|
Hello group,
I would like to share a troubleshooting story I am in the middle of.
I bought a broken TX802 on the internet. I knew it was broken when I bought it, I felt like taking a chance. The seller said it worked but has "digital noise".
It turns out this TX802 is in pretty good shape. Everything works except that the sounds are wrong. There appears to be something wrong with the output levels of the operators.
If I make an init sound, I am supposed to get a single operator sine wave with an organ envelope. The user interface shows the expected parameters. But when I play it, I get more or less the sound of algorithm 1 with all operators at 99. A harsh overmodulated FM sound. When I then change the algorithm to 32, I get a clean sine wave sound.
It's a fascinating puzzle for me. Almost all the explanations I can come up with would make the TX sound more broken than it actually does. The EGM and OPS2 chips cannot be broken because they can make clean sounds. (I have a DX7II which is my reference.) Data is getting corrupted only in a very specific place, which happens to affect the output levels.
My best guess at the moment is that there is a problem with the chip enable line of the EGM chip that causes it to latch bad output level data when it should not be reading data in the first place. But I have not seen evidence of that yet on my oscilloscope.
I don't know if anyone else finds this interesting, but if there is someone who finds it interesting, they are probably on this list. :)
Cheers, Jacob
|
Hi,
I did a full recap of the TX802, and it sounded awesome afterwards.
Electrolytic capacitors are dead on most 80s Yamaha Roland Korg gear.
To change all capacitors is a PITA, but the machine gets new expected life.
Just my suggest. Not every tech guy is capable of doing that without damaging the unit.
Good luck, and take care.?
JJ
Hello group,
I would like to share a troubleshooting story I am in the middle of.
I bought a broken TX802 on the internet. I knew it was broken when I bought it, I felt like taking a chance. The seller said it worked but has "digital noise".
It turns out this TX802 is in pretty good shape. Everything works except that the sounds are wrong. There appears to be something wrong with the output levels of the operators.
If I make an init sound, I am supposed to get a single operator sine wave with an organ envelope. The user interface shows the expected parameters. But when I play it, I get more or less the sound of algorithm 1 with all operators at 99. A harsh overmodulated FM sound. When I then change the algorithm to 32, I get a clean sine wave sound.
It's a fascinating puzzle for me. Almost all the explanations I can come up with would make the TX sound more broken than it actually does. The EGM and OPS2 chips cannot be broken because they can make clean sounds. (I have a DX7II which is my reference.) Data is getting corrupted only in a very specific place, which happens to affect the output levels.
My best guess at the moment is that there is a problem with the chip enable line of the EGM chip that causes it to latch bad output level data when it should not be reading data in the first place. But I have not seen evidence of that yet on my oscilloscope.
I don't know if anyone else finds this interesting, but if there is someone who finds it interesting, they are probably on this list. :)
Cheers, Jacob
|
I would troubleshoot with a midi editor like SoundQuest or
whatever to see exactly what gets changed realtime.
With an init patch you may be missing something.
If you get the other operator levels back to zero do they stay at
zero after saving?
Can you make a patch? from scratch set levels back to zero and
then set them like 10 20 30 then save and recall that patch?
If you're able to recall that patch I would compare the original
and the modified one (modified by the machine I mean) to verify my
findings.
If it's indeed *only* the OP output levels that get scrambled
only then I would proceed further chasing that error.
In fact one could take it even further and take a screenshot of a
patch with ALL values set in a custom way and check what else gets
changed before going into a single direction.
On 2/3/2023 10:31 PM, José Juan wrote:
toggle quoted message
Show quoted text
Hi,
I did a full recap of the TX802, and it sounded
awesome afterwards.
Electrolytic capacitors are dead on most 80s
Yamaha Roland Korg gear.
To change all capacitors is a PITA, but the
machine gets new expected life.
Just my suggest. Not every tech guy is capable
of doing that without damaging the unit.
Good luck, and take care.?
JJ
Hello group,
I would like to share a troubleshooting story I am in the
middle of.
I bought a broken TX802 on the internet. I knew it was
broken when I bought it, I felt like taking a chance. The
seller said it worked but has "digital noise".
It turns out this TX802 is in pretty good shape. Everything
works except that the sounds are wrong. There appears to be
something wrong with the output levels of the operators.
If I make an init sound, I am supposed to get a single
operator sine wave with an organ envelope. The user
interface shows the expected parameters. But when I play it,
I get more or less the sound of algorithm 1 with all
operators at 99. A harsh overmodulated FM sound. When I then
change the algorithm to 32, I get a clean sine wave sound.
It's a fascinating puzzle for me. Almost all the
explanations I can come up with would make the TX sound more
broken than it actually does. The EGM and OPS2 chips cannot
be broken because they can make clean sounds. (I have a
DX7II which is my reference.) Data is getting corrupted only
in a very specific place, which happens to affect the output
levels.
My best guess at the moment is that there is a problem with
the chip enable line of the EGM chip that causes it to latch
bad output level data when it should not be reading data in
the first place. But I have not seen evidence of that yet on
my oscilloscope.
I don't know if anyone else finds this interesting, but if
there is someone who finds it interesting, they are probably
on this list. :)
Cheers, Jacob
|
Nothing gets changed in the patch. This is not like a memory corruption problem.
What is happening is that what the display says does not match the
sound the machine makes. For example, the display says all output
levels are zero except OP1 with 99. I can save this just fine, that is
not the problem. The sound that comes out the machine is just not what
the display says.
The main CPU is responsible for editing and patch storage. The
secondary ("slave") CPU is responsible for sending the currently
playing sounds into the EGM, and the algorithm into the OPS2. This is
where the problem must be.
Using a software MIDI editor tells me nothing about what goes on with
the EGM. The MIDI editor can only talk to the main CPU and the main
CPU thinks everything is fine.
I suspect that the output levels and the L1, L2 and L3 levels are
combined because changing L1, L2 and L3 does nothing, just like
changing output levels do nothing
Changing L4 does work; if I set it to 99 on a carrier I get the drone
sound you would expect.
The other thing that _does_ work is setting envelope rates. If I take
an init patch and set R1 of OP2 and OP3 to 0, then I get the proper
sound (single operator sine wave).
My best guess is still that some but not all of the 32 bytes that go
into the EGM get corrupted. And it's always the same bytes that get
corrupted.
Op vr 3 feb. 2023 om 21:45 schreef belzrebuth s via groups.io
<belzrebuth@...>:
toggle quoted message
Show quoted text
I would troubleshoot with a midi editor like SoundQuest or whatever to see exactly what gets changed realtime.
With an init patch you may be missing something.
If you get the other operator levels back to zero do they stay at zero after saving?
Can you make a patch from scratch set levels back to zero and then set them like 10 20 30 then save and recall that patch?
