Tascam M-___ Story...

...somebody, sometime sort of promised some PSU schematics?

Check your PMs.

I would also like to mention (again?) Douglas Self's book, Small Signal Amplifier Design. He goes into good detail about mixer PSUs (up to some 5A delivery) especially wrt noise and protection.

I do need to get that book sometime.

One way to get RIPPLE down is a 2 stage filter after the rectifier so..Cap-R-cap where R is about 2.2 Ohms or whatever does not drop the regulator differential too low.

This concept is already incorporated in the PS-3500 +/-15V supply...22,000uF main filter, 100R series resistor, 100uF secondary filter and then additionally a small value bypass cap. There are also 470uF output filters with small value bypass caps.

...Also some noise reducing tricks around the regulators.

The +/-15V supply doesn’t utilize typical 7815/7915 regulators. Not sure if the noise reducing tricks are based around those kind of parts, but Teac did incorporate noise and ripple reducing tips found in the application examples on the datasheet for the Mitsubishi M5230L dual regulator. Datasheet here:

https://aliot.com.ua/pdf/m5230.pdf

Refer to cap C037 on the PS-3500 schematic.

Protection: 80,000 mfd of downstream capacitance could certainly put the regulation series devices in danger! Solution is a hefty diode rev' biased across the device. 'Mutual shutdown: Many ICs, the NE series for sure do not like having just ONE supply rail and will die. Self gives a simple diode circuit for protection but it probably could not be incorporated in those PSUs (can't tell. No map!) One bolt on solution is a relay that will only pull in for all the rail voltage. (can be bypassed with bleed Rs to cut the inrush current.)

These are *exactly* the kinds of solutions I’m looking at...or a 555 based delay circuit...anyway...*some* sort of soft-start circuit.

Do not (grannies and eggs here!) be tempted to up the res' caps. That would bugger the 'power factor' and give the mains traff a harder time. (The R then cap avoids this problem)

No temptation here...based on my calculations after conducting the load testing on the main transformer, 22,000uF main filters are probably overkill as it is, and justify the use of the small value bypass ahead of the regulator.
 
So the donated PS-3500 power supply from Miroslav arrived yesterday. Huge thanks to Miro again for doing that...I feel very fortunate. You took a lot of time to dig it out, make copies of documents, carefully pack and send it...and it wasn’t free to send, and something like that has value and you gave that value up to contribute to the M-__ project. Thanks again.

You're welcome to it...put it to work.
Happy to help out.



So it’ll work!


 
SUCCESS!! At least with one of the current issues. :)

Small victories feel huge at this point.

The Control Room buss outs are now working correctly...clean clear audio and the level boost, mono, dim and mute functions are all working correctly...and no pops or clicks when actuating any of the function switches...that’s a first!

I discovered in incorrectly installed 4 replacement FETs ahead of the Control Room output booster amp. I mis-identified the original parts, and the PCBs aren’t screened with the pin function, and I didn’t study the circuit to confirm which pin was the gate. So I thought my replacement parts were pin-to-pin equivalents but they werent. I replaced the 4 incorrectly installed J112 FETs with 4 new ones with pins 2 & 3 reversed (to reconfigure the pins to DGS):

5C9850D0-A182-4857-9EDB-1EDFF67C6DD3.jpeg


I also replaced (for the second time) the 4069 logic chip for good measure since the one that was previously installed is the one that was there when one side of my dual bench supply went gunny-sack while the module was powered. I also replaced the output driver opamp fed by the FETs. I’m pretty sure what was there was fine, but the signal path of that opamp has been getting blasted with a lot of DC.

As with any frustrating/excruciating experience in electronics I've learned a lot, and I have a much better understanding of how the resistor networks, logic chips and FETs work throughout the mixer. And now when I engage different switching functions I understand what that logic chip is doing and how it works. Yes!

