(...Yeah, I know it's a piece of crap but I've got a library of tapes that need digitizing)
If anybody here knows it (Fostex Model 80 1/4" 8 Track), it had a notorious issue with segments of the level meters dying. I think mine went in twice for it. Twenty years later I've got another one on the way, which supposedly has the same issue with a couple channels. Does anybody know what it was that used to blow in those? I've got a Mouser order I'm building and it would be great if it was some common LED driver chip I could grab a bunch of, but I'm fearing it'll end up some proprietary Fostex thing that I can't get.
Think I turned up my answer after many hours of fishing for M80 info. A user in the homerecording forums mentioned "ordering the LEDs and soldering them in" and that he was lucky that only four had gone out across three tracks, so I guess that means the lights are what pops and that they aren't full bar modules (whew!).
Still wonder why it would have such a problem blowing out LEDs though, or if there's a limiting issue, why won't the next ones burn out too? (maybe they just got hold of a big batch of really really crappy ones heh )
A user in the homerecording forums mentioned "ordering the LEDs and soldering them in" - - so I guess that means the lights are what pops and that they aren't full bar modules (whew!).
Well, looks like I'm not quite so lucky. They are individual segments, but not only are they surface mount, they make regular surface mount look like full sized thru-hole. They also aren't the type with the tinned contacts on either end. They're tiny squares with a small wire hanging off the top, and the other contact is apparently on the bottom. This, I believe, is the problem. They don't appear to burn out, but rather just pop loose from the board giving them an open connection. They're so small that I can only tell they aren't connected by looking through a 60x microscope. The dead ones are sort of floating off the board with the wire holding them to it.
I tested them by tapping a 5V supply through a resistor across the two traces feeding them (hope I didn't feed anything dangerous back to the driver chip). The dead ones can actually be revived by pressing them down with a needle or something.
The question: Have any of you guys dealt with those type of lights, and how would you proceed (all info implies that Fostex no longer supplies the boards)? I'm considering holding them down with a needle and smearing a dab of quick drying glue or something at the edge to keep them there, but I'm afraid of using something I can't remove and being stuck with it if it doesn't work or fails later on.
PS- Info also seems to point to the 16 track Fostex and the R8 as having similar problems, so maybe someone else could benefit from a working fix. If I can arrive at a glue-related fix, I may do all the good ones as preventive maintenance.
Last edited by Jidis; 25th January 2009 at 11:20 PM..
Reason: forgot my picture
It looks a bit like the inside of the Telefunken bargraph modules i.e. surface mount and very small!
You should be able to get the LEDs and then all you need to do is solder them in....
OR, how about a small blob of silver loaded paint used for repairing PCB traces?
The silver paint type of solution was what I was hoping for. Or some type of conductive glue.
Unfortunately (x2 actually): 1. I'm thinking now that the pressing down of the light may not work in all cases. I think some of them had a small square shaped chunk lift up from the top of the LED, which also has the connection wire attached. and 2. I'm not getting much of anything from the one meter I tore open now. Either I zapped the main chip with reverse voltage while checking the lights, or all the segments are now non-functioning (it isn't hard to break that connection). The module has also been out for a while, so it's probably had other opportunities to get hurt as well.
I wish they had used a common module like the Telefunken you mentioned. With the LED flaw hitting so many people, they should have at least maintained a stock of replacements. The driver chip is also buried under one of those black epoxy blobs, so you can't swap that if it goes out either, and there's nothing else on those boards but a resistor and a cap. I still haven't heard back from Fostex, but I'll keep trying.
On the plus side, the measurement from pad to pad on most of those segment spaces seems to be able to accommodate a standard surface mount LED like you see on hard drives,etc. I may try to get a voltage across a working one on the Fostex and order a batch of those. They don't have LEDs with current limiting or anything built in do they? I don't want to risk overloading it with a different type.
