News:

Forum may be experiencing issues.

Main Menu

BOSS CE1 PROJECT: MN3002 or TDA1022P: DOC UPLOADED

Started by drog_trog, March 14, 2018, 01:09:54 PM

Previous topic - Next topic

RDL68

More than another week later, I've finally got around to looking at my non working CE1 again.
I stripped it back to bare bones, took out the 2 mini mod PCB's (for blend & hi/low toggle), disconnected the stereo out jack & then socketed in a new Recom power module.
The bypass LED was still working but the flashing rate LED/s weren't. A quick voltage check on those told me they were getting the correct fluctuating voltage, so considering it's a situation where the original power module burnt out (& stunk me out) a few weeks ago, it's looking like it probably caused a power surge & wiped a few things out with it.
I'll have to continue later by checking the IC's in my IC tester, possibly replacing all the transistors etc. I'll get there eventually, it's just been quiet in the CE1 thread so I threw a post in for those watching it.
I did plug a guitar in this time, got no wet signal through it, but was getting a ticking noise in one of the modes (chorus or vibrato) whose ticking noise rate was adjustable by the speed pot, so at least some of it is working. The Recom wasn't overheating this time around either, nor did anything else then or now, but it definitely looks like there will be a few parts to replace when I determine which ones exactly. I'll have any replacement parts needed on hand, so no worries there.
I hope Skyled had better luck first go, his list of mods was much more complex than mine. :)

RDL68

I had another quick look at it before calling it a day. All the op-amps tested OK. Turns out there was nothing wrong with the rate LED/s either, they just weren't working because the ground lug on the bypass 3pdt was no longer happening for some yet to be determined reason. Hopefully my fears of a surge when the power module failed aren't likely now though.
So, for now, I moved the rate LED ground back to the main PCB to get it going (even though it keeps flashing in bypass mode from there). There's still an issue either with the bypass 3pdt possibly being faulty or something elsewhere causing a problem with that ground lug. The other odd thing with the rate LED is it's only flashing in vibrato mode now & permanently on in chorus mode.
One other serious thing I discovered was in effect mode the output jack is shorted.
Far out, I've made dozens of effects, but I haven't borked one this bad before, I'll have to give myself an uppercut.
Definitely a few things for me to work through later.
Next time I look at it I'll disconnect the middle lug in that row of the bypass 3pdt & check it's switching properly & go from there, but it sounds like that will be the least of my worries. I'm not even at the audio probe stage yet...




RDL68

Got back to debugging this misfiring CE1 today.
I first spotted & cleaned up a short on the PCB that seemed to fix any issues of the bypass 3pdt. To be 100% sure there wasn't a secondary problem with the bypass 3pdt I replaced it.
The output jack was no longer shorting when the effect was on by then.
Brilliant, it was audio probe time. Signal made it to the BBD input but not the output of it, so I dropped in a new one & it worked, signal to the output jack.
The rate LED's were flashing for both vibrato & chorus mode again, not just vibrato mode as it was when I last looked at this thing a couple of days ago.
At least by the time I called it a day today, I had a guitar in & hearing chorus.
However, there's still a few issues to sort out before I can move onto reintroducing the stereo out jack & any mods one by one.
Firstly there's a pulsing hum present in both chorus or vibrato mode. Also, not really getting a vibrato mode.
The Level pot has a scratchy weirdness to it, but works. But for a new one doesn't seem right.
And the Recom power module was starting to get a bit warm after a while, which may or may not be normal.
I'll have a think about these last few things for a while & then get back to it when I can.



