NGY

Over the years I have collected some astonishing photos. I thought I would open a thread for them. No, not meant as "tips and tricks" that the title of this section would suggest. Quite the contrary - rather mementos of the "don't do it". If you have similar photos and/or stories, please share them here. I hope these can then help saving some MD devices. The apropos of this topic was this very photo I saw today on an on-line fleamarket page. The guy advertises himself as one who repairs minidisc gear and sells parts. He draws our attention to that the drive on the photo has a "special feature" - a hole drilled into the side of the bracket, to (quoted:) "make it easier adjusting the laser" ... The "don't do it" part (and it is not about drilling the hole, though it could also be): the variable resistor on the side of the laser head is NOT for setting/adjusting the laser. (It is there only for the OP manufacturer - during the production process they set the operating point of the photoelectric circuit that will monitor how the laser diode is doing when in use. Any changes to this setting can result in an abnormal operation of the laser, that can drastically shorten its lifecycle.)

Yeah ... the MDM-3 flavours can be a real PITA sometimes ... You did quite a few checks/tests, and based on the results, besides the 'deck suffers from "TOC writing/blank disc" problem' I did not feel anything particular to be worried about (but the dark display obviously). The "TOC writing/blank disc" problem is mostly related to either the OP or the OWH. If your deck plays a loaded disc without skips, that tells the machine is "nearly" fully functional. Whether the OP is dying or the OWH is busted is fairly easy to separate, because you can 1) do a visual check on the OWH, then 2) measure the continuity of the magnetic coil + the yellow ribbon cable of the OWH, when disconnected from the BD board. If all good, the OP is to deal with. But that's a whole lot of a different story then :-) .

OK, let's summarize, how far we got: - on the main board copper traces of FLCLK/FLDT/FLCS are intact between the MCU and CN303 (so are the solder joints of the respective jumper wires) - the FLCLK line is not shortcircuit to ground on the main board (for the maximalists: how about the two other lines?) Based on this and previous findings, we exluded all possible causes for the (otherwise good) Clock signal getting from the MCU to the display IC. My logic tells me it is then the MCU that is responsible for the dark display. The battery gets its regulated charging voltage (3.3V) from IC309 (pin 6). This voltage gets thru a protection diode (~0.6V forward drop) and a current limiting resistor (R316, also drops a tiny bit of voltage when charging), therefore, the 2.68 V you measured seems fully normal to me. We pretty much depleted all possible ideas here. I could only have one very last (and again "weird") shot on the battery. We know it for sure, that a deck is fully functional (but keeping user settings and unwritten TOC data when powered down) without the battery, or even with a dead battery. What if you desolder the new battery? Would that change anything "'back" then?

Yes, it is one last check I would do - as I wrote above (added as an EDIT): ”what voltage can you measure on pin 23 of the MCU”. This is to clarify everything is OK around the new battery, and the MCU sees it properly too.

I admit we diverted your topic a bit - nevertheless, are you still interested in a possible solution?

True, "cold solder joints" (or "dry joints") are possible - I also posted such errors (like the one here). However, and it also relates to the possible SM error ... ... I prefer using my meters to do such checks on the board itself, for a couple of reasons. While Sony SM-s are generally very good quality, some of them are scans only (like the one for 510), not digitized blueprints, means they cannot be zoomed properly and my eyes are too old for following these tiny lines. I have rarely seen errors in Sony MD SMs, the very seldom ones I ever saw were rather in the schematics, not on the PCB prints. And if I have to look for a connection (or the lack of it) beetween point A and B, measuring is the best way, meters won't overlook. That's why suggested above to perform those two checks, to see if the serial comms lines' traces from the MCU pins do reach the ribbon connector's contacts properly. But just for you ;-) (no, there is no error in the SM, pin 11 of CN303 goes directly to pin 71 of IC316, via the sole jumper wire of JW143): also And, to exclude one "weird" but possible thing, I wanted to see, that after ... ... the MCU sees the new battery voltage properly. Because if there was an issue there (i.e., reversed polarity, or shorted battery), the MCU would take it as a Reset (see the reset jumper wires, for a forced reset they simply short the battery to ground via R16, being 68 or 100 ohms only), and when the MCU resets what it need to, it probably disables the display too.