If you're able to recall that patch I would compare the original and the modified one (modified by the machine I mean) to verify my findings.
If it's indeed *only* the OP output levels that get scrambled only then I would proceed further chasing that error.
In fact one could take it even further and take a screenshot of a patch with ALL values set in a custom way and check what else gets changed before going into a single direction.
On 2/3/2023 10:31 PM, José Juan wrote:
Hi,
I did a full recap of the TX802, and it sounded awesome afterwards.
Electrolytic capacitors are dead on most 80s Yamaha Roland Korg gear.
To change all capacitors is a PITA, but the machine gets new expected life.
Just my suggest. Not every tech guy is capable of doing that without damaging the unit.
Good luck, and take care.
JJ
El vie, 3 feb 2023, 21:24, <contact@...> escribió:
Hello group,
I would like to share a troubleshooting story I am in the middle of.
I bought a broken TX802 on the internet. I knew it was broken when I bought it, I felt like taking a chance. The seller said it worked but has "digital noise".
It turns out this TX802 is in pretty good shape. Everything works except that the sounds are wrong. There appears to be something wrong with the output levels of the operators.
If I make an init sound, I am supposed to get a single operator sine wave with an organ envelope. The user interface shows the expected parameters. But when I play it, I get more or less the sound of algorithm 1 with all operators at 99. A harsh overmodulated FM sound. When I then change the algorithm to 32, I get a clean sine wave sound.
It's a fascinating puzzle for me. Almost all the explanations I can come up with would make the TX sound more broken than it actually does. The EGM and OPS2 chips cannot be broken because they can make clean sounds. (I have a DX7II which is my reference.) Data is getting corrupted only in a very specific place, which happens to affect the output levels.
My best guess at the moment is that there is a problem with the chip enable line of the EGM chip that causes it to latch bad output level data when it should not be reading data in the first place. But I have not seen evidence of that yet on my oscilloscope.
I don't know if anyone else finds this interesting, but if there is someone who finds it interesting, they are probably on this list. :)
Cheers, Jacob
|
This is true. Plus, do not hesitate to change out the tantalum caps on the digital boards, too. Many electrolytic have been changed only to find problems in data addressing and especially at boot.? The little, amber, tantalum caps are usually to blame. This gear is 35 years old at least. ? L. ?
toggle quoted message
Show quoted text
From: [email protected] [mailto:[email protected]] On Behalf Of José Juan
Sent: Friday, February 03, 2023 2:31 PM
To: [email protected]; contact@...
Subject: Re: [YamahaDX] TX802 troubleshooting? Hi, I did a full recap of the TX802, and it sounded awesome afterwards. Electrolytic capacitors are dead on most 80s Yamaha Roland Korg gear. To change all capacitors is a PITA, but the machine gets new expected life. Just my suggest. Not every tech guy is capable of doing that without damaging the unit. Good luck, and take care.? ? Hello group, I would like to share a troubleshooting story I am in the middle of. I bought a broken TX802 on the internet. I knew it was broken when I bought it, I felt like taking a chance. The seller said it worked but has "digital noise". It turns out this TX802 is in pretty good shape. Everything works except that the sounds are wrong. There appears to be something wrong with the output levels of the operators. If I make an init sound, I am supposed to get a single operator sine wave with an organ envelope. The user interface shows the expected parameters. But when I play it, I get more or less the sound of algorithm 1 with all operators at 99. A harsh overmodulated FM sound. When I then change the algorithm to 32, I get a clean sine wave sound. It's a fascinating puzzle for me. Almost all the explanations I can come up with would make the TX sound more broken than it actually does. The EGM and OPS2 chips cannot be broken because they can make clean sounds. (I have a DX7II which is my reference.) Data is getting corrupted only in a very specific place, which happens to affect the output levels. My best guess at the moment is that there is a problem with the chip enable line of the EGM chip that causes it to latch bad output level data when it should not be reading data in the first place. But I have not seen evidence of that yet on my oscilloscope. I don't know if anyone else finds this interesting, but if there is someone who finds it interesting, they are probably on this list. :) Cheers, Jacob
|
It is not recommended to replace all caps, only the bad ones. Service experts say: If it works, leave it so.
Daniel Forró
toggle quoted message
Show quoted text
Hi,
I did a full recap of the TX802, and it sounded awesome afterwards.
Electrolytic capacitors are dead on most 80s Yamaha Roland Korg gear.
To change all capacitors is a PITA, but the machine gets new expected life.
Just my suggest. Not every tech guy is capable of doing that without damaging the unit.
Good luck, and take care.?
JJ Hello group, I would like to share a troubleshooting story I am in the middle of. I bought a broken TX802 on the internet. I knew it was broken when I bought it, I felt like taking a chance. The seller said it worked but has "digital noise". It turns out this TX802 is in pretty good shape. Everything works except that the sounds are wrong. There appears to be something wrong with the output levels of the operators. If I make an init sound, I am supposed to get a single operator sine wave with an organ envelope. The user interface shows the expected parameters. But when I play it, I get more or less the sound of algorithm 1 with all operators at 99. A harsh overmodulated FM sound. When I then change the algorithm to 32, I get a clean sine wave sound. It's a fascinating puzzle for me. Almost all the explanations I can come up with would make the TX sound more broken than it actually does. The EGM and OPS2 chips cannot be broken because they can make clean sounds. (I have a DX7II which is my reference.) Data is getting corrupted only in a very specific place, which happens to affect the output levels. My best guess at the moment is that there is a problem with the chip enable line of the EGM chip that causes it to latch bad output level data when it should not be reading data in the first place. But I have not seen evidence of that yet on my oscilloscope. I don't know if anyone else finds this interesting, but if there is someone who finds it interesting, they are probably on this list. :) Cheers, Jacob
|
If it is software problem, than replacing OS Eproms (three pieces) can help. But this can be hardware problem.?
Daniel Forró
toggle quoted message
Show quoted text
I would troubleshoot with a midi editor like SoundQuest or
whatever to see exactly what gets changed realtime. With an init patch you may be missing something.
If you get the other operator levels back to zero do they stay at
zero after saving?
Can you make a patch? from scratch set levels back to zero and
then set them like 10 20 30 then save and recall that patch?
If you're able to recall that patch I would compare the original
and the modified one (modified by the machine I mean) to verify my
findings. If it's indeed *only* the OP output levels that get scrambled
only then I would proceed further chasing that error. In fact one could take it even further and take a screenshot of a
patch with ALL values set in a custom way and check what else gets
changed before going into a single direction.
On 2/3/2023 10:31 PM, José Juan wrote:
Hi,
I did a full recap of the TX802, and it sounded
awesome afterwards.
Electrolytic capacitors are dead on most 80s
Yamaha Roland Korg gear.
To change all capacitors is a PITA, but the
machine gets new expected life.