RDL68

In addressing the issues with my CE1...
Replaced the Level pot, as it was sounding scratchy & weird when turning it.
Chorus mode was working, not Vibrato, so using the schematic, following the route from the Ch/Vib 3pdt I replaced C35,C36,D13,D14 & put in a new LM1458.
There was a loud pulsing hum in both Ch or Vib modes, so considering I already replaced the Recom, I replaced whatever else was in the power supply section, which was C7,C11,C12,C16,C17, also the 3 diodes, D1,D2 & D6.
Powered it up & this is the story. Level pot is no longer sounding abnormal when turned. The loud pulsing hum in either mode is gone, but a lower volume constant hum is present in both modes.
And here's the bummer, after all that messing about, the MN3002 chip seems to have died, no longer getting Chorus & audio probe is now stopping at the MN3002 input pins. As I don't have another MN3002, I'm unable to continue working on this one indefinitely.

Someone on Ebay back in September advertised 100 x MN3002's, in lots of 10 for $100, but one buyer came along & bought them all for $1000. The only other MN3002's on Ebay (at the time of writing) are not Panasonic BBD's. Apparently (from reading around the web) there's a new chip called MN3002 that has an entirely different function. On top of that there's bound to be fake Panasonics & remarked chips around as well.

As for the Dod Flanger 640, some vintage RCA brand CD4013AE chips turned up. They tested OK in my IC tester, so I decided to replace the 4 old pots in the 640 while I was at it. They were 3x D100k & 1x D500k. I just used 3x B100k & 1x A500k. Got all that done & put the CD4013AE in. Powered it up & although I now get a slightly filtered clean signal through in effect mode (as apposed to nothing at all), the audio probe is still stopping at the input pins of all the SAD1024's I've tried in it. When the Accent knob is on max, no signal through in effect mode. Looks like all the SAD1024's I have (4 of them) must be bad. I've eliminated every other possibility I can see. So this is another one I have to put away again indefinitely.




Scruffie

Unless the BBD's were purchased as fakes, 4 (5 if you include the MN3002) being dead is very unlikely, from a lot of BBD repair experience, I'd estimate they're the problem only 3-5% of the time unless they've been extremely poorly handled.

A BBD will usually be dead at the output, not the input as well. The CE-1 has a complex bias set up hooked to the input that could be causing your issue, have you tried probing the input without the BBD's installed?
Works at Lectric-FX

RDL68

Hi Scruffie,

I've only audio-probed the BBD's while they're socketed. The signal makes it to the input pins but not the output pins, so that is not a good sign.

With the CE1, it was chorusing as is, all that was changed between powering it up was a pot, some capacitors, diodes & a fresh LM1458 chip. So that BBD has gone from signal passing through the BBD to no longer reaching the output pins (signal only making it to the input pins).

With the Dod 640, literally everything has been replaced except the resistors. I've spent a lot of time making sure there are no shorts & that every trace goes where it should. I can never rule out 100% that there may be something somewhere causing this to happen, but my gut feeling is all the sad1024's are bad. They're very sensitive to age, static, handling & so forth. When used in Flangers they can fail a lot quicker just by using the feedback/accent knob at higher levels (perhaps in delay units that are tweaked for lots of repeats too), although that isn't the cause on this occasion.

I'm confident the ones I have aren't fakes, just more inclined to believe they have all failed for whatever reason. The last 2 reticons I bought were from a repair guy that had recently retired & closed his repair shop. He was just selling off a few odds & ends he had left over. They were from a large batch he sourced through official suppliers in the 80's & he'd used many from that batch over the years to repair vintage gear. They were stored professionally & mailed to me shielded & packed professionally, no bent pins, just NOS. It doesn't mean they were still working after so long & it doesn't mean I can rule out 100% that I may have missed something causing them to appear dead either, but they're sure looking dead. The other 2 I had were bought years ago, one new from an audio place in LA & one used & tested working before being mailed to me at the time by one of those synth guys, so my confidence in their survival isn't high after so long.

I've asked drog trog if he still has any of the TDA1022 version CE1 PCB's left, because at least I have all the parts here to whip up another PCB, including several TDA chips (which are at least still buyable to future proof it). If/when more MN3002's surface, I can get back to sorting that version out & also happy to have both versions or to give one away later.