I suspect there was a problem with the logic chip and/or one or more of those FETs when the console came into my possession. I replaced the chip at one point and I think symptoms were better, but not gone. Then I think I roasted the logic chip when the dual bench supply failed, and then I put in those J112 parts in the wrong orientation...and so began the tail-chasing...and all made harder when I have no schematics and didn’t understand how the circuit worked, but now I have a partial schematic, and in the process discovered it’s ideologically similar to the monitor buss switching logic in the M-50/M-512/M-520. Anyway...

Now back to the malfunctioning oscillator!
 
At the risk of upsetting Granny with eggs...Can I give a short test procedure for jfets?

You need a meter on Ohms and some croc leads. If pinouts unknown pick two. Does it read 100R or less? If so good! You have S & D (unless device is ffkd) . Now bring a plastic comb close to the third pin. Fet should turn off a bit. If not comb hair and try again. Try not to touch the gate but in fact they seem to survive if you do!

MOSFETs on the other hand are less rugged, make up a test rig with PP3s.

Another dodge for kit full of 'state indicator LEDs" A DMM on 'diode' setting will usually light these as you poke about.

Dave.
 
Thanks for the info...I’ll add it to my bag of tricks!

I verified the pin configuration and part type of the “304” transistor using this technique:

“Measure a FET pulled out of the board with an ohmmeter. Set the meter to Diode Test. Attach the positive meter lead to the pin you think is the gate, and the negative lead to one of the other pins. If the reading is around 0.7 volts, you've found the gate diode. Now move the negative lead to the remaining pin. If you get the same reading, the FET is N-channel, the positive meter lead is connected to the gate, and the other 2 pins are source and drain. Usually S and D are interchangeable. If you didn't get the same reading when moving the negative meter lead, connect the positive lead to another pin and try again. Same procedure for P-channel FETs, except interchange the positive and negative meter leads (negative lead goes to gate).”

:cool:
 
Thanks for the lesson gentleman...... This place is an awesome repository of information!!

Glad you knocked one out SB....... I’m really curious (as I’m sure you are too....) what that oscillator issue is.......
 
Glad you knocked one out SB....... I’m really curious (as I’m sure you are too....) what that oscillator issue is.......

So the oscillator depends on a certain amount of noise in the signal path to get kick started. It’s like feedback, but frequency and amplitude controlled. But it’s got to have enough noise to get started. First step is replacing the small value caps around the opamp with parts that are in-spec, and go from there. I also got this zany idea today to pull the local power rail filter caps and see if that made a difference. The caps I used with the recap are lower ESR caps than the stock ones, and would do a better job of shunting HF noise to ground...better filters. So...yank ‘em and see if it makes a difference with the oscillator functionality. Problem is those filter caps also take care of the talkback mic amp, but that’s not really critical if that becomes noisier. If there were modern opamps installed for the oscillator and talkback mic amps, the filters being gone might not make a difference (because of better CMRR spec) but those amp stages use the 4558...filters being gone might make a difference. Hmmm.
 
Okay...well...that didn’t work. But at least I didn’t break anything else. :D

So I need to finish getting my Mouser order together so I can replace those small value caps around the oscillator driver and then return to the drawing board.
 
So here's a chunk of useless info.

