A newly working Commodore 720 - many questions...

classic Classic list List threaded Threaded
17 messages Options
Reply | Threaded
Open this post in threaded view
|

A newly working Commodore 720 - many questions...

Rob Clarke
Following on from my success with the 610 two weeks ago, I have now got
the 720 (mostly) working. Thanks again to all on CBM-H who've helped me out.

Glorious green monochrome pic here, running the burn-in tests:

http://inchocks.co.uk/commodore/B_Series/720Working.jpg

The observant will notice that banks 3 & 4 are not populated. Although
it's an immaculate 720 case, it's actually a 710 board. The board looks
unused (there is not even any signs of scratches or oxidation on the
IEEE edge connector) but as the case and monitor ar mint, I think I owe
it to them to put another 128k on board.

http://inchocks.co.uk/commodore/B_Series/720Board.jpg

Q1) As I understand it, I just need to fit the 16 missing drams and
decoupling caps?

Q2) WRT to the differences in the PLA between the lo and high profile. I
think I read somewhere that it was supposedly to be able to accomodate
the coprocessor board. However, looking at the schematics, the PLA just
routes RAS and CAS for the 4 memory banks, using all 8 outputs but just
on different pins. Could someone explain where the differences are?

Q3) The burn-in tests correctly identify whether the machine is a high
or low profile model. ANyone know how it programatically does this?

If you look at the pic below, I've had to jury rig the power, borrowing
the PSU from my 610 to power the board while using the original PSU to
power the monitor. As per my previous post, the PSU for the 720 is shot.
Even adjusting the trimmer, it will only output 4.6v before cutting out.
I'm happy to put a more modern PSU inside but this raises a couple more
questions.

http://inchocks.co.uk/commodore/B_Series/720PowerRig.jpg

Q5) The case of the 720 PSU says it's rated for 80w but the back of the
720 says 120w, can someone else with a high profile machine clarify
this? Does mine have the "wrong" PSU?

Q6) I think modern ATX supplies have now dropped the -12V rail. From
looking at the schematics this is only used for the RS232 port. As I'm
not likely to use this any time soon, will I have any issues if I omit
the -12V rail?

Q7) The CBM-II PSU's all supply a 50/60hz signal from the PSU, which I
think is just used for the TOD clock on the CIA. As I'm unlikely to find
this on a modern PSU, is there a really simple way of generating this
signal?

Sorry for so many questions!

Rob








       Message was sent through the cbm-hackers mailing list
Reply | Threaded
Open this post in threaded view
|

RE: A newly working Commodore 720 - many questions...

Martin Hoffmann-Vetter
Hello,

> Q1) As I understand it, I just need to fit the 16 missing drams and
> decoupling caps?

Yes, that's all!

> Q2) WRT to the differences in the PLA between the lo and high
> profile.

The logic is the same. Only the pinout is changed! All the PLAs have the
logic to drive four banks of ram.

> Could someone explain where the differences are?

Only the pinout and the layout. And that's the problem with a 8088 board in
a 610!

> Q3) The burn-in tests correctly identify whether the machine is a high
> or low profile model. ANyone know how it programatically does this?

There are two pins at the keyboard 6525. One is for identify the high or low
profile board. The other is for 50 or 60 hertz.

> Q6) I think modern ATX supplies have now dropped the -12V rail. From
> looking at the schematics this is only used for the RS232 port. As I'm
> not likely to use this any time soon, will I have any issues if I omit
> the -12V rail?

Or you look for an old AT supply.

> Q7) The CBM-II PSU's all supply a 50/60hz signal from the PSU, which I
> think is just used for the TOD clock on the CIA. As I'm unlikely to
> find this on a modern PSU, is there a really simple way of generating
> this signal?

Yes, use an optocoupler.

Martin


       Message was sent through the cbm-hackers mailing list
Reply | Threaded
Open this post in threaded view
|

Re: A newly working Commodore 720 - many questions...

MikeS
In reply to this post by Rob Clarke
----- Original Message -----
From: "Rob Clarke" <[hidden email]>
To: <[hidden email]>
Sent: Wednesday, February 15, 2012 6:34 PM

> Q6) I think modern ATX supplies have now dropped the -12V rail. From
> looking at the schematics this is only used for the RS232 port. As I'm not
> likely to use this any time soon, will I have any issues if I omit
> the -12V rail?
--
AFAIK ATX supplies still supply -12V (blue wire, pin 14 of the 24-pin
version, pin 12 of the 20-pin version).

