TI 99/22 Issues

[Shift838]

So, I wanted to start a new thread on the 99/22 issues that can be found.  I have already put this one on the issues tracker of the github site.  But for others that are building one of these you will want to take this into account.

 

So, for the standard GROM port on the board.  If you have socketed your board as most of us have then IC U18 (74LS00) is in the way of allowing the cartridge to be inserted fully.   Removing the IC socket for that IC will allow it to insert fully. See photo.

 

 

[Gary from OPA]

19 minutes ago, Shift838 said:

So, I wanted to start a new thread on the 99/22 issues that can be found.  I have already put this one on the issues tracker of the github site.  But for others that are building one of these you will want to take this into account.

 

So, for the standard GROM port on the board.  If you have socketed your board as most of us have then IC U18 (74LS00) is in the way of allowing the cartridge to be inserted fully.   Removing the IC socket for that IC will allow it to insert fully. See photo.

 

 

What about socketing the cartridge port connector having two 18 pin female headers that edge card cartridge port plugs into, that should raise it up high enough. Of course that would add stability issues with possible loose connection unless you glue it somehow. Or is there wire wrap edge card cartridge port available that would have longer pins.

 

The other option is to use a cartridge raiser addon board I think Dan has gerber's for that on his GitHub and use the 36 pin secondary connector just above the onboard cartridge eprom.

 

I plan on building something for that secondary connector with 4 cartridge slots with automatic RML switching. Like a widget.

[Shift838]

Just now, Gary from OPA said:

What about socketing the cartridge port connector having two 18 pin female headers that edge card cartridge port plugs into, that should raise it up high enough. Of course that would add stability issues with possible loose connection unless you glue it somehow. Or is there wire wrap edge card cartridge port available that would have longer pins.

 

The other option is to use a cartridge raiser addon board I think Dan has gerber's for that on his GitHub and use the 36 pin secondary connector just above the onboard cartridge eprom.

 

I plan on building something for that secondary connector with 4 cartridge slots with automatic RML switching. Like a widget.

The cartridge riser that Dan create is actually for the J15 header and not the grom port.

 

What is really needed if you socket all the chips is a GROM port extender, but of course you would need a Straddle Card Edge connector and those can get pricey.  I was looking for one earlier and have not found anywhere that has them in stock yet.  Prices so far vary from almost $9 to $12 each.

[Gary from OPA]

1 minute ago, Shift838 said:

The cartridge riser that Dan create is actually for the J15 header and not the grom port.

 

What is really needed if you socket all the chips is a GROM port extender, but of course you would need a Straddle Card Edge connector and those can get pricey.  I was looking for one earlier and have not found anywhere that has them in stock yet.  Prices so far vary from almost $9 to $12 each.

I done a upside chip trick a few times, solder the socket on the bottom of the board and then carefully reverse the pins fully on the 74ls00 and plugging into the bottom of the PCboard.

[Shift838]

Just now, Gary from OPA said:

I done a upside chip trick a few times, solder the socket on the bottom of the board and then carefully reverse the pins fully on the 74ls00 and plugging into the bottom of the PCboard.

i thought about that too, but just decided to take the socket off.

[RickyDean]

I've already soldered two of those sockets, so I'll have to determine the best course of action, for the two boards I'm doing. Just finished soldering talll the sockets last night

[nuxi]

Well, fortunately it's a TTL chip: chances of it going bad is practically nil, so I'll just go ahead and solder it in. Thanks for the warning though! Good thing to fix in v.97

[RickyDean]

I'm leaning toward installing two rows of machined sip sockets, then solder the connector to that. Should raise it high enough and still be strong for inserting/removing cartridges. Opinions?

Edited September 26 by RickyDean

[Shift838]

I have designed a small grom extender that will raise it plenty.  It's 23.25mm tall

 

If anyone wants one, I can do an order to get some in.  

[Gary from OPA]

28 minutes ago, RickyDean said:

I'm leaning toward installing two rows of machined sip sockets, then solder the connector to that. Should raise it high enough and still be strong for inserting/removing cartridges. Opinions?

Yes, what I was thinking of my first post, but hard to socket the cartridge port to the sip socket not much space to work with without melting some plastic, maybe glue them together.

[RickyDean]

I was thinking solder paste and some heat would be sufficent.

[RickyDean]

6 minutes ago, Gary from OPA said:

Yes, what I was thinking of my first post, but hard to socket the cartridge port to the sip socket not much space to work with without melting some plastic, maybe glue them together.

