Page 8 of 15
Re: An Improved MENSCH™ Microcomputer
Posted: Sun Sep 14, 2025 3:38 pm
by plasmo
JLCPCB is the one I used most often. I also have used Seeed Studio. Both are excellent shops.
Since this is your first design, you may want to add a pair of ground near center of the board. It serves 2 purpose, instrumentation ground for your scope, and ground points to solder additional signal returns if needed since this is 2-layer board with fast components. You may also consider not having solder mask over the vias because it is easier to probe or solder wires to it if you have to cut and patch. Good luck!
Bill
PS, I would recommend you using the cheapest shipping option, but I also know the sufferings waiting for the slow boat from China. So $25 for shipping is cheap compared to to all the anti-anxiety drugs you’ll need!
Re: An Improved MENSCH™ Microcomputer
Posted: Sun Sep 14, 2025 3:45 pm
by GARTHWILSON
On to the next step. Where and how and where do I get this fabricated?
See our topic, "Cheap PCB stories..." (although it could stand to be updated), and the relevant page of the 6502 primer, "What About Custom PC Boards?"
Re: An Improved MENSCH™ Microcomputer
Posted: Sun Sep 14, 2025 4:57 pm
by BigDumbDinosaur
On to the next step. Where and how and where do I get this fabricated?
I recommend JLCPCB as well. They are reasonable in cost and make a good product. Most orders seem to be turned around fairly quickly.
Re: An Improved MENSCH™ Microcomputer
Posted: Sun Sep 14, 2025 5:58 pm
by Martin_H
I generated the gerber and drill files and uploaded them to PCB Way. It failed review with the following error message:
Code: Select all
Failed reason:
Lack of the Soldermask layer.;
Lack of the Keepout layer (Board Outline), please add it in GKO/GM1/OUTLINE Layer.;
How do I generate those?
Re: An Improved MENSCH™ Microcomputer
Posted: Sun Sep 14, 2025 6:13 pm
by gilhad
Did you marked them in File / Fabrication Outputs / Gerber ? (or there, what you used for export)
Re: An Improved MENSCH™ Microcomputer
Posted: Sun Sep 14, 2025 7:05 pm
by Martin_H
The problem is I am not sure of the linkage between checkboxes on that dialog and what PCB Way needs. I found a page on the PCB Way website showing the export dialog with what they want checked, re-exported, and re-submitted. But it still gives me an error "Lack of the Keepout layer (Board Outline), please add it in GKO/GM1/OUTLINE Layer.;". Looking at the dialog it's not obvious what adds that.
Update: I think it's the edge cuts layer they want. So I will try again.
Update to the update: It was edge cuts and I was able to submit the order. But the shipping and fees added up to $49.00. I hope it works.
Re: An Improved MENSCH™ Microcomputer
Posted: Mon Sep 15, 2025 1:54 am
by Martin_H
@plasmo, I was wondering if test points were something I should be concerned about. The via options were also a puzzle when I ordered the PCB and I went with the default which are covered by the solder mask.
My debug plan is as follows:
Code: Select all
Build the bare minimum for the PCB to work. Basically, the FTDI header, DS1813, pullup resistors, decoupling caps, crystals, loading caps, and the w65c265. At that point it's only a bit more than a breakout board and is a simpler problem to debug.
Before any power is applied verify connectivity of the w65c265 ground and power pin pairs, with no shorts to anyplace unexpected. The larger manual solder pins should help with this task, while the fine pitch of the QFP will work against me.
Apply regulated power through the FTDI header and test if FCLKOB and CLKOB oscillating at the expected frequencies.
If it works, I can then attach the serial out to my scope and see if the w65c265 is announcing itself on the serial line. I have a w65c265qbx so I know what it should look like.
Move on to an FTDI to PC serial link to see if I can use the onboard ROM monitor.
Finally build out the rest of the PCB in stages: additional RAM, the EEPROM, and finally the 65c22. The chip select LED's will help debug, as will the ROM monitor.
@all, I didn't see the additional recommendations for JLCPCB before I began the order from PCBWay. Perhaps next time. In any case, I will let people know how fast they make and ship them, and my impression of their quality.
I'm compiling a spreadsheet of links to order the components from Mouser. Unfortunately, the w65c265 is currently out of stock! They say they have 13 on order but no ETA. I should probably have checked that before ordering the PCB. Here's a link to the BOM in progress.
https://docs.google.com/spreadsheets/d/ ... sp=sharing
Most of the parts are easy to find on their website, but the crystals are confusing me. Does the load capacitance need to match the loading capacitors on the PCB? Take the 32.768 KHz crystal for example. The loading capacitance on the PCB is 27 and 47 pF, but that's not an option on the website.
Re: An Improved MENSCH™ Microcomputer
Posted: Mon Sep 15, 2025 2:19 am
by plasmo
A bench power supply with current readout and adjustable voltage is ideal. Absence of that you should have wall wart with regulated 5V. Power it with FTDI header is really the last resort since 5V from USB is not well regulated, sometimes only 4.0-4.5V and you can only draw limited current.
Ground test points are useful to ground your scope to minimize ringing and provide more accurate readings. Power the board up with just W65C265 and support parts is a good idea. You should not have ground bounce issue at this stage. When other devices are added, you potentially may have ground bounce causing intermittent failures. This is where a wire or two from the ground of other devices to the central ground points may fix the intermittent failures. This is just in case; chances are you won't have any problem with ground. Access to unplated vias are helpful for cuts and patches, especially when fine pitch SMT devices are involved. Then again, with so many eyes on your design, you are likely not have any problems.
