While cleaning the storage I found a very interesting power supply made by Sanwa and marked SNW-7E which works with 110 or 220V
The model SNW-7F is the same model working with only 110V
It’s a very nice and compact power supply ( with a tendency of getting really hot!) and I learned that it was used on the most famous japanese superguns .
When I powered it up it worked but on the -5V line I got -7,5V .
I decided to try to repair it and fortunately the -5V was given by a 7905 regulator with few parts around.
I immediately noticed that the regulator didn’t regulate at all since it passed exactly the same voltage on the input on its output.
Replacing it with a new 7905 fixed completely the problem
Note:
it turns out that this power supply is just a rebadged version of the RT-85A model made by Taiwanese company Meanwell
I recently aquired a few Robocop PCBs, and this particular one seemed to be working fine at first glance. Eventually however, I realised that certain sounds seemed to be missing from the attract mode.
In particular I noticed Robocops speech as he rolls off the Prime Directives was not there. A closer look playing the game revealed further sounds missing, like the groans from the enemies as you kill them and others.
I then realised that is was just the ADPCM samples that were missing.
So I looked at the CPU board where most of the final digital and analogue circuits are. First I checked the DACs and found no samples present, so went further back to the large YM2203 sound chip. I probed pin 19 of the IC and would have expected to hear the samples here, however I heard nothing except static.
Pin 19 circled in redPin 19 circled in red
I was able to determine that the ADPCM samples are stored in the EN02 ROM, so I popped this into an eprom reader and checked the content.
The eprom was corrupted, so I burned a new one and the samples were restored.
Today I’ve uploaded the latest version of my BINman software.
This version is v4.3.2 and has had some rewrites and a few extra features added.
First is the ability to deinterleave the bytes of a file into nibbles.
Second feature added can be found under the “Special” menu. It was added as a request from a user. This swaps the nibble values from one byte and the adjacent byte then deinterleaves that file. For example an original file of: “12 34 AB CD” will output two files, one containing “13 AC”, the other “24 BD”.
Its not something I’ve ever had the need for but obviously someone wanted it so here it is.
As usual it can be found in the “Downloads/PC Software” section.
GeneralComments Off on Jaleco ‘JK-03’ reproduction successfully tested
Dec312018
Some time ago I made a reproduction of the Jaleco ‘JK-03’ custom SIL used for inputs handling on some Mega System 1 PCBs but at that time I was not able to test it due to the lack of a proper PCB hence I asked for some volunteers.You can read the post here:
Today I can report the reproduction is fully working thanks to the test carried out from the user ‘Astro X’ of AP forums on his Saint Dragon PCB with missing ‘JK-03’ (actually an hack was installed as workaround) :
I received for repair a quite rare original Spelunker PCB (on Irem M62 hardware).Board is a three stack one made of a top board (which carries most of sound hardware) :
SOUND board
A middle CPU board (specific for each M62 game) :
CPU board
And a bottom VIDEO board :
VIDEO board
The PCB had severe GFX faults, the sprites were only lines vertically stretched all over the screen:
All the sprites circuit lies on bottom VIDEO board which is the same for all games that run on M62 hardware.Looking at Kung-Fu Master schematics I could figure out that data bits from sprite ROMs are fed into the custom marked ‘KNA6034201’ :
Custom ‘KNA6034201’ in DIP40 packageCustom ‘KNA6034201’ schematics
The inputs were all active but most of outputs floating:
So the custom was most likely internally faulty.Luckily I have done a reproduction of this component some time ago.You can think of it like a 24-bit parallel to serial shift register:
I removed the custom and installed the reproduction:
‘KNA6034201’ repro
The sprites were back but not perfectly as they were lacking of lines and misplaced too:
Sprites not yet perfect…
The sprite line buffer consists in two 2k x 8-bit static RAMs (Toshiba TMM2018 used here) :
Sprite line buffer
Probing them revealed on both a floating address line (pin 1, A7) :
Address lines are generated by the surface mounted custom ‘KNA6074601’ :
Custom ‘KNA6074601’
Its pinout/implementation from Kung-Fu Master schematics :
I had no other choice than replacing this part so I asked the owner to look for a donor board.He found and sent me a Vigilante PCB which carried the ‘KNA6074601’ on bottom board :
Vigilante donor board
I removed the faulty one and soldered the spare back :