I was thinking solder paste and some heat would be sufficent. Could also use a header connector, then solder individual pins.

Edited September 26 by RickyDean
added content

[RickyDean]

My thinking process is imagining 2 header strips soldered in with one of these, modified to fit close and soldered to the top of the header pins. Then offset the connector from them one or two rows away. Opinions?

 

Or this cut down?

Edited September 27 by RickyDean
added content

[Gary from OPA]

30 minutes ago, RickyDean said:

My thinking process is imagining 2 header strips soldered in with one of these, modified to fit close and soldered to the top of the header pins. Then offset the connector from them one or two rows away. Opinions?

 

Or this cut down?

That could work. Also if anyone planning on having a plastic case enclosure, the cartridge port most likely going to have up be raise up a bit anyway. So the height should be adjusted correctly to match the final case design.

Edited September 27 by Gary from OPA

[Artoj]

5 hours ago, Gary from OPA said:

Yes, what I was thinking of my first post, but hard to socket the cartridge port to the sip socket not much space to work with without melting some plastic, maybe glue them together.

 

56 minutes ago, Gary from OPA said:

Also if anyone planning on having a plastic case enclosure, the cartridge port most likely going to have up be raise up a bit anyway. So the height should be adjusted correctly to match the final case design.

Excuse the poor photo Yes I agree, as an interim before the case is finalised to your own specs, some headers will suffice. I will be making choices in the next few weeks.  

[nuxi]

I dunno. This seems a lot of fuss over a single 7400. Am I missing something?

 

ETA. Ah, missed Gary's comment about the housing. Ok, I get it now.

Edited September 27 by nuxi
duh

[Shift838]

I decided to power the 99/22 up after triple checking all connections for solder bridges.  Having my fingers crossed and nothing.  Just a black screen, no tone, nothing.

 

So, for my first test, I tested all my voltages to make sure they were all reading good on all the chips for their correct voltages.  On the 9900, -5v on pin 1, +5v on pins 2 & 59 and +12v on pin 27, all read good.  Verified none of the logic chips are getting hot which would indicate a bad or fake IC.  Next, I pulled out a logic probe to start testing some things and I have found that pin 6 (reset) of the CPU is being held low.  The reset line of the CPU should stay low for a split second then go high when you turn the unit on.  This is not the case, it stays low, this will basically cause the processor to halt from my understanding and not allow it to do anything at that point.

 

I know the processor is good.  I have pulled from a working unit the TMS9900, 9901, 9904 and 9918.  

 

Now I'm at a loss.  What could be causing the CPU reset line to stay low at power up?  According to the troubleshooting guide I need to check U601, which I believe is the 9904 (maybe someone can verify that for me).  the 9904 tests good and as I said came from working unit and it was already socketed so it was an easy pull for testing.

 

suggestions anyone?

 

[Stuart]

Which 9904 specifically have you fitted? I think the 99/22 spec called for a TIM9904 and a 48MHz crystal. If you've pulled and fitted a TIM9904A used with a 12MHz crystal, that might be the problem if you're not getting a clock.

[Shift838]

4 hours ago, Stuart said:

Which 9904 specifically have you fitted? I think the 99/22 spec called for a TIM9904 and a 48MHz crystal. If you've pulled and fitted a TIM9904A used with a 12MHz crystal, that might be the problem if you're not getting a clock.

i purchased the crystals from the BOM, the 9904 I have is a 9904ANL

[Gary from OPA]

43 minutes ago, Shift838 said:

i purchased the crystals from the BOM, the 9904 I have is a 9904ANL

you will have to change it to 12mhz crystal then if its 9904A and the two caps and resisters for it to work correctly. -- look at the original 4A schematics it should show the two different setups, the 48mhz version (which was mainly on the /4 motherboards and the first few runs of the /4A, later on all consoles used the 12mhz setup. - Future revisions of the 99/22 PCBoard should be setup to use both types of crystal timing setups.

 

Personally, in my setup, I going to setup two oscillators and and tie it to unused CRU pin on the 9901 so I can software switch from 3mhz to 4mhz setup, and also add a manual reset button wired to the 9904A chip. the good thing of using a pair of oscillators is that it doesn't matter if you use 9904 or 9904A chip.