The OWH is broken, and the tip with the magnetic coil is missing. This spare part is practically non-existing these days - even if you find one, might cost 100+ USD or more... You can try to look for donor machines from your list (might also be rare and expensive), but beware, this very part is often busted in these decks :-( . If you allow one comment: the S30 is such an old machine, and is not particularly excellent when it comes to recording (see how ATRAC evolved over the years). For the money you would spend on a spare part or a donor machine ("uncertainty included"), you could buy a way better deck - say an S50, but even an S40 or an S38 is already better in many terms, than the S30. Just a thought.

I am breaking my head over this ... Here is what we know: - the deck is operable, but the display - the VFD tube is intact, so are the related supply voltages - the front panel ribbon and its sockets are spotless, at both ends - the MCU is not giving the Clock signal on its FLCLK pin, that is directly wired to CN303 on the main board. Let's take this last one further, if we can do some more checks there. (For the first three, the only thing I could/would do is to perform all previous checks again, to exclude anything I possibly overlooked.) The "trivial" stuff: - is there any damage to the main board's copper traces? With the front panel ribbon disconnected (deck powered down), can you measure continuity between pins 71/72/73 of the MCU chip (IC316) and pins 11/9/7 of CN303 respectively? (Please be very careful touching only the proper contacts.) - is there anything on the main board that can cause the Clock line shortcircuit to ground? With the front panel ribbon disconnected (deck powered down), can you measure some low value resistance between 1) pin 71 of the MCU and ground, also 2) pin 11 of CN303 and ground? (Please see my note above on the "how to" - as you will measure directly on the MCU.) If all OK, to exclude the MCU being damaged with 100% confidence, we would need another 510 available for a quick swap of the front panels. Possible outcomes then: - "guest" front panel + your 510: A) display OK = your 510's MCU is allright, dark display = your MCU is damaged - your 510's front panel + "guest" 510: C) display OK = your front panel is allright, D) dark display = your front panel is damaged Cases A) and/or D) would then need further investigation for where we did possibly go wrong, but I doubt if either one was the case. Nevertheless, I am certainly always prepared to see the unexpected, and learn something new :-) . EDIT: just a weird idea ... what voltage can you measure on pin 23 of the MCU? First with the deck is ON, then with the deck powered down and waited ~30 seconds (looking for +3.0V ... +3.3V to ground). (Again, please be careful when touching the IC's pin, not to accidentally shortcircuit to the neigbouring one.)

For the records: the Clock signal is supposed to be there even with a static display. Just did a quick test on a JE500:

No, you're very probably right. All indications are that the MCU is damaged here.

