the data in the ROM itself was bad you'd get a checksum error but it's possible that one of address or data lines between the FPGA and the ROM is bad.  I think that if a data line was bad, it would be toast but one of the upper address lines might just cause it to not load a portion of its program or put it in the wrong place.  It's a long shot but I thought I'd mention it.
Logged
Frank G.
Newbie
*
Offline Offline

Posts: 7

unregistered visitor


Email
« Reply #7 on: January 22, 2013, 09:25:01 AM »

Bob - I checked the signals on the pins outlined in your previous response and I was able to confirm that square wave signals were present. I went further to verify clock signals coming from the FPGA XC3042. Per the users manual, the XC3042 divides _MC (512 FS) from the lexichip to make 256FS, 128FS, _128FS, 8FS, FS and FSQUAD. I checked all these clock signals and verified that all were functioning except for 8FS. 8FS showed a consitent 0VDC reading regardless of how I manipulated the controls on the G2's front panel. Should an 8FS clock signal be rpresent at all times? If so, does this mean that I have a faulty XC3042 or is there a potentially faulty input into the XC3042 that I should trace?
Logged
bsellon
Chief Administrator
Administrator
Hero Member
*****
Offline Offline

Posts: 770


Time reveals truth.


WWW Email
« Reply #6 on: January 16, 2013, 12:28:35 AM »

Yeah, you should be seeing some kind of data from the 2186 pin 31 output (DT0)  since that would contain the input data for the delays and reverbs.  It's interesting that the 2186 is doing all the other stuff it needs to do but won't output serial audio data to the Lexichip.  It may be that one of the clock signals feeding the 2186 is used exclusively for the Lexichip audio data so take a look at pin 35, 33, 32 and 42 on the 2186. Pin 35 is the SCLK0 so that's a good bet for the DT0 used for DSP_LEX_DATA.  All these clocks should just be square waves at various frequencies.

The fact that it was an intermittant problem and is not especially sensitive to temperature makes me hopeful that it's just a bad connection and not a bad part.
Logged
Frank G.
Newbie
*
Offline Offline

Posts: 7

unregistered visitor


Email
« Reply #5 on: January 13, 2013, 11:56:50 PM »

Thanks Bob. I did a little more hot/cold testing with and without pressure applied but could not perceive any change in behavior. I used a scope and checked the DSP LEX DATA at pin 31 on the 2186 and pin 63 on the Lexichip and all I was able to see is a continuous 5 VDC (straight line). I checked LEX DSP DATA on pin 34 of the 2186 and pin 61 on the Lexichip and  I verified serial data communication (0  to 5 VDC square wave pattern) between the two chips. I presume that I should  not be measuring continuous 5 VDC between pins 31 on the 2186 and pin 63 on the Lexichip correct?  What kind of waveform should I be looking for?  The above was true regardless of whether or not I set the G2 with verb and delay buttons on or off.
Logged
bob
Newbie
*
Offline Offline

Posts: 4


Email
« Reply #4 on: January 10, 2013, 02:42:53 PM »

No verb is even more indication that it's the serial audio data links between these chips.

Did you try the heat/cold test? 

I don't know if I mentioned this before but you want to listen with headphones if you can.  The changes might be subtle and more-easily missed through speakers.  Cutting in and out will produce an audible clicks which you'll hear more clearly if you use a low frequency tone or sound such as a timpani.  We've started putting some test mp3 files here: http://www.stecrecords.com/?RecordId=581

You can also try gently applying pressure to each of the chips while listening to audio through it to see if you can hear the reverb or delay kick in.  This can weaken any other borderline solder contacts so press very lightly.  If you hear it kick in at all you'll know it's just a bad connection. If you never hear anything it could be a bad chip.

The next step I'd take is to put a scope on the in and output of each chip mentioned earlier. Both chips and an input and both have an output, any of which could be bad so the best you can do is try to eliminate things.  If you see data being output from the Lexichip, its probably not the problem. Same for the 2186.  If you see data on all outputs and inputs I'd suspect the 2186.

Logged
Frank G.
Newbie
*
Offline Offline

Posts: 7

unregistered visitor


Email
« Reply #3 on: December 30, 2012, 03:25:55 AM »

I've reflowed the solder on the Lexi chip the 2186 and the FPGA and I still get the same error. I made sure the eprom is properly seated in its socket. I've tested for continuity on all relevant circuits including EPROM  and everything checks ok. One other thing I should mention is that when I get the "LEX 2186" error, there is no "E-" code on the display. I've also verified supply voltages and all my test points check OK. Any guidance anyone can offer would be greatly appreciated.
Logged
Frank G.
Newbie
*
Offline Offline

Posts: 7

unregistered visitor


Email
« Reply #2 on: December 20, 2012, 11:33:41 PM »

Bob - I'm not getting any reverb either.
Logged
Frank G.
Newbie
*
Offline Offline

Posts: 7

unregistered visitor


Email
« Reply #1 on: December 20, 2012, 03:04:57 PM »

Greetings all.

