IIGS not resetting

Have you checked the power supply voltages?

When you say reset isn't being sent out, what is the voltage on the CPU pin 40?

Is the board a rom 01 or 03?

Regards
Geoff B

Is there any display visible on the screen either text or graphics when powered on?

While the ADB chip can generate a reset, the Mega II also has a power on reset circuit and generates a reset signal when first turned on.

With all cards removed, can you capture the reset on pin 40 of the CPU with your oscilloscope and upload a picture.
Is the reset signal different with or without the ADB keyboard/mouse connected?

I think when your creating message you select media browser to upload file(s)

If you REALLY tied the reset trace to +5 and pin 40 of the CPU is still fluctuating between 0 and +5, the clearly you either have a broken trace or (more likely) the CPU socket is bad!

If the CPU pin 40 is stuck at 0 V or nearly 0, then something in the computer is pulling down /RESET.

As long as /RESET is asserted, there's no way anything can happen!

The /RESET line is connected to all of the major chips in the IIgs, but it has no pullup or RC timer connected to it. So while it's possible for any of the chips to pull the line down to force the computer to reset, it can only become high (and thus, allow the machine to start working) if it is pulled high by one of the chips. The chip that pulls /RESET high must be the ADB microcontroller M50740, for reasons that it is the only chip with its own power-on-reset timing circuit attached. After the ADB microcontroller itself comes out of reset, it can begin executing instructions which probably involve a self-test or sanity check, possibly involving communication with ADB GLU, after which it pulls its P25 pin 49 high to bring the rest of the computer out of reset.

So the reasons for the /RESET to be stuck low are that either the ADB micro is broken internally, its connections to the board are broken, or that it sees some fault that prevents it from completing its self-test, or that one of the other chips is pulling /RESET low and preventing this line from coming to a high state. You said you measured an intermediate voltage on this line, something like 200 mV, which means that the last case is likely happening (ADB is pulling the line high but another chip is pulling it down, fighting each other). Determining which chip it is may be difficult although there are techniques using inductive probes to find which chip pin is sinking this current. If two chips are fighting each other there will always be current flowing from one to the other.

pcbishaf wrote:

I think the next step would be to remove all the chips in sockets that are connected to the reset line. 

This sounds a bit odd as the approach, may be better to see if you can find if reset has a short to ground somewhere on the board. This should be easy check.

Similar to what jeffmazur said, if you tied reset to 5V directly and the system didn't work that's a concern and I would not try that again before checking reset doesn't have a short to ground. You can kill your supply if there is a short to ground and you connet 5V to ir.... that would be bad.

Any chance you tried the self-test before it stopped booting? 

Did you do or notice anything betweeen when it was working and when it started doing this? Do you have cards installed in the system, if so get them out for the troubleshooting. 

roespierre covered some good points on the reset line but know this, the GS is quite a complex system to troubleshoot. There's so much going on between the mega, the 816, fast and slow RAM, the ADB controller and ADB glu... this can be a very difficult task even when you know a lot about the GS. 

When you said the voltages are good, where are you checking this? Is your scope 2 or 4 channel? If you monitor all voltages at the same time (4 channels) this would be best but if you only have two just wastch 5 and 12V because those are the main ones ones used when starting up.  The speaker click you're hearing could be something but also not uncommon for ROM01 motherboards.

You are correct RAM could easily be your problem, but only after power is ruled out.  The most important thing about power is stable voltages, meaning not a lot of movement they should read flat all the time. Even mometary dips are not good.  There may be a minor ~0.75V dip on 12 when the drive spins up ( you may see the same same dip on -12 and -5) but that's nothing to worry about and that's the only significant movement you should see. 

For pictures, the media browser should be here:

media browser.PNG

Unfortunately it's not as easy as drag and drop, you need to load the file, upload the file, click next, next, submit, save or something along those lines... it's clunky at best. The image will be inserted where the cursor was and don't worry about the size in the  editing window it will be sized better when you post.

A reset signal is not a pause or wait signal. Its purpose is to initialize state machines to a known state, from which operations can proceed deterministically. As a simple example, a flip-flop can assume a state of '0' or '1' when power is applied. A computer might contain thousands or millions of flip-flops, so you can see that it is no good if it were simply to start running after being switched on. It is necessary to have a reset signal to clear all of these flip-flops so that their behavior can predictably proceed.