> Q7) The CBM-II PSU's all supply a 50/60hz signal from the PSU, which I
> think is just used for the TOD clock on the CIA. As I'm unlikely to find
> this on a modern PSU, is there a really simple way of generating this
> signal?
--
As mentioned previously, an opto-isolator across the PS's AC input in series
with an appropriate resistor or capacitor to drop the voltage/limit the
current as required and a pull-up resistor if needed.


       Message was sent through the cbm-hackers mailing list
Reply | Threaded
Open this post in threaded view
|

Re: A newly working Commodore 720 - many questions...

Ullrich von Bassewitz
In reply to this post by Rob Clarke

On Thu, Feb 16, 2012 at 12:34:07AM +0100, Rob Clarke wrote:
> Q1) As I understand it, I just need to fit the 16 missing drams and
> decoupling caps?

Yes. And, with some effort, the machines can also be extended to 1MB (960K
usable). Not that there are any programs using this huge amount of memory:-)

> Q2) WRT to the differences in the PLA between the lo and high
> profile. I think I read somewhere that it was supposedly to be able
> to accomodate the coprocessor board. However, looking at the
> schematics, the PLA just routes RAS and CAS for the 4 memory banks,
> using all 8 outputs but just on different pins. Could someone
> explain where the differences are?

As far as I know the differences are only in the pinout. Some signals are on
different pins, but the basic equations are the same.

> Q3) The burn-in tests correctly identify whether the machine is a
> high or low profile model. ANyone know how it programatically does
> this?

There are two input ports (PC6, PC7) on the 6525 that is connected to the
keyboard. Their meaning is as follows

    PC6  PC7
   ----------------------------------------
     0    0    CBM 6x0  PAL
     1    0    CBM 610  NTSC
     0    1    unused
     1    1    CBM 7x0

> Q6) I think modern ATX supplies have now dropped the -12V rail. From
> looking at the schematics this is only used for the RS232 port. As
> I'm not likely to use this any time soon, will I have any issues if
> I omit the -12V rail?

I don't think so.

> Q7) The CBM-II PSU's all supply a 50/60hz signal from the PSU, which
> I think is just used for the TOD clock on the CIA. As I'm unlikely
> to find this on a modern PSU, is there a really simple way of
> generating this signal?

Optocoupler has already been named. In the old days there have also been
people who fed the signal by an 555 oscillator. Clock precision is low, but
you don't have to create electronics that are connected to the high voltage
mains line.

Regards


        Uz


--
Ullrich von Bassewitz                                  [hidden email]

       Message was sent through the cbm-hackers mailing list
Reply | Threaded
Open this post in threaded view
|

Re: A newly working Commodore 720 - many questions...

Christian Dirks
In reply to this post by Rob Clarke
Am 16.02.2012 00:34, schrieb Rob Clarke:
> Q1) As I understand it, I just need to fit the 16 missing drams and
> decoupling caps?

You also need the BASIC 256 ROMs, or the added 128k will not be used by
BASIC.
On startup, the kernal tests the installed RAM, but there is no feedback
how many RAM was found.
Even if RAM is faulty, there is no display.
> Q3) The burn-in tests correctly identify whether the machine is a high
> or low profile model. ANyone know how it programatically does this?
>
As already mentioned, it is done by reading PC6 and PC7 from the 6525
which handles the keyboard.
These are also read by the kernal on startup an thic can be used to get
the better 8x14 matrix display on a 610. When doing so, the character
ROM from a 720 is needed.

> Q7) The CBM-II PSU's all supply a 50/60hz signal from the PSU, which I
> think is just used for the TOD clock on the CIA. As I'm unlikely to
> find this on a modern PSU, is there a really simple way of generating
> this signal?
>
Can simply be generated with a schmitt-trigger like in the c64.
Take a look at the schematics at
http://zimmers.net/anonftp/pub/cbm/schematics/computers/c64/250469-rev.A-left.gif
how the signal for the cia pin 19 is generated.
For the schmitt-trigger (U23 in the schematics) a 74LS14 can be used.



--
Christian Dirks
[hidden email]



       Message was sent through the cbm-hackers mailing list
Reply | Threaded
Open this post in threaded view
|

Re: A newly working Commodore 720 - many questions...

MikeS
----- Original Message -----
From: "Christian Dirks" <[hidden email]>
To: <[hidden email]>
Sent: Thursday, February 16, 2012 8:21 AM
Subject: Re: A newly working Commodore 720 - many questions...