Just my suggest. Not every tech guy is capable
of doing that without damaging the unit.
Good luck, and take care.?
JJ
Hello group, I would like to share a troubleshooting story I am in the
middle of. I bought a broken TX802 on the internet. I knew it was
broken when I bought it, I felt like taking a chance. The
seller said it worked but has "digital noise". It turns out this TX802 is in pretty good shape. Everything
works except that the sounds are wrong. There appears to be
something wrong with the output levels of the operators. If I make an init sound, I am supposed to get a single
operator sine wave with an organ envelope. The user
interface shows the expected parameters. But when I play it,
I get more or less the sound of algorithm 1 with all
operators at 99. A harsh overmodulated FM sound. When I then
change the algorithm to 32, I get a clean sine wave sound. It's a fascinating puzzle for me. Almost all the
explanations I can come up with would make the TX sound more
broken than it actually does. The EGM and OPS2 chips cannot
be broken because they can make clean sounds. (I have a
DX7II which is my reference.) Data is getting corrupted only
in a very specific place, which happens to affect the output
levels. My best guess at the moment is that there is a problem with
the chip enable line of the EGM chip that causes it to latch
bad output level data when it should not be reading data in
the first place. But I have not seen evidence of that yet on
my oscilloscope. I don't know if anyone else finds this interesting, but if
there is someone who finds it interesting, they are probably
on this list. :) Cheers, Jacob
|
Have your tried to replace OS Eproms?
Have you done diagnostics test?
There were some reports about noise at the outputs in some series of TX802 - solved by replacing opamps.
Problems with sound quality can be also caused by muting transistors. Somebody deactivated them or pull out, somebody replaced them.
Check the signal from the output to DAC, to find if the problem is in analog circuit or digital circuit.?
Check all crystals, too.
Daniel Forró
toggle quoted message
Show quoted text
Hello group, I would like to share a troubleshooting story I am in the middle of. I bought a broken TX802 on the internet. I knew it was broken when I bought it, I felt like taking a chance. The seller said it worked but has "digital noise". It turns out this TX802 is in pretty good shape. Everything works except that the sounds are wrong. There appears to be something wrong with the output levels of the operators. If I make an init sound, I am supposed to get a single operator sine wave with an organ envelope. The user interface shows the expected parameters. But when I play it, I get more or less the sound of algorithm 1 with all operators at 99. A harsh overmodulated FM sound. When I then change the algorithm to 32, I get a clean sine wave sound. It's a fascinating puzzle for me. Almost all the explanations I can come up with would make the TX sound more broken than it actually does. The EGM and OPS2 chips cannot be broken because they can make clean sounds. (I have a DX7II which is my reference.) Data is getting corrupted only in a very specific place, which happens to affect the output levels. My best guess at the moment is that there is a problem with the chip enable line of the EGM chip that causes it to latch bad output level data when it should not be reading data in the first place. But I have not seen evidence of that yet on my oscilloscope. I don't know if anyone else finds this interesting, but if there is someone who finds it interesting, they are probably on this list. :) Cheers, Jacob
|
They always say stuff like that. They work on a timeclock.? They gotta make money with every job.? I can take 10X as long if I desire. Point being, the digital components are most probably still good. The linear components are always suspect at this age. ? L. ?
toggle quoted message
Show quoted text
From: [email protected] [mailto:[email protected]] On Behalf Of Daniel Forró via groups.io
Sent: Friday, February 03, 2023 8:38 PM
To: [email protected]; josejuangallego@...
Subject: Re: [YamahaDX] TX802 troubleshooting? It is not recommended to replace all caps, only the bad ones. Service experts say: If it works, leave it so. ?
? Hi, I did a full recap of the TX802, and it sounded awesome afterwards. Electrolytic capacitors are dead on most 80s Yamaha Roland Korg gear. To change all capacitors is a PITA, but the machine gets new expected life. Just my suggest. Not every tech guy is capable of doing that without damaging the unit. Good luck, and take care.? ? Hello group, I would like to share a troubleshooting story I am in the middle of. I bought a broken TX802 on the internet. I knew it was broken when I bought it, I felt like taking a chance. The seller said it worked but has "digital noise". It turns out this TX802 is in pretty good shape. Everything works except that the sounds are wrong. There appears to be something wrong with the output levels of the operators. If I make an init sound, I am supposed to get a single operator sine wave with an organ envelope. The user interface shows the expected parameters. But when I play it, I get more or less the sound of algorithm 1 with all operators at 99. A harsh overmodulated FM sound. When I then change the algorithm to 32, I get a clean sine wave sound. It's a fascinating puzzle for me. Almost all the explanations I can come up with would make the TX sound more broken than it actually does. The EGM and OPS2 chips cannot be broken because they can make clean sounds. (I have a DX7II which is my reference.) Data is getting corrupted only in a very specific place, which happens to affect the output levels. My best guess at the moment is that there is a problem with the chip enable line of the EGM chip that causes it to latch bad output level data when it should not be reading data in the first place. But I have not seen evidence of that yet on my oscilloscope. I don't know if anyone else finds this interesting, but if there is someone who finds it interesting, they are probably on this list. :) Cheers, Jacob
?
|
Dear All
Help please !!!
