In this thread we build the projects that others have shared - Sony 129x clone card for PVMs

[bodgit]

This is such a treat - always a pleasure to see the work of a master, so many sweet little touches in there. I particularly like that you moved the edges of the ground plane in from the edge of the board. If the solder mask rubbed through against the edge of the guide there won't be conductivity. I'm sure there are a million more that I haven't found yet

I've probably done something like this:

View: https://youtu.be/ahKH19iN2SM


I like and hate routing PCBs in equal measure. It's a perverse logic puzzle trying to route everything without resorting to hopping back and forth between both sides of the PCB. I am an amateur though so I've probably committed some sort of routing faux pas somewhere though...

My cards have been swapped over a few times and I noticed that the little threaded tabs in the PVM chassis that hold the card secure had a tendency to catch and scratch the solder mask off the PCB so I just added a no-fill zone on the top and bottom edges which should hopefully fix that. Although with any luck I shouldn't need to swap cards so often, if at all!

The only other significant change I made is to move the two switches back from the edge of the card slightly. I noticed with a cap fitted to the switch you can still see exposed plunger/spring between the cap and the bracket so I just moved the switches inboard by a few mm so the base of the cap will sit inside the bracket. It will probably mean the bracket will need the holes around the switches made slightly bigger and standardise on either round or square caps. 1RBLK is the part # for a black round cap I was leaning towards using, they're 5mm in diameter.

 

[hatmoose]

Boards on order now (along with neobiosmasta, blueSCSI, SAT2Neo, a 42 pin adapter for the 866, and some others that I've forgotten already.
Now I have 20 working days to wait for delivery and to reflect on the significant natural beauty of living in the middle of nowhere.

 

[bodgit]

I've got my PCBs now. I'm just waiting for the components to arrive. When I built the original clones my component order was getting messed about with line items disappearing from stock, and every time I amended it a) it took 2/3 days and then b) something else would disappear from stock. So I placed another order with everything definitely in stock with the aim to cancel the original order. Except I forgot to, so it's been in pending since August last year, literally now just waiting for those pesky ADG1611 ICs to finish being produced. I worked out what additional components I need to complete the BoM for the new design and ordered those separately so I'm just waiting for this order to ship.

For the standoffs, I've now tried both 23 & 24 mm; 23 mm seem to be a better fit IMHO.

The only revision I might make to the PCBs are the cutouts for the two securing tabs on the SCART connector. When I ordered the original clone boards, JLCPCB cut them as square holes, but on this batch they're just circular and the tabs won't go through easily. I'll try and square the holes with a rasp this time, and I'll fix up the SCART footprint so its less ambiguous.

 

[bodgit]

Still waiting for components to arrive so I’ve had a go at designing a suitable bracket in OpenSCAD. Printed the first couple of mm to check all of the cutouts are in the right place and it’s mostly there:

Not bad for my first attempt! I’ve got a couple of sub-mm tweaks to make; BNC cutouts could be a tiny bit smaller and I didn’t quite get the angle right on the SCART socket but that was mostly by eye as every technical drawing out there omits that measurement.

I added some labels to each port and they’re readable, but a bit blobby around the edges, I’m not sure if I can tweak some settings or need a finer nozzle for that level of detail? I’ve literally just stuck with printer/slicer defaults so far.

 

[hatmoose]

looking awesome! And great work in OpenSCAD there. I’m trying to transition my entire tool chain away from rental software and onto FOSS too.
FreeCAD (hard!)
KiCAD (awesome!)
Slic3r (an oldie but a goodie)
My Ender3 pro is about the cheapest 3D printer on the market, it can achieve good results with some tuning of the settings. From that photo it looks like your flow rate is a bit low, I followed YouTube vids for how to dial it in.

The boards have arrived, my wife will be giving them to me for Christmas. Can’t wait for Christmas Day!

 

[bodgit]

looking awesome! And great work in OpenSCAD there. I’m trying to transition my entire tool chain away from rental software and onto FOSS too.
FreeCAD (hard!)

I used FreeCAD to create some 3D models, ironically for a KiCAD library for some JST connectors, and I found the UI a bit confusing but eventually managed to create what I needed. I was going to use FreeCAD again to design this bracket and found myself getting frustrated in the UI looking for the tools/options I knew existed, so I thought I'd stop and give OpenSCAD a go instead. I quite like that it's basically a reproducible script.

I like CAD in general and did some AutoCAD qualifications about 25 years ago on what is now an ancient DOS version so I think my brain works quite well in "CAD mode".

KiCAD (awesome!)

I've only ever used Eagle previously and then jumped to KiCAD when AutoDesk borged it and started charging a subscription.

From that photo it looks like your flow rate is a bit low, I followed YouTube vids for how to dial it in.