It's just one of those things, it's not that I couldn't just buy an original CE1 or one of the clones floating around, it's more for the challenge & the whole DIY thing. A challenge that KO'd me on this occasion,hehe, but that's the way it goes sometimes. Also, I'm not one to post build reports, don't have the time or space for painting labels etc, there's so many people better at it than me, but I just liked the effort DT put into this one & wanted to support it as there wasn't much additional info or many people posting on it.

On a more crazier note, I've got another couple of sad1024's on the way, this time from the same place I got the working MN3002 from.
I know, I'll never learn, hehe...

Scruffie

#96
Quote from: RDL68 on November 01, 2018, 08:42:44 PM
Hi Scruffie,

I've only audio-probed the BBD's while they're socketed. The signal makes it to the input pins but not the output pins, so that is not a good sign.

Ah, I misunderstood, I thought you were saying the signal died at the input pin, not that it reached there and then stopped.

QuoteWith the CE1, it was chorusing as is, all that was changed between powering it up was a pot, some capacitors, diodes & a fresh LM1458 chip. So that BBD has gone from signal passing through the BBD to no longer reaching the output pins (signal only making it to the input pins).

That's quite a few parts changed, lots of things could have happened during that to cause the surrounding circuitry to fail but why would those things kill a BBD? Unless you're doing this in your best wool suit in a room full of balloons, a socketed BBD is usually pretty safe.

Quote
They're very sensitive to age, static, handling & so forth. When used in Flangers they can fail a lot quicker just by using the feedback/accent knob at higher levels (perhaps in delay units that are tweaked for lots of repeats too), although that isn't the cause on this occasion.

The first part is true, they are sensitive to static and handling but where did you get the idea that they're sensitive to feedback? Feedback simply increases the level to the input and in any design worth its salt will have no effect on anything that could harm the BBD.
Age related death, perhaps one chip by fluke (though I've never heard of it happening before) but without an external reason such as moisture or static, there's no reason 4 would die, I know they're considered sensitive but they're not that sensitive.

People always assume the BBD is the culprit as it's the most expensive and rare part but this usually ends up not being the case, it's just one of those confirmation bias/internet rumours that refuses to die. Have you actually tested the clock in the DOD? Replacing parts doesn't always fix things.

I don't know your skill level but regardless, I would consider putting up some debugging info such as voltages for others to take a peak at before you go writing off 5 BBD's. Also if you don't have a scope, buying even one of the cheap $30 Chinese models would be invaluable in this situation.
Works at Lectric-FX

RDL68

I could understand if replacing a few caps on the CE1 may have loosened a wire somewhere causing an issue, but the dreaded signal not passing through the BBD issue wasn't one I wanted to see. It's looking like it didn't survive the tinkering. I was mindful of not touching the pins etc, so it's disappointing it happened.

With the SAD's failing with Flanger feedback up high, that's just something I read recently googling around, so it's grain of salt stuff, if it's wrong it's wrong & appreciate the correction. It's not related to what's happened with these 2 circuits, just meant to emphasize they can fail much easier than other chips.

I'm certainly nowhere near an expert with these things, that's for sure. I haven't done this professionally, so it's all learn as I go over the years. More often than not, they work first go & if not I can eventually get them working.
After building my own oscilloscope a while ago, I wasn't happy with it because apart from being a single probe thing, I found the 2 mini calibration trimmers both failed after just a few turns. I never bothered trying to fix it, although it could be done to get it tuned in a little better, I just ended up buying a dual probe scope recently instead. It's not a small USB thing, it's a standalone Gratten. It's actually larger than I anticipated, but it is what it is. It came in handy setting up the Ibanez FL99 from Dino & Phil recently, getting the 2 waveforms on screen & adjusting the trimmers to match the waves. A couple of other times I've tried using it things didn't make sense to me, hence my inexperience with it.