I have a DX 7 initial version
Problem ...No sound generated at all ....is it a battery change that is required or else
Please extend your help . Many Thanks
With kind regards, Raj
From Mauritius
On Saturday, February 4, 2023, 02:15:04 AM GMT+4, <contact@...> wrote:
Nothing gets changed in the patch. This is not like a memory corruption problem. What is happening is that what the display says does not match the sound the machine makes. For example, the display says all output levels are zero except OP1 with 99. I can save this just fine, that is not the problem. The sound that comes out the machine is just not what the display says. The main CPU is responsible for editing and patch storage. The secondary ("slave") CPU is responsible for sending the currently playing sounds into the EGM, and the algorithm into the OPS2. This is where the problem must be. Using a software MIDI editor tells me nothing about what goes on with the EGM. The MIDI editor can only talk to the main CPU and the main CPU thinks everything is fine. I suspect that the output levels and the L1, L2 and L3 levels are combined because changing L1, L2 and L3 does nothing, just like changing output levels do nothing Changing L4 does work; if I set it to 99 on a carrier I get the drone sound you would expect. The other thing that _does_ work is setting envelope rates. If I take an init patch and set R1 of OP2 and OP3 to 0, then I get the proper sound (single operator sine wave). My best guess is still that some but not all of the 32 bytes that go into the EGM get corrupted. And it's always the same bytes that get corrupted. Op vr 3 feb. 2023 om 21:45 schreef belzrebuth s via groups.io <belzrebuth= [email protected]>: > > I would troubleshoot with a midi editor like SoundQuest or whatever to see exactly what gets changed realtime. > > With an init patch you may be missing something. > > If you get the other operator levels back to zero do they stay at zero after saving? > > Can you make a patch? from scratch set levels back to zero and then set them like 10 20 30 then save and recall that patch? > > If you're able to recall that patch I would compare the original and the modified one (modified by the machine I mean) to verify my findings. > > If it's indeed *only* the OP output levels that get scrambled only then I would proceed further chasing that error. > > In fact one could take it even further and take a screenshot of a patch with ALL values set in a custom way and check what else gets changed before going into a single direction. > > On 2/3/2023 10:31 PM, José Juan wrote: > > Hi, > > I did a full recap of the TX802, and it sounded awesome afterwards. > > Electrolytic capacitors are dead on most 80s Yamaha Roland Korg gear. > > To change all capacitors is a PITA, but the machine gets new expected life. > > Just my suggest. Not every tech guy is capable of doing that without damaging the unit. > > Good luck, and take care. > > JJ > > El vie, 3 feb 2023, 21:24, < contact@...> escribió: >> >> Hello group, >> >> I would like to share a troubleshooting story I am in the middle of. >> >> I bought a broken TX802 on the internet. I knew it was broken when I bought it, I felt like taking a chance. The seller said it worked but has "digital noise". >> >> It turns out this TX802 is in pretty good shape. Everything works except that the sounds are wrong. There appears to be something wrong with the output levels of the operators. >> >> If I make an init sound, I am supposed to get a single operator sine wave with an organ envelope. The user interface shows the expected parameters. But when I play it, I get more or less the sound of algorithm 1 with all operators at 99. A harsh overmodulated FM sound. When I then change the algorithm to 32, I get a clean sine wave sound. >> >> It's a fascinating puzzle for me. Almost all the explanations I can come up with would make the TX sound more broken than it actually does. The EGM and OPS2 chips cannot be broken because they can make clean sounds. (I have a DX7II which is my reference.) Data is getting corrupted only in a very specific place, which happens to affect the output levels. >> >> My best guess at the moment is that there is a problem with the chip enable line of the EGM chip that causes it to latch bad output level data when it should not be reading data in the first place. But I have not seen evidence of that yet on my oscilloscope. >> >> I don't know if anyone else finds this interesting, but if there is someone who finds it interesting, they are probably on this list. :) >> >> Cheers, Jacob >> >> >> >> > > > > > >
|
It can be anything, your description is not enough to do proper diagnostics.
Is it totally dead, no display?
Start with checking fuses, PSU voltages, replacing bad electrolytic caps in PSU, replacing battery, cleaning and reseating all connectors and IC’s in sockets.
When you get some display, do diagnostics test, load patch data and play.
Later you can consider replacing OS Eprom for the latest one.
Daniel Forro
toggle quoted message
Show quoted text
On Feb 4, 2023, at 13:10, raj juggapah via < rajdata01@...> wrote:
Dear All
Help please !!!
I have a DX 7 initial version
Problem ...No sound generated at all ....is it a battery change that is required or else
Please extend your help . Many Thanks
With kind regards, Raj
From Mauritius
On Saturday, February 4, 2023, 02:15:04 AM GMT+4, < contact@...> wrote:
Nothing gets changed in the patch. This is not like a memory corruption problem. What is happening is that what the display says does not match the sound the machine makes. For example, the display says all output levels are zero except OP1 with 99. I can save this just fine, that is not the problem. The sound that comes out the machine is just not what the display says. The main CPU is responsible for editing and patch storage. The secondary ("slave") CPU is responsible for sending the currently playing sounds into the EGM, and the algorithm into the OPS2. This is where the problem must be. Using a software MIDI editor tells me nothing about what goes on with the EGM. The MIDI editor can only talk to the main CPU and the main CPU thinks everything is fine. I suspect that the output levels and the L1, L2 and L3 levels are combined because changing L1, L2 and L3 does nothing, just like changing output levels do nothing Changing L4 does work; if I set it to 99 on a carrier I get the drone sound you would expect. The other thing that _does_ work is setting envelope rates. If I take an init patch and set R1 of OP2 and OP3 to 0, then I get the proper sound (single operator sine wave). My best guess is still that some but not all of the 32 bytes that go into the EGM get corrupted. And it's always the same bytes that get corrupted. Op vr 3 feb. 2023 om 21:45 schreef belzrebuth s via <belzrebuth= [email protected]>: > > I would troubleshoot with a midi editor like SoundQuest or whatever to see exactly what gets changed realtime. > > With an init patch you may be missing something. > > If you get the other operator levels back to zero do they stay at zero after saving? > > Can you make a patch? from scratch set levels back to zero and then set them like 10 20 30 then save and recall that patch? > > If you're able to recall that patch I would compare the original and the modified one (modified by the machine I mean) to verify my findings. > > If it's indeed *only* the OP output levels that get scrambled only then I would proceed further chasing that error. > > In fact one could take it even further and take a screenshot of a patch with ALL values set in a custom way and check what else gets changed before going into a single direction. > > On 2/3/2023 10:31 PM, José Juan wrote: > > Hi, > > I did a full recap of the TX802, and it sounded awesome afterwards. > > Electrolytic capacitors are dead on most 80s Yamaha Roland Korg gear. > > To change all capacitors is a PITA, but the machine gets new expected life. > > Just my suggest. Not every tech guy is capable of doing that without damaging the unit. > > Good luck, and take care. > > JJ > > El vie, 3 feb 2023, 21:24, < contact@...> escribió: >> >> Hello group, >> >> I would like to share a troubleshooting story I am in the middle of. >> >> I bought a broken TX802 on the internet. I knew it was broken when I bought it, I felt like taking a chance. The seller said it worked but has "digital noise". >> >> It turns out this TX802 is in pretty good shape. Everything works except that the sounds are wrong. There appears to be something wrong with the output levels of the operators. >> >> If I make an init sound, I am supposed to get a single operator sine wave with an organ envelope. The user interface shows the expected parameters. But when I play it, I get more or less the sound of algorithm 1 with all operators at 99. A harsh overmodulated FM sound. When I then change the algorithm to 32, I get a clean sine wave sound. >> >> It's a fascinating puzzle for me. Almost all the explanations I can come up with would make the TX sound more broken than it actually does. The EGM and OPS2 chips cannot be broken because they can make clean sounds. (I have a DX7II which is my reference.) Data is getting corrupted only in a very specific place, which happens to affect the output levels. >> >> My best guess at the moment is that there is a problem with the chip enable line of the EGM chip that causes it to latch bad output level data when it should not be reading data in the first place. But I have not seen evidence of that yet on my oscilloscope. >> >> I don't know if anyone else finds this interesting, but if there is someone who finds it interesting, they are probably on this list. :) >> >> Cheers, Jacob >> >> >> >> > > > > > >
|
I was about to say the same thing almost .. more information please Raj, what do you have on the display for a start? ? Good to hear you’re still with the group Daniel, funny, I was just thinking about you this morning and wondering what had happened to the DX group. This was before I saw the original post about the TX802 ? Regards, Brian. ? ?
toggle quoted message
Show quoted text
From: Daniel Forró via groups.ioSent: Saturday, 4 February 2023 3:21 PM To: [email protected]; rajdata01@...Subject: Re: [YamahaDX] TX802 troubleshooting ? It can be anything, your description is not enough to do proper diagnostics. Is it totally dead, no display? Start with checking fuses, PSU voltages, replacing bad electrolytic caps in PSU, replacing battery, cleaning and reseating all connectors and IC’s in sockets. When you get some display, do diagnostics test, load patch data and play. Later you can consider replacing OS Eprom for the latest one. ?
