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

[hatmoose]

The Minigun Superguns went well - you can read all about them here.
https://www.arcade-projects.com/thr...d-the-projects-that-others-have-shared.17296/
I should probably sell/trade the extras, let me know if there is interest

So clearly time to take on something more challenging. This time i'll be replicating Martin Hejnfelts awesome work on the 129x clone card for late model sony PVM's. I've been in contact with Martin on this one, he was super helpful with all my noob questions. Thank you Martin, you are both a scholar and a gentleman.

You can read about the project here
https://immerhax.com/?p=459
It's worth noting that Martin has produced a series of designs. An identical clone, a simple clone (the one I will be making), A VGA+SCART model. and others

How this came about for me was that my trusty PVM 9040 is a rotten bench test monitor because it's composite only and PAL only.
I wanted a new(?) and fancy bench test monitor that could do composite, Svid, component, and RGBS. and worked with PAL and NTSC inputs
An PVM 8045 or a PVM 9045 would have been ideal, but I couldnt find one for sale at a reasonable price, they were all three hundred bucks plus(madness!)
This PVM 9L2 came up for sale locally for sixty bucks. Given that it has been used as a field monitor the hours are low. The reason it was so cheap is that the geometery is all funky - but that holds no terror for the accomplished (or in my case reckless) CRT fiddler. I'll fix that last.

Of course the 9L2 also does not have Component or RGBS. These can be added with the rare and expensive 129x interface card. Clearly the only sensible thing to do is make one myself following Martins awesome project. It's also worth noting that although there are two blanking plates on the back here there is only room inside for one interface card.

Old lion (PVM 9040) Meet young lion (PVM 9L2)

So through the course of this thread I will be celebrating Martin's awesome work, showing him that his efforts are valued and appreciated, sharing tips and tricks, and (hopefully) encouraging him to continue his research into a possible x68 clone card. I'll be making 5 of these because that was the minumum board order. So if there is any interest I will probably sell the remaining 4 to help raise funds for the x68 research

 

[hatmoose]

Ordered the PCB's off pcbway today - this was super simple because Martin has kindly posted the gerbers here
https://immerhax.com/?p=459

I'll be doing the "simple" board - in hindsight I should have done one of the more complex ones because thy are technically sweeter and would have been better practise. Never mind.

PCBway takes a few days to manufacture, then a few weeks to ship to my part of the world.

In the meantime I will gather all of the parts for the BOM. This is much harder than it sounds because not all of the parts in the BOM are in production or can be ordered from digikey/mouser/whatever. Martin helped me out with the details and I promised I would post the summary here, which I shall in the next few days

 

[hatmoose]

So the BOM for this was harder than i expected. Had to hunt around quite a bit for some of these parts.

My prefrence (as always) is for quality parts with a tracable supply chain - basically Digikey, Mouser or if desperate RS. But some of these parts just werent available through those channels. There were a couple of parts, like the Arduino and ISP programmer where the cost of the Ali part was too low to ignore. normally I'm not a fan of buying copies where the original is available for purchase - but 10x the cost....

5 hours of internet scavenger hunting later, with a tonne of help from Martin, I give you a completed and orderable...

BOM for the "Simple" BKM-129x clone card

THS7374
https://www.digikey.co.nz/product-detail/en/texas-instruments/THS7374IPWR/296-41661-1-ND/5224405
ADG1611
https://www.aliexpress.com/item/1005001617175577.html?
74VHC125 (TSSOP14)
https://www.digikey.co.nz/product-d...nd-storage/74VHC125FT/74VHC125FTCT-ND/7251513
Arduino Nano v3 clone (5V/16MHz)
https://www.aliexpress.com/item/1005001636568464.html
ISP programmer for the Nano v3 clone - you might not actually NEED this, but just in case
https://www.aliexpress.com/item/4000307498536.html
MIC3490-3.3 discontinued and hard to get, But you can substitute the 5v version LM3490IM5X-5.0
https://www.aliexpress.com/item/32905717761.html
DTC144EKA
https://www.digikey.co.nz/product-d...uctor/DTC144EKAT146/DTC144EKAT146CT-ND/650727
64-pin connector
https://www.aliexpress.com/item/4000113049761.html
BNC plugs
https://www.aliexpress.com/item/32888289661.html