I was rooting around in my Tascam 58 parts machine last night looking for metalwork to mount the add-on power transformer to the PS-3500 power supply I got from Miroslav...the power transformer for the +/-35V power rails for the uberjuju semi discrete ultra high headroom line amps in the M-__. The parts 58 was an "OB" balanced version, so laying in there was the input/output balance amp board. A number of posts ago I gave an account of my acquiring the service manual for the mkI version of the LA-40, and found the output balance amp to be kind of like a baby brother to the amp circuit in the M-__...well, I pulled the PCB out of the parts 58 and stared at it for a couple minutes and thought "hmmm...this looks awfully similar circuit-wise to the LA-40...I wonder...", so I pulled up the schematic for the 58 balance amp assembly and got out the LA-40 mkI service manual...sure enough, they are very, very similar. The LA-40 mkI has some neat additional features like a little PCB mounted trimmer to adjust the CMRR on each of the output channels...and the level is switchable between +4 and -20...the input balance amp on the LA-40 mkI is quite a bit more intricate than the 58 balanced input...the 58 circuit is about as simple as you can get, which has its plusses and minuses...so anyway, I then remembered thinking in the past the 58 balance amp board looked an awful lot like the balance amp board in the M-512/M-520...so I pulled out that manual and sure enough, the output balance amp is identical between the 58 and and the M-512/M-520, and, again, all like a baby brother to the M-__. The bare PCB in the M-500 consoles is almost the exact same part as the 58's, but because the M-500 balance amp doesn't have balanced input to unbalanced output functionality like the 58 or the LA-40, there's about a quarter to a third of the board that isn't stuffed with parts. Another difference is the M-500 output line drivers operate on +/-18V, while the 58-OB and LA-40 mkI output line drivers run on +/-20V. that affords relatively high headroom, and something I don't think you'll find on much anything else contemporary to those units from their competitors back in the day...certainly won't find anything running on +/-35V like the line drivers in the M-__...Teac was thinking outside the borders of the box they had defined for themselves...its interesting to see the trickle down from the no-holds-barred M-__. I was also reminded the 58-OB and M-500 output amps were switchable +4/+8dBm...like the M-__. +8dBm is the broadcast industry standard...Teac was gunning for the video and broadcast industry...they were stretching to other areas besides home and project studio recording. Anyway, its really interesting to me to see these themes across different products from that era.
 
The oscillator works!!!

I got the oscillator working!!

New WIMA box film caps did the trick! There are three 4,700pF caps (C1, C3 and C5) and one 390pF cap (C2) in the oscillator circuit...replaced them all with new low tolerance parts and the oscillator is firing strong on all five bands straight away at power up!

Yahoo!!!

Took awhile to perform this surgical procedure as it took me forever to put my Mouser parts order together; I was ordering for multiple projects, several of which took a lot of research and collaboration to determine what parts were needed, but I finally got the order the day before yesterday (my biggest Mouser order ever :eek:), swapped in the parts and I am now checking the oscillator repair off the to-do list! There are a couple other parts I want to replace as preventative maintenance in the Control Module, but then I’m going to run one more comprehensive function test, put the cosmetics back on and call it “done”...100%. I’ve had the Control Module apart so many times I lost count long ago. Hoping that’s coming to an end for now.

And yes...go ahead and point out I’m spending my New Year’s Eve nerding out on the M-__. It’s been a busy couple weeks with the holidays, and my wife had to work night shift tonight...just me and the animals here...and I’m glad I gathered the gumption to swap those caps...I’ll be honest I wasn’t confident it would work. But it did!!

:D

BED38AF5-E72D-4506-ADFC-2DF247079811.jpeg
 
Well done Beats! Oscillators can be funny things. Yonks ago I had a Ford estate car and the radio would not work below about 2C. I would set off to work and get R4 news just as I rolled into the car park. Local RF oscillator just on bottom gain tolerance I would guess but I could not be A'ed to have it apart and investigate, 355 days of the year it was fine!

BTW, I have just recalled that in the days of all valve tape decks the manual gave the instruction to short the erase head to kill the bias osc' so you could measure audio signals. DON'T do this for a transistor deck. Yes, you WILL kill the bias osc' but likely the transistors in it as well! (sorry if this is 'grannies and eggs' to to Beats)

Happy new year friend.

Dave.
 
Lol...love that.

Too true...because...when the perseverant keep at it like a three year old on crack *nobody* wants to be the one withholding victory...toss that like a hot potato. :D
 
:D ..... yeah....and don’t pay any attention to Waterloo and Napoleon’s eventual exile....I’m sure that’s all just fake news....... :D
 
Well, folks...after many months of agonizing frustration and disappointments, the Control Module of the Tascam M-__ is verified 100% functional and reassembled. I’m “done” with the Control Module unless something goes kaput on it (and something will eventually I’m sure), or until I’ve finalized upgrade details to the high-headroom +/-35V powered balanced main buss output amps. Otherwise all the selective “upgrades” and repairs are finished, it’s recapped, and fini. The only thing left to do is to fine tune the level balance between the L and R channels of the main buss (trimmers accessible through the dress panel directly above the master fader), and calibrate the main buss meters. Will take a little chunk of time this eve to do that. So the next thing is to return to the comprehensive function testing of the I/O modules. Wow what a meandering journey this has been...I’m finally back to what I was doing mid 2014...! :eek:

Geez.