PS- It was a disappointment to find that the post about replacing them may not have been accurate. I know my rework skills are minimal, but holding a needle next to this LED chip's wire while under 60x magnification looked about like holding a tennis ball can up to a pencil. He'd have to have worked on it under a microscope and even then I can't see any iron tip being that size, let alone Fostex selling you that sort of chip LED.
Get some LEDS in stock and get someone younger with good eyesight to have a go.
Silver epoxy paint I got was for repairing car window heaters and the like. From Farnell in the UK (Newark in USA) I think.
I probably will unless Fostex surprises me somehow. This really sucks. I now have two segments on another meter out, which I believe were fine before. Not only that, but I checked the traces to those lights with a full scale signal coming in and got zilch, so I guess that mystery IC may be more involved than suspected. The output to the working ones is barely over 2 volts, so I guess regular lights would work (if the driver was alive).
I was reading about that black epoxy glob. I thought those were just to hide chip numbers, but found that there are low cost surface mount chips out there with no actual body. The glob handles that part. So maybe legs are popping loose from the board under the glob or something.
I also checked the pins for the whole module. They're on regular four pin headers labeled Vin,Vref,Vcc, and 0V. I can also easily check the voltages on them. That one cap/one resistor deal reminds me of what the minimum circuit was in the app notes for those VU meter ICs (LM3915?). I'd love to just etch new boards, but the meters have 12 segments (2 too many).
PS- Yeah my eyesight does suck, but I don't know anybody younger that I could trust with this sort of thing. I usually work with strong reading glasses or one of those desk-mount magnifier lamps.
You can get chips from Sanyo (I think) LB---- to do lightmeters, Fostex may have beren usin 'naked' versions of these which are similar to LM3915 series. Telefunken used 2 X interleaved 5 segment chips (267 / 2XX) to get a 10 segment meter. The choice is yours!
Was tied up on non-audio crap most of the day and hadn't been online.
Yeah, I ran into those Sanyos you mention (LB1412M) while doing searches for meter driver chips yesterday and decided they were probably it unless Fostex had made something themselves. Unfortunately, they look to be old now. The only ones I could find were DIP versions. It's surprising nobody else stepped up with one for >10 segments. You can chain together the LM39xx things too, but these modules are already probably too tight on space (they're only about 5/8" wide by 4"). Switching to ten segments and doing a surface mount LM chip might be an option. I'll have to decide on what it's worth to me though.
PS- I discovered at least their R8 machine had switched to 7-segment VUs after the Model 80. Maybe the Sanyos were already too hard to get at that point.
Hi www.profusion.co.uk had them in the past. I think I have a couple somewhere as I was thinking about using some but didn't.
You may be able to do something with remotely wired LEDs from a more standard board if it is only 2 channels (multitrack would be ridiculous!).
Man, thanks! Looks like you were right and the price is reasonable too unless shipping to the US is high. I've got a pile of decks I'm semi-restoring right now, so I may do a couple others before I decide what to do on that Fostex. If I can etch a working board for it, I'll probably do all eight, so it may be a bit of an undertaking.
BTW- Just got a package from Teac, including the service manual for a rackmount four track here. Hadn't even considered what was in that, but it turns out they're LED 12-segs as well, and use something which may be more recent. It's an IR2E41. Couldn't find the exact datasheet or anything else, but there was a pdf here which refers to it (looks to be for the LED modules). That and Tascam's diagrams also indicate a thru-hole chip, so there may have been no small ones anyway.
Reviving an old thread here. I recently fixed the VU metes in a FOstex 450 mixer, which uses the same boards as the 80. I just removed all the old LEDs and replaced them with 0603 surface mount ones. It's working great now! I documented the process here if anyone is still interested:
I ended up getting another 80, as well as the R8 you have, but may try one or two of the problem boards. Way back when, I had the 450 too, but don't think I ever had any lights fail. The 80 on the other hand, went in at least twice for it.
I may be wrong, but I think they came up with something better for the R8, as the dead LED thing was pretty well known by that point.