A computer typically has a watchdog reset circuit, a counter which is constantly counting down towards zero. Periodically, some task in the machine writes to the counter, to "fill it up" again so that it won't reach zero. But if the computer is stuck or hung, nothing will fill up the counter, it will continue counting down to zero, and when it reaches zero it will assert the reset signal to bring the computer back to a known state where it isn't stuck.

Earlier you said that the /RESET line dropped "after about 20 seconds". I would expect a watchdog reset to time-out in a much shorter time, but there you go. Without any CPU activity (because you removed the CPU) the firmware inside the ADB microcontroller gets stuck waiting for a response that never arrives, and its watchdog reset activates which brings it to a known state. It stops pulling P25 high because the output pin is reset to its initial state at that time. Most of the time, the known initial state of output pins is "hi-Z" or no output, effectively disconnecting the chip from that wire.

With nothing connected to a wire, it will hold the same voltage for a while just from its parasitic capacitance, but it will soon bleed away. A scope probe on the wire will pick up noise, just like it does when you wave it around in the air. Depending on how the probe is grounded there could be a lot of noise.

So what you discovered when removing those chips is what would be expected to happen.

The RESET on Apple II's is normally pulled high via pullup resistor to 5V. Then any source that wants to reset can connect collector of an NPN transistor to RESET and emitter to GND, then when the transistor is turned on it will pull the collector down towards ground (RESET low). If no device is pulling it low then RESET would be pullup to 5V.

As far as I am aware of the Mega II chip, the reset pin is both an input and output.
During power on it pulls the Reset low until power is considered stable by internal Power on circuit and will reset internally if reset is pulled low externally to Mega II.
The ADB micro can also pull the Reset low due to the correct keyboard combination being pressed.

You could try shorting Pin 16 of ADB micro temporarily to GND to discharge C39 and see if this changes reset going to the 65816.
This would cause the ADB Micro to reset and start up correctly if it had not when the IIgs was first powered on.
Also check voltage on ADB Micro Pin 48, which would be connected to IIe keyboard header Pin 15, which is RESET from IIe keyboard if it was connected.

Three chips that could be removed from sockets on ROM1 to see if they impact RESET, are 65816, Socketed VGC, DOC5309.

pcbishaf wrote: The output was from a new CPU that I purchased. I also monitored BE on the CPU and this is high until the wider section of the graph drops towards zero. The narrower section of high output on the graph did not affect this pin. I am thinking this means the software is starting up and something else is not working. This leads me back to my original thinking that a FAST RAM chip is not functi

I see the scope trace says 200ns but I can't make out what that time is, it it a division or full frame?Do you have timing for all the states? It looks like ther's a reset loop that includes both a system reset (guessing this is the short low which was 8 cycles on the other IIs and I didn't check the GS reference. But then there's the longer "high" which I am guessing is the system starting up until something decided it needs to reset so pulls it low and then the process repeates.  The question still seems to be which component is pulling it low. I also haven't checked if /RST is pulled up to 5V but would assume that may be the case, but I don't see the plot gettng to 5V, looks like it hangs below 4.8V for all those pulses, if pulled up that would be near the lower end of acceptable and may be something to look at.  Are all cards still removed?  Could you set all slots to "your card" to disable everything but ADB controller? 

Can you clarify what these unknowns are?

I am intrigued by the small rise to the reset line and assuming that's at 5V, or does that rise indicate some component has pulled itself offline which would cause the voltage rise. 

Please clarify.PNG

 

pcbishaf wrote: The output was from a new CPU that I purchased. I also monitored BE on the CPU and this is high until the wider section of the graph drops towards zero. The narrower section of high output on the graph did not affect this pin. I am thinking this means the software is starting up and something else is not working. This leads me back to my original thinking that a FAST RAM chip is not functi

I see the scope trace says 200ns but I can't make out what that time is, it it a division or full frame?Do you have timing for all the states? It looks like ther's a reset loop that includes both a system reset (guessing this is the short low which was 8 cycles on the other IIs and I didn't check the GS reference. But then there's the longer "high" which I am guessing is the system starting up until something decided it needs to reset so pulls it low and then the process repeates.  The question still seems to be which component is pulling it low. I also haven't checked if /RST is pulled up to 5V but would assume that may be the case, but I don't see the plot gettng to 5V, looks like it hangs below 4.8V for all those pulses, if pulled up that would be near the lower end of acceptable and may be something to look at.  Are all cards still removed?  Could you set all slots to "your card" to disable everything but ADB controller? 