If you think it's worth persevering with the Dod 640, I'll keep going, but no output signal from the BBD's gave me the impression they have problems, ie failed. I agree, it doesn't mean they have all failed, it just really inconveniently looks that way despite hoping at least one if not more are still OK. I haven't tested the clock, nor am I a modulation engineer, I've only used a voltmeter to check if voltages were fluctuating on some of the various IC's pins & it looked that way, but it's not a clock test by any means. I can only go by every replacement part testing OK before going on the PCB. So it should work, it just isn't getting past the BBD's & apart from your suggestion of checking the clock, I'm not sure how to progress for now, so I'll try to figure out how to test the clock if nothing else.

Thanks for the advice too. By the way, I've put together quite a few of the Lectricfx builds & all worked well first go. The Zirconia was the only one that sounded weird & unusable but Keefe emailed me back within an hour saying increase the voltage from 9v, which I did & it worked perfectly after that.


Scruffie

#98
Ah okay well usually the clock is actually the more likely chip to die than the BBD itself (as I mentioned, that's just one of those persistent internet rumours) and no clock means... no signal through the BBD!

You can do a simple test on the BBD clock pins (or on the clock its self if you know which pins are providing the pulses) they should have 50% of the supply voltage on them, it's not a perfect test but if the two pins have different voltages, you know something is definitely wrong with the clock.
Next, if the voltages are correct and your multimeter has a frequency counter (Hz setting) you can test what frequency the clock is providing by probing one of the two clock pins, again, not as good as a scope but simple to try. If there is a good sweeping frequency (the 640 is similar to the MXR-117 so something like 50kHz to 1MHz at full range and a slow sweep would be a ball park) then you can assume the clock is probably okay, although that's not the end of the story.

A BBD will also not produce an output if a) the input bias is out of range or b) there is no source follower resistor or something is wrong with it (for the SAD1024 that's a resistor going to ground from the output).
Works at Lectric-FX

RDL68

Besides the BBD, the only other chips on it are an LM324N & RC4136N, both op-amps, which test OK in my IC tester (the part number pops straight up on the screen). That just leaves the 4013, which is a flip flop chip (also tests OK), so that must be the clock of the thing.  I'll check out the 4013 pinout & get my head around what is what & get the multimeter out. It does have a frequency counter, had to use it to calibrate a few others along the way. I just wish the IC tester could check BBD's...

If everything checks out clock related, then resistors are all that's left that I haven't replaced on the 640. Generally if they aren't burnt out or contaminated with anything that's causing them to disintegrate, they usually stand the test of time. However, if it turns out it's a resistor gone bad all this time I will kick myself. It'll be a first for me, a bad resistor that is, not kicking myself, never said I was perfect...

Scruffie

Don't forget broken traces and bad solder joints, two things that can show up over time and can cause plenty of havoc and be very tricky to spot.

Yes the 4013 is the clock, the pulses will either be produced by pins 1, 2&5 or 13, 9&12, you just need to trace which are going to the BBD.
Works at Lectric-FX

RDL68

Scruffie, I didn't have any luck with frequency measurement. Maybe I was doing it wrong. No frequencies were fluctuating.
I've done it before with one or two other flangers, the reading goes increasingly up, then slowly down, just a matter of noting the top & lowest frequencies & adjusting the trimmers accordingly (or in this case seeing if there's any movement at all).
I looked over the pins of the 4013 on the trace side of the PCB. There were no breaks in traces so it was puzzling.

I remembered what happens to TV fly-leads when they've been in use for a while (the cable from TV to the wall). Even if they haven't been touched or moved for a few years they can suddenly stop working or give a bad reception. You can check them with a continuity tester, they'll check out OK, but still wont do their job. HDMI cables can suddenly go flaky too. And those LED torches, don't you just hate how they start flickering & you have to keep tapping them to keep the light on, you can't clean the battery contacts enough to keep them stable...
So I thought while I had the 640 guts out, I'll replace the whole socket the 4013 was in, as it was over a quarter of a century old.
I took out the BBD to keep it safe & then also the 4013. I used some trusty solder wick & the old 4013 socket popped right out with no damage to any traces. I soldered a new socket in place, put the chips back in & powered it up. The first BBD I put in got a result, so yes it's now sort of working almost but not quite as it should, but at least it's chorusing & flanging a bit now.

