The Lynx FDC Project and issues

Avatar de Usuario
Womble
Mensajes: 75
Registrado: 19 Nov 2020 18:59
Agradecido : 15 veces
Agradecimiento recibido: 77 veces

Re: The Lynx FDC Project and issues

Mensajepor Womble » 28 Nov 2020 01:16

Último mensaje de la página anterior:

I use Eagle, but only because the engraver is set up for it. I should start learning one of the new open source ones like KiCad.
I used to use EdWin which was pretty nice for the time in the mid 90s.

There are a couple of different schematics for the controller board I've seen, one is French and uses a 1770, whether this means a different ROM in the addon module I don't know but the schematic is very simple when compared to the original.

Of course I have no way of knowing whether these interfaces work, I do not know the author of these designs.

http://www.retrogubbins.co.uk/downloads ... c-french-1

http://www.retrogubbins.co.uk/downloads ... c-french-2

oracle_jedi
Mensajes: 34
Registrado: 09 Nov 2020 23:23
Agradecido : 13 veces
Agradecimiento recibido: 60 veces

Re: The Lynx FDC Project and issues

Mensajepor oracle_jedi » 28 Nov 2020 02:30

I am not an electrical engineer. I have absolutely no clue what I am doing. I just want to be clear on that.

I had a conversation with Frank a couple of years back about the FDC. Frank sent me a Gerber file that DipTrace could not read, and neither could OshPark. He also sent me a JPG picture of the top side of his (re)design for the FDC. Frank was super helpful, but for a number of reasons, we never seemed to quite get to usable PCB design file that I could send for fabrication.

Hence during this lockdown I decided to have a go, using my real FDC as a template, Frank's JPEG, whatever I could scrounge up from the web and some help from a great friend on the Atari forums who is an amazing PCB design engineer.

I checked Frank's JPEG and it also shows a connection from SKT1:15 to IC2:10. The real FDC does not have this link.

The original FDC has a connection from IC2:10 to SK1:14 and IC1:5 (A0).

Frank's design also seems to show IC15:7 as NC. On my FDC it is GND.

Frank's design seems to show IC15:14 as NC. On my FDC is it +5V. (Probably irrelevant as IC15:13 is also +5V).

Frank's design seems to show IC16:15 as connecting to C3(-) and GND. On my FDC it only connects to C3(-)

Frank's design seems to show IC16:1 as GND. On my FDC it is NC.

I don't know if I am reading the JPEG wrong. It's more than likely. Its also possible these changes are improvements that help. I don't know. Three boards from OshPark (minimum order) is $100 with shipping. I'd really like to get the design right before ordering them up. My LynxDOS attempt was an epic fail although much cheaper as the boards were vastly simpler.

I've attached the JPEG Frank shared with me so you can see what I have been working off.

disk_Controller - L1.jpg
disk_Controller - L1.jpg (340.45 KiB) Visto 228 veces


I've also attached the current state of my attempt to recreate it. Since it has more than 300 drill holes you will need the Hobbyist version of DipTrace or better. The free download version tops out at 300 drill holes.
Adjuntos
Lynx Disk Controller 8A [spaced wire].zip
(31.61 KiB) Descargado 3 veces

Avatar de Usuario
Womble
Mensajes: 75
Registrado: 19 Nov 2020 18:59
Agradecido : 15 veces
Agradecimiento recibido: 77 veces

Re: The Lynx FDC Project and issues

Mensajepor Womble » 28 Nov 2020 04:05

>>>>Frank's design also seems to show IC15:7 as NC. On my FDC it is GND.

That's' one of the power pins, def wont work without that being grounded

>>>Frank's design seems to show IC15:14 as NC. On my FDC is it +5V. (Probably irrelevant as IC15:13 is also +5V).

That's the other power supply pin :) so yes very relevant.

So IC15 has no power connections at all is one problem.

>>>>>Frank's design seems to show IC16:15 as connecting to C3(-) and GND. On my FDC it only connects to C3(-)

fdcdetail.jpg
fdcdetail.jpg (54.59 KiB) Visto 221 veces


As shown here pin 15 should connect to C2. It looks to me as though it does connect there in Franks layout but also it (looks) like it connects to the ground via the decoupling capacitor to the left of IC16. Can't really tell without seeing each separate layer.

>>>>>Frank's design seems to show IC16:1 as GND. On my FDC it is NC.

From the datasheet its supposed to connect to one side of the timing capacitor (of the second monostable in the IC).

And indeed on a NOS FDC controller PCB it is indeed NC as you say.

