Random patch resetting - was yellow oled issue

[Road_Kill]

Greetings! I have a metal cased, R5C PreenFM2 complete with the wonderful Newhouse yellow OLED. I've had the synth for years and it's never worked right. However, what playing I've done has convinced me that this is - or could be - my all time favorite FM synth even with all the newer entries now on market. Alas, I get to working on it, resolve a problem or two, then life causes the work to be sidelined, repeat cycle, and here we are in 2019 and I still want a working Preen! When I first received and assembled the kit, there was a faulty trace that I fixed that was a manufacturing error on some of the r5c's. Shortly there after, I started getting display errors once the synth had been running for a while. Turns out the OLED was running really hot with way too much current draw. While the evidence indicated the OLED was defective from the gitgo, I purchased a replacement. Then the display errors returned, but this time the OLED was running cool and within spec. Flash forward to the present. I have performed the 22ohm resistor/shielding plate modification on the synth and now my display is rock solid and crystal clear...

HOWEVER, now when I run the synth, it will randomly unload/reset/reinitialize any loaded patches I have. For example, if I load a combo, a few seconds later it's changed to Preset 004, 005, and/or 006. Even that isn't stable as sometimes I'll watch a voice occasionally toggle from one value to the next (e.g. Back and forth between Preset 005 and 006). This seems to occur with the same random nature that the OLED display error occurred and has got me wondering if this has been going on the whole time I've had the new OLED, but went undetected because I was fixated on the display errors at the time.

Could this be noise on the USB_DP & USB_DN lines? Anybody else experience this? I'm VERY interested in 'eurofying' my Preen, but I need to get this synth up and running and rock solid first. It's been too long! 

TIA

[Road_Kill]

FYI - it's currently running OS 2.07. (I've recently moved and my shop and computer are still getting unpacked.) Also, the problem is much more apparent and occurs far faster when midi is being received.

[Xavier]

I never heard of such a problem before.
All that sounds like bad solder joint or shortcut somewhere.

Maybe around the 74H165 and the encoder/button area with the 2 network resistors.
It can also be a ground problem that would make the encoder/button to randomly trigger values and the display not to be stable.

The only solution is to take time with a multimeter and verify carfefully all solder joints.

You can start by checking the voltages of the first schema here :
http://ixox.fr/forum/index.php?topic=63509.msg65602#msg65602

Then use the second one o verify the tracks that start at MCLK, MQH and MLD on the preenf405 to the 74HC165 et resistor network.

Xavier

[Road_Kill]

Hi Xavier,

I think I might know the issue. I haven't had time to disassemble the synth yet, but I went back to my original post about the bad trace on the r5c board I had to see which pins I had to jump together to complete the damaged path (https://ixox.fr/forum/index.php?topic=63547.msg65848#msg65848). Sure enough, it's pin 6 MCLK on the 405 over to a 74H166. The wire used to effect the repair runs directly under the insulated & grounded plate used for the yellow OLED fix. Maybe I'll go back and find the damaged trace and see if I can bridge it where it's damaged so I can then remove the wire under the OLED shield..? In the mean time, I'm open to any and all suggestions!

[Xavier]

I'm not sure at all about your conclusion 

Could you post high quality pictures of your PCB without OLED, both side ?

[Road_Kill]

Xavier - Apologies for not replying sooner as this was a busy week. While I initially said I wouldn't be opening up the synth again until about this time, in fact I immediately turned around after my last reply and opened the synth back up! (Because who needs sleep when you could get a PreenFM up and running?)

The first thing I realized was that things were a lot more cramped with the insulated ground plate under the OLED than I initially thought. One consequence was that the wire coming from MCLK patching the defective trace was mashed such that it was possible intermittent shorts could occur with adjacent pins. And while I didn't see any evidence, I couldn't rule out random shorts occurring with solder joints piercing the insulation of the grounded shield. So I covered the circuit board under the shield with a layer of electrical tape, paying extra attention to really isolating the wire tacked to pin 6 of the 405.

I have been running the synth continuously since then, with and without sequencers running for many hours, and with as much hands-on as the past few days have afforded me. I think it's safe to declare this synth operational! So yes, it took you telling me to look at pin 6 to actually consider that might be the problem despite the big ugly wire hanging off of it. But now I've hopefully made the synth robust and if not, I know where to concentrate future improvements/modifications.

Once again,
Thank You!