Passives
C7, C8, C13 = 100uf 16v 6.3 x 5.3 electrolytic
https://www.digikey.co.nz/product-detail/en/nichicon/UCM1C101MCL1GS/493-14525-1-ND/5800225
C10 = 2.2uf 25v 0805
https://www.digikey.co.nz/product-detail/en/kemet/C0805C225K3RAC7800/399-11941-1-ND/5267659
C11 = 1uf 25v 0805
https://www.digikey.co.nz/product-detail/en/kemet/C0805C105M3RACTU/399-8008-1-ND/3471731
C1, C2, C3, C4 ,C5, C6, C9, C12 = 0.1uf 25v 0805
https://www.digikey.co.nz/product-detail/en/kemet/C0805C104M3RACTU/399-8000-1-ND/3471723
R1, R18 = 10k 1%
https://www.digikey.co.nz/product-d...2ATTE1002F/2019-RK73H2ATTE1002FCT-ND/11476595
R2,R3,R4,R5,R14,R15,R16,R17 = 1R 1%
https://www.digikey.co.nz/product-d...2ATTD1R00F/2019-RK73H2ATTD1R00FCT-ND/10234940
R6,R7,R8,R29,R30,R31 = 150R 1%
https://www.digikey.co.nz/product-d...2ATTD1500F/2019-RK73H2ATTD1500FCT-ND/10234931
R9 = 0R or Jumper
Use a jumper for this
R10,R11,R12,R13 = 75R 1%
https://www.digikey.co.nz/product-d...RTTD75R0F/2019-RK73H2ARTTD75R0FCT-ND/12548709
R20,R22,R25,R27 = 100k 1%
https://www.digikey.co.nz/product-d...2ATTD1003F/2019-RK73H2ATTD1003FCT-ND/10235427
R21,R24,R26,R28 = 100R 1%
https://www.digikey.co.nz/product-d...RTTD1000F/2019-RK73H2ARTTD1000FCT-ND/12548737
R23,R19 = 4k7 1%
https://www.digikey.co.nz/product-d...RTTD4701F/2019-RK73H2ARTTD4701FCT-ND/12548766

The Ali parts will take 13-20 days to arrive at a minimum, probably more with covid

so this project is on hold until the parts arive

 

[hatmoose]

PCBway exceeded all reasonable expectations and delivered all the boards already. Digikey delivered the vast bulk of the components with their normal ourstanding quality and attention to detail. Aliexpess delivered the Arduino clones in THREE days.
Just waiting for the last few weird oddballs off Aliexpress, this might go faster than I thought.

I also ordered a bunch of boards for some other projects too (i'll write about those later). If you're wondering why they are all different colors I have not been loving the solder mask on the OSHpark stuff. It scratches off too easily. So I ordered a bunch of different colors off PCBway to see if they were any better.

 

[hatmoose]

Digikey are (as always) amazing - 5 days door to door and free shipping too!
Exemplary speed from Aliexpress standard shipping - only 14 days
Not such good speed from China Standard Airmail - 14 days and still waiting to depart the country...

So I have everyting I need to complete this project EXCEPT U1 (ADG1611). I'll be doing the whole lot with hot air, so I'm going to do everything I can without crowding the U1 pad, then finish off when that part arrives.

 

[hatmoose]

I did all of these at 390 degrees with a big nozzle which worked well. I'm guessing that this must be 390 degrees at the element, because its nowhere near that when it comes out of the tip. This seems to be about the right amount of solder paste for no bridges and decent coverage


I actually had to do some detective work for U2 and U4 because pin 1 is not marked on the silkscreen, luckily it was easy to trace the pins and check the datasheet. If anone else makes one of these the correct orientation is as follows

U4 (top left)


U2 (also top left)


So nearly all done on the test board, now I just need that ADG

I also found this video talking about a minor change to the legs on the 7374 to disable to the low pass filter for 720p sources

View: https://www.youtube.com/watch?v=fVdsC-JxgNM


So if you are using the gerbers from here
https://immerhax.com/?p=459
be aware that it's much easier to make this change BEFORE the 7374 goes on
(edit) alternative description from Martin himself here
https://immerhax.com/?p=697

 

[hatmoose]

Been a while since I updated this one - still waiting for my buddies in china to deliver hard to find chips that cant be bought anywhere else

But while I have been waiting for Aliexpress Standard Shipping, the awesome Martin Hejnfelt has reverse engineered the much more complicated (and also much more rare and expensive 68x card)

Blog post here
https://immerhax.com/?p=697
and youtube interview with Bob from retrorgb here

View: https://youtu.be/CVSulq0pL1M


Unfortunately the clones are currently ruinously expensive due to the global chip shortage, but hey, one day, one day...

