4 Volts at the GROM port

[Kchula-Rrit]

Do the system GROMs get their +5V power separate from the rest of the motherboard?  The schematic in the back of the blue TI tech manual seems to indicate everything gets the same +5V, except for the VDP and its RAM, which has an inductor on it.

 

I've been trying to figure out why my non-QI 99/4A is unstable.  I \measured the +5V pin at the cartridge slot, and found it has 4 volts on it.  This morning I opened-up the console and put a clip lead on the +5V pin that feeds the motherboard, then used an ohm-meter probe to see if there were any breaks in the power traces.  At the GROM port I measured 5.5 ohms between the input +5V line and the cartridge +5V pin.  Then I tried to follow the trace on the motherboard, reading about 5.5 ohms along the way, including the rail-thing along the CPU chip.  I followed the trace to the motherboard GROMs, and still found 5.5 ohms.  The half of the CPU chip next to the rail-thing (seems to be fed by the rail) also measured 5.5 ohms.

 

It doesn't help that the rail seems to be a power (and maybe ground) conductor.

 

The side of the CPU chip nearest the power connector, and the power pins on several of the TTL chips, all measured 1.1 ohms.  The meter leads themselves were about 1 ohm (cheapie $8 meter).

 

Any thoughts or hints where to look?  Anything that might not be in the schematic?

 

K-R.

 

[matthew180]

The white "bus bars" used on the 99/4A motherboard are indeed for power distribution.  AFAIK they are each unique and have various combinations of the voltages.  They also break easily, so you should check that all the legs on each one are intact and not cracked.  I always thought they were just shielding and treated them with little respect (pushing them over and out of the way).  I learned the hard way (console stopped working) that they indeed carry power.

 

The resistance on the 5V rail is certainly a concern.  What it feels like is that the 5V rail is broken and what you are measuring is some passive path through a chip somewhere.  Have you tried measuring with the meter leads reversed?  Or it could also be a bad or broken solder joint on one of the bus-bars.  There was a post not too long ago, I think, where someone mapped out the bus-bars.  Also, since they are for power and ground distribution only, I suspect they are not directly represented on any schematic.

[Gary from OPA]

a bad grom chip can pull the current down, try removing the TI Basic groms 1 and 2 and see how it behaves. check all those blue jumper wires as well, they can be at faulty sometimes also. have you tried swapping out the actual power supply itself. and remember the grom chips need minus power as well.

Edited February 6 by Gary from OPA

[Kchula-Rrit]

1 hour ago, matthew180 said:

The white "bus bars" used on the 99/4A motherboard are indeed for power distribution.  AFAIK they are each unique and have various combinations of the voltages.  They also break easily, so you should check that all the legs on each one are intact and not cracked.  I always thought they were just shielding and treated them with little respect (pushing them over and out of the way).  I learned the hard way (console stopped working) that they indeed carry power.

 

The resistance on the 5V rail is certainly a concern.  What it feels like is that the 5V rail is broken and what you are measuring is some passive path through a chip somewhere.  Have you tried measuring with the meter leads reversed?  Or it could also be a bad or broken solder joint on one of the bus-bars.  There was a post not too long ago, I think, where someone mapped out the bus-bars.  Also, since they are for power and ground distribution only, I suspect they are not directly represented on any schematic.

I thought the rails were to stiffen the motherboard; a computer system I worked on in the 1980s had metal stiffeners that kept the boards from sagging.  They were connected to the internal ground plane, which made it handy to clip a ground lead for a meter, scope, or logic analyzer.  I suspected they might carry power when I started following the +5V traces.  What you said above confirms it.

After posting the question I put a clip lead back onto the +5V console-power connector, and then put another clip lead on a cartridge break-out board that was plugged into the console riser.  It read 5 ohms, as I expected.  When I wiggled the bus bar that runs alongside the CPU chip the resistance went down to less than 1 ohm and then varied between 1 and 5 ohms.  I think the bus bar is internally broken.  If it, I assume I can put a jumper on the bottom of the motherboard?

