DL1540 Repair

[cattlot]

Hi all, first time posting here, Currently repairing this non-booting DL1540, on the main cpu board i probed these abnormal signals being sent to the analog board, in general does anyone know what could be causing these signal instabilities?
The board and daughterboards were all powered by an external supply?
Tracing the pins back they all go through different hex inverters or diffrerent trancivers so one chip cannot be the root cause of the problem,
thanks for any replies

[indeterminate]

Your probably barking up the wrong tree.
Start with what is a  DL1540 when was it maid a photo of the cpu card.
are the power rails correct & noise free
is there a clock signal on the cpu
is the cpu out of reset
is the eprom enabled
is there activity on the data bus
is there activity on the address bus

[Swake]

Yokogawa scope?

[cattlot]

yes it is, I forgot to mention that it is an oscilloscope. Whoops. It is a yokogawa oscilloscope DL 1540

[Swake]

You write it is not booting. Is there a logo shown on the screen or absolutely nothing?

As it is a CRT if there is nothing on the screen it might be that the issue is with the CRT itself, not with the rest of the scope. Let's start with that.

[Swake]

Service Manual. No schematic but there is a repair flowchart and block diagram.

Edit: for the CL version, with an LCD screen.

[cattlot]

Hi, I believe the crt works, on the crt PCB there are no bulging caps, or burned components and when the device is turned off there is static shown on the screen, other than that I've done no other testing.

 What I think is happening is that it's not booting past a certain point because there is a problem with one of the boards, I think this is because when it turns on a buzzer beeps and all the front panel LEDs turn on and stay on maybe suggesting that the main board is waiting for a response from another.

I've found that service manual online, it's for the lcd version of the scope and does not really provide any helpful information other than suggesting the power supply is faulty, I've tested it and it provides all the correct voltages apart from the +24v line which gives ~+25v when powered on with everything connected.

The only real thing that ive found is that the ic components of the analog board get very hot after a short amount of time even with only the power going into it. The analog board has a -2v regulator and I tested it and it's ok so idk what else to look at

[Swake]

The self test suggests it can handle dead boards and report about these. Probably to a certain extend only.
I'd say disconnect everything you can except the GCP (graphics) board.
If there is still nothing on the screen, try to find out if the CPU board talks to that board and if those signals might make sense.
Also execute the checks that indeterminate has suggested in the second post.

[cattlot]

Hi, after disconnecting all boards except the CGP board nothing was displayed, showing the same behaviour as before. I did some of the suggested things and this is what I found:

The power supply is fine and has no noise that i can see.

All of the clocks (16mhz and 4mhz) on the CPU board are giving the correct frequency, notible things are that the 4mhz clock has a peak voltage of 5v and is a square wave and the 16mhz has a peak voltage of around 3.2v and is more of a triangular wave. the 16mhz goes into the main MPU (TMP68303F-16), (I think this clock is ok because the MPU controls the interface and since it boot chimes and leds light up if there was a clock issue i dont think it would even reach that point).
However the waves are very different and might be a cause for concern. On the PCB there are 2 crystal oscillators, a 16MHZ and 20MHZ, they both give the correct frequency. Since no 4mhz oscillator, it must be generated elsewhere

while poking around on the CPU board I found a data bus that looked very weird, it connects to the MPU (TMP68303F-16) , 2xRam (HM514800C), 2xFlash (LH28F800), Floppy drive interface (D72069GF), rtc chip(62423 rtc), and a parralel interface (D71055GB). The bus is very noisy and has weird voltage spikes, It is 8 bits wide and pulled up to 5v, i think one of these components have failed

The address lines of the Flash (LH28F800) and  Ram (HM514800C) ram are also connected together and connect to the MPU (TMP68303F-16) and 2 yokogawa custom chips (A6152LG and A6151LG) probably the IDP shown on the block diagram (Page 28 in the manual)

its a good sign that there is activity but the fact that the traces being distorted like that suggests a failed chip? The data lines' the weird ramping when low could be the problem? I might be going in the wrong direction tho

I have attached the clocks and bus traces on my post, I realise this is probably a much harder repair than i initaly thought it was, I can upload pictures of all the boards to make it easier to understand what im talking about
Thank you for any help you can provide

[Swake]

That 'digital' signal pulling upwards over time indeed is bizarre. I have never seen this before.
In the 3rd picture of your first post it seems also to oscillate while raising. It is pulled 2.5 V below zero somehow.