TIM9904.pdf

Edited Thursday at 11:43 PM by Gary from OPA
added TMS9904A datasheet

[Stuart]

1 hour ago, Shift838 said:

i purchased the crystals from the BOM, the 9904 I have is a 9904ANL

I can't see where the BOM is at the moment, but looking at photos of the boards, it looks like the crystal frequency is marked as 48MHz on the PCB, and photos of the designer's board appear to show a 9904 (without the "ANL").

 

So two options might be to get yourself a TIM9904 without the "A" (aka 74LS362), or change the crystal to a 12MHz one, and you also need to change the accompanying tank circuit inductor and capacitor across the 9904 pins 1 and 2 (3.3uH and 50pF would probably work).

 

NOTE that I'm not saying that that is definitely the cause. But the power-up reset signal goes through a D-type flip-flop in the 9904 which is clocked internally by the phase 3 clock, and if the clock isn't working because you've got the wrong crystal and tank circuit components, the power-up reset going high won't make it out of the 9904 to the processor.

 

Also check the 9904 pin 5 with your logic probe - check that that is initially low then goes high as you would expect for a power-up reset signal. Pin 4 (connected to the processor /RESET input on pin 6) should follow it but it is not according to your earlier post.

 

@stevee671 appears to be in the same situation, looking at a photo he posted earlier. So you might want to hold off and see what his experience is.

 

Edit ... Just found the schematics for the board. The arrangement of the reset switch J1 would make me a little nervous, as it directly grounds the FFQ output of the 9904 on pin 4 which the 9904 will be trying to drive high. The reset switch would be better connected to the FFD input on pin 5 IMHO.

Edited Friday at 12:17 AM by Stuart

[stevee671]

@Stuart the BOM can be found here: https://github.com/danwerner21/TI99_22

[nuxi]

4 hours ago, Gary from OPA said:

 

Personally, in my setup, I going to setup two oscillators and and tie it to unused CRU pin on the 9901 so I can software switch from 3mhz to 4mhz setup, and also add a manual reset button wired to the 9904A chip. the good thing of using a pair of oscillators is that it doesn't matter if you use 9904 or 9904A chip.

TIM9904.pdf 9.7 MB · 0 downloads

Ok, tell me why that's better than a hard switch.  I've got a console or two with a physical switch, and most of the time it doesn't even cause an issue if I hit the switch mid-calculation.

[Gary from OPA]

25 minutes ago, nuxi said:

Ok, tell me why that's better than a hard switch.  I've got a console or two with a physical switch, and most of the time it doesn't even cause an issue if I hit the switch mid-calculation.

Software switch is better if you coding new software and need the normal 3mhz for certain tasks, most of the time 4mhz you can run everything, but for certain operations it's nice to be able to have the system automatically switch down without the user needing to do so.

[Shift838]

22 hours ago, Stuart said:

I can't see where the BOM is at the moment, but looking at photos of the boards, it looks like the crystal frequency is marked as 48MHz on the PCB, and photos of the designer's board appear to show a 9904 (without the "ANL").

 

So two options might be to get yourself a TIM9904 without the "A" (aka 74LS362), or change the crystal to a 12MHz one, and you also need to change the accompanying tank circuit inductor and capacitor across the 9904 pins 1 and 2 (3.3uH and 50pF would probably work).

 

NOTE that I'm not saying that that is definitely the cause. But the power-up reset signal goes through a D-type flip-flop in the 9904 which is clocked internally by the phase 3 clock, and if the clock isn't working because you've got the wrong crystal and tank circuit components, the power-up reset going high won't make it out of the 9904 to the processor.

 

Also check the 9904 pin 5 with your logic probe - check that that is initially low then goes high as you would expect for a power-up reset signal. Pin 4 (connected to the processor /RESET input on pin 6) should follow it but it is not according to your earlier post.

 

@stevee671 appears to be in the same situation, looking at a photo he posted earlier. So you might want to hold off and see what his experience is.

 

Edit ... Just found the schematics for the board. The arrangement of the reset switch J1 would make me a little nervous, as it directly grounds the FFQ output of the 9904 on pin 4 which the 9904 will be trying to drive high. The reset switch would be better connected to the FFD input on pin 5 IMHO.

Took your advice and it worked.  Replaced the crystal with the 12mhz, 3.3uh and a 50pf cap.

 

I have also tested my USB adapter that plugs right into the keyboard connector and it works with no probably.

 

Also, for those that will be using an F18A as video.  Make sure you have the sound jumper (J7) shorted on pins 2 and 3 otherwise it will not boot and you will just get the ugly F18A error screen.