65XX SBC general help and color display help needed

[backspace119]

GARTHWILSON wrote:

backspace119 wrote:

GARTHWILSON wrote:

That's where the MC145406 triple line driver and receiver is nice if you already have ±12V for other things on the board, because you won't need any of those capacitors, since it doesn't need the charge pump.

I may really need to look into getting -12v on the board. I saw your "non-standard power supplies" in the circuit potpourri, and I'm resistant to it only because iirc it uses a lot of discrete components (diodes I believe) that would take up quite a bit of space.

Yeah, it would kind of defeat the purpose. Without re-reading 24 pages of this topic, I was thinking you were using a power supply that already included a negative voltage.

Quote:

Something I asked a while ago and just remembered it never got answered, I have the CTS and RTS connected for the USB-TTL header, through resistors like the Rx and Tx lines, my assumption was this would give me hardware handshaking on it, but I seem to remember that...Daryl's? page didn't use those lines for the usb-ttl header.

Can I use these lines? Will they provide hardware handshaking?

I have one of the FTDI USB-to-RS232 adapters, and I have it set up to do the handshaking on CTS and RTS. The FTDI unit I have is very clumsy softwarewise compared to using a real COM port on my dedicated DOS PC though. I know BDD gave a link somewhere to a product that gives a set of RS-232 ports to modern PCs that came without them, and has the driver software. It might have been startech.com or sealevel.com . It probably needs a desktop PC that you can put cards into, rather than a laptop.

I doubt it would go in my PC, as it has no PCI busses, and I'd be damned if they put that on PCI-E

I'll probably just grab a 6 cable USB to TTL adapter, that has RTS and CTS. hell, I could probably build one if I needed to.

EDIT: forgot to ask the question I had, or maybe I already asked it in my last post, will running the wait states off a variable 1699 work? Also, I'm hoping that the 1699 is well behaved when switching frequencies, I assume so, but that might be a gotcha

[Arlet]

I've never used RTS/CTS in any of my projects. If it's used as a command line interface, then the prompt works as natural flow control. If it's a packet based interface, I make sure I can receive a full packet in a single burst.

[backspace119]

Arlet wrote:

I've never used RTS/CTS in any of my projects. If it's used as a command line interface, then the prompt works as natural flow control. If it's a packet based interface, I make sure I can receive a full packet in a single burst.

I've had some woes with software handshaking before, especially over FTDI, so I'd rather have it and not need it than the other way around

[backspace119]

So here's the newest layout, minus the wait state stuff, which I haven't added in yet, I think it came out a lot cleaner this time, and only 447 vias, vs last time after cleaning it was still 600 vias

and the 3d view to go with it:

[Arlet]

I see you have some slotted holes. Have you checked if the PCB manufacturer can make those, and if there's an extra cost ?

In most cases, you can substitute them with big round holes instead.

[backspace119]

Arlet wrote:

I see you have some slotted holes. Have you checked if the PCB manufacturer can make those, and if there's an extra cost ?

In most cases, you can substitute them with big round holes instead.

The pads are slotted, the holes are round.

It makes it a little bit easier to not burn the mask, while not taking a lot of extra space.

EDIT: OH, I see what you mean, I'm not sure, I don't think the manufacturer will charge extra for it, but we'll see when I get it quoted (it's already dirt cheap)

[Arlet]

backspace119 wrote:

EDIT: OH, I see what you mean, I'm not sure, I don't think the manufacturer will charge extra for it, but we'll see when I get it quoted (it's already dirt cheap)

Also make sure that the manufacturer gets the correct data. If you're sending gerber files, you need to include the slots in the milling or outline layer. Also, they need to understand that you want them plated (I assume you do). Some manufacturers automatically assume you want plated slots if you put them in a pad, but they may not all do that. It can be helpful to add some text in the appropriate layer with an explanation what you want, e.g. "3x plated slot, length * tool diameter".

There are free gerber viewers that you can download to inspect the gerber files before you send them away.

[backspace119]

Arlet wrote:

backspace119 wrote:

EDIT: OH, I see what you mean, I'm not sure, I don't think the manufacturer will charge extra for it, but we'll see when I get it quoted (it's already dirt cheap)

Also make sure that the manufacturer gets the correct data. If you're sending gerber files, you need to include the slots in the milling or outline layer. Also, they need to understand that you want them plated (I assume you do). Some manufacturers automatically assume you want plated slots if you put them in a pad, but they may not all do that. It can be helpful to add some text in the appropriate layer with an explanation what you want, e.g. "3x plated slot, length * tool diameter".

There are free gerber viewers that you can download to inspect the gerber files before you send them away.

FlatCAM should do the trick for me, I use it to create GCode from gerber files for my CNC.

Something I just mentioned in another thread that should probably go here.

