SONY HB900, Help

Page 2/3
1 | | 3

By Wlcracks

Champion (315)

Wlcracks's picture

03-04-2019, 16:12

I don't know this power supply. I guess its a linear power supply (big transformer). Maybe the FULL BRIDGE RECTIFIER is kaput and AC is on the cap. Oscilloscope will tell you everything.

By RetroTechie

Paragon (1551)

RetroTechie's picture

04-04-2019, 01:02

This is a switch-mode supply - rectifier is in 100V AC section. Can't find a schematic in the service manual, only a pcb layout for the power board. But seems like a fairly straightforward circuit. Mains AC fuse, line filters, rectifier + big cap, some oscillator circuit (all discrete components?), probably with frequency or duty cycle controlled via the secondary -> primary optocoupler.

May I suggest: disconnect from mainboard / drives etc. Replace fuse with a slow-blow one of much lower value (say, 1/10 to 1/4 of nominal current), connect to AC and check the output voltage(s).
-Output voltages fine: might be a short circuit (or perhaps another blown capacitor) on mainboard.
-Output voltages missing: some more parts on the power board failed. In this case it would -for starters- be interesting to know if high voltage DC is present on the big primary capacitor. Yes -> primary power switch transistor may have failed, or some issue in control/feedback circuit. This may require a 'pro' to figure out the problem. No high voltage on cap (or fuse blows again) -> maybe rectifier failed as well. Easy to check & repair.

Whatever you do, DON'T replace mains fuse with a 'stronger' one or even bridge it! That will just kill more stuff on the next power up. Cool

When you saw "sparks flying", did you see where in the machine? Power supply area or elsewhere? If power supply area, primary or secondary side?

By Gloriou

Expert (118)

Gloriou's picture

04-04-2019, 06:15

RetroTechie wrote:

This is a switch-mode supply - rectifier is in 100V AC section. Can't find a schematic in the service manual, only a pcb layout for the power board. But seems like a fairly straightforward circuit. Mains AC fuse, line filters, rectifier + big cap, some oscillator circuit (all discrete components?), probably with frequency or duty cycle controlled via the secondary -> primary optocoupler.

May I suggest: disconnect from mainboard / drives etc. Replace fuse with a slow-blow one of much lower value (say, 1/10 to 1/4 of nominal current), connect to AC and check the output voltage(s).
-Output voltages fine: might be a short circuit (or perhaps another blown capacitor) on mainboard.
-Output voltages missing: some more parts on the power board failed. In this case it would -for starters- be interesting to know if high voltage DC is present on the big primary capacitor. Yes -> primary power switch transistor may have failed, or some issue in control/feedback circuit. This may require a 'pro' to figure out the problem. No high voltage on cap (or fuse blows again) -> maybe rectifier failed as well. Easy to check & repair.

Whatever you do, DON'T replace mains fuse with a 'stronger' one or even bridge it! That will just kill more stuff on the next power up. Cool

When you saw "sparks flying", did you see where in the machine? Power supply area or elsewhere? If power supply area, primary or secondary side?

*Already did disconnecting it from the FDDs and main board, connected the PSU with slight load and the output was fine (11.9V)
*As of capacitors on main board, nothing seem to be bad, all in good shape despite of the age.
*spark was in the area between the big capacitor and the additional socket for the digitizer, (Primary)

By Wlcracks

Champion (315)

Wlcracks's picture

04-04-2019, 08:22

Maybe this can help.
https://www.youtube.com/watch?v=NoGl7oeh7eU

I repaired a lot of stuff, but switched power supplies are out of my league. They work on an feedback system and sometimes custom ic's and high voltage. Some switched power supply's need secondary load to work correctly. Having a mains decoupling transformer and a bench load resistors are a must. No schematic, no other supply to compare too, not my cup of tea, maybe replace it by an other switched supply. There a lot of (cheap) alternatives available from china. Its more a mechanical thing finding the right size.

By ali6x944

Master (196)

ali6x944's picture

16-06-2019, 11:16

I and @Gloriou were working on this machine for quite some time, for the power supply I would say it may be the switching transistors because when those fail they tend to blow up the primary smoothing caps, however, we have long moved from the power supply to the main board, if my assumption is correct if the power supply fails with a shorted switching transistor then the output of the power supply will not be available, thus the board wouldn't be affected as severely as if one of the rail regulators fail...

so powering the main board with the necessary voltages from an ATX power supply would solve the problem, and would show a working machine, but to our surprise, the machine didn't turn on properly i.e a blue screen appeared for a fraction of a second before it blacked out.
sadly the service manual is nowhere to be found, so I have to go with gut-feeling and common sense, I have read an artical in Alexandre Souza page about his MSX Sony HB-F900 that did not accept cartridges and how he solved the problem by changing a buffer:

Alexandre Souza wrote:

This micro [HB-F900] is all protected (something that MSX is not). You can see that all of the address, data, and control bus signals are buffered, that is, there is a little bit that "reinforces" the signal when it passes through it, and it helps to get in trouble and something to ' - protect what lies behind it. And yes, I know that part of the data buffers is not in this scheme: o) So the first thing I did when I saw that the cartridges were not being recognized, was to pass the scope on the input and output of the ICs that buffer the signals. I found two signals at the input of the buffer, with 0 at the output.

from what I understood that the buffering also apply to the VDP and MSX-engine and Z80, right?
if so I should look for a buffer or a tristate latch that is not working form Z80 to VDP, right?
I should also check if the Z80 is working, so I am making a Z80 based on Thomas Scherrer minimum Z80 tester.

By RetroTechie

Paragon (1551)

RetroTechie's picture

17-06-2019, 02:25

Did you check ALL supply voltages while machine was powered by ATX psu?

A link to the service manual can be found on the model's Wiki page on this site. Just for clarity: this is a HB-G900 we're talking about here, not a HB-G900AP right? Question At least the digital part is very different between those. Dunno about power supply. Power supply looks the same between these models. Service manual for the -AP has a schematic of it on page 90.

Don't bother with 'wild goose chases' like a Z80 tester or data/address buffers. Basics first: optical checks -> power supply -> clock signals -> reset circuit. Seems you're still stuck in 2nd part. Logic probe (or better: oscilloscope) would be helpful for further checks.

By ali6x944

Master (196)

ali6x944's picture

18-06-2019, 12:05

@RetroTechie, All the voltages have been checked from the ATX, with only the -12v being -9v because we happen to use a really old ATX we had laying around...

-for the onboard voltage check, everything seems to check out, I haven't tested the big chips like the VDP or the MSX-engine or FDC yet, but I will get to them soon.

-optically only the rectifier cap seems to be damaged with a slight bulge on top, and a crystalizing RTC battery, the battery was removed immediately, no damage to the board is visible.

-clock and reset weren't checked yet, will hopefully. check them with my scope...

By ali6x944

Master (196)

ali6x944's picture

18-06-2019, 12:20

And for the machine, it is the HB-F900 -Japanese version- similar to Alexandre Souza repaired.
sadly HB-G900 is not that similar to its Japanese counterpart, neither in the main board nor display board.

By ali6x944

Master (196)

ali6x944's picture

22-06-2019, 11:59

Solved

By Manuel

Ascended (15449)

Manuel's picture

22-06-2019, 14:03

Nice to read that, but it would be better if you explained how....

Page 2/3
1 | | 3