Great to see you are doing a thorough work here. I think your VFD is allright then. With that, we can narrow it down to the communication problem between the MCU and the VFD IC. It seems the Clock line does not work. It could be at the MCU side, either a silicon damage on that pin, or a firmware blockage (possible reason for the latter, if at all, is not clear to me). It also could be on the display driver IC side, i.e., if there is an internal shunt (burnt circuit) on the chip, that pulls the clock line to ground. Without the tools mentioned above, it is a bit tricky to figure out which one, but we can get an idea with some creativity. First you need to somehow disconnect the display driver from the MCU, while keeping the rest of the front panel still connected. One way is to disconnect pins 61-62-63 of IC701 by desoldering and lifting 1-1 legs of R722-723-724. Once done, measure again the clock line from the MCU on CN303 as you did before. If still nothing, your MCU is not clocking the serial comms line, means it cannot send data to the display chip. If you see the ~3V, then IC701 is the one to check further. Now, you need to measure if pin 62 is not shortcircuit to ground. Normally, I am using an ESR meter to do such a check, not a DMM (more on the "why" here). As a workaround, you can try using a 1k resistor in series with your DMM probes, and measure the total resistance (with the AC cord pulled off, obviously). If you get ~1k, then the internal circuit behind that pin is probably burnt (then you can go and re-check it without the 1k resistor, you probably will not make it worse). If you see a min. one magnitude higher value, then it should be OK. Whatever you find, it means the machine will need some replacement parts and a bit of careful SMD soldering job, to get the display back. Or, a donor device, and just swap the needed board. Regarding the "copper sheet" - apologies if I was misleading you. I wrote off the top of my head (as I usually do), and remembered wrongly that particular way of grounding was on the 510's too. It is certainly there on its midi equivalent S38 (see on the photo below), but eventually it looks like there is nothing such on the 510's anymore. The copper strip is soldered to the ground foil of the front panel PCB, then it goes between the chassis and the front bezel. I saw some decks in the past when this little piece was floating above the chassis, not touching it, that caused other problems. EDIT: one more thought, as I am not familar with how the MCU and the display driver IC communicate. It is probable, that when only a static message is shown on the VFD (like "NO DISC"), there is nothing happening on the comms line. Therefore, while measuring the Clock line again, "do something" that would normally change what is displayed - say load a disc, or press the Repeat (or Play Mode) button a few times, etc..

Great job. You sorted out most of the possible errors but the dark display. And a - possibly - laser wear or misadjustment thing waits for being fixed later: Anyhow, let's continue working on the display. First, let's see if the VFD filaments get their voltage, and if they do, they are not broken. You can measure the filament voltage on the VFD - there must be some AC 3V between the leftmost and rightmost pins (actualy, 3-3 pins). If that voltage is there when the deck is on, unplug the AC cord as well as the front panel ribbon, and measure the filaments in continuity mode (between same pins). Also you can do a visual check - look for those tiny wires across the tube: You already measured that the system 3.3V gets to the display driver IC, the negative voltage for the cathodes is there too - we are happy with these. Next thing would be to see if the serial comms lines from the MCU work, but that would call for a data analyzer (or min. a scope). Without any, you can measure at least the Clock line on pin 62, looking for some 3V. Also, a quick check on the Reset line, if the MCU is not disabling the display IC for any reason - expect DC 3.3V on pin 60, if that is OK. (Not much probable, but if the internal oscillator of the driver IC happens to be not working, can also lead to a dark display - a high frequency square wave must be on pin 59). If everything above are in order, maybe a last check on the front panel ribbon, at both ends, as well as on the ribbon connectors, looking for bent/broken contacts, or any foreign particles stuck inside the socket, etc.. And a weird one: check if all earthing straps, as well as that copper sheet at the bottom side of the front panel are properly connected to the chassis. If all OK, then I have to stop here, this is all I can do remotely. That 320 would be a great help now, having a (nearly exact) same front panel as the 500/510 (some buttons/resistors missing/different), and a quick swap would tell us if the problem lies on the main board or on the front panel.

Instead of the 530, an S50 might be your choice then, if you have limited space. It is a great machine, midi equivalent of the 640: Type-R, MDLP, SF-Edit, pitch-control, etc., plus it is PC-Link compatible.

Interesting - the display of my (EU version) JA20ES is also somewhat dimmer compared to my 640 next to it, although the 20ES was hardly used (~220 hrs only) when I bought it. Could that be, that in those top-end models Sony intentionally decreased the anode-voltages, to allow longer lifetime for their VFD-s?

The latter. Regular component decks are not designed to work standing on their side. Their mechanisms rely partly on "normal gravity" - springs, magnetic hubs, disc doors, sleds, etc., to name a few parts. Edit: apologies Kevin, we wrote at the same minute. I did not mean "smashing" on your post, I replied blindly to the op.