 

[hatmoose]

Good progress - Aliexpress came through and delivered my ADG1611 chips - these are extrodinarly hard to get at the moment. Not sure if it's because of the global chip shortage or because they are obsolete.

The silk screen doesnt show which way up the chips need to go. I found this picture and verified by checking the datasheets, it is as follows
U1 = bottom right
U2 = top left
U4 = top ;eft


Because I had never made one of these before I put it under the microscope to check it carefully and found a sneaky solder bridge hidden UNDER one of the chips, took me ages to drag it out and clear it...


So hardware fabrication is done - Tomorrow I'll learn how to program a cheap chinese arduino nano clone from the BOM I posted a while back

 

[hatmoose]

Spoilers - it's done, and it works! Getting late here, I'll describe all the "stuff" I had to do to burn the code tomorrow. But if you're wondering if this thing actually works and if it is actually possible to make one at home - it actually is.

Before - option = blank


After - option = BKM-129x
You'll notice that the BKM-129x defaults to COMP(composite) when it it first installed, the colors are all funky until you change it to RGBS



Here is Street Fighter Zero 2 on a Japanese Sega Saturn in RGBS - my photography skills have not improved...

Note 9040 sitting on the bottom shelf, green with envy and, jealous of his brothers RGBS goodness

I'm actually pretty pleased with myself - worked first time!

 

[hatmoose]

Got sidetracked printing 3D brackets instead of writing up my notes on buring the code to the Arduino clone

There is an awesome 3-D printable bracket for this available here
https://www.thingiverse.com/thing:4178104/files
The bracket is pretty much required, without it the card wobbles all over the show and will probably snap off the backplane connector eventually

But the ones on thingiverse looked a bit plain-jane - I wanted something that was more in keeping with the spirit of the original card.

So I added some words and labels, obviously my 3D printer wont print the colors, so I'll try the nail-polish and acetone technique to see if I can get the words looking good.


Here is the stock card for comparison


if anyone knows what font Sony used for these things (and where I can get it!) please let me know and I'll update with the right font. I only chose this one because it is easy to read and prints well.

 

[hatmoose]

The first print of the new bracket came out OK, but as you can see the top corner peeled off the bed a bit - this is an exceptionally long and thin piece. Also the writing is very tiny and fiddly and some of that didnt print the way I wanted it to. I'll give it another go with hairspray on the bed of the 3D printer and see if that helps with 1st layer adhesion.



While I waited for the new bracket to print I started work on another board so that I can test with one while I use the other one "in production"

For this second one I implemented the Low Pass Filter fix as suggested by Martin - this only does anything for 720p. I dont dare hope that I will ever own a PVM/BVM capible of 720p - but if i ever get the chance, I'll be ready

Here is how it looked when done


Part of the fix is to bridge the 5v and GND pins together. The way I did that in this photo was with a T12 BC2 tip - normally I use the flat side on the board for drag soldering SMD pins, but if you flip it upside down and load the flat surface with a (tiny) amount of solder you can touch the two pins at the bend and bridge pins really neatly like in the photo.

 

[hatmoose]

I had never in my life even SEEN an arduino before I started work on this project. For extra challenge I chose to use the cheapest clone available on Ali and the most unreliable-poorly documented programer

General "stuff"
AVR - this is the thing that's going to run our code. The clone AVR that I used is based on the ATMEGA328P - there are about a million of these available, ad far as I can tell they are all pretty much the same
ISP - this is the In System Programer that pushes the code from the PC to the AVR.
It is possible to address the AVR directly by plugging it into a USB port as long as you get one "with the bootloader". But most of the projects that I want to do need to be directly programed with an ISP, so I skipped the whole bootloader thing

So for hardware I had the AVR connected to the ISP through the supplied adaptor, and the ISP attached to the USB port of my PC

In order to address the ISP properly I had to update the WinUSB driver. Updating this driver is harder than just downloading a new version - I used a 3rd party software tool called zadig to do it for me
https://zadig.akeo.ie

The I downloaded the Arduino IDE from here
https://www.arduino.cc/en/software

Martins code is dependent on some external libraries which I downloaded and installed into the Arduino IDE by following the instructions here
https://github.com/NicksonYap/digitalWriteFast

And to get the Arduino IDE to interface with the clone ISP and the clone AVR I added Minicore to the Arduino IDE by following the instructions here
https://github.com/MCUdude/MiniCore

Then I got Martins code for the 129x from here
https://github.com/skumlos/bkm-129x-mcu/

the Arduio IDE is a lot smarter than some other IDE's i've used. when you open a new project file (.ino) for the first time it prompts to set up all the folders and dependencies