Its a bit difficult to read but those numbers on IC16 don't look right - pins 2,3,5 are for the monostable outputting on pins 5 & 7
however the pin used for output as shown in the schematic detail above is (hard to see) but looks like 10, the other trigger pins for that mono are on 12,13,14 .

So as it stands I'm a bit confused by what's shown, one half of IC16 isn't used, but the schema shows the wrong pins being used.

I just checked each pin on the schematic against a fresh FDC PCB. The schematic shown above is almost right... or the board is definitely wrong...

Pins 3,4,5 on the schema above are shown tied to +5 and as this is a cmos chip that is indeed where inputs should be tied (or grounded, depends on the chip) and not left floating.

In the actual PCB these are left NC which is frowned on so I suspect the actual layout of the original board didn't follow the schematic.

Avatar de Usuario
Womble
Mensajes: 75
Registrado: 19 Nov 2020 18:59
Agradecido : 15 veces
Agradecimiento recibido: 77 veces

Re: The Lynx FDC Project and issues

Mensajepor Womble » 28 Nov 2020 04:16

Its very difficult to make out from the JPG (I downloaded Diptrace but balked at a 2 gig install), is there any chance you could post a JPG of each layer?

(( EDIT - I realised you only have the JPG bah humbug)

From your JPG above it looks like 3,4,5 are tied together (top white trace) and to pin 11 (which is only supposed to be connected to pin 5), there's also an output from pin 12 going somewhere. None of this looks like the original schematic or layout.

oracle_jedi
Mensajes: 34
Registrado: 09 Nov 2020 23:23
Agradecido : 13 veces
Agradecimiento recibido: 60 veces

Re: The Lynx FDC Project and issues

Mensajepor oracle_jedi » 28 Nov 2020 05:37

From your JPG above it looks like 3,4,5 are tied together (top white trace) and to pin 11 (which is only supposed to be connected to pin 5), there's also an output from pin 12 going somewhere. None of this looks like the original schematic or layout.

Right, I only have a JPG of this. But checking against my real FDC, pins 3, 4 and 5 are tied together on the foil side of the PCB.

fdc_ic16_bottom.jpg
fdc_ic16_bottom.jpg (131.75 KiB) Visto 215 veces


Pin 5 is connected to Pin 11 and Pin 3 is connected to Pin 16 on the component side, so I think this is right, at least as the actual PCB has it.

fdc_ic16_top.jpg
fdc_ic16_top.jpg (171.76 KiB) Visto 215 veces


Pin 12 goes to IC14:11 and also IC8:11 - again this seems to match a real FDC PCB.

jltursan
Mensajes: 3290
Registrado: 20 Sep 2011 13:59
Agradecido : 313 veces
Agradecimiento recibido: 868 veces

Re: The Lynx FDC Project and issues

Mensajepor jltursan » 28 Nov 2020 10:05

So seems the DipTrace schematic is wrong or at least quite different from the original schematic or even the real PCB :-(. The schematic I have seems to be the same as the JPEG posted by oracle_jedi.

So IC15 has no power connections at all is one problem.

Tested over my assembled board, there're no GND or VCC in IC15!! -banghead

Well, as I don't have this last, I can't compare;...but I can export the DipTrace schematics to Eagle format. I have checked the result and seems that DipTrace has done a really good job.

Lynx DOS FDC.zip
(24.13 KiB) Descargado 5 veces

Three boards from OshPark (minimum order) is $100 with shipping. I'd really like to get the design right before ordering them up.

About OshPark, forget it and try JLPCB, you won't regret :-)

Avatar de Usuario
Womble
Mensajes: 75
Registrado: 19 Nov 2020 18:59
Agradecido : 15 veces
Agradecimiento recibido: 77 veces

Re: The Lynx FDC Project and issues

Mensajepor Womble » 28 Nov 2020 14:30

Ooops I am a fool, I didn't see those bottom linking traces on 3,4,5

Here are pictures of an unpopulated PCB

FDCtopsmall.png
FDCtopsmall.png (1.03 MiB) Visto 192 veces


FDCbottomSmall.png
FDCbottomSmall.png (1.03 MiB) Visto 192 veces



larger here:

http://www.retrogubbins.co.uk/downloads ... b-pictures

Avatar de Usuario
Womble
Mensajes: 75
Registrado: 19 Nov 2020 18:59
Agradecido : 15 veces
Agradecimiento recibido: 77 veces

Re: The Lynx FDC Project and issues

Mensajepor Womble » 28 Nov 2020 14:32

Yes JLPCB has done well for me in the past too.

jltursan
Mensajes: 3290
Registrado: 20 Sep 2011 13:59
Agradecido : 313 veces
Agradecimiento recibido: 868 veces

Re: The Lynx FDC Project and issues

Mensajepor jltursan » 28 Nov 2020 15:17

I've fixed the following in my PCB with some cutting and wirewrapping:

I checked Frank's JPEG and it also shows a connection from SKT1:15 to IC2:10. The real FDC does not have this link.