On Feb 4, 2023, at 13:10, raj juggapah via <rajdata01@...> wrote: ? I have a DX 7 initial version Problem ...No sound generated at all ....is it a battery change that is required or else Please extend your help . On Saturday, February 4, 2023, 02:15:04 AM GMT+4, <contact@...> wrote: Nothing gets changed in the patch. This is not like a memory corruption problem.
What is happening is that what the display says does not match the
sound the machine makes. For example, the display says all output
levels are zero except OP1 with 99. I can save this just fine, that is
not the problem. The sound that comes out the machine is just not what
the display says.
The main CPU is responsible for editing and patch storage. The
secondary ("slave") CPU is responsible for sending the currently
playing sounds into the EGM, and the algorithm into the OPS2. This is
where the problem must be.
Using a software MIDI editor tells me nothing about what goes on with
the EGM. The MIDI editor can only talk to the main CPU and the main
CPU thinks everything is fine.
I suspect that the output levels and the L1, L2 and L3 levels are
combined because changing L1, L2 and L3 does nothing, just like
changing output levels do nothing
Changing L4 does work; if I set it to 99 on a carrier I get the drone
sound you would expect.
The other thing that _does_ work is setting envelope rates. If I take
an init patch and set R1 of OP2 and OP3 to 0, then I get the proper
sound (single operator sine wave).
My best guess is still that some but not all of the 32 bytes that go
into the EGM get corrupted. And it's always the same bytes that get
corrupted.
Op vr 3 feb. 2023 om 21:45 schreef belzrebuth s via
<belzrebuth=[email protected]>:
>
> I would troubleshoot with a midi editor like SoundQuest or whatever to see exactly what gets changed realtime.
>
> With an init patch you may be missing something.
>
> If you get the other operator levels back to zero do they stay at zero after saving?
>
> Can you make a patch? from scratch set levels back to zero and then set them like 10 20 30 then save and recall that patch?
>
> If you're able to recall that patch I would compare the original and the modified one (modified by the machine I mean) to verify my findings.
>
> If it's indeed *only* the OP output levels that get scrambled only then I would proceed further chasing that error.
>
> In fact one could take it even further and take a screenshot of a patch with ALL values set in a custom way and check what else gets changed before going into a single direction.
>
> On 2/3/2023 10:31 PM, José Juan wrote:
>
> Hi,
>
> I did a full recap of the TX802, and it sounded awesome afterwards.
>
> Electrolytic capacitors are dead on most 80s Yamaha Roland Korg gear.
>
> To change all capacitors is a PITA, but the machine gets new expected life.
>
> Just my suggest. Not every tech guy is capable of doing that without damaging the unit.
>
> Good luck, and take care.
>
> JJ
>
> El vie, 3 feb 2023, 21:24, <contact@...> escribió:
>>
>> Hello group,
>>
>> I would like to share a troubleshooting story I am in the middle of.
>>
>> I bought a broken TX802 on the internet. I knew it was broken when I bought it, I felt like taking a chance. The seller said it worked but has "digital noise".
>>
>> It turns out this TX802 is in pretty good shape. Everything works except that the sounds are wrong. There appears to be something wrong with the output levels of the operators.
>>
>> If I make an init sound, I am supposed to get a single operator sine wave with an organ envelope. The user interface shows the expected parameters. But when I play it, I get more or less the sound of algorithm 1 with all operators at 99. A harsh overmodulated FM sound. When I then change the algorithm to 32, I get a clean sine wave sound.
>>
>> It's a fascinating puzzle for me. Almost all the explanations I can come up with would make the TX sound more broken than it actually does. The EGM and OPS2 chips cannot be broken because they can make clean sounds. (I have a DX7II which is my reference.) Data is getting corrupted only in a very specific place, which happens to affect the output levels.
>>
>> My best guess at the moment is that there is a problem with the chip enable line of the EGM chip that causes it to latch bad output level data when it should not be reading data in the first place. But I have not seen evidence of that yet on my oscilloscope.
>>
>> I don't know if anyone else finds this interesting, but if there is someone who finds it interesting, they are probably on this list. :)
>>
>> Cheers, Jacob
>>
>>
>>
>>
>
>
>
>
>
>
? ?
|
Many Thanks Daniel
Hints are perfect
Display
When I select cartridge ...it says.........not ready which is fine as I did not insert the cartridge.
Back on normal it says something which i will share with you and possible you can furher direcgt me
Many Thanks again Bro
With kind regards Raj
On Saturday, February 4, 2023, 08:21:59 AM GMT+4, Daniel Forró via groups.io <danforcz@...> wrote:
It can be anything, your description is not enough to do proper diagnostics.
Is it totally dead, no display?
Start with checking fuses, PSU voltages, replacing bad electrolytic caps in PSU, replacing battery, cleaning and reseating all connectors and IC’s in sockets.
When you get some display, do diagnostics test, load patch data and play.
Later you can consider replacing OS Eprom for the latest one.
Daniel Forro
On Feb 4, 2023, at 13:10, raj juggapah via < rajdata01@...> wrote:
Dear All
Help please !!!