>> Q7) The CBM-II PSU's all supply a 50/60hz signal from the PSU, which I
>> think is just used for the TOD clock on the CIA. As I'm unlikely to find
>> this on a modern PSU, is there a really simple way of generating this
>> signal?
>>
> Can simply be generated with a schmitt-trigger like in the c64.
> Take a look at the schematics at
> http://zimmers.net/anonftp/pub/cbm/schematics/computers/c64/250469-rev.A-left.gif
> how the signal for the cia pin 19 is generated.
> For the schmitt-trigger (U23 in the schematics) a 74LS14 can be used.
--
Unless you can find 9VAC or some other low voltage line frequency AC signal
in an ATX PS somewhere, I don't think an 'LS14 would withstand 110/220V very
well... ;-)


       Message was sent through the cbm-hackers mailing list
Reply | Threaded
Open this post in threaded view
|

Re: A newly working Commodore 720 - many questions...

Gerrit Heitsch
In reply to this post by Ullrich von Bassewitz
On 02/16/2012 01:41 PM, Ullrich von Bassewitz wrote:

>
> On Thu, Feb 16, 2012 at 12:34:07AM +0100, Rob Clarke wrote:
>
>> Q7) The CBM-II PSU's all supply a 50/60hz signal from the PSU, which
>> I think is just used for the TOD clock on the CIA. As I'm unlikely
>> to find this on a modern PSU, is there a really simple way of
>> generating this signal?
>
> Optocoupler has already been named. In the old days there have also been
> people who fed the signal by an 555 oscillator.

Another possibility would be the ELM440 (60Hz) or ELM446 (50Hz). They
are 8pin ICs that use a 3.58MHz crystal and supply the 60 or 50 Hz plus
a 1 Hz signal. Datasheet can be found here:

  http://elmelectronics.com/ebench.html

Or you build this:


                                                    ___ R2
                                               +---|___|--- +5V
                      ___ R1                   |
   ------ -----------|___|-------*-----+       *------+
      |/|||/|                    |  ...|.......|..    |
      |/|||/|                   --- . ---    |/  .    |
      |/|||/| TR1            D1 \ / . / \ => |   . O1 |
      |/|||/|                   --- . ---    |\  .    |
      |/|||/|                    |  ...|....   V .    |
   ------ -----------------------*-----+   ....|..    |
                                               |      |
                                              ---     |
                                              GND     |
                                                      |
           +------------------------------------------+
           |
           |   |\         |\
           +---| )O---*---| )O------- TICK
               |/     |   |/
                      |
                    ----- C1
                    -----
                      |
                     ---
                     GND


   TR1 = Transformer 230V => 6V 50mA (smallest you can get)
    D1 = 1N4148
    R1 = 820 Ohm
    R2 = 1.8 KOhm
    C1 = 10 nF
    O1 = 4N35

I once built this to replace the busted PSU for an Amiga 2000 with a
standard PC PSU for a friend. Even though the capacitor is not
recommended to be hooked up like that, it worked for years without
problems and produced a reliable 50 Hz signal without glitches.

  Gerrit


       Message was sent through the cbm-hackers mailing list
Reply | Threaded
Open this post in threaded view
|

RE: A newly working Commodore 720 - many questions...

Rob Clarke
In reply to this post by Martin Hoffmann-Vetter
Thankyou all so much for the replies.

>> Q1) As I understand it, I just need to fit the 16 missing drams and
>> decoupling caps?
>
>You also need the BASIC 256 ROMs, or the added 128k will not be used by
>BASIC.
>On startup, the kernal tests the installed RAM, but there is no feedback
>how many RAM was found.
>Even if RAM is faulty, there is no display.

Thanks Christian. Makes the test rather pointless doesn't it...

>> Q6) I think modern ATX supplies have now dropped the -12V rail. From
>> looking at the schematics this is only used for the RS232 port. As I'm
>> not likely to use this any time soon, will I have any issues if I omit
>> the -12V rail?
>
> Or you look for an old AT supply.

MikeS was correct, it was the -5V rail that had been dropped from the ATX
specification. The -12V still exists. I've ordered a 150W ATX supply used
for powering mini-itx boards. It'll supply about 4.2 watts on the 12V,
which should be enough for the monitor.