Then I was able to burn the 129x code to the AVR using the ISP with the "Upload Using Programer" function in the IDE. These are the settings I used
Board ATmega328
Clock external 8mhz
BOD 2.7v
EPROM retained
Compliner LTO disables
Variant 328P/328PA
Bootloader no
Programer UABasp

For my use case (Sony PVM-9L2) it makes absolutely no differnce if the boot loader is there or not. But for some more modern monitors the time that the bootloader takes to run (approx 1.5 seconds) can cause the monitor to fail it's detect sequence and not find the card.
Martin published some research on it here
https://immerhax.com/?p=480

So with the 720p fix applied, and the bootloader fix done, this project is nearly complete, just need to finish the bracket

 

[bodgit]

For this second one I implemented the Low Pass Filter fix as suggested by Martin - this only does anything for 720p. I dont dare hope that I will ever own a PVM/BVM capible of 720p - but if i ever get the chance, I'll be ready

If you use the Gerbers from Martins Github repository, they should already have this lowpass fix incorporated.

Patiently waiting for my PCBs to arrive, I'm going to build both the full BKM clone with in & out BNC connectors as well as the VGA & SCART combo.

 

[hatmoose]

If you use the Gerbers from Martins Github repository, they should already have this lowpass fix incorporated.

Patiently waiting for my PCBs to arrive, I'm going to build both the full BKM clone with in & out BNC connectors as well as the VGA & SCART combo.

I did my boards in April, before the May commit with the 7374 fix - so I'm on the old version (rev c)

But, but but, and this is a super cool learning opportuntity - I think there is a "bug" in the May commit (rev d)

When I open the kicad in the master the 7374 fix shows correctly (from kicad 3d viewer)

But when I open the gerber in the master the 7374 fix is not correct (from online gerber viewer)

I think he may have forgotten to generate new gerbers from the new kicad when he did the may commit

This is super exciting, I might get to write my first genuine github pull request to fix a bug! I wonder if martin will commit it?

 

[bodgit]

I did my boards in April, before the May commit with the 7374 fix - so I'm on the old version (rev c)

But, but but, and this is a super cool learning opportuntity - I think there is a "bug" in the May commit (rev d)

When I open the kicad in the master the 7374 fix shows correctly (from kicad 3d viewer)

But when I open the gerber in the master the 7374 fix is not correct (from online gerber viewer)

I think he may have forgotten to generate new gerbers from the new kicad when he did the may commit

This is super exciting, I might get to write my first genuine github pull request to fix a bug! I wonder if martin will commit it?

I think you might be right, at least with the BNC version. The Gerbers for the combined SCART & VGA version look fixed to me, but looking at the Github repository for the BNC version, the Gerbers were last modified 11 months ago, so they're older than the KiCad files. I'll have to fix that PCB up when it arrives, although, am I likely to acquire a 720p PVM/BVM that uses one of these cards?

I found a discrepancy for the MIC voltage regulator between the schematics and the BOM on the PCB silkscreen and Martin responded pretty quick, so you should be good.

 

[hatmoose]

Now that the cards work I needed some way to convert SCART to BNC. Of course I could have just built the SCART version of the card, but I have (about 7) other PVM's and BVM's too, so I wanted a more general purpose solution.

Voultar has an awesome SCART to BNC converter that looks like this
https://twitter.com/voultar/status/1032684033656975361?lang=en

It was a bit hard to find the files, Voultar does not Github and all the RetroRGB links were broken, eventually I found them here
https://oshpark.com/shared_projects/JUCQntka

And it was also quite hard to find parts - mouser and digikey and RS can't seem to maintain reliable supply.
Aliexpress equivalents were as follows
Right angle RCA https://www.aliexpress.com/item/1760788561.html
Right angle SCART https://www.aliexpress.com/item/4000674825515.html?
Right angle BNC https://www.aliexpress.com/item/32888289661.html

So I made a few because the minimum PCB order was 5


I've really tried to step up my product photography game on these, without much success


Now I just need some BNC to BNC cables. Plain black would work OK, but I like to think that the blue electrons will flow best along a blue cable... Stay tuned for more pointless attention to detail...

 

[hatmoose]

One of the cool things about this hobby is that things that I expect to be really hard (like SMD soldering) turn out to be really easy
Sometimes things that I think will be really easy (like crimping BNC cables) turn out to be really hard.