I have a DX 7 initial version
Problem ...No sound generated at all ....is it a battery change that is required or else
Please extend your help . Many Thanks
With kind regards, Raj
From Mauritius
On Saturday, February 4, 2023, 02:15:04 AM GMT+4, < contact@...> wrote:
Nothing gets changed in the patch. This is not like a memory corruption problem. What is happening is that what the display says does not match the sound the machine makes. For example, the display says all output levels are zero except OP1 with 99. I can save this just fine, that is not the problem. The sound that comes out the machine is just not what the display says. The main CPU is responsible for editing and patch storage. The secondary ("slave") CPU is responsible for sending the currently playing sounds into the EGM, and the algorithm into the OPS2. This is where the problem must be. Using a software MIDI editor tells me nothing about what goes on with the EGM. The MIDI editor can only talk to the main CPU and the main CPU thinks everything is fine. I suspect that the output levels and the L1, L2 and L3 levels are combined because changing L1, L2 and L3 does nothing, just like changing output levels do nothing Changing L4 does work; if I set it to 99 on a carrier I get the drone sound you would expect. The other thing that _does_ work is setting envelope rates. If I take an init patch and set R1 of OP2 and OP3 to 0, then I get the proper sound (single operator sine wave). My best guess is still that some but not all of the 32 bytes that go into the EGM get corrupted. And it's always the same bytes that get corrupted. Op vr 3 feb. 2023 om 21:45 schreef belzrebuth s via <belzrebuth= [email protected]>: > > I would troubleshoot with a midi editor like SoundQuest or whatever to see exactly what gets changed realtime. > > With an init patch you may be missing something. > > If you get the other operator levels back to zero do they stay at zero after saving? > > Can you make a patch? from scratch set levels back to zero and then set them like 10 20 30 then save and recall that patch? > > If you're able to recall that patch I would compare the original and the modified one (modified by the machine I mean) to verify my findings. > > If it's indeed *only* the OP output levels that get scrambled only then I would proceed further chasing that error. > > In fact one could take it even further and take a screenshot of a patch with ALL values set in a custom way and check what else gets changed before going into a single direction. > > On 2/3/2023 10:31 PM, José Juan wrote: > > Hi, > > I did a full recap of the TX802, and it sounded awesome afterwards. > > Electrolytic capacitors are dead on most 80s Yamaha Roland Korg gear. > > To change all capacitors is a PITA, but the machine gets new expected life. > > Just my suggest. Not every tech guy is capable of doing that without damaging the unit. > > Good luck, and take care. > > JJ > > El vie, 3 feb 2023, 21:24, < contact@...> escribió: >> >> Hello group, >> >> I would like to share a troubleshooting story I am in the middle of. >> >> I bought a broken TX802 on the internet. I knew it was broken when I bought it, I felt like taking a chance. The seller said it worked but has "digital noise". >> >> It turns out this TX802 is in pretty good shape. Everything works except that the sounds are wrong. There appears to be something wrong with the output levels of the operators. >> >> If I make an init sound, I am supposed to get a single operator sine wave with an organ envelope. The user interface shows the expected parameters. But when I play it, I get more or less the sound of algorithm 1 with all operators at 99. A harsh overmodulated FM sound. When I then change the algorithm to 32, I get a clean sine wave sound. >> >> It's a fascinating puzzle for me. Almost all the explanations I can come up with would make the TX sound more broken than it actually does. The EGM and OPS2 chips cannot be broken because they can make clean sounds. (I have a DX7II which is my reference.) Data is getting corrupted only in a very specific place, which happens to affect the output levels. >> >> My best guess at the moment is that there is a problem with the chip enable line of the EGM chip that causes it to latch bad output level data when it should not be reading data in the first place. But I have not seen evidence of that yet on my oscilloscope. >> >> I don't know if anyone else finds this interesting, but if there is someone who finds it interesting, they are probably on this list. :) >> >> Cheers, Jacob >> >> >> >> > > > > > >
|
Hi Brian
Good to hear fron you as well
For sure am with you and the group as this keyboard is here to stay for the next century for sure !!
I will send the display messages
Thanks for your support
With kind regards Raj
On Saturday, February 4, 2023, 08:27:43 AM GMT+4, Brian <briantolson@...> wrote:
I was about to say the same thing almost .. more information please Raj, what do you have on the display for a start? ? Good to hear you’re still with the group Daniel, funny, I was just thinking about you this morning and wondering what had happened to the DX group. This was before I saw the original post about the TX802 ? Regards, Brian. ? ?
From: Daniel Forró via groups.ioSent: Saturday, 4 February 2023 3:21 PM To: [email protected]; rajdata01@...Subject: Re: [YamahaDX] TX802 troubleshooting ? It can be anything, your description is not enough to do proper diagnostics. Is it totally dead, no display? Start with checking fuses, PSU voltages, replacing bad electrolytic caps in PSU, replacing battery, cleaning and reseating all connectors and IC’s in sockets. When you get some display, do diagnostics test, load patch data and play. Later you can consider replacing OS Eprom for the latest one. ? On Feb 4, 2023, at 13:10, raj juggapah via <rajdata01@...> wrote: ? I have a DX 7 initial version Problem ...No sound generated at all ....is it a battery change that is required or else Please extend your help . On Saturday, February 4, 2023, 02:15:04 AM GMT+4, <contact@...> wrote: Nothing gets changed in the patch. This is not like a memory corruption problem.
What is happening is that what the display says does not match the
sound the machine makes. For example, the display says all output
levels are zero except OP1 with 99. I can save this just fine, that is
not the problem. The sound that comes out the machine is just not what
the display says.
The main CPU is responsible for editing and patch storage. The
secondary ("slave") CPU is responsible for sending the currently
playing sounds into the EGM, and the algorithm into the OPS2. This is
where the problem must be.
Using a software MIDI editor tells me nothing about what goes on with
the EGM. The MIDI editor can only talk to the main CPU and the main
CPU thinks everything is fine.
I suspect that the output levels and the L1, L2 and L3 levels are
combined because changing L1, L2 and L3 does nothing, just like
changing output levels do nothing
Changing L4 does work; if I set it to 99 on a carrier I get the drone
sound you would expect.
The other thing that _does_ work is setting envelope rates. If I take
an init patch and set R1 of OP2 and OP3 to 0, then I get the proper
sound (single operator sine wave).
My best guess is still that some but not all of the 32 bytes that go
into the EGM get corrupted. And it's always the same bytes that get
corrupted.
Op vr 3 feb. 2023 om 21:45 schreef belzrebuth s via
<belzrebuth=[email protected]>:
>
> I would troubleshoot with a midi editor like SoundQuest or whatever to see exactly what gets changed realtime.
>
> With an init patch you may be missing something.
>
> If you get the other operator levels back to zero do they stay at zero after saving?
>
> Can you make a patch? from scratch set levels back to zero and then set them like 10 20 30 then save and recall that patch?
>
> If you're able to recall that patch I would compare the original and the modified one (modified by the machine I mean) to verify my findings.
>
> If it's indeed *only* the OP output levels that get scrambled only then I would proceed further chasing that error.
>
> In fact one could take it even further and take a screenshot of a patch with ALL values set in a custom way and check what else gets changed before going into a single direction.
>
> On 2/3/2023 10:31 PM, José Juan wrote:
>
> Hi,
>
> I did a full recap of the TX802, and it sounded awesome afterwards.
>
> Electrolytic capacitors are dead on most 80s Yamaha Roland Korg gear.
>
> To change all capacitors is a PITA, but the machine gets new expected life.
>
> Just my suggest. Not every tech guy is capable of doing that without damaging the unit.