It's either that or the infill overlap I think, I'll fiddle with the settings for the next test print once I've adjusted all of the ports.

 

[bodgit]

After a couple more partial prints I got all the ports in the right place so I thought what the heck, print the whole thing out in its intended black filament:

I then test fitted everything together with some of the various ports in situ (they’ll be the last things I solder in normally):

I’m pretty happy with that.

While I’m still waiting for components I’ll get on with starting to make the changes to the Arduino firmware.

 

[hatmoose]

Boards arrived a little while ago - will get to work fabricating...


Also note to myself - the LED indicator goes this way

 

[bodgit]

I've literally just approved Mouser to ship my component order, minus some ceramic caps that appear to be unobtanium but I can manage without.

 

[hatmoose]

I've literally just approved Mouser to ship my component order, minus some ceramic caps that appear to be unobtanium but I can manage without.

Got distracted, back into it now

Just looking at the BOM, perfection does not come cheap...
U3 = ADG733BRUZ
U4 = ISL59885ISZ
U6 = ISL4089IBZ - this one in particular is a spicy meatball

Were you able to find a cheaper supplier than Digikey/mouser?

 

[hatmoose]

Started soldering it all together today.

The ISL4089IBZ is about a $30 part, came in its own bag, in its own box, with clear instructions to place on a flat surface before opening the box - and inside this tiny little chip. Fun!

 

[hatmoose]

Getting close now - If I make another one of these I'll probably shuffle the bottom 5-pin JST connector back 1/4 an inch. The big array of caps makes it a tight fit at the moment.

I do absolutely love JST-NH; The connectors are nice, the pins are thick, sockets always crimp really nicely, the mechanical strain relief is awesome - what a great system.

 

[hatmoose]

So nearly finished!

I carefully checked and double checked the BOM, then left a couple of parts off my order. So another few days while I wait for the missing parts to arrive.


In the meantime I will print the bracket that @bodgit has kindly shared here
https://github.com/bodgit/bkm-129x-ultimate/tree/main/bracket

 

[bodgit]

Got distracted, back into it now

Just looking at the BOM, perfection does not come cheap...
U3 = ADG733BRUZ
U4 = ISL59885ISZ
U6 = ISL4089IBZ - this one in particular is a spicy meatball

Were you able to find a cheaper supplier than Digikey/mouser?

I think I found a handful of ISL59885ISZ on eBay, and the ISL4089IBZ was available from RS.

 

[bodgit]

Getting close now - If I make another one of these I'll probably shuffle the bottom 5-pin JST connector back 1/4 an inch. The big array of caps makes it a tight fit at the moment.

Yes, I realised that mistake. A defect to be fixed in the next revision.

 

[hatmoose]

Hardware = done. It really is a thing of beauty - thank you!

And heavy. It feels like a serious piece of enterprise-grade piece of kit.

I'll try it with Martins stock 129x firmware and see if I can get some basic functionality and see what happens


Speaking of enterprise-grade hardware - I'm not totally in love with the Aliexpress BNC ports.

 

[hatmoose]

Came back to this one after a long hiatus, progress!

Basic functions seem to be working. Card detects as expected, cycles inputs with the input button


No green, not sure if this is a fabrication fault, somehow tied to the SoG features, or a result of the firmware hackery I did to get it to compile.


But - it works! Great progress

Other notes to self, firmware sometimes needs to be reflashed at random, card sometimes fails to detect on boot and needs reboot (suspect this is related to the bootloader timeout that Martin noted a while back), other oddness (suspect from the button high/low)

 

[hatmoose]

Troubleshooting the green

The sensible way to start troubleshooting would have been to power this from an external power source, probe around to see if I can figure out where the green signal is getting lost.

The time consuming, difficult, and expensive thing to do is design a custom "L-series z" board extender. So clearly that's the right approach here.


I designed this to poke out of the back of the monitor quite a long way. I'll print some kind of mechanical strain/locking bracket so the DUT doesn't come unplugged while being probed at.

Since my Kicad skills are comically awful I'm almost ashamed to post the outcome, but here it is. Looking at this picture I probably should have made the voltage traces a bit thicker... anyhoo

 

[hatmoose]

After making this thing I realised that I had missed a couple of key things
1) Probe points
2) correct port alignment.

Its good enough for this test - now to design a bracket to hold everything in place so I can lie the monitor on it side and give it a jolly good probing.

 

[hatmoose]

After much jiggery-pokery (the best kind of pokery) I have made enough test equipment to begin troubleshooting.



My not very sophisticated test procedure shall be
1) Probe for continuity to check that green line matches others
2) 240p test suite solid colour screen - logic probe for pulse on each colour
3) 240p test suite - oscilloscope for waveform on each colour
4) begin randomly reflowing and replacing chips