The original FDC has a connection from IC2:10 to SK1:14 and IC1:5 (A0).

Frank's design also seems to show IC15:7 as NC. On my FDC it is GND.

Frank's design seems to show IC15:14 as NC. On my FDC is it +5V. (Probably irrelevant as IC15:13 is also +5V).

I've also added a 22nF cap to IC15

Frank's design seems to show IC16:15 as connecting to C3(-) and GND. On my FDC it only connects to C3(-)

Frank's design seems to show IC16:1 as GND. On my FDC it is NC.

This seems mostly correct, pins 1,8 and 15 must be GND; but I suppose one must be left unconnected.

I'll keep checking and soon I'll try again the controller. Hope my Lynx don't get fried... :roll:

Avatar de Usuario
Womble
Mensajes: 75
Registrado: 19 Nov 2020 18:59
Agradecido : 15 veces
Agradecimiento recibido: 77 veces

Re: The Lynx FDC Project and issues

Mensajepor Womble » 28 Nov 2020 15:39

>>>>I checked Frank's JPEG and it also shows a connection from SKT1:15 to IC2:10. The real FDC does not have this link.

correct, this looks like an error

>>>>The original FDC has a connection from IC2:10 to SK1:14 and IC1:5 (A0).

correct

>>>>Frank's design also seems to show IC15:7 as NC. On my FDC it is GND.

it has to be ground

>>>>Frank's design seems to show IC15:14 as NC. On my FDC is it +5V. (Probably irrelevant as IC15:13 is also +5V).

pins 14 and 13 are both +5 correct

>>>Frank's design seems to show IC16:15 as connecting to C3(-) and GND. On my FDC it only connects to C3(-)

just double checked here on the real PCB, , IC16 pin 15 ONLY goes to the -ve end of C2, nowhere else, as in the original schematic


>>>>Frank's design seems to show IC16:1 as GND. On my FDC it is NC.

correct

>>>>>This seems mostly correct, pins 1,8 and 15 must be GND; but I suppose one must be left unconnected

Pin 1 is the connection to a capacitor for the second monostable in the chip, this is unused, hence NC.
Pin 8 must be ground - power pin, correct
Pin 15 must be ONLY connected to C2

Avatar de Usuario
Womble
Mensajes: 75
Registrado: 19 Nov 2020 18:59
Agradecido : 15 veces
Agradecimiento recibido: 77 veces

Re: The Lynx FDC Project and issues

Mensajepor Womble » 28 Nov 2020 15:42

btw for those without Eagle, you can view the schematic posted by jltursan here:

https://www.altium.com/viewer/

jltursan
Mensajes: 3290
Registrado: 20 Sep 2011 13:59
Agradecido : 313 veces
Agradecimiento recibido: 868 veces

Re: The Lynx FDC Project and issues

Mensajepor jltursan » 28 Nov 2020 18:49

Pin 1 is the connection to a capacitor for the second monostable in the chip, this is unused, hence NC.
Pin 8 must be ground - power pin, correct
Pin 15 must be ONLY connected to C2

Seems to be harmless to have all of them connected with GND, internally the IC has all of them tied; so finally no changes in this section of the schematic.

Well, my Lynx is still alive and...the FDC is also alive!, it works!! -thumbup

Already tested:
- 5.25" drive, 40 tracks formatting and verifying. Result: "12 blocks used, 788 blocks free" -drinks
- Saved and loaded a BASIC sample.

Need to test: 3.5" drive and formatting type (80 tracks really?), MB8877 FDC IC.

Summarizing the changes I've finally done on the PCB:

>>>>I checked Frank's JPEG and it also shows a connection from SKT1:15 to IC2:10. The real FDC does not 
have this link.

correct, this looks like an error

>>>>The original FDC has a connection from IC2:10 to SK1:14 and IC1:5 (A0).

correct

>>>>Frank's design also seems to show IC15:7 as NC. On my FDC it is GND.

it has to be ground

>>>>Frank's design seems to show IC15:14 as NC. On my FDC is it +5V. (Probably irrelevant as IC15:13 is
also +5V).

pins 14 and 13 are both +5 correct

I've also fixed the DipTrace schematic so it can be mass-produced easily :-)