Is it some capacitance/inductance playing games?
Would think that it might somehow be found back in the power supply current consumption.

Assuming that with 'static' you mean a screen full of noise pixels. Static on the screen of an old TV with a tuner meant 'no signal'. Would you be able to inject some other video signal in the CRT to exclude that part? Of course you'll need to find were to inject it first.

[cattlot]

That 'digital' signal pulling upwards over time indeed is bizarre. I have never seen this before.

it really is strange, could there be like a capacitor charging on the pin or something? I dont really think its a problem tho and just more a quirk of the system because it only acts like that when its idle and the mpu just doesnt control it? I probably am wrong tho

Assuming that with 'static' you mean a screen full of noise pixels. Static on the screen of an old TV with a tuner meant 'no signal'. Would you be able to inject some other video signal in the CRT to exclude that part? Of course you'll need to find were to inject it first.

I dont really have a way to put anything on screen as i dont have any tv equipment, but i know where to put in the signal as one of the pins directly connect to a signal amplifier . What i mean by static is like when power is turned off the crt lights up full screen and as the caps drain I can see the indvidual traces of the crt die down, kind of like how old tvs act when turned off (best way i can describe it), Ive seen other DL1540's act the same way on videos online.

Would think that it might somehow be found back in the power supply current consumption.

wdym?

What i found out is that the bus does actually works and does some form of initalisation on the components on the CPU board, because if a pin was held high by me shorting a data pin, the boot chime didn't happen so i think that the CPU board's main components do work.

in the interconnect of the CPU board and AQC board the bus does connect to it but through a HC245 bus tranciver and on the AQC side of the bus it showed really weird behavior, maybe there is a problem with it? but i want to do a bit more reaserch before i do anything

Ive also got a seperate question, on the analog board the +5v and -5v have low resistances (49 ) and (71 ), could this be a sign of a faulty chip? or is it a red herring, as ive mentioned prevously the components on that board get very hot so like idk

[Swake]

because if a pin was held high by me shorting a data pin, the boot chime didn't happen so i think that the CPU board's main components do work.

Best tip of the day. This is such a simple test that it is borderline genius.

if the CPU goes idle that can mean that
- it hangs for some reason, like the firmware is corrupt.
- it does not receive any 'boot data' and just stops till the next reset.
- some protection kicks in. (like it knows somehow that the video card is broken)
- it is not being reset continuously by the reset chip, else you would see the startup sequence over and over again.

Ive also got a seperate question, on the analog board the +5v and -5v have low resistances (49 ) and (71 )

That gives about 100 mA on the positive side and 70 mA on the negative side. 0.85 Watt, not really cooking power if you ask me. If you had something shorted like a dead tantalium cap it would be a different story.

could this be a sign of a faulty chip? or is it a red herring, as ive mentioned previously the components on that board get very hot so like id

Hot is relative, if steam raises up when you put a wet finger on it, then yes it is too hot    There is likely another source of power on top of the + & - 5 V as less than a watt really is not all that much.

wdym?

That mini staircase effect must have a reason. I've only seen a somewhat similar waveform probing DDR3 memory, but that is 1 or 2 nS timing not 500 nS, and likely had an entirely difference cause. I'd like to understand where this is coming from. It is a very long shot that you might be able to measure the current delivered by the power supply and find back that specific staircase in the current trace.

Remarkable is that when the signal comes down it jumps below 0 V, that is extra weird, as if there was a negative voltage available. An opamp could play that game, only this is TTL stuff. Is some near dead component feeding the -5V somehow back into digital (data) lines? Now that I re-think it, this might be a probing issue, but 2 V negative is a lot.

Somehow it makes me think of a composite video signal. But that should not make any sense.

Could you disconnect the power supplies from the rest and connect it to an electronic load, or some resistors as load. The idea is to measure ripple with a constant load and find out if the power supply is not having some lower than normal limit somewhere.