First thing I noticed was it was very difficult to find the sweet spot for the 2 trimmers, so at this stage I'm close to zeroing in on the right spots, but I'm not getting the seriously over the top flanging I know this thing is capable of, it's more mild & chorusy. From memory, it was never that difficult adjusting the trimmers, they had a much wider working range. In fact one of the trimmers I could adjust to use the 640 as more of a flanger or adjust it the other way for one of the best chorus sounds ever. Now it's like I'm trying to find the 2 sweet spots of both trimmers just to get it working at all.

Also it's making some random popping noises & after messing about a bit since freshly wiring it up, moving or tapping the PCB seems to set it off somewhat as well, also has the effect of the effect cutting in & out or going clean straight through, so I'm thinking loose wire(s) or flaky something somewhere.
It's great to make some major progress though, at least now it's working again, but for the moment it still sounds like any old pedal that needs a service & tune up, so I'll have to put a bit more time into it to get it stable, clean its carbies etc.

Thanks again for the encouragement to stick with it. I thought I had it narrowed down to being the 4013 as the problem & when I replaced it with the exact brand & model AE chip & it still didn't work, I was really disappointed because I had already re-flowed the pins of that 4013 socket & checked the traces etc. Yet, by replacing the socket itself today, the thing decides to start working. Unbelievable.
Now I'm wondering if I should replace the other 3 sockets as well, they might all have corrosion or whatever else impeding contact from the other IC pins. Some of the pins could be working as resistors now. I shudder to think...



RDL68

With the CE1, I haven't given up on it, I'll just have to go over the area around the LM1458 where I replaced a couple of diodes & caps as there's a chance I missed something, dry joint or whatever, so there's hope for the MN3002 yet.
Due to the rarity of those chips & having no spares now, I still intend doing the CE1 TDA version (fortunately drog trog still had a TDA PCB to sell me) as I already have a few of those chips on hand, along with everything else needed. It's just to future proof my own CE1 so I can keep it going if the main chip does/or has failed on the MN3002 version. It'd also be quicker & more cost effective to populate another PCB than try searching for spare MN3002's.

With the Dod 640, I obviously missed the elephant in the room = the old IC sockets, corrosion on the top side etc, so getting rid of the remaining old ones is bound to clean up the signal a heck of a lot. Re-flowing their pins just wasn't good enough, lesson learnt. Another point is even the cheap IC sockets around these days do a better job maintaining their grip on IC pins than what was available circa 1980. The "turned pin sockets" even better. The resistors should still be OK & I reckon the worst thing that could've happened to a few over time is they may have started getting a little bit noisier. I'll see how it goes after getting the remaining old IC sockets out of there first, as I already re-flowed all the resistors recently. If new IC sockets get it stable & quiet, I'm not keen on replacing all resistors just for the sake of it.

The 640 was a fairly good flanger design (apart from the AC transformer causing some hum & risking power surges), but I really liked the chorus it could do as well. Most chorus circuits tend to use a single MN3007, being a single 1024 stage BBD, but there was something about the Dod using the 2 x 512 stages in parallel & out of phase that made it special. I'm sure there'd a be a few chorus designs around that do something similar, I just haven't searched for any, just going by what I have & only the 4 button Boss dimension springs to mind as a possibility of doing things that way (having built the blueshift last year with its 2 BBD's), albeit with buttons & not pots.
Same with Flangers, most don't use 2 x BBD's in parallel & out of phase, but a lot in the past used the 2 x 512 stage halves of the sad1024's in series which is being replicated exactly by a single MN3007 in remakes by the more experienced guys offering their PCB's these days. The BYOC Flanger is the only DIY one I'm aware of that did it with 2 x BBD's in parallel & out of phase. The Ibanez FL99 used 2 BBD's as well, but I haven't looked into if it's in series, parallel etc, sure sounds great that one though.