Can you clarify what these unknowns are?

I am intrigued by the small rise to the reset line and assuming that's at 5V, or does that rise indicate some component has pulled itself offline which would cause the voltage rise. 

Please clarify.PNG

 

pcbishaf wrote: The output was from a new CPU that I purchased. I also monitored BE on the CPU and this is high until the wider section of the graph drops towards zero. The narrower section of high output on the graph did not affect this pin. I am thinking this means the software is starting up and something else is not working. This leads me back to my original thinking that a FAST RAM chip is not functi

I see the scope trace says 200ns but I can't make out what that time is, it it a division or full frame?Do you have timing for all the states? It looks like ther's a reset loop that includes both a system reset (guessing this is the short low which was 8 cycles on the other IIs and I didn't check the GS reference. But then there's the longer "high" which I am guessing is the system starting up until something decided it needs to reset so pulls it low and then the process repeates.  The question still seems to be which component is pulling it low. I also haven't checked if /RST is pulled up to 5V but would assume that may be the case, but I don't see the plot gettng to 5V, looks like it hangs below 4.8V for all those pulses, if pulled up that would be near the lower end of acceptable and may be something to look at.  Are all cards still removed?  Could you set all slots to "your card" to disable everything but ADB controller? 

Can you clarify what these unknowns are?

I am intrigued by the small rise to the reset line and assuming that's at 5V, or does that rise indicate some component has pulled itself offline which would cause the voltage rise. 

Please clarify.PNG

RAM would not be my first suspect, because it looks like something is pulling reset low and AFAIK that's not a RAM thing... system keeps running with bad RAM and may crash but not usually reset. 

In addition, both fast and slow RAM are soldered in place an getting them out is not easy so should be done by someone with good rework skills. It's way too easy to destroy stuff and I've seen some horrible examples recently... so sad.

Considering that's pin 49 (Reset), I think you found your problem. Did you perhaps dislodge that pin with a probe when you were checking it?

You need some fine tools to work on these small-pitched smd devices. Try to gain the skills working on a spare board you don't care about. Or get someone with the skill to repair it. Definitely fixable though; just need a very small soldering iron tip and try to nudge the pin back in place. Don't try to add solder. You should be able to just tack it down with what's there.

Check for a short between pin 49 and 48 or 50.  

Looking at the pre-work picture, and I'm sorry the quality isn't good enough to see the detail I'm looking for, it looks like the pad may have been lifted/torn. So reflow may not be enough. You got a microscope? Where are you at? You may need some help with this one because if the pad is lifted (and it only goes to that trace right there under the soldermask) you could use some help getting that repaired well. IMHO this isn't something you do if you don't know what you're doing. 

If the pad is stil intact with the board, and the trace isn't torn, what you did for the reflow may have beeen the best option. But I wouold have hit it with flux first and then dragged the  dry tip across the pins in that section. I know you're not a rework guy but heat is both your best friend and worst enemy. Run hot and quick, but not too hot nor too long because heat+excess time always produces bad results. 

 Since you didin't notice a change, I'm guessing the pad and trace are experiencing a break up and now next door neighbors with a tiny air-gap fence. 

The schematic doesn't always match where componsents on the cirucit are, but looking at the schematic you can see next to the ADB glue chip RESET.L line is pulled up to 5V through a 200k resistor so that means if RESET is pulled  low by a chip very little current flows to ground, 25uA of current... and only for as long as reset stays low.  Hopefully every device connnected to reset has high impedence on that input, so you may be fine here. Because touching 5V to a pin for even an instant is more than enough time to do serious damage when there's no resistance. Would have been better to run it through a large resistor just like is done in the circuit but I would not do that wihtout fully studying the circuit and all components connected to that singal first. 

A faulty component maybe getting hot and as it does the resistance decreases. When power is removed it starts to cools down and as it cools down the resistance starts increasing back to 57.8K.
This could be any of the components connected to RESET.L Net coming into each page of the schematic.
Not connected example - ADB MIcro Pin 16 is labeled RESET.L but isn't connected to RESET.L Net on page 5 of the schematics.