I like your approach, Kevin, and yes, on the low voltage DC side the regulator circuits should be OK with even a 15...20% higher input voltage. However, the main risk here is at the transformer side. Not sure about Sony, but in the commercial goods world it is common to squeeze out the most possible performance using the least possible material. When it comes to an AC power transformer, we speak about a lot of iron and copper to be saved. Then, when you consider how a transformer works, a 10...20% increase in the AC voltage can result a much higher increase in the energy "pumped" into the transformer, and that "extra" converts to different kind of losses (iron core loss, copper wire loss, etc.), and ultimately, heat. Normally, such a transformer is designed to work near the limit of possible magnetic saturation of the core, to keep the size (= cost of the material) small. If you go over this limit, that's when problems happen. Same for the copper windings: the diameter of the wire is designed to bear the maximum current that particular device is expected to draw. But if you increase the voltage, with the same copper resistance, the current will increase, so will the heat generated inside the copper. On top of all, this is a non-linear function :-( . I.e., 10% more current (or voltage) means 20+ % more energy, and 20% more current means 40+ % more energy. We did not speak about electric and magnetic (also, audible) noises of an over-saturated transformer, or voltage swings on the AC net, that can take the input AC voltage way higher, than the nominal value. Example: it is not uncommon on the 230V grids, that the voltage occassionally goes up to 238...242V. I believe it is similar on the 100/110/120V grids too. Imagine, if a device meant to be used on a 100V line gets a - say - 125...126V AC input voltage on a 120V net.

While I fully agree with Stephen and would not butcher such a nice deck, I might have a non-destructive solution for you. Without having the SM, first you would need to carefully lift the PSU board from the chassis, and take some good photos of both the component side and the soldering side. It would be also helpful to have the component side from a few different angles besides from above, as well as the solder side with a ruler held alongside next to the panel. I need to check a few things then I can tell if my idea works or not. Again, what I have in mind is non-destructive and fully reversible, so no worries. (To remove the PSU board you obviously need to unplug the AC cord first, wait about a minute or two, then disconnect both the AC cord's and the main board's PSU cable connectors, and finally unscrew all screws that hold the board/main transformer.)

That's no good news then. Just for a quick countercheck, is this the same when you try to enter regular Service Mode? If you can enter Service Mode we can do some further checks. If you cannot, we must do some guessing and trial-and-error. [In that case - as always when we try to fix something remotely, some questions or hints may make no sense to you (that is, not a fault at your end :-) ), but please try to follow the path. I am not saying my method is "the" way, but what I write here is exactly how I would try to work if the faulty device was in front of me. As I don't see the machine itself, I can only work from what feedback I get from you, therefore it is important we are on the same page. And please forgive me if this small "preword" is completely unnecessary for you - I have already run into cases here when the person at the other end of the line was moving on a totally different path (of his thougths), that resulted a bit of chaos in analyzing the results and possible causes, that led to an unsuccessful repair process. Your questions, ideas and other important feedback are always welcome though.] Back to your 510: I am trying to separate these possible faults: - MCU - the main processor has some of its I/O ports directly wired to those ribbon connectors. As you wrote above, misalignment of a cable under power might cause some trouble. Having said that, these I/O ports are protected to a certain extent (with pull up resistors, and/or internal latches, etc.), so there is a good chance the MCU is still undamaged. - display - as you noticed, the VFD is driven by a special IC. This IC receives data from the MCU on a serial line, but even if that line is damaged, the rest of the deck might still be operable, independently from a dark display. - drive mechanism - partly controlled by the BD board logic, partly by the MCU, let's see what works and what does not. - OP - one of the most sensitive parts is the optical pick-up, highly sensitive to ESD damages. A broken OP - among others - can for example cause the loaded disc not spinning up. - rest of the MB - there are power regulators, analogue audio and digital signal switching circuitries there - I would not expect any of them being damaged, but may need some checks too. So our next step: you wrote you could eject a loaded disc. If the Eject button works, so should the FF/FR, Stop and Record buttons do, because they are on the same MCU control line. Now, if you can enter service mode, even with a dark display, you should be able to load a disc, and try to move the sled back and forth, as well as to lower the OWH onto the disc as if it was in rec mode, even if the disc is not spun up. Then pressing the Stop button would then lift the OWH. Please let us know if these work. Next: in normal mode (Power led green) load a disc, set the Timer switch to Rec, set the source selector switch to Analogue, then unplug the deck (or use the switch on the back panel). After 10...20 seconds, plug the AC cord back in, and observe what the drive does: i.e., it moves the head into rec position, or just ejects the disc after a while, etc. Let us know what you find. So much for now. If there are no results I am expecting here, we can only possibly move forward if you can have access to another, similar deck - more on that later.