RDL68

I put some more attention into the Dod 640 over the last couple of days.
I replaced the remaining old IC sockets, then sprayed on some PCB cleaner, as I'd left so much flux residue everywhere. There was enough still there from the early days to warrant a good clean. Most wasn't budging, so I left it overnight to soak a bit & then gave it a 2nd spray today. A sacrificial toothbrush & pin (for scraping between closer pads) later, it was clean enough to eliminate the possibility of any old residue having turned partially conductive. I let it dry a few hours, then I put all the IC's back in & it's now running as good as new. There's no noise artifacts or pops, just a great unique vintage flange. It all seems stable for now.

The range on the clock trimmer is back as well. I just set it at the end for best flange, then backed it off a little to stop it going mental, so I assume turning it towards the chorus end would be increasing the delay time, hence being better for chorus.
I'm not sure what the other trimmer was called (maybe "color") but that one only had a very small range to tune in the effect, so as usual, just a matter of hitting on the spot in the middle of that narrow zone with the least distortion. It took a few goes that one, but once in the right spot it sounds great.

Replacing the 3x D100k pots with B100k's works perfectly. I'm just not happy with the curve of the A500k as a replacement for the D500k (speed knob), so I think I'll have to swap that one for a B500k. It's feeling too bunched up at the fastest speeds.
Anyway, it's all done & working, no noise or any issues whatsoever, the 70's carbon resistors quietly live on, just have to swap a speed pot later to get the curve feeling better.

Most of my pedals from those early 80's days were Boss & no Boss I've owned has ever needed a repair. They're all still working as good as new, even the oldies. I had an Ibanez analog delay fail on me during those days (the pink one with the small square chrome switch), but not the Boss DSD (digital sampler delay).
Quick story, I once used the DSD for a prank on some friends that I knew were on the way over. The house had a momentary button outside for the doorbell inside, so I used some jumper leads inside from the actual bell end to a regular guitar lead & used that as the sample trigger mechanism plugged into the socket on the DSD. Then I set up a seriously huge stack of speakers & amps behind the front door....and they were expecting "ding dong" haha. Good old Boss pedals, what can't they do....



RDL68

After 6 weeks off all things electronics, I got back to finishing the CE1 this week.
The MN3002 version was OK in the end. First thing I saw when getting back to it was it just had 2 wires that had worked loose from the mode 3pdt after all the dithering about replacing the caps/diodes mentioned previously. Instantly fixed.
But, as I wanted to future proof this & as drog trog had also sent me a TDA1022 version PCB, not to mention having several of those TDA chips on hand, I soldered up that TDA version.

It worked first go, but it had the same annoying hum present as my MN3002 version. I first thought it may have been because I wired the 2 x 3pdt's instead of soldering them directly to the PCB's. It wasn't caused by that though, I tested it with a 9v battery & it worked perfectly without the hum, so the power adapter was the cause of the hum. Oddly enough it works fine with a 9v battery in vibrato mode but seems to lack the current from a battery to kick in the chorus mode, but that's OK, I'll get a better power adapter later. I read somewhere that these things don't like switch-mode power supplies, so that's probably all it is.

I still have to wire up the mode LED's & think about whatever other mods later, but as a bare bones thing it is working exactly as it should. The enclosure already has spare holes for a toggle & a couple more pots, so I'll get all that sorted later.

The only change I made from the first board I did was the 3 power caps next to the recon power module. Instead of neat & small 35v caps, I used larger 50v Nichicon low ESR's (with better heat tolerance, longer life), which meant I had to lay them down on wacky angles, but they fit OK in the enclosure.

I'm not sure how to set the CE1 trimmer, but I've left it somewhere in the middle of its range & it seems to be fine as is. I think it would need an oscilloscope & the Boss service notes to sort that out properly though...