>> Q7) The CBM-II PSU's all supply a 50/60hz signal from the PSU, which I
>> think is just used for the TOD clock on the CIA. As I'm unlikely to
>> find this on a modern PSU, is there a really simple way of generating
>> this signal?
>
> Yes, use an optocoupler.

Wow - so many options. I like the option of a single chip solution from
Gerrit as it doesn't require opening the PSU or attaching or hooking up to
mains voltages. Incidently, the MAX machine used a 555 to generate the TOD
signal, as Uz speculated.

Rob

       Message was sent through the cbm-hackers mailing list
Reply | Threaded
Open this post in threaded view
|

Re: A newly working Commodore 720 - many questions...

Ethan Dicks
On Thu, Feb 16, 2012 at 12:37 PM, Rob Clarke <[hidden email]> wrote:
> Wow - so many options. I like the option of a single chip solution from
> Gerrit as it doesn't require opening the PSU or attaching or hooking up to
> mains voltages. Incidently, the MAX machine used a 555 to generate the TOD
> signal, as Uz speculated.

There used to be an 8-pin DIP that was commonly used in the 1970s for
generating 60Hz using a 32768Hz watch crystal.  These days, it's often
done with an 8-pin PIC or AVR.  I don't remember which part I used,
but I made one about 5 years ago, since I needed to run an 8051-based
clock off of DC and the design had a diode and a 74HC14 off of the
low-voltage AC side of the rectifier.  I remember it being moderately
accurate.  For that matter, unless you are using your PET as a desk
clock, a 555 is probably accurate enough for casual use, and it's hard
to get cheaper than that.

-ethan

       Message was sent through the cbm-hackers mailing list
Reply | Threaded
Open this post in threaded view
|

Re: A newly working Commodore 720 - many questions...

MikeS
----- Original Message -----
From: "Ethan Dicks" <[hidden email]>

> On Thu, Feb 16, 2012 at 12:37 PM, Rob Clarke <[hidden email]>
> wrote:
>> Wow - so many options. I like the option of a single chip solution from
>> Gerrit as it doesn't require opening the PSU or attaching or hooking up
>> to
>> mains voltages. Incidently, the MAX machine used a 555 to generate the
>> TOD
>> signal, as Uz speculated.
>
> There used to be an 8-pin DIP that was commonly used in the 1970s for
> generating 60Hz using a 32768Hz watch crystal.
--
Well, there's the MM5369, although it uses a 3.58 MHz TV crystal like the
ELM part that Gerrit pointed to.

A lot of other unique and interesting ICs at Elm as well, albeit a little on
the pricey side (especially with shipping); thanks for the link, Gerrit.


       Message was sent through the cbm-hackers mailing list
Reply | Threaded
Open this post in threaded view
|

Re: A newly working Commodore 720 - many questions...

Ingo Korb
"MikeS" <[hidden email]> writes:

> A lot of other unique and interesting ICs at Elm as well, albeit a little on
> the pricey side (especially with shipping); thanks for the link, Gerrit.

Based on the pinout of their chips I'd guess that they're actually
selling pre-programmed PICs, specifically something from the
12C5xx/12F5xx series.

-ik

       Message was sent through the cbm-hackers mailing list
Reply | Threaded
Open this post in threaded view
|

Re: A newly working Commodore 720 - many questions...

Jim Brain
On 2/16/2012 2:25 PM, Ingo Korb wrote:

> "MikeS"<[hidden email]>  writes:
>
>> A lot of other unique and interesting ICs at Elm as well, albeit a little on
>> the pricey side (especially with shipping); thanks for the link, Gerrit.
> Based on the pinout of their chips I'd guess that they're actually
> selling pre-programmed PICs, specifically something from the
> 12C5xx/12F5xx series.
>
> -ik
>
>         Message was sent through the cbm-hackers mailing list
Note 1 on the ELM440 confirms that:
Notes:

1. This integrated circuit is produced with a Microchip Technology
Inc.'s PIC12C5XX as the core embedded microcontroller.



--
Jim Brain
[hidden email]
www.jbrain.com


       Message was sent through the cbm-hackers mailing list
Reply | Threaded
Open this post in threaded view
|

Re: A newly working Commodore 720 - many questions...

Jim Brain
In reply to this post by Gerrit Heitsch
On 2/16/2012 10:49 AM, Gerrit Heitsch wrote:
>
>    D1 = 1N4148
Why is D1 needed?  In the reverse mode, the optocouple won't conduct, so
they're no concerns.
>    C1 = 10 nF
I assume it's there to stretch the cycle.  Given the 6.3V transformer
will have enough forward voltage for most of the half cycle, is the cap
really even needed?