Good luck!
Bill
Re: An Improved MENSCH™ Microcomputer
Posted: Wed Sep 17, 2025 12:28 am
by Martin_H
My new soldering iron tips and surface mount practice kit arrived today. The tips look good, and the practice kit looks like helpful. Its instructions are in Chinese, and nothing has markings on it, so it should be an interesting puzzle too. The PCB is divided into functional and nonfunctional circuits. The functional circuit uses 0805 surface mount resistors, LED's, and a SOIC IC for blinkenlights. The nonfunctional has two unmarked QFP-44, SOIC-14, and progressively smaller SMD resistors. I think the idea is you keep going smaller until you reach your limit, or the atomic scale.
The star of the show are the unmarked QFP-44's, as that's what I need to practice on. For the really small stuff I have no idea how you would remove the resistors from their package, let alone solder them. It only cost $8.00, so if I mess it up, I order another until I can confidently solder them. My illuminated magnifying desk lamp should earn its keep.
The PCB's shipped today. I went through my parts inventory, and I have all the headers, jacks, electrolytic capacitors, IC sockets, DS1813, 1N5817, reset button 128kB RAM, EEPROM, and a 32.768 kHz crystal. But the W65C265S8QG is still out of stock. I plan to place my order and hope to get in line.
Re: An Improved MENSCH™ Microcomputer
Posted: Wed Sep 17, 2025 1:33 am
by gilhad
I use tweezers like this
https://www.aliexpress.com/item/1005007589440904.html or this
https://www.aliexpress.com/item/1005007427140451.html or similar to manipulate SMD resistors/LEDs.
I take the SMD package (paper line with holes, covered by plastic) into left hand, tweezers into right, catch the edge of plastic by tweezers and bend/slide it toward left hand, until one hole with SMD is opened, then I rotate the package near the desk (ideally on the desk) so the SMD fall out (and not lost, so the shorter fall path, the better). If it does not want to came out, I help it with the tweezers, usually I force the tweezers gently from back throught the package, pierce it and force the SMD out. Then I put the package safely away, use the tweezers to orientate the SMD right (if it is upside down, I take it by tweezers rotate and lay back, or let fall from like 5 mm in hope it faal upside up), take it by tweezers, place on PCB and while still holding it I use left hand to solder it in place from one side, then a open tweezers and put them away and solder it from other size.
For SMD ICs I do it basically the same, just solder one pin first, then diagonally opposite pin, then the rest of pins.
There is like million tutorials on youtube if you search soldering SMD, like this
https://www.youtube.com/watch?v=GgDqgixKTpU
( and I simply love this one
https://www.youtube.com/watch?v=i5MNLTc7YhY 
)
Re: An Improved MENSCH™ Microcomputer
Posted: Wed Sep 17, 2025 2:42 am
by Martin_H
Thanks for the links. I have a pair of precision cross tweezers, they are normally closed, and you squeeze to open them. Hopefully they'll work for this.
The last video was madness, but in a good way. I have also gone down the SMD soldering tutorial rabbit hole. I also found out the digital microscopes are surprisingly cheap.
Re: An Improved MENSCH™ Microcomputer
Posted: Wed Sep 17, 2025 5:18 am
by plasmo
It is easier to prototype SMT by flipping the chip over like a “dead bug”. The ground leads can bend down to solder to ground foil and VCC leads can bend up to connect to power ring. The signal leads can bend at different angles to have more clearance. Photo shows a salvaged 68302 wired to DIP RAM and ROM. So it is quite possible to prototype QFP100 W65C265, but you do need more soldering practice.
Bill
Re: An Improved MENSCH™ Microcomputer
Posted: Thu Sep 18, 2025 1:53 am
by Martin_H
I spent forty minutes tonight with the SMD soldering practice kit. I stopped because of fatigue, but here's my overall experience:
My 2x lighted magnifying desk lamp is more than adequate for 0805 parts.
I used too much flux for the first few resistors. A little goes a long way.
The 0805 resistors are straightforward but a bit fatiguing. Slightly harder than through hole resistors.
The SOIC-14 was easier than a DIP-14 because I didn't have to flip the board over, or putting an IC into a socket later. I have bent pins a few times doing that.
The QFP-44 was surprisingly easy to do. I put flux on the pad and tacked one corner down, nudged it a bit to square it up, and tacked an opposite corner down. I started doing a pin at a time, got bored and drag soldered it, after a cleanup with desoldering braid it looked respectable. I examined it and didn't see any shorted pins.
A pad pulled off while doing a resistor. I may have overheated it, or it may be a cheap PCB.
Re: An Improved MENSCH™ Microcomputer
Posted: Thu Sep 18, 2025 11:21 am
by gilhad
Congratulation

Now you are ready to go

Re: An Improved MENSCH™ Microcomputer
Posted: Mon Sep 22, 2025 4:39 pm
by Martin_H
DHL just dropped off a package from PCBWay. It would have come on Friday, but I was out of town, and delayed it one business day. Overall speedy turnaround for something from halfway around the planet.
The PCBs are well packed, lots of padding, and the PCBs were vacuum packed in bubble wrap. The latter being tricky to open.
The PCB material is thick and everything looks good. Seeing it in real life makes me wish I put a bit more space between some of the SMD components. But I just need to be careful.
The Mouser order arrives on Wednesday with everything but the w65c265. I can assemble some of the board then, but finishing it will require the backorder to ship.
Above them is the SMB practice kit, which I haven't cleaned off the flux yet.