Yamaha SPX90 Repair notes.

Yamaha_spx90.txt

a "repair " note for a failed to repair one. I even worked out the schematic

but could not find the reason for the slightly linear ps characteristic, not

full switching, and then thermal runaway. Even tried rewinding the secondary

of a generic smps toroid and could get it reliably working but only with a

fan. There was some shorted turns in the original, presumably due to excess

heat build up, but before blowing the TOP66 chopper. In retrospect just

enough space for 2 squat linear tx and retaining the ac throughput circuitry

built onto another board

Yamaha SPX 90, 1986 , Digital Multi Effects Processor

Cold checking , have found 2 things wrong with the ps so not powered up yet

No SMPS controller IC on the ps board , a TOP66 switcher driver and 3 small

transistors.

Raw + and -18V go to main board +/-15V regulators , raw 5V , ie rectifier

off the switching transformer followed by one L and one C goes straight to

the 5V TTL rail. Not even an obvious fat 5V zener across it.

No SMPS controller IC , that I can see on the main board and the 2 lines

from the main board marked PE and PC that go to the mains side of the ps via

opto isolator .

There are 15 assorted wirewound resistors on this ps , again rather odd.

If I get it working , a ventillation hole and grill cut in the top cover

would help.

It was that hand over period , 80s , when phenolic pcbs went over to

polyester.

This Yamaha has polyester main board and phenolic light brown ps pcb that

has a lot of heat dark browned/brittlised? areas. Not surprising with

assorted 15 W/W plus sm driver etc with no ventilation holes anywhere.

Got it back to working to its earlier non-working condition before it was

seriously

mangled by a "repairman".

Can now see that the original problem was an 85 dec C electro that did not

like being cooked in there.

I determined the schematic for this SMPS

The 5V regulation, unlike the +/-15V, is on the ps itself, in discrete form,

consisting of 300mW zener,preset and TO92 transistor.

Working at 2/3 mains coltage but temp of the chopper transistor goes too

high, if I let it, and power consumption goes well above 20W.

Looks as though something has gone wrong with the mark/space control.

It slowly increases so the "5V" line goes 2...3V and then jumps to 5V with

equal mark/space which is presumably not necessary.

I checked there is no excessive load on the +/-18V or 5V rails despite

all functions working as far as front panel display.

To get a rough idea of the amperage capability of a small 5V SM ps

before attaching it to its load, to simulate, just in case the regulation is

kapput.

It uses a TOP66 of sufficient rating to

drive a few hundred watts across but the ps is relatively low power , but

unknown rating.

The one in question is only about 20mm cube

I've fathomed out how the ps works. The 5V regulator uses only a 100mA TO92

transistor because the error is fed forward via an optocoupler to the sm

oscillator.

The 5V line defines the output to +/-18V as well.

The original trouble was wavering 5V, which has returned, prior to replacing

some more caps . In the broken mode if say 2.5V on the 5V o/p then only

+/-10V on the unloaded +/-18V lines.

Scaling linearly, an at hand 70 W smps, with its 45mm cube transformer.

I would estimate volume for vol then this one with 20mm cube would be about

6W total. Allowing 1W for the few op-amp supplies and 2 relays then only 1

amp at 5V, its suppling 30 odd 74HC and 3 LSI.

When operating properly the oscillation is about 59 KHz with equal

mark/space ratio,

when at half-cock then about 40KHz short mark/ space

The consumption plate states 20W, assuming average SMPS 70 percent

efficiency then 14W at DC.

uses 2SC2792, 2SC2634 /2034?, 68uF/400V, 2x 1R FR

3 x 100K w/w ,6.8,2x 68K,22,2x 330, 2x 120K, 6.8, 1.5, 22, 27

2 fuses marked micron63 FN191R0J

6.8,100R on main board

The second optocoupler is the other way round.

Feeding rectified but unsmoothed mains signal info to the main board.

Phase info for mains hum/noise reduction?

6.2V zener on ps for 5V regulation section.

In working order but short duration of 65 percent mains

load was 85mA on -18V line, 128mA on +18V, and 360mA on 5V

so 5.6W (placing 0.5Rs and 0Rs between plugs and sockets)

At 70 percent mains was getting 2.5V on '5V' and -10.5V on '18V' slowly

building up to 5V, feeds forward via pc1.

Replaced main chopper transistor, D1207, 2x6,8 to 12R, 2x 330 to 2x1000, 22

to 68

and 1.2K 1/3 watt setting chopper bias to 390R and unnecessarily replaced

PC511 just

in case.