If you made a list of each component connected to the RESET.L Net, you could check each component in the list if any are getting very warm/hot to touch when powered on, take precautions not to get burnt if any are extra hot.

I'm not clear on what you're measuring. "Ohms" are a unit of resistance, and resistance measurements are only valid for unpowered circuits (power OFF). You aren't likely to damage anything by trying to measure resistance in a powered circuit, but it won't yield a valid measurement.

pcbishaf wrote:

 

As I said my electronic knowledge is limited.  I didn't know resistance measurements are not valid in a powered up circuit.  I knew something was wrong with my reading but didn't know why.  Thanks for clearing up this knowledge gap I have.

Because of this, I'm going to suggest you may want to stop.

I say that because I'm currently working to repair the work done by an similar (anxious and excited not an EE type owner) who asked for help with his new II system. Someone suggested it was likely a Rev 3 board which inticed the owner to ask how to verify, and this is where the snowball starts rolling down the hill.... I'm almost 100% sure his system would have worked had he not been told to pull the CPU to check the revision. At the guidance of the all knowing internet hive he learned how to pull the CPU, but pulled the CPU AND socket from the board. Good Job, well done! He then tried to reinstall the CPU and power up the system. But when the system didn't start up correctly they again seeked help from the hive and stated pulling ROMs and RAMs (or what they though were RAMs but were actually 7400 logic parts used for the paddle reading because they doesn't know what a RAM is!)  So again, without knowing what they were doing, they were just creating problem upon problem upon problems. ICs inserted backwards or in the wrong socket.... you name it I saw it! And you should have seen my horror when I just lifted the CPU and socket from the board with my fingers! Not saying you are in the exact same boat, but shorting 5V to a reset pin is a good idication there is potential for unknowingly doing something bad which someone who knows better may approach in a different (safer) way . 

Hopefully you haven't accidently done the same level of damage as this other guy  has (basically destroyed the CPU socket, possibly the CPU, all the original ROMs, clock generator, and several 7400 logic parts!). I know this is a learning adventure for you which if you're fine. But is losing the board OK for you? If you really want to board working again... the best option may be to get to someone with the knowledge and skills to provide a decent chance of getting it back. That said, if a SMD like the ADB GLU or ADB controller needs to be replaced the group of us that can do that work greatly shrinks to few. 

The worst part of that story was it was the result of nothing more than pulling chips from sockets!! Crazy on so many levels it's hard to believe. 

I know I'm sounding kinda of "dad-ish" with my comments but we're well beyond the "basic" trouble shooting and diagnosing level. We're in the deep end of the pool and this comes with risks. I just wanted to point out there are times when that extra bit of experience can make the world of difference and this is not the ideal way to get that help. The other thing to understand is the GS is a pretty complex design, it's nothing like troubleshooting the standard II's. You don't find a lot of people that really understnad the design at a very low level. Most of us know bits and peices... that knowledge only goes so far before we hit a brick walll. 

I'm sure others may comment, but for me.... I can't tell you the last time I cut a trace except to repair traces/pads. Your board you do what you want. The other comment I'll share on that... you can actually break it more harm and each of these "experiments" does make it harder to recover. Everything you do does make it more difficult for someone else to fix. Just like the guy who just pulled chips, the work done by him is really an unknown black hole to anohter. You can describe but you're really speaking in a different language. I'm sure you've already written off this board, and now it's an learning project which is fine. As long as you unerstand there's not much help that will follow because so many assumptions that could be made have to be tossed out and it's really like brigning up a circuit from scratch. This is what I'm doing with the Reve3 II board, it's not fun.   

The sawtoothing is intersting and is that a very slow ramp down? Almost sounds like a BOR response which would suggest the supply is offline (and that should never happen). Do you have a scope capture of that with both Reset/ and 5V? The overheating could very well mean an internal short, but if that's the case you may be able to detect that by probing the pins. Mabye or maybe not. You also need a special puller to remove the VGC and that seems like a suspect far down the line. ADB may be a better candidate because the joints on the keyboards weere known to have issues. The circuit at the input to the motherboard is farily simple and may be worth looking into. Sadly this is one of those cases where you follow the schematic and see where things go south. that's tough to do remotely.  Did you try starting up with ADB port open?

