Calling out for Electronic's wizard help

[LinkoVitch]

Hopefully someone with far greater electronics skills than me, or the diagrams will be able to answer this one

 

I have a couple of dead (ish) JagCD's (really?? moi?? ), two of the units are not detecting the power connected to the units. I have gone through the Manufacture voltage tests and found that TP88 reads a bit low (0.93v instead of 1.35v +/- 10%). My own deductions have this as the "reference input for reference voltage generator" (Pin 4 U3) on digital servo processor.

 

Now I have the laser transport completely disconnected to be able to access the test points, I am assuming that this is how it would be tested in the fab. The JagCD is connected to a powered up Jag and I have confirmed both primary voltage regs are on and supplying the correct voltage, (not de-soldered jumpers to check output current from them tho).

 

I cannot see where this pin is getting it's current from, or if it is being supplied from the IC itself (which I suspect not as it's a ref voltage?)

 

Not sure if this is another symptom of the actual fault or the actual fault itself. So my question ultimately would be, does anyone know where the unit detects the power is present on the PCB, or if this is likely to be the actual issue? Was hoping for a nice obvious blown cap or bubbling IC, but nothing is leaping out at me

 

Any ideas?

[+Stephen Moss]

Can you explain what you mean by "not detecting the power connected to the units" as I would take that to mean either in main incoming power from the CD power unit or the Jaguar power which I would expect to be detected by monitoring the Vcc pin on the Jaguars cartridge port (at least that is how I would do it being the most obvious method).

 

That said, as you indicate the primary regulators U5 & U13 are outputting the correct voltage it would appear the unit is receiving power from its main supply. It also indicates that it is detecting the Jaguar is powered as from the 7 page pdf circuit diagram file I have I believe (although it is by ne means clear) with there being on other obvious source within the CD unit that the Jaguar cartridge port is the source of the Vcc driving T5 which in turn drives the on/off (standby) pins of U5 & U13, if that was not being detected U5 & U13 would not be producing an output.

 

As for TP88/VRH using the find vacillity these only appear once on the circuir diagram suggesting it is not a driven input. The data sheet I have found for U3 is not exactly clear as it states that VRH is an input pin however it is software controlled as to whether it is used as the reference source or if 1 of 4 internally generated reference voltages (0.9V - 2.75V) are used.

As there appears to be no voltage source on the CD schematics to drive it as an input I can only assume that the voltage you were measuring is either noise pick up or the pin becomes a monitor point for the selected internal reference voltage when not used as an input. It would make sense to use it as internal reference monitor pin however the data sheet only refences the pin as an input so either it is an undocumented feature (at least in that version of the data sheet) or the source of the measured voltage if it is driving the pin is not defined on the schematics I have.

 

It is unclear to me if the selection of the reference voltage was factory set or reset at every power on by the CD BIOS, if we were to assume that the voltage you read is the result of the pin being used as a monitor point for the internal reference voltage and not an externally derived signal then the possibilities at to why is it not the expected voltage are...

1) The document you are referencing for the test voltages are not specific to the setup used for the Jaguar CD unit

2) The document you are referencing for the test voltage is specific to the Jag but the spec was changed at some point resulting in some CD units using a different reference voltage

3) Data corruption of either the factory set or power on set value is resulting in the wrong voltage being selected

4) Assuming U3 is getting the correct supply voltage and there is not voltage selection data corruption the internal reference generator is defective.

 

I don't know it that will be of any help as a lot of the TP88 related stuff is a best guest based on what little information I have to go on.

Do you know if the CD PCB is multilayered or just double sided, if it is just double sided the source of any external voltage driving TP88 may be traceable, in which case if TP88 only appears to be connected to pin 4 U3 then you are definitely measuring noise or the selected internal reference voltage.

[LinkoVitch]

Thanks for the response Stephen. I'll try to answer your questions

 

What I mean by not detecting power, is the Jag displays the "Power cable not connected" graphic when the unit is connected (and the power cable is connected).

 

My tests with the TPs are all following the Hardware testing document Atari sent to the fab for production testing. Things may have changed over time, I need to strip a mostly working unit and test TP88, but it's a bit of an arse (shielding off, laser out etc). I have some units that need transports changing but otherwise detect power so I plan to test TP88 on one of these units to see if it's different.

 

You are correct RE the ON voltage for the regulators comes from the Jag.. With no Jag connected the 2 regulators are in off mode (pin 2 high IIRC), with the Jag connected and powered on, these are switched on, which makes sense, rather than have them there burning power with the unit off.

 

All my TP voltage tests were carried out between ground and the TP with the JagCD PCB powered and connected to a powered up Jag.

 

I was wondering if this could be simply connected so that it could be tested to ascertain validity of a unit in production. The track I can see only goes to pin4 and the TP88 point. I cannot see if it travels under the IC itself (pesky SMD! ) and as I said this may not be the cause of the fault.

 

What my primary question is regarding is where on the PCB the software is able to determine power is connected to the unit, is there a bit set in a hardware reg that indicates all powered and correct, and how is this bit set. It could be that this device is also involved in that process, or it's something elsewhere and this is another symptom.