Once I’ve got the I/O modules working 100%, I’ll do some selective upgrades to each of those over time, but before I embark on that I will certainly declare the M-__ “done” before the next phase. There is a power supply upgrade in the future too.

But here’s the Control Module all dolled up:

226A3A80-CE34-4E10-A125-C91750BB479D.jpeg

87A7D27C-5941-45B8-B106-26DD6CC198EF.jpeg
 
And.....the new year is off to a great start.

Incredible progress you've made, especially considering there is no manual, nor anyone that knows anything about this console
 
Lots...

...and lots...

...and LLLLLOTS of help is what it has taken. Boy has this been a community effort. And honestly I know for fact I’ve frustrated some of my helpers along the way because I couldn’t “let it go”.

I was reminded how obsessive I can be this afternoon when I was putting the Control Module back together: there are two self-tapping sheet-metal screws that fasten the dress panel to the chassis at the top of the panel. One of them was missing. I searched for 45 minutes, and in the process even looked in my pile of hardware from parted out Tascam gear. I found another screw with the same finish, same size head and thread...you’d never know it was a different screw once fastened, but it was a different length. I knew that would bug me. So I kept looking, and looking, and looking. I finally found I had put both the original screws in the baggie with the knobs. Phew!! :o So I think I’ve demanded a lot of patience from my helpers, and it doesn’t make it any easier this project has dragged on for over 9 years now. It would have taken a lot less time if my life wasn’t so full, and if I could leave well enough alone and settle for “good enough”, but my “good enough” is pretty ridiculous sometimes...most of the time I guess. But I have learned A TON through this project, and am getting pretty damn capable of troubleshooting and fixing at least this device, which enhances its practicability, which is is huge to me. I’m thankful for what I’ve learned. It’s nice not being intimidated so much by complicated puzzles. Like I know, as of today, I’ve got one I’ll need to sort out with I/O module #12.

Some of you may recall I/O Modules 1-11 are pretty refined and jank-free, but module #12 has got hand-written labeling all over the chassis, and a fair amount of janky soldering and resoldering on the PCBs. It was clearly the test-bed and once they had it working or at least close they patterned the others after it. So I’ve already had to make several repairs and modifications to it to match it up to the others, and today when I was testing the global programmable Monitor Buss source switching on the Control Module it wasn’t working right. The Control Module has the ability to control the source of the Monitor Buss on the I/O modules...each I/O module Monitor Buss source select switchrack has a REMOTE enable switch that, when latched, disables that module’s source switches and defaults source select control to the switchrack in the Control Module...there are also two banks of input source select controls in the REMOTE section of the Control Module, so the source of each I/O module main input, MIC, LINE 1 or LINE 2, can be controlled at the Control Module, if the I/O module is so enabled, and there’s bank A and bank B so you could have two distinct sets of modules’ inputs controlled by the Control Module, and maybe some not enabled and independent...source controlled at the module...cool thing is REMOTE banks A and B controls also include mute functions, so you can even have the global source control off on the Control Module for, let’s say, bank A (which leaves the REMOTE-enabled I/O module’s source select control still active at the module), but latch the MUTE button in the bank A switchrack on the Control Module and any modules with REMOTE A enabled will now be muted...voila...mute groups!! Anyway...I should, at some point, put up a YouTube video about this feature of the M-__...it would make more sense, and it’s a really cool and powerful feature of the M-__. In fact I believe it is the very reason the Control Module is named as such rather than something more common like Master Module...so, back to my point, the Monitor Buss REMOTE switching is not working right with module #12, and it surely has to do with the CMOS logic switching circuits on module 12. I’ll fix it. Who knows...maybe it’s never worked right. I’ll make it work.
 
Back
Top