ADB port open means nothing connected to the port. ADB ports were known to have issues over time and while I'm not sure exactly what could cause what you're seeing it would not be a surprise to hear you keyboard is a factor.I'm looking at the opposite (diagonal line) corner of the chip in the schematic. The stuff going on over there if there's a short going out, coming back the controller may trigger a reset. No matter what you do with a resistor will change what's happening internal to the chip, if the chip wants the singal at ground it will take the singal to ground including the 5V you put on the other side if that resistor is just current to drop that voltage to ground. You're basically creating a pull up, but if it's ground on the other side the pull up resistor just dissipates power in the form of current running through it. 

pcbishaf wrote:

...there is no voltage coming off this pin going to the RESET.L line. Therefore, the +4.8V being seen on the line at startup is coming from someplace else.

The RESET.L line has a 9.1K pull-up resistor. All other connections are either inputs or control circuits with tri-state or open-collector (or equiv) outputs. Thus no one device is "driving" the RESET.L line high. The +4.8V is "coming from" the pull-up; when it goes low it is because something is pulling it down.

That being said, your resistance measurement of the RESET.L line being 58K is troublesome. With the pull-up, it should be no more than 10K through the power supply. You should measure this again and swap the meter leads to see if the value decreases. Or better yet, measure the RESET.L line to +5. If you don't get 9-10K then I would suspect the pull-up is faulty.

Untitled.png

pcbishaf wrote:Also, I am not sure I have made it clear, but pin 49 of the ADB micro controller does not seem to be connected to the chip. Or at least, there is no voltage coming off this pin going to the RESET.L line. Therefore, the +4.8V being seen on the line at startup is coming from someplace else.

I believe you may shared a picture which generated some questions for me and I believe I asked, but don't remember any response. IIRC the pin looked bent, possibly sorting with neighbor, lifted, pad damage, etc.  I don't recall much other than it looked like it had been hit with a chisel and wondered if that had been addressed. At the time, I felt i needed to be. I don't recall where that ended (too many of these conversations happening at the same time and this one is far from the most troubling one (the II board is uggh)). 

If there's a short anywhere on the adb bus it could heat the chip, the chip could also have thermal protection (ie reset) so in a siutation where the chip senses a problem it does what it can (reset) to protect itself. I'd have to look through 2 pages of details and stuff, but what really provides more details than speculation and experiments is the pictures of the signals. I'm assuming you've already verified the transistor and other components between the adb port and controller are good. Those would be my first look because one thing that's not really understood is adb with the GS it that it was NOT designed for hot-swapping (and maybe I said this already) it wasn't until a second or thrid generation mac that the circuits got the protections needed for hot-swapping on the bus. That coupled with the way the ADB components age and the way these things are known to have issues makes shorts/opens/random crap totally plasuable. 

As is alway the case with troubleshooting... there's a logical path and progression.  Start the SIMPLE as bare bones as you can get.... check the all the simple and easy things to rule them out, work up and out to the more complex. Transistors are like the wild card in all these cases because when one failes the results can present in so many different ways. They do short internally and if that's the case, the adb controller may sense and reset to protect. 

When you say you measure resistance on RESET_L is close to 58k how are you measuring that, ie  where are your probes? When you describe turning the power on and resistance changing what does it change to low resistance or high resistance that slowly drops back to 58K? Is there a cap along that path and could you be measuring the resistance through the cap? Measuring resistance through a cap (intentional or not) charges said cap which produces a changing resistance.  Powering on a system charges the caps which... is also a really high reistance measured across the cap.  Sorry, but this is why this is so difficult.... saying these things you say may be clear to to you but from my perspective I can see 10 different things you MAY be doing and still have no idea if any of those are even close to what you're really doing. Hard to make progress like this... just my $0.02.  Someone may know folks in your area that could help but seriously I think we'll untimatley find this is either someithing extremely basic or pretty difficult to fix. How far is Chicago from Springfield? There''ll be a bunch of Apple II folks there next week and I don't know who brings what for equipement, but they have a solder night where you may be able to find someone who could help. Equipment will be important and I don't know what people bring with them, but maybe someone will have the right tools and skills.... just a thought.