• Please review our updated Terms and Rules here

8-Bit IDE Controller

Three to four weeks is a serious commitment. I'm hoping this machine does a little better than that. Otherwise, I might draft a machine at work into use - I don't have to worry about people hitting the power switch on those machines.

I hope we don't have any late revisions - starting over would be painful.
 
lynchaj said:
In other words IO=300h, ROM=D0000 would be

SW1

1 2 3 4 5 6 7 8
0 1 1 1 0 1 0 0

which is the exact opposite of what I have on the silk screen table. Argh.

/snip

It appears the table needs some rework. Would adding the a not to the table "0=ON, 1=OFF" be too confusing? Probably.
Click to expand...

Yeah, saying 0 = on and 1 = off makes it very difficult to set it properly. Can't you just flip all the bits in the table and leave 0 = off and 1 = on?

The silk screen edits look awesome btw :D
 
Hi! Yes, I will remake the table and hopefully clarify it a bit more. I'll give that a try tonight assuming I get home from work at a reasonable hour. Things at work are ramping up again so I am less available than usual.

Please, as many people as possible, download the schematic and PCB layout and review it as in depth as you can. Look for what you are interested in and ask the questions even if we have discussed it before. We need to know now and I'd rather find the problems *before* we go to PCB production than after. Especially those with electronics and design experience!

Depending on how the second round of testing goes this one design could be the one so I wouldn't save your concerns for the next round because there may not be one. Seriously, constructive criticism is not going to bother me and if you find an issue prior to spending $$$ on a PCB order will make my day.

Thanks in advance!

Andrew Lynch
 
Andrew,

I've had a solid week of runtime. I started with ripup at 100, went down to 1, went up to 100,000 and now I'm back down to 10. It has not removed any vias or changed the overall trace length in the last 4 or 5 runs.

I've got it posted at http://www.brutman.com/XT-IDE_v2_optimized.dsn . Can you download it and do a sanity check?

Should I keep going, or is it done? I know you said go through two complete cycles, but I don't think it is going to make any progress.


Mike
 
Hi Mike,
The optimizer trails off towards the end but the reason for two full cycles is to allow it to shorten the overall trace length as much as possible and also improve routing "orthogonality". The autorouter/optimizer is weighted to put all horizontal traces on component side and vertical traces on copper side. It will do some juggling back and forth to improve the balance during the final cycle.

In the past I've noticed that during these final passes occasionally the autorouter/optimizer will make a breakthrough and eliminate a bunch of vias and make major changes to the layout in the process. Then I'll restart the final cycles again due to the resulting disruption of the PCB layout.

During the last couple of weeks things have been very busy at work so I am way behind on everything at home. I'll check out the latest PCB routing. There are design checks in KiCAD which will verify that all the pads are connected and clearances met.

Thanks and have a nice day!

Andrew Lynch
 
Tandy support!!

Tandy support!!

I see there have been some recent postings about problems with Tandy 1000 series computers. I have a working 1000SX machine and while I can't do the kind of programming/logic analyzing you guys need I know my way around computers and hardware and I can use debug to dump code if someone points the way. I had previously posted, long ago, my interest in this project as one of my old mfm/rll drives bit the dust. My primary drive is still ok but a replacement drive for the dead one turns out to be pretty expensive. Moving to IDE drives would make maintenance a lot easier.

Off-topic: Since there seems to be a lot of hardware techs on this board I'll ask. I need to find a replacement 3.5inch 720K floppy drive for my 1000SX (1.44 drives aren't supported, even as a 720K drive). True 720K are scarce and big bucks however my research shows some models of 1.44mb drives can be "hacked" or jumpered into 720K mode for just this purpose. Can anyone suggest where I might get such a drive or even which drive to look for?

Greg Lentz
 
I'm getting ready to ship a 1000 to hargle. It is boxed, and should be at the post office Monday. It took me a while to get everything together.

As for a 720k drive in an SX, I just pulled a 1.44 meg drive off the shelf and made sure to cover the media sense hole in the diskette. Worked perfectly (for me).
 
Greg,

Please start a new thread for that question - we've over 90 pages in this thread without trying to go off topic. :)


Mike
 
Hi! I am finally catching up. The good news is I reviewed Mike's DSN file. It is OK. I had to make a couple of tweaks on the trace layout on the edge connector and there is a common radius error on one of the connector pins when I imported the files back into KiCAD. That's fixed.

I updated the copper side silk screen with the correct (I hope) switch settings. 0=off, 1=on like you'd expect. I do not trust that switch settings table and every time I think about it I get confused. The table needs your extra detailed scrutiny.

Also ran the DFM checks and DRC tests and they all are coming back clean.

Here are some files on DFM. Please download the PDFs from the N8VEM wiki XT-IDE folder. That's all for now.

PLEASE, PLEASE, PLEASE, do a thorough review and bring out all your issues and concerns. Even if its a re-hash lets flush out all the problems before this goes to PCB manufacturing.

Once there is hardware on the dock, its TOO LATE!

Thanks and have a nice day!