49 minutes ago, Gary from OPA said:

a bad grom chip can pull the current down, try removing the TI Basic groms 1 and 2 and see how it behaves. check all those blue jumper wires as well, they can be at faulty sometimes also. have you tried swapping out the actual power supply itself. and remember the grom chips need minus power as well.

My console uses an external power supply that puts out +/-5V and +12V and connects directly to the motherboard.  A branch off of it, with what I call a "navel" jack, powers my NanoPEB.  I forgot to mention that the power, measured at the console was +5V, and that the -5V, measured at the cartridge breakout board is -4.98V.  Right now, after wiggling the bus bar, the resistance is really low.  I'll try jumpering the bus bars first.

 

K-R.

Edited February 6 by Kchula-Rrit

[matthew180]

28 minutes ago, Kchula-Rrit said:

I think the bus bar is internally broken.  If it, I assume I can put a jumper on the bottom of the motherboard?

 

I doubt it is broken internally, they are just strips of metal.  I suspect any breaks are going to be with the legs that come out of the strips.  That's where mine broke, and they are a PITA to solder since, well, the metal strips are just like heat-sinks, and they are attaching to power and ground planes on the PCB (which are also just big heat sinks).

 

Sure, you can use jumper wires, but you need to map out the bus-bar to see which legs are connected, and then jumper all the points that are connected to each internal rail.  Like I mentioned, all of them are unique (I suspect, and you should make sure if you plan to replace them), and they can carry any of the voltages as well as ground.

 

[RickyDean]

1 minute ago, matthew180 said:

 

I doubt it is broken internally, they are just strips of metal.  I suspect any breaks are going to be with the legs that come out of the strips.  That's where mine broke, and they are a PITA to solder since, well, the metal strips are just like heat-sinks, and they are attaching to power and ground planes on the PCB (which are also just big heat sinks).

 

Sure, you can use jumper wires, but you need to map out the bus-bar to see which legs are connected, and then jumper all the points that are connected to each internal rail.  Like I mentioned, all of them are unique (I suspect, and you should make sure if you plan to replace them), and they can carry any of the voltages as well as ground.

 

I do believe there was someone here, sometime back, that posted a diagram of the rails and the paths that they handled, as well as the voltages. Maybe they'll chime in here about where that diagram is?

[Gary from OPA]

Ahh, the beauty of 2 sided PCBoard, if the TI99 console was made today it would have multiple-layers, and you would not see those bars bringing ground and power to different areas of the board, but it does make it look better i think, like little boxes each different area are in, cool looking!

[Kchula-Rrit]

The problem is fixed, and my console comes up to the title screen and seems stable.

 

At first I was going to use my 100W resistance solderer to re-solder the rails.  After some fiddling (that thing heats up really fast) I decided against it after trying to get the probes into those tiny spaces.  I also thought about how it works, by passing a current through the piece to be soldered and thought, a slip-up could fry this irreplaceable computer!

 

I thought about putting jumpers along the legs, on the solder side of the board and decided against it because some of the legs appeared to be wired to the ground plane, and some appeared to connect to the power plane!

 

After thinking about it a bit I decided to connect a jumper from a place near the power-input connector and Pin 9 on one of the GROM sockets.  After checking several times with the meter to make sure there were no power shorts I started the computer and got 4.9V at the cartridge port, instead of the 4V before.  The Blasto cartridge came up normally.

 

However...  When I put the computer back together, the keyboard cable, that brittle thing that people here have talked about, snapped off at the keyboard end.  Fortunately, I had installed a USB keyboard adapter when I tried the battery conversion.  Now I have another problem to pester the forum about.

 

But the cartridge-port power has been solved.  Thanks to those who helped!

 

K-R.

Edited February 7 by Kchula-Rrit