Avatar de Usuario
Womble
Mensajes: 75
Registrado: 19 Nov 2020 18:59
Agradecido : 15 veces
Agradecimiento recibido: 77 veces

Re: The Lynx FDC Project and issues

Mensajepor Womble » 28 Nov 2020 19:15

>>>Pin 1 is the connection to a capacitor for the second monostable in the chip, this is unused, hence NC.
>>>Pin 8 must be ground - power pin, correct
>>>Pin 15 must be ONLY connected to C2
>>>>Seems to be harmless to have all of them connected with GND, internally the IC has all of them tied; so finally no changes in this section of the schematic.

except for Pin 15, this only goes to C2. Pin 1 probably should not be grounded either, its not an 'input'. I don't understand 'connected internally', they aren't :)


Well, my Lynx is still alive and...the FDC is also alive!, it works!! -thumbup

EXCELLLENT NEWS :)

EXCELLLENT NEWS :)

EXCELLLENT NEWS :)

oracle_jedi
Mensajes: 34
Registrado: 09 Nov 2020 23:23
Agradecido : 13 veces
Agradecimiento recibido: 60 veces

Re: The Lynx FDC Project and issues

Mensajepor oracle_jedi » 28 Nov 2020 20:47

jltursan escribió:
Well, my Lynx is still alive and...the FDC is also alive!, it works!! -thumbup

Already tested:
- 5.25" drive, 40 tracks formatting and verifying. Result: "12 blocks used, 788 blocks free" -drinks
- Saved and loaded a BASIC sample.

Need to test: 3.5" drive and formatting type (80 tracks really?), MB8877 FDC IC.

Summarizing the changes I've finally done on the PCB:

I've also fixed the DipTrace schematic so it can be mass-produced easily :-)


Excellent! So great that this now works!

oracle_jedi
Mensajes: 34
Registrado: 09 Nov 2020 23:23
Agradecido : 13 veces
Agradecimiento recibido: 60 veces

Re: The Lynx FDC Project and issues

Mensajepor oracle_jedi » 28 Nov 2020 20:54

Frank's design is also missing the bypass cap on the +12V rail to IC1. Is that a problem or not? Having blown up my FDC in the past by sending 12V down the 5V line I am nervous about connecting anything to a +12V supply.

jltursan
Mensajes: 3290
Registrado: 20 Sep 2011 13:59
Agradecido : 313 veces
Agradecimiento recibido: 868 veces

Re: The Lynx FDC Project and issues

Mensajepor jltursan » 28 Nov 2020 21:25

I'm right now feeding the FDC with +5V and +12V (I'm still using the 1793) without glitches, seems to work rock-solid :-)
Btw, I've just tested a 3.5" drive and works flawlessly. Obviously the drive has the usual mods, RDY signal and DS0 enabled. It formats to the same 40 tracks capacity, no black magic here. Just need to test the MB8877 and get rid of the +12V power line.

except for Pin 15, this only goes to C2. Pin 1 probably should not be grounded either, its not an 'input'. I don't understand 'connected internally', they aren't

...but they must be, I've tried to isolate pin 1 by all means (GND of block 1) without success so it must be connected to the common ground in pin 8. I haven't tried with pin 15 but expected the same behaviour. Anyhow, with the three pins to ground, the FDC works great so better leave them as they're now in the schematic :-)

oracle_jedi
Mensajes: 34
Registrado: 09 Nov 2020 23:23
Agradecido : 13 veces
Agradecimiento recibido: 60 veces

Re: The Lynx FDC Project and issues

Mensajepor oracle_jedi » 28 Nov 2020 22:35

oracle_jedi escribió:
Also, the Lynx will NOT write anything to the HFE disk unless I force pin 34 low (short it). The HxC device works fine with the Amiga, ST and an old XT, so I think the HxC is fine. The Lynx FDC just seems supper fussy on the RDY line.


Having pulled all the ICs off the Lynx FDC to verify Frank's design, I realized I missed soldering pin IC13:11.

That was the RDY problem. Now it writes fine without shorting pin 34 of SKT2/3.

The HxC format problem is most likely due to the Lynx trying to format track 41. A real 200K drive will fail and the Lynx will determine it is a 200K drive. I think the HxC accept the track 41 format but since it is a 200K disk image it doesn't actually work. I might need an HxC firmware update to make this work, but it is not really a problem since I can make new blank virtual disks by simply copying the blank Lynx HFE disk back to the SD card.


Volver a “Camputers Lynx”

¿Quién está conectado?

Usuarios navegando por este Foro: No hay usuarios registrados visitando el Foro y 1 invitado