I am a new member to this group and I wanted to share a conversation I am having with Mr. Bob Sellon regarding the subject  start up problem I am having with my MPX G2. Please read our correspondence to date starting at the bottom of this entry and add your comments / recommendations as you see fit.

I'd like to thank Bob again for all his help thus far and I look forward to sharing my experiences and ultimate resolution of this issue with this group!

Best Regards,

Frank
__________________________________________________________________________________________________________________________________________

You didn't mention reverb.  Is the reverb also not working?  Reverb and delay are both done by the Lexichip.

A bad Lexichip is certainly a possibility but I haven't had a lot of experience with chips becoming intermittent. Once they develop a problem that's usually it but it's possible that the chip itself is coming apart.  You can sometimes localize a problem by using a heat gun and freeze spray.  I would sometimes use the cardboard center of a roll of toilet paper to keep the heat or cold on a single chip while listening to a simple "bong" or other pure tone on the reverb to see if I could hear a change when a particular part of the circuit was sensitive.  When chips fail, they will sometimes change as they change temperature.

If you can, it might be worth trying to reflow the solder on just those pins with a fine-tip soldering iron.  If there is bad solder, the pressure of the multimeter probe will make it work while you are doing your test.  You could also set up a reverb or delay program and see if it cuts in/out when you press on those pins or the chips.

I hope you don't mind but I'm going to put this discussion on our forum so others can benefit from it.

__________________________________________________________________________________________________________________________________________
Hi Bob,
 
Thank you again for your reply. I checked the pins/circuits below and I have continuity between the two chips at the leads. Off the top can you think of anything else I can check? Are there other pins that handle serial communication between the two chips? Is it possible that my Lex chip is bad? As I indicated before the unit works well except for delay functions.
 
Thanks for your help,

Frank
__________________________________________________________________________________________________________________________________________

Hi Frank,
You are on the right track inspecting and trying to reflow solder.  My buddy Keith at TheJukeYard.com repairs a lot of G2s and he's mentioned problems with solder, especially with the Lexichip (U85).  I'm pretty sure he has a special tool for his soldering station for reflowing large chips but a heat gun might do the job.  Definitely want to be careful not to overheat the chips or the board which can create whole new classes of problems.

I'm pretty sure the Lex-2186 test checks the serial communications between the Lexichip which handles reverb and delays, and the Analog Devices 2186 DSP chip which handles all the other audio processing.  Pin 61 is the output and pin 63 is the input on the Lexichip and pin 31 and 34 are the out and ins respectively on the 2186 (U88).  Take a close look at these pins but you can also put an ohm meter between them to see if there is continuity and even add a tiny wire to connect the pins manually if there is an open you can't see (crack in the etch under the chip, etc.).

I hope this helps,
Best regards,
Bob Sellon
Stec Records

__________________________________________________________________________________________________________________________________________
Sorry Bob.. I neglected to mention below that I ran all the other built in diagnostic tests (i.e. DRAM, WCS, SRAM, ROM Checksum, Digipot, Sample Rate etc.) and everything checks out OK. The only error I am getting is the Lex-2186.
 
Best regards,
Frank
 
__________________________________________________________________________________________________________________________________________
Hi Bob,
 
I have been doing a lot of research on the web with regards to an issue I am having with my Lexicon MPX G2 unit and I see you are a well-known and well respected authority on this subject. I  was wondering if you could provide a little guidance. During start up I encounter a LEX-2186 failure. This started as an intermittent problem but is now a consistent issue. When I push the “program” button after this error is encountered the unit starts up and is fully functional except for delay functions. I have read the MPX G2 service manual and have not been able to resolve this issue. I have inspected the board visually for poor connections / solder joints and have found none. I have even tried to reflow the board in the area of the Lex and 2186 chips using a heat gun but still no resolution. I am reaching out to you as  a last resort to determine if there is anything further I should try to resolve this problem.
 
Thank you very much in advance for any help you can provide and I apologize if I am disturbing you in any way by sending this e mail.
 
Kind Regards,

Frank Golemis
Logged
The Stec Records Forum  |  Audio Gear  |  MPX G2  |  Lex-2186 Error During Start Up « previous next »
Pages: [1] Go Up Print 
Jump to:  
Powered by MySQL Powered by PHP SMF 2.0.2 | SMF © 2011, Simple Machines
Apollo design by Bloc
Valid XHTML 1.0! Valid CSS!
Page created in 0.128 seconds with 20 queries.