Andrew Lynch

Click to view Quote for design XT-IDE-full-02.zip: https://www.my4pcb.com/Net35/freedfm/0011697402352279/quote.aspx

New! Click to view PLOTS (beta release) for design XT-IDE-full-02.zip: https://www.freedfm.com/freedfm/0011697402352279/results/plots.htm

Click to view DFM results for design XT-IDE-full-02.zip: https://www.freedfm.com/freedfm/0011697402352279/results/summary2.htm
 
lynchaj said:
Hi! I am finally catching up. The good news is I reviewed Mike's DSN file. It is OK. I had to make a couple of tweaks on the trace layout on the edge connector and there is a common radius error on one of the connector pins when I imported the files back into KiCAD. That's fixed.

I updated the copper side silk screen with the correct (I hope) switch settings. 0=off, 1=on like you'd expect. I do not trust that switch settings table and every time I think about it I get confused. The table needs your extra detailed scrutiny.

Also ran the DFM checks and DRC tests and they all are coming back clean.

Here are some files on DFM. Please download the PDFs from the N8VEM wiki XT-IDE folder. That's all for now.

PLEASE, PLEASE, PLEASE, do a thorough review and bring out all your issues and concerns. Even if its a re-hash lets flush out all the problems before this goes to PCB manufacturing.

Once there is hardware on the dock, its TOO LATE!

Thanks and have a nice day!

Andrew Lynch

Click to view Quote for design XT-IDE-full-02.zip: https://www.my4pcb.com/Net35/freedfm/0011697402352279/quote.aspx

New! Click to view PLOTS (beta release) for design XT-IDE-full-02.zip: https://www.freedfm.com/freedfm/0011697402352279/results/plots.htm

Click to view DFM results for design XT-IDE-full-02.zip: https://www.freedfm.com/freedfm/0011697402352279/results/summary2.htm
Click to expand...
There is an error: 8765 should be 5678, and the same with 4321; it should be 1234. The values are correct.

After this is fixed, the tables could also be swaped, so the I/O Range is on the left (or rigth as it will appear in the inverted version from the PDF) and ROM-address on the rigth (or left as it will appear in the inverted version from the PDF).
 
those web based layouts look fantastic. I like how simple it is to view the stuff too.

If I'm understanding that quote properly, it looks to be about $150-200 more expensive than the stuff we did in china. (for 10 parts)
I'll do it if you think it's a requirement that we use this fab house, but I'd like to check with our friends overseas too. I think they did a really good job on the last run.

------

I've got a tandy HX en route to me right now, should be here in a day or two.
There may be some things that need to be changed depending on how this machine works, so we can't quite wrap it up just yet.
 
Just a warning ... the trace optimization is still running. I'm still going through the last cycle, so don't do anything more until it's done and I give Andrew the last version.

(Andrew jumped the gun .. he convinced me to keep going, then started working with the partial results.)


Mike
 
hargle said:
those web based layouts look fantastic. I like how simple it is to view the stuff too.

If I'm understanding that quote properly, it looks to be about $150-200 more expensive than the stuff we did in china. (for 10 parts)
I'll do it if you think it's a requirement that we use this fab house, but I'd like to check with our friends overseas too. I think they did a really good job on the last run.

------

I've got a tandy HX en route to me right now, should be here in a day or two.
There may be some things that need to be changed depending on how this machine works, so we can't quite wrap it up just yet.
Click to expand...



Hi! Yes, I would recommend 4PCB if there are features only they can provide but the manufacturer used last time should be just fine. There is nothing fancy about this design except for maybe the gold fingers on the card edge connector. The rest is bog standard PCB stuff. I don't think 4PCB is necessary here but I do get quotes from them since the free DFM service is convenient and helpful. Their prices are pretty high though.

The drawings I posted are in "near final" condition with the exception of the updated trace layout for the PCB from Mike as soon as it is available. Please download the drawings and verify as thoroughly as possible.

Thanks and have a nice day!

Andrew Lynch
 
And two more vias just got removed in the past hour. We're down to 53. :)
 
per said:
There is an error: 8765 should be 5678, and the same with 4321; it should be 1234. The values are correct.

After this is fixed, the tables could also be swaped, so the I/O Range is on the left (or rigth as it will appear in the inverted version from the PDF) and ROM-address on the rigth (or left as it will appear in the inverted version from the PDF).
Click to expand...

Hi Per! Thanks! You are right and the table was completely flipped!

I changed it all around and reposted the PDF on the N8VEM wiki.

Thanks! I appreciate your great catch! Have a nice day!

Andrew Lynch
 
lynchaj said:
Hi Per! Thanks! You are right and the table was completely flipped!

I changed it all around and reposted the PDF on the N8VEM wiki.

Thanks! I appreciate your great catch! Have a nice day!

Andrew Lynch
Click to expand...

Now, the tables are fine, but isn't it better if the whole thing is centered on the design?
 
per said:
Now, the tables are fine, but isn't it better if the whole thing is centered on the design?
Click to expand...

Hi Per! Thanks! There has to be some space reserved for the installation of brackets which pushes everything to the right on the front of the board (left on the back) to reserve room for the drill holes and and machine screws, etc. I can move the tables to the left on the back a bit but not much.

Thanks and have a nice day!

Andrew Lynch
 
lynchaj said:
Hi Per! Thanks! There has to be some space reserved for the installation of brackets which pushes everything to the right on the front of the board (left on the back) to reserve room for the drill holes and and machine screws, etc. I can move the tables to the left on the back a bit but not much.

Thanks and have a nice day!

Andrew Lynch
Click to expand...

That's problably fine. I just don't like the text being that close to the edge of the PCB.
 
You must log in or register to reply here.
Back
Top