Atari I,Robot PCB testing, troubleshooting, and WORKING!

Be aware that the MFG of the PROM chip and it's location are important on this game. Do not choose any random PROM when programming a replacement. The speed idiosyncrasies of these chips is baked into it's location/operation.

I would be surprised if Atari was able to manage this BITD, i.e. specific manufacturer part only allowed for a specific location on the board for the same common part code. Unless they were single-sourced, which is doubtful due to risk of procurement. :unsure:

I've never seen this claim before for other boards using PROMs. Is I,Robot that dependent?

Unfortunately, the Atari AVL, which I have, does not list the approved manufacturers of 136029-XXX bipolar PROMs.
 
I did a lot of PROM swapping on 3 of the sets I worked on.. I'd be very confident in TBP24S41N and any equivalent that is rated at the same speed personally.
All 13 PROMs are the same part# - just different bits on them..

My real question is.. is that PROM adapter/pcb available? One of those would be very helpful.
 
Last edited:
That would be great. a way to read PROMs fast like that is nice..
If you want a board I can mail one out. These ones are a bit flawed because I didn't think it through. I have a couple left. They work but need a bit of modification.
 
That would be great. a way to read PROMs fast like that is nice..

Yeh, better than reversing the pins/socket on an individual PROM-to-W27C512 adapter and reading them as 512.
One board to read them all. Nice.

S135_3dt.jpg
 
Yeh, better than reversing the pins/socket on an individual PROM-to-W27C512 adapter and reading them as 512.
One board to read them all. Nice.

S135_3dt.jpg
I made a bunch of adapters like that for the atmel 27c256 45ns otp eproms. I really should finally have these all made and tested. I was much dumber 4 years ago than I am now, so I should also check the design...

 
I would be surprised if Atari was able to manage this BITD, i.e. specific manufacturer part only allowed for a specific location on the board for the same common part code. Unless they were single-sourced, which is doubtful due to risk of procurement. :unsure:

I've never seen this claim before for other boards using PROMs. Is I,Robot that dependent?

Unfortunately, the Atari AVL, which I have, does not list the approved manufacturers of 136029-XXX bipolar PROMs.

It's not hard to imagine at all. They are PROMs after all, being programmed with a specific file in a production envt. They assigned a specific part to the file being burned to it. And then that part gets put into a specific location in a board. If they couldn't do that then they couldn't handle ROMs either? Not sure why you see that being all that difficult because nobody could build anything without that basic SOP.

I recall this based upon conversations I had with @mhavoc when he was fixing IRobot boards and working on the ICY. There were opportunities for glitching because all the state logic had to match up and small variations in timing resulted in weird outputs.

Yes, I understand that IRobot was right on the edge. At the time Dave S. designed that HW it was cutting edge and faster than anything out there. It had a lot of very tight timing considerations.
 
If it works after mods - I'd like one.. thanks :)
I'll PM info.

This is what we had to mod on it, just tied the upper 4 data line low. When they were floating it would return random data on those upper bits.

IMG_3837.JPEG

IMG_3838.JPEG

IMG_3836.JPEG


There other thing I found is that the files my EMP10 spit out were doubled. I had to delete lines 0400-07F0 then they verified against Romident. (I both verified them in the EMP10 and read them to have Romident doubled check)
 
Perhaps, based on the uniqueness of the PROM usage. So, can you confirm that TI TBP24S41N can be used universally in all relevant locations on an I,Robot CPU board? These seem to be 40nsec (ADD) / 20nsec (SEL) parts.


Can you confirm which ones don't work?

It would be educational to know what MFR parts are present on owner boards in the various locations.

IR_CPU_PROM.jpg


Do the same restrictions apply also to the 82S123/129/XXX etc?

IR_VID_PROM.jpg
 
I don't think this is going to be a simple situation of testing a few PROMs and having a definitive answer. I'm no expert on silicon but my general knowledge tells me that any given part number will have a varying degree of variance in it's speed. The part number is simply some basic form of "guarantee" by the MFG that the chip will meet and or exceed that spec. That doesn't mean it's not faster in some lots vs others. Thus one lot of a specific MFG part/speed might still behave differently than another in the same situation.