>
> Good luck, and take care.
>
> JJ
>
> El vie, 3 feb 2023, 21:24, <contact@...> escribió:
>>
>> Hello group,
>>
>> I would like to share a troubleshooting story I am in the middle of.
>>
>> I bought a broken TX802 on the internet. I knew it was broken when I bought it, I felt like taking a chance. The seller said it worked but has "digital noise".
>>
>> It turns out this TX802 is in pretty good shape. Everything works except that the sounds are wrong. There appears to be something wrong with the output levels of the operators.
>>
>> If I make an init sound, I am supposed to get a single operator sine wave with an organ envelope. The user interface shows the expected parameters. But when I play it, I get more or less the sound of algorithm 1 with all operators at 99. A harsh overmodulated FM sound. When I then change the algorithm to 32, I get a clean sine wave sound.
>>
>> It's a fascinating puzzle for me. Almost all the explanations I can come up with would make the TX sound more broken than it actually does. The EGM and OPS2 chips cannot be broken because they can make clean sounds. (I have a DX7II which is my reference.) Data is getting corrupted only in a very specific place, which happens to affect the output levels.
>>
>> My best guess at the moment is that there is a problem with the chip enable line of the EGM chip that causes it to latch bad output level data when it should not be reading data in the first place. But I have not seen evidence of that yet on my oscilloscope.
>>
>> I don't know if anyone else finds this interesting, but if there is someone who finds it interesting, they are probably on this list. :)
>>
>> Cheers, Jacob
>>
>>
>>
>>
>
>
>
>
>
>
? ?
|
Op za 4 feb. 2023 om 05:12 schreef raj juggapah <rajdata01@...>: I have a DX 7 initial version Could we move this discussion to its own thread? It has nothing to do with the TX802 problem this thread started with
|
Op za 4 feb. 2023 om 03:39 schreef Daniel Forró via groups.io <danforcz@...>: It is not recommended to replace all caps, only the bad ones. Service experts say: If it works, leave it so. I'm with Daniel on this. I have no trouble replacing capacitors myself but I don't want to solder and desolder on the board without good reason. There are no tantalum capacitors in the TX802. Because it's mostly digital there are also very few electrolytics. There is no point in me replacing DC blocking capacitors on the headphone output if there is a problem in the digital domain. There is one capacitor next to the EGM which I've already replaced with no result.
|
Op za 4 feb. 2023 om 03:44 schreef Daniel Forro <danforcz@...>: Have your tried to replace OS Eproms? I have re-seated them, if that is what you mean. I have an EPROM reader so I could read their contents and compare the checksums with those listed on . I don't find it a promising direction though. If an EPROM was corrupted there would be crashes, or more other things that go wrong. Have you done diagnostics test? Yes. There were some reports about noise at the outputs in some series of TX802 - solved by replacing opamps. As I mentioned in my first post, if I take an init patch and change the algorithm to 32 then I get a perfectly clean sound. That rules out any problems in the output stages. Problems with sound quality can be also caused by muting transistors. Somebody deactivated them or pull out, somebody replaced them. If the muting transistors were bad they would be bad all the time. They don't know what FM algorithm I'm using. Check the signal from the output to DAC, to find if the problem is in analog circuit or digital circuit. As I wrote above, I already know the DAC is fine. It would not be able to produce clean sine waves with algorithm 32 if it was broken. Check all crystals, too. If the crystals were misbehaving many other things would be wrong. MIDI reception would be messed up. There would be weird sound artifacts on the outputs. The outputs can produce clean sine waves with algorithm 32 and the crystals don't know about the algorithm.
|
Two things to avoid: dismissing input from Daniel Forro and making assumptions. This is first and foremost electronic troubleshooting.
Check all the Power Supply voltages. I have seen an increasing number of Yamaha PSU failures which are happening because a LOT of the Yamaha designs use a Switch Mode PS. I have been able to repair some of the PSUs, but not all. There are many threads online discussing this or that synth or effects unit needing PSU repairs and various successes and failures.
Look at the waveforms being discussed here with a scope. Don't have these tools? Why not? a decent DMM can be had for less than $100. A brand new excellent scope for under $400. At some point one realizes they don't have the knowledge, skills or tools. This is fine for learning, but don't expect infinite patience from online help.
I am not pro-cap replacement, but as far as the PSU is concerned this is a common problem with them. Definitely check the OS checksum, even better would be a CRC or a direct BIN file comparison.? The BIN files are everywhere.
Reseating the EPROM was a good move. Sockets get dirty. Bad connections cause a myriad of problems. Look at the schematics and Service Manual. Most of the Yamaha FM products have a huge similarity and some of the Service Manuals are very helpful, others less so.
toggle quoted message
Show quoted text
On 2/4/2023 1:24 AM, contact@... wrote: Op za 4 feb. 2023 om 03:44 schreef Daniel Forro <danforcz@...>:
Have your tried to replace OS Eproms? I have re-seated them, if that is what you mean. I have an EPROM
reader so I could read their contents and compare the checksums with
those listed on .
I don't find it a promising direction though. If an EPROM was
corrupted there would be crashes, or more other things that go wrong.
Have you done diagnostics test? Yes.
There were some reports about noise at the outputs in some series of TX802 - solved by replacing opamps. As I mentioned in my first post, if I take an init patch and change
the algorithm to 32 then I get a perfectly clean sound. That rules out
any problems in the output stages.
Problems with sound quality can be also caused by muting transistors. Somebody deactivated them or pull out, somebody replaced them. If the muting transistors were bad they would be bad all the time.
They don't know what FM algorithm I'm using.
Check the signal from the output to DAC, to find if the problem is in analog circuit or digital circuit. As I wrote above, I already know the DAC is fine. It would not be able
to produce clean sine waves with algorithm 32 if it was broken.
Check all crystals, too. If the crystals were misbehaving many other things would be wrong.
MIDI reception would be messed up. There would be weird sound
artifacts on the outputs. The outputs can produce clean sine waves
with algorithm 32 and the crystals don't know about the algorithm.
|
Hi Grant, Op za 4 feb. 2023 om 14:54 schreef grantpbt <grantbt@...>:
Two things to avoid: dismissing input from Daniel Forro and making
assumptions. This is first and foremost electronic troubleshooting.