 

Do you have any details on where the ref voltage being used is set on the PCB? any likely candidates for failure? (I'd imagine the driver processor for the laser shouldn't be a likely part to fail?)

 

Thanks for your help.

(have attached the test document I found in case you don't have it.)

 

 

testpro4.pdf

[+Stephen Moss]

What I mean by not detecting power, is the Jag displays the "Power cable not connected" graphic when the unit is connected (and the power cable is connected).

 

What my primary question is regarding is where on the PCB the software is able to determine power is connected to the unit, is there a bit set in a hardware reg that indicates all powered and correct, and how is this bit set. It could be that this device is also involved in that process, or it's something elsewhere and this is another symptom.

 

Do you have any details on where the ref voltage being used is set on the PCB? any likely candidates for failure? (I'd imagine the driver processor for the laser shouldn't be a likely part to fail?)

 

It is going to take me a while to digest that document, particularly without a disassembled CD unit to reference.

 

In the mean time it can be easy when changing jumper position to miss putting one back into operational mode, while I doubt that is the cause of the problem probably a good idea to triple check, particularly J223 & 224 on the regulator outputs.

 

To help narrow down the problem I presume you get the same error message when no power unit is connected to the CD unit, if so then we need to consider two things...

1) How does the Jag know there is a CD unit attached?

I suspect the VCC from the Jag powers EPROM U10 and it runs some kind of checksum on that to identify it as the CD unit in a similar way it runs a checksum on a cart.

 

2) If the CD unit is unpowered what is generating the message?

Presumably the answer would be the Jag BIOS, so how is it detecting a lack of power?

It would be reasonable to think the Jag is looking for a monitor voltage coming from the CD unit in a similar way that the CD unit uses the VCC from the Jag to enable its voltage regulators.

Consequently I would try removing the CD unit and measuring the voltages on all the VCC pins and the 9V pin of the Jaguars cartridge connector, the connection that is absent of any voltage is probably the incoming supply detect (if it is using this detection method), if I was designing it I would be monitoring the raw incoming 9V supply.

If the 9V pin on the Jaguars cartridge connector does indeed have no voltage on it I would then plug a power unit into a working CD unit and check to see if the 9V from it is being fed to the 9V pins on the connectors, if it is then compare that to the non functioning units.

 

Do you have any details on where the ref voltage being used is set on the PCB? any likely candidates for failure? (I'd imagine the driver processor for the laser shouldn't be a likely part to fail.

 

I think there are only two logical options as to where the ref voltage is stored, the first is that it was factory set into U3 and stored there, however as the device does not appear to have any type of memory that seems unlikely. Consequently the more likely option considering the only connection to U3's serial data pins comes from the Microcontroller U7 is that the values are written into the Micros program code and written by it into U3 at power up before the microcontroller enters its main loop.

But I have no clear evidence or documentation to support this, it just seem the most likely scenario.

[LinkoVitch]

The CD Units act like a jag cart, they have a ROM in them that contains VLM and BIOS, this boots like a normal cart, the Jag doesn't "know" or detect special hardware, that would be a very short sighted design plan leaving no possibilities for anything new to come out after the initial launch, and all the code for those devices would need to be written in advance.

 

The code is obviously checking something (I imagine a magic memory address, or perhaps trying to talk to butch or something, and from the result determining if the unit has power. I am wanting to know how this happens (software) and where the voltage test is done (hardware). I doubt there is a programmed ref voltage stored in an IC, more "does pin n have power, if so, set memory address $xxxx to 1 or something.

 

Of course it could be more complex and perhaps communication to the CD subsystem happens to check it is happy etc, there may be error codes etc, but I would assume the test is something simple.

[+Stephen Moss]

 

The code is obviously checking something (I imagine a magic memory address, or perhaps trying to talk to butch or something, and from the result determining if the unit has power.

You could be correct however that would then suggest that the "connect power cable" error message is the only error code displayed for any problem with the CD unit, which would be a rather poor and misleading choice as it suggest there is no incoming power from the power unit rather than alluding to the actual problem as there could be main power coming in, just no power coming from the internal regulators or something like a defective Butch.

 

Hence to me (although not necessarily Atari as they have their quirks) it would make more sense for that message to be solely used in reference to the main incoming 9V supply at the power socket, particularly as you have established the correct voltage is coming from the regulators. Thus my suggestion to check the 9v & 5v voltages on the connectors incase the voltage is being monitored within the Jaguar and not within the CD unit as I cannot find any point in the schematic for the CD unit where the power appears to be monitored although that does not mean it is not being internally monitored somehow.

 

For now I cannot think of anything else to try on the hardware side other than to suggest checking the supply to every GND & +5 pin on Butch as the error message provides no frame of reference as to where the problem could be other than at the power in pin or how that is being determined, so unless the test document you provided subsequently shines some light I think the only way you will find where & how the voltage is being monitored will be to trawl the BIOS source codes.

 

I doubt there is a programmed ref voltage stored in an IC, more "does pin n have power, if so, set memory address $xxxx to 1 or something.

Sorry I obviously was not clear, the part of my previous post to which you are referring post was relating to the selection/setting of the reference voltage used by U3/at TP88, not the power monitoring issue.