Jim

       Message was sent through the cbm-hackers mailing list
Reply | Threaded
Open this post in threaded view
|

Re: A newly working Commodore 720 - many questions...

MikeS
----- Original Message -----
From: "Jim Brain" <[hidden email]>
Sent: Wednesday, February 22, 2012 1:57 PM

> On 2/16/2012 10:49 AM, Gerrit Heitsch wrote:
>>
>>    D1 = 1N4148
> Why is D1 needed?  In the reverse mode, the optocouple won't conduct, so
> they're no concerns.
--
Maximum reverse voltage of LEDs is generally very low, 6V in the case of
4Nxx optocouplers.
--
>>    C1 = 10 nF
> I assume it's there to stretch the cycle.  Given the 6.3V transformer
> will have enough forward voltage for most of the half cycle, is the cap
> really even needed?
--
Just guessing, maybe to filter out line noise?

mike


       Message was sent through the cbm-hackers mailing list
Reply | Threaded
Open this post in threaded view
|

Re: A newly working Commodore 720 - many questions...

Gerrit Heitsch
In reply to this post by Jim Brain
On 02/22/2012 07:57 PM, Jim Brain wrote:
> On 2/16/2012 10:49 AM, Gerrit Heitsch wrote:
>>
>> D1 = 1N4148
> Why is D1 needed? In the reverse mode, the optocouple won't conduct, so
> they're no concerns.

A LED can only take a very limited reverse voltage. The number I once
learned was 5V, apply more and you might destroy it. The antiparallel
1N4148 limits the reverse voltage to 0.7V.



>> C1 = 10 nF
> I assume it's there to stretch the cycle. Given the 6.3V transformer
> will have enough forward voltage for most of the half cycle, is the cap
> really even needed?

The idea with the cap was not to stretch the cycle but to filter out
short voltage spikes (line noise). The value was a guess back then, but
if you omit it, your clock will be fast (that part was tested).

  Gerrit


       Message was sent through the cbm-hackers mailing list
Reply | Threaded
Open this post in threaded view
|

Re: A newly working Commodore 720 - many questions...

Jim Brain
On 2/22/2012 1:42 PM, Gerrit Heitsch wrote:

> On 02/22/2012 07:57 PM, Jim Brain wrote:
>> On 2/16/2012 10:49 AM, Gerrit Heitsch wrote:
>>>
>>> D1 = 1N4148
>> Why is D1 needed? In the reverse mode, the optocouple won't conduct, so
>> they're no concerns.
>
> A LED can only take a very limited reverse voltage. The number I once
> learned was 5V, apply more and you might destroy it. The antiparallel
> 1N4148 limits the reverse voltage to 0.7V.

I thought they were well about 6V.  Makes sense.
>
> The idea with the cap was not to stretch the cycle but to filter out
> short voltage spikes (line noise). The value was a guess back then,
> but if you omit it, your clock will be fast (that part was tested).
That's my digital side peeking out.  Never thought about line noise on
the TTL.

JIm


       Message was sent through the cbm-hackers mailing list
Reply | Threaded
Open this post in threaded view
|

Re: A newly working Commodore 720 - many questions...

Gerrit Heitsch
On 02/23/2012 02:42 AM, Jim Brain wrote:
> On 2/22/2012 1:42 PM, Gerrit Heitsch wrote:
>>
>> The idea with the cap was not to stretch the cycle but to filter out
>> short voltage spikes (line noise). The value was a guess back then,
>> but if you omit it, your clock will be fast (that part was tested).
> That's my digital side peeking out. Never thought about line noise on
> the TTL.

Well, the problem is kind of on the digital side... All involved
circuits from the optocoupler to the clock counter are fast enough to
consider a 1 ms or shorter spike on the input as a valid pulse. Those
happen quite often on the power grid.

The first 74LS14 gate will make the sorta half sine wave into a proper
square wave, the capacitor will slow the rise of its output to some
degree since it needs to be charged. So if the pulse on the input is
short enough, it won't be long enough to lift the input of the second
74LS14 over the threshold and the pulse won't be seen by the clock.

Again, directly hooking up a capacitor of that size to a TTL output
stage is not recommended, but in my case it worked for years without any
problems.

  Gerrit


       Message was sent through the cbm-hackers mailing list