I am not dismissing Daniel's input, I am reacting to his suggestions. I really appreciate that Daniel refers to actual problems that others have had with the TX802, such as the output muting transistors. Check all the Power Supply voltages. I have seen an increasing number of
Yamaha PSU failures which are happening because a LOT of the Yamaha
designs use a Switch Mode PS. I have been able to repair some of the
PSUs, but not all. There are many threads online discussing this or that
synth or effects unit needing PSU repairs and various successes and
failures. Logical analysis clearly shows that the problem is in the digital domain. Everything from the EGM chip on down functions correctly. So the only power line that matters here is the 5V digital supply. I have checked that, with an oscilloscope, and it looks fine. Take a look at the block diagram of the TX802 on page 4/5 of the service manual. Everything from the EGM on down is working correctly. That also implies that the "master oscillator" works correctly: if it didn't, you would never get clean sine waves (at the correct pitch!) out of the outputs. If you look on the left of the block diagram you see that the "master MPU" runs on its own clock crystal. It is responsible for receiving MIDI. If that clock was not running, or not running at the right rate, incoming MIDI data would be garbled. MIDI works fine so I know that clock crystal is fine. I am not pro-cap replacement, but as far as the PSU is concerned this is
a common problem with them. I agree. But this just doesn't look like a PSU problem. Remember that the main CPU works fine. The synth responds to MIDI, I can use all the buttons in the interface, the display works, there is no garbled text in it. It does not crash. The BIN files are everywhere. I wish that were true. I have spent a lot of time looking already. The only source I can find for the BIN files is , where you have to pay $10 a pop. Do you know where to find them? The ROMs are called XB223, XB224 and XB536. I agree that a byte for byte comparison is the right thing to do but I only have those checksums. And unfortunately, that page does not have the checksums for TX802 v1.2, which is what I have. Reseating the EPROM was a good move. Sockets get dirty. Bad connections
cause a myriad of problems. Look at the schematics and Service Manual. Thanks, I have spent a lot of time looking at the service manual before I even posted here. That is how I know about the EGM and OPS2 chips. Most of the Yamaha FM products have a huge similarity and some of the
Service Manuals are very helpful, others less so. Thanks, I agree. The DX7II is a close relative of the TX802. I have spent a lot of time reading its service manual too. In spite of them having the same sound generator ICs, the architectures of the two are different. In the DX7II the primary CPU drives the EGM/OPS2 and the secondary CPU handles keyboard inputs etc. In the TX802, the secondary CPU drives the EGM/OPS. This makes the flow of data quite different. You can clearly see this if you compare the block diagrams in the respective service manuals. The DX7/DX9 service manual has some high level information about the EGS and OPS sound generator IC's. The IC's in the TX802/DX7II are different however. For example, the old EGS has an 8 bit input address bus. The new EGM has only a 5 bit address bus. We have to guess what the inputs to the EGM exactly are. The weird thing is that the EGM is getting input data that is mostly correct (amplitude envelope rates are correct, amplitude L4 is correct, pitch EG works correctly) and only incorrect in some places: the output level and L1/L2/L3 of the operator amplitude envelopes. I see no evidence on my scope that they are broken, but I think I will try replacing the RAM chip of the secondary CPU and the 74HC138 address decoder that enables the EGM and OPS2 IC's. If I can get hold of the bin files I will also replace the ROMs because it's easy to do. Jacob
|
I thought I would write a quick note of encouragement for Jacob. I’m sorry you’re fighting such an odd problem, but as someone who reads the list, I appreciate your posting the details. This is a learning experience for all of us watching.
toggle quoted message
Show quoted text
Hi Grant,
Op za 4 feb. 2023 om 14:54 schreef grantpbt <grantbt@...>:
>
Two things to avoid: dismissing input from Daniel Forro and making assumptions. This is first and foremost electronic troubleshooting.
I am not dismissing Daniel's input, I am reacting to his suggestions. I really appreciate that Daniel refers to actual problems that others have had with the TX802, such as the output muting transistors.
Check all the Power Supply voltages. I have seen an increasing number of Yamaha PSU failures which are happening because a LOT of the Yamaha designs use a Switch Mode PS. I have been able to repair some of the PSUs, but not all. There are many threads online
discussing this or that synth or effects unit needing PSU repairs and various successes and failures.
Logical analysis clearly shows that the problem is in the digital domain. Everything from the EGM chip on down functions correctly. So the only power line that matters here is the 5V digital supply. I have checked that, with an oscilloscope, and it looks
fine.
Take a look at the block diagram of the TX802 on page 4/5 of the service manual. Everything from the EGM on down is working correctly. That also implies that the "master oscillator" works correctly: if it didn't, you would never get clean sine waves (at
the correct pitch!) out of the outputs.
If you look on the left of the block diagram you see that the "master MPU" runs on its own clock crystal. It is responsible for receiving MIDI. If that clock was not running, or not running at the right rate, incoming MIDI data would be garbled. MIDI works
fine so I know that clock crystal is fine.
I am not pro-cap replacement, but as far as the PSU is concerned this is a common problem with them.
I agree. But this just doesn't look like a PSU problem. Remember that the main CPU works fine. The synth responds to MIDI, I can use all the buttons in the interface, the display works, there is no garbled text in it. It does not crash.
The BIN files are everywhere.
I wish that were true. I have spent a lot of time looking already. The only source I can find for the BIN files is
http://www.musictechnologiesgroup.com/mtg_eproms.htm, where you have to pay $10 a pop.
Do you know where to find them? The ROMs are called XB223, XB224 and XB536.
I agree that a byte for byte comparison is the right thing to do but I only have those checksums. And unfortunately, that page does not have the checksums for TX802 v1.2, which is what I have.
Reseating the EPROM was a good move. Sockets get dirty. Bad connections cause a myriad of problems. Look at the schematics and Service Manual.
Thanks, I have spent a lot of time looking at the service manual before I even posted here. That is how I know about the EGM and OPS2 chips.
Most of the Yamaha FM products have a huge similarity and some of the Service Manuals are very helpful, others less so.
Thanks, I agree.
The DX7II is a close relative of the TX802. I have spent a lot of time reading its service manual too. In spite of them having the same sound generator ICs, the architectures of the two are different. In the DX7II the primary CPU drives the EGM/OPS2 and
the secondary CPU handles keyboard inputs etc. In the TX802, the secondary CPU drives the EGM/OPS. This makes the flow of data quite different. You can clearly see this if you compare the block diagrams in the respective service manuals.
The DX7/DX9 service manual has some high level information about the EGS and OPS sound generator IC's. The IC's in the TX802/DX7II are different however. For example, the old EGS has an 8 bit input address bus. The new EGM has only a 5 bit address bus. We
have to guess what the inputs to the EGM exactly are.
The weird thing is that the EGM is getting input data that is mostly correct (amplitude envelope rates are correct, amplitude L4 is correct, pitch EG works correctly) and only incorrect in some places: the output level and L1/L2/L3 of the operator amplitude
envelopes.
I see no evidence on my scope that they are broken, but I think I will try replacing the RAM chip of the secondary CPU and the 74HC138 address decoder that enables the EGM and OPS2 IC's. If I can get hold of the bin files I will also replace the ROMs because
it's easy to do.
Jacob
|
José Juan