I have experienced a similar situation a while ago. It was a regular USB mouse, and the PC was not a VAIO, but the exact same thing happened. I tried several times to disable/uninstall/reinstall the mouse driver - but it did not help. Finally I concluded, that one of the applications (or, the OS itself) might have had some interference in the system settings. As I had no time to debug each and every software on the machine, I simply made a system restore from an earlier back up I knew had been working all good, then the error disappeared. I am not saying here this is "the" solution, but in my case it worked well.

Hi Stan, welcome to the forums. These are all good signs of the device not yet being fully dead. That the power button works shows your MCU is alive - it is sensing some of the buttons and can control the standby/power on led, therefore, it seems it is not blocked by a fault somewhere, that would cause the MCU forcing the device stay in Standby. That the disc load/eject works shows that your BD board is alive (might have some faults though - to be figured out later on). Let's begin with the trivial stuff: - I would first check the two ribbons that connect the drive to the main board. On the 500/510 decks it is fairly easy to seat one of the ribbons the wrong way, at the main board end. The wider one is pretty straigthforward (i.e., short enough, and folded sharply), but the other one can be plugged 180 degrees twisted around (I faced this issue a couple of times, and luckily those decks always survived this mistake). Both ribbons' contacts should face the rear of the deck, not the front. In other words, the ribbon side with the writing on it is facing the rear, the plain white side is facing the front. - second I would check the front panel's ribbon, for contacts peeling off like this, or being damaged otherwise: - then there is a display/controls check mode you can try, to see whether or not your display driver is alive, and/or the MCU can sense all other buttons/switches, as well as the IR sensor. Turn the deck off with the switch on the back panel, press and keep the AMS knob and the record button pushed, then switch the deck back on. The display should lit all segments, then by pressing each buttons one by one, you can turn off the segments one by one. Once all done, other tests can be performed, following the instructions appearing on the display. Let us know the outcomes of these checks, and we can take it from there.

Is this what you are looking for? (These head-unit connectors are pretty much standardised these days, therefore it is very probably compatible with the AX5000 - and other models/makes - too. Wire colors might differ though.)

Looks like this feature is available on selected QS (920, 930, 940) and on the ES decks only. Yet another reason for using a QS/ES deck for recording :-) .

The day has come - your MDM-7 jig boards are thrown together and ready to ship out.

The 920 is an SP-only machine, while the 480 is MDLP capable. Edit - a bit more information: the 920 is ATRAC 4.5 and the 480 is ATRAC Type-S. There is also Type-R in between, and anything below Type-R cannot play MDLP discs (with LP2 and/or LP4 recordings). Even some of the Type-R machines are still SP-only (i.e., the x30 family - 530, 630, 730, 930), while others are MDLP-capable (i.e., the x40 and x70 families - 440, 640, 940 and 470, 770).

Sure. Those items I purchased several years ago are not visible in my account history anymore, but there are many others as I see, and now even shipping directly within the UK, so you might not need to wait weeks, for a little extra money. As listings/sellers come and go on eb@y, for future readers here is my search term: "battery heat shrink tube", that gives tons of items. One > here < seems to be a good seller, for example, having a lot of products and providing good product information. A word on which size to choose: this tube can shrink to about half of its original, flattened size. You measure your new shell and calculate the circumference, then choose the closest tube width. For example, I used 23 mm wide for AA cells and 30 mm wide for 18650 cells - 18650 diameter ~18 mm, circumference ~57 mm, tube width 30 mm (2*30=60).