[edit]
The PROMs in IRobot are all combinatorial logic and as such various changes in propagation due to part behavior are going to result in glitches that likely were not completely engineered out but were rather part selected out.
 
Looking in the manual again, the info on the "MB ERROR XX"

1745857495366.png


But I don't see where it says what the signal on pin 10 of 9L (a LS174) should look like.


I still haven't tried any other bit slicers yet. I did swap them around to see if anything changed on the "BAD DATA Y"

I suppose I should pull out Battlezone and test those in it since I know those for sure work and see if anything changes. Need the replacement PROMs before doing anything though...
 
I would be surprised if Atari was able to manage this BITD, i.e. specific manufacturer part only allowed for a specific location on the board for the same common part code. Unless they were single-sourced, which is doubtful due to risk of procurement. :unsure:

I've never seen this claim before for other boards using PROMs. Is I,Robot that dependent?

Unfortunately, the Atari AVL, which I have, does not list the approved manufacturers of 136029-XXX bipolar PROMs.
I think to clarify on this point from @bakerhillpins is that many of these PROM locations REQUIRE the higher speed TI PROMS or the B (high speed) Harris PROMs... and to @VectorCollector point, Atari spec'd high speed PROMS for ALL the locations.... I think they were all TI (Plastic) or Harris(Ceramic). What I have seen over the years is that when PROM's have died, previous repairs over the years have thrown in the lower speed PROM's because folks were mostly looking a capacity+configuration vs specifically the speeds because it is a bit of a rabbit hole finding the speed suffix information (especially for aformentioned Harris because they changed their speed labeling through the early 80s)... So, I guess my actual point is that you can't necessarily go with replacing what is there because it may not be original and may have been replaced with an out of spec PROM... the lower speed PROMs *will work* in some locations but I expect by doing that, it is probably working your PCB set closer to a 'You have hit a black hole' message.
 
As I showed above from the I,Robot manual/BOM, Atari did indeed specify 82HS "high speed" parts for the CPU board and typical 82S for the VID.
So, I assume the dependency on speedy bipolars only relates to the CPU board.

I must admit this is the first time I've seen 82HS137 instead of 82S137.
There are only three KLOV threads mentioning the HS part. Which is fascinating.

The most relevant thread being this one from 2012 ->


Which implies I,Robot is unique to needing the faster PROMs.
And perhaps the A version of the S would work as well. YMMV

82S137 - 60ns max access time
82S137A - 45ns max access time

82HS137 - 45ns max access time
82HS137A - 35ns max access time
 
Last edited:
Yea.. The schematics showed the 82HS137's also..
1745863522080.png

One of the sets I worked on had all plastic TBP24S41. The other sets had a couple of the ceramic versions sprinkled in here/there.
I had looked up the speeds and they seemed to be the same..?

I actually didn't see any chips labeled 82HS137 as it turns out..
 
Thanks for the heads up! That kinda sucks because some of these are tricky to read what the original chip were.

1745867047585.png



@lilypad19 when you replaced bad PROMs where did you get them from?


I ordered a few of these from Mark but his programmer is down so I'll have to find someone who can program them.

View attachment 815596

I really need to find a DATA I/O,..


Not sure what I'm actually going to get from Mark, but if its the HM1-7643-5 listed it should be ok. *Edit: even though the '-5' matches the original, its not the same speed.


We'll find out. Not many options for getting these proms. Eventually not being able to get them will be the death of these original boards.
 
Last edited:
Where are you guys getting the info that TBP24S41N is equivalent? I'm not seeing it on any of the cross ref lists....

Never mind, I see it now on the pic of the damn PROM I posted. lol 🤦‍♂️
 
Where are you guys getting the info that TBP24S41N is equivalent? I'm not seeing it on any of the cross ref lists....

Never mind, I see it now on the pic of the damn PROM I posted. lol 🤦‍♂️
All the PROMs on the 5 sets I worked on were that prom, except a few ceramics here/there. Board #1 - all 13 PROMs were the TBP24S41N, no ceramics
 
Back
Top Bottom