I also learned a lot about the history of 50 Ohm BNC and 75 Ohm BNC connectors.
Once upon a time these used to be two different physical connectors.
50 Ohm was mechanically compatible with 75 Ohm, but 75 Ohm was not mechanically compatible with 50 Ohm.
in about 1985 they retained the two standards for cables, but standardised on 50 Ohm for the connectors
So you can still buy 50 and 75 Ohm BNC cables, and 50 and 75 ohm BNC terminators, but they all have 50 Ohm BNC connectors on them.
Unfortunately because Co-ax cable is slightly related to "prosumer" audio there is so much misinformaiton on this out there that I could hardly believe it.
Including gems like this on wikipedia (not technically inaccurate, just totally irrelevant)
https://en.wikipedia.org/wiki/File:BNC_50_75_Ohm.jpg

Things started out well - the "Sony" colors for RGBS are Red, Green, Blue, Black so I found some nice RG59 cable in those colors.


Unfortunately stripping the ends to get them ready to crimp is an incredibly fiddly operaton. And a bad strip always results in a bad cable - it would be almost impossible to do these by hand. It took me nearly 2 hours to learn how to use the stripping tool, including completley disassembling it to adjust the blade height, blade depth, blade gap and blade distance.

This video was amazingly helpful

View: https://www.youtube.com/watch?v=nAXt6Nj99Nw




So after 4 hours I managed to complete one cable that I was completely happy with. On the bench you may be able to see the detritus of 30(?) failed attempts before I got it perfect. This one is actually too short to be of any use, I just wanted to finish one off.


Future cables will be much faster now that I have set up my tools correctly. But this was one of those interesting times when a really easy job turned out to be really hard.

 

[hatmoose]

my key learning in making the cables was that striping the wire to the correct dimensions is the most important thing. I nearly didn’t buy a cable stripping tool but not having one would have made this 10x harder. But once I got the tooling set up I was able to rip out a bunch of cables pretty quickly


I made these ones quite short because I want to attach Voultars awesome SCART to BNC adapter to the size of a 9” PVM, the Sony 9L2 in the picture.


here it is displaying the 240p test suite in glorious RGB at last!


As you can see from the pic the geometry needs a lot of work.

 

[hatmoose]

Well - the whole country is now in what they call "level 4 lockdown" which basically means police in the streets, all schools and businesses shut.

No-one leaves the house unless it is to buy food or seek urgent medical attention. In my part of the country there are police checkpoints. If you go through a checkpoint they will want to see proof of a medical appointment or accompany you to the doctor/supermarket just to make sure.

So the good news is that I have a lot more time for arcade projects

Today I started work on a case for the Voultar RGBS breakout board - this is the hardest design i've worked on because the inclusions (holes) are all at odd heights and spacing and etc. Normally my approach is to make a first draft that fits, then start manipulating the planes and etc to make it look good.

I've set myself a goal of no more than 3 prints from prototype to final.

This is how it looks so far - on the printer now

 

[hatmoose]

While I was waiting for the cases to print I did some experiments with BNC crimping tools

One of these costs TWENTY TIMES as much money as the other one, can you guess which is which?


L-com was a maker of top-quality tooling and dies for the fiber optic industry - they were acquired by infinitE who experimented with "best shoring" in the early 2000's making the HT-330K coax tool in china. The chinese govt promptly stole the IP and chinese factories have been churning out copies ever since.

Strangely they are reluctant to print the words "Patent USA TAIWAN" on the copies, so the fakes are easy to spot. It always puzzled me that they were so particular in this (one) regard.

My fake set came from Aliexpress for ten bucks
My real set came from a retiring electrician who was dumping all his tools, It was lightly used and looks to be 30+ years old (I got some other fun stuff too, will do another post later). If you look carefully you can see his name written on them. There is some monor surface rust but a bath in WD-40 and a toothbrush will sort that right out.


I think most of the cost of the genuine set is not in the tool itself, it is in the dies. The american ones are milled and the tolerances are extreamly tight. The chinese ones are cast and the tolerances are as loose as a goose

Does 20x the cost actually make a difference to the quality of the crimp?

The genuine cable stripper is much easier to adjust and cuts perfectly every time, there is a lot of rework with the clone
But the big difference is in the quality of the crimp that the genuine tool/die can produce

This is a Radial crimp onto Sharkwire 75 ohm RG59 co-ax with cheap crimp tool - mechanically the joint is sound, but it's not what I would call pretty. I can see why people cover them in heat-shrink


This is the same thing but with the genuine tool - it looks like I expect it to look, perfect.


So mechanically the main difference is in the amount of time and rework, and the appearance of the finished product. Is there an electrical difference? I'll do some testing on the oscilloscope to find out