I saw ABT and FCT devices mentioned in the big thread on 816 timing. I looked them up and they both look pretty fast, the ABT seems to be a bit faster and cheaper. Is there a catch with these devices? If not, why aren't they recommended in places like the primer? (I should mention, that thread seems to suggest they only come in SMD packages, which is not the case, here's a link to a DIP 573

[GARTHWILSON]

backspace119 wrote:

I saw ABT and FCT devices mentioned in the big thread on 816 timing. I looked them up and they both look pretty fast, the ABT seems to be a bit faster and cheaper. Is there a catch with these devices? If not, why aren't they recommended in places like the primer?

The primer is intended to be just that—a primer—so it's kept rather simple.  The faster logic families are more likely to spell trouble for first-time builders.  I would like to add more info about the various 74__xx logic families (probably in a separate article); but IIRC, ABT cannot pull up high enough to guarantee a valid CMOS logic '1'.

[backspace119]

GARTHWILSON wrote:

backspace119 wrote:

I saw ABT and FCT devices mentioned in the big thread on 816 timing. I looked them up and they both look pretty fast, the ABT seems to be a bit faster and cheaper. Is there a catch with these devices? If not, why aren't they recommended in places like the primer?

The primer is intended to be just that—a primer—so it's kept rather simple. The faster logic families are more likely to spell trouble for first-time builders. I would like to add more info about the various 74__xx logic families (probably in a separate article); but IIRC, ABT cannot pull up high enough to guarantee a valid CMOS logic '1'.

Ah, that's probably the catch with it, I assume that's probably why the "T" is tacked onto the name, for TTL levels like the HCT. I'll check out the data sheet and report back what it's logic 1 output is.

[backspace119]

It seems the ABT may be targetted at reducing ground bounce btw, as it's mentioned a lot in the docs.

The output 1 is 3v @5v with current loading of -3ma. 2.5v @ 4.5v with -3 ma, and 2v @ 4.5v with -18ma or above of current loading.

EDIT: datasheetof the 573 I'm looking at

[backspace119]

On an unrelated note, I've added in the wait state generator circuit, it's expanded the board a bit because I didn't want to put it in some of the areas that are routed that may have kept the board size the same. The good thing about this is I can use the extra board space for mounting holes, which I need to start considering. (it's only an extra 5-10mm, so only minimally more cost).

Here's the layout:

and the 3D view:

[backspace119]

A question I have about my layout:

You'll notice that the resistor+capacitor in line for RDY are oriented differently. The resistor stands off the board while the capacitor lays on it. I laid the capacitor down so the current flow would be relatively parallel, and I'm considering doing the same for the resistor, is this necessary? Or should I just make them both vertical?

[GARTHWILSON]

backspace119 wrote:

You'll notice that the resistor+capacitor in line for RDY are oriented differently. The resistor stands off the board while the capacitor lays on it. I laid the capacitor down so the current flow would be relatively parallel, and I'm considering doing the same for the resistor, is this necessary? Or should I just make them both vertical?

The resistor's resistance will dwarf the small added inductance you get from standing it up; so don't worry about it.

Quote:

Ah, that's probably the catch with it, I assume that's probably why the "T" is tacked onto the name, for TTL levels like the HCT. I'll check out the data sheet and report back what it's logic 1 output is.

The "T" normally refers to the input thresholds, not the output.  So for example 2.5V at the input of a 74HCTxx part will be in solid '1' territory, whereas for 74HCxx, it would be undefined.  WDC's processors' data sheets say they need 70% of Vcc to get a valid '1', which at 5V is 3.5V.  Some here have challenged that assertion, saying their observation is that it can handle lower-voltage 1's ok.  I have not tested it myself, and even if I found it to be true in the unit(s) I test, I might not want to trust that all their units will always perform that way.

[backspace119]

GARTHWILSON wrote:

backspace119 wrote:

You'll notice that the resistor+capacitor in line for RDY are oriented differently. The resistor stands off the board while the capacitor lays on it. I laid the capacitor down so the current flow would be relatively parallel, and I'm considering doing the same for the resistor, is this necessary? Or should I just make them both vertical?

The resistor's resistance will dwarf the small added inductance you get from standing it up; so don't worry about it.

Quote:

Ah, that's probably the catch with it, I assume that's probably why the "T" is tacked onto the name, for TTL levels like the HCT. I'll check out the data sheet and report back what it's logic 1 output is.

The "T" normally refers to the input thresholds, not the output. So for example 2.5V at the input of a 74HCTxx part will be in solid '1' territory, whereas for 74HCxx, it would be undefined. WDC's processors' data sheets say they need 70% of Vcc to get a valid '1', which at 5V is 3.5V. Some here have challenged that assertion, saying their observation is that it can handle lower-voltage 1's ok. I have not tested it myself, and even if I found it to be true in the unit(s) I test, I might not want to trust that all their units will always perform that way.

Ok, so then the ABT series may not work...I may get a set of AHC and ABT and try it out, to see at least if the one I get can handle it.

I've been looking into cplds today and I'm amazed by how simple it is to design and burn one. This must lend to their popularity here. I'm hoping to fine one that's pin compatible with the one used for the 65SPI so I can design my own replacement as a pet project, and to keep the chip available for people in the future. I'm not sure if a pin compatible cpld exists though, because none of the ones I've looked at so far are.