No clock signal to the Shifter

[russellnash]

Thanks for the replies. The board certainly appears to be designed to use two chip TOS, but I could try it with six chip also.
The L18 inductor which does look dark grey in the middle is not open circuit, but I could replace it anyway.
Most of the transistors are new because they were missing. I haven't soldered them close to the board like the originals for easy identification.
Pin 2 is stuck high without the shifter chip.
Replacing the U41 chip doesn't make any difference.
The measured values on the chip are:

1 4.87V
2 2.41V
3 0.4V (pulses on the logic probe)
4 4.87V
5 3.15V
6 1.67V
7 0V
14 4.87V

[stephen_usher]

Looking at the schematic again, those value on the 74HC74 seem reasonable. Hard to tell without an oscilloscope if they are correct.

If the oscillator is oscillating then you should see that on the base of Q8. This should be controlling the voltage on the collector of Q8, mirroring the oscillation. The oscillation should then be seen on the base of Q6, which will control the output (32MHz) seen at the Shifter. As this is constantly high it suggests that the Q6 is constantly enabled. If the inductor L19 isn't passing any current then the base of Q6 will be dragged low eventually.

[russellnash]

Thanks for the short explanation Stephen, it helps because I don't always understand what's going on when I look at a schematic. My electronics knowledge goes back to 1987 doing "O" level with relatively little experience since. I will have another look and fiddle around in the next days.

[rubber_jonnie]

Thanks for the replies. The board certainly appears to be designed to use two chip TOS, but I could try it with six chip also.
The L18 inductor which does look dark grey in the middle is not open circuit, but I could replace it anyway.
Most of the transistors are new because they were missing. I haven't soldered them close to the board like the originals for easy identification.
Pin 2 is stuck high without the shifter chip.
Replacing the U41 chip doesn't make any difference.
The measured values on the chip are:

1 4.87V
2 2.41V
3 0.4V (pulses on the logic probe)
4 4.87V
5 3.15V
6 1.67V
7 0V
14 4.87V

If swapping U41 doesn't change anything, I'd be looking closer at the circuit (See the earlier schematic) to make sure all the points in it are connected, and if you haven't tested the transistors you might want to. I include any originals in that.

Also, the scorched inductor looks to be L22 which is a key component in the circuit for Y2 (32Mhz xtal), but not only that, the inductor at L20 (Just to the right of the HC74) also looks to be slightly scorched too.

With regards to the transistors, if they are not direct replacements, so identical parts to the originals, but equivalents, don't assume that the pinout (Collector, base, emitter) is the same and use the silk screen as a fitting guide. I had this catch me out once before where the transistor needed to be reversed 180 degrees or needed the legs crossing.

Also, if they have been taken from another machine, if you haven't tested them you may have a dead one. I always test even new transistors with a component tester, especially since it will always tell me what legs are CBE too.

[stephen_usher]

To add to @rubber_jonnie 's comments, even transistors with the same part number but from different manufacturers can have their legs changed around. Not only this but if there's a sub-model marking this can be significant too. The whole transistor marking business is a mess.

[rubber_jonnie]

To add to @rubber_jonnie 's comments, even transistors with the same part number but from different manufacturers can have their legs changed around. Not only this but if there's a sub-model marking this can be significant too. The whole transistor marking business is a mess.

It certainly is...

[Smonson]

Hullo, just wanted to put my 2c in. The 32MHz clock going into the shifter is very weak and it can be hard to detect it if your scope is too low impedance. A better test is the 16MHz coming out on a different pin, which is buffered.

The shifter also plays an integral role in generating the clock with positive feedback so it is possible that it's non functional.

[russellnash]

I have only a logic probe and multimeter. I have some pulsing on the output of the shifter. The shifter is confirmed functioning. I have no way of knowing which way round my transistors collector and emitter are so, I just tried swapping round Q8. I ended up with pulsing at pin 2 but nothing out at pin 39. I then swapped Q6 and ended up with nothing. So I guess they were the right way round (flat side emitter-base-collector). Since turning them back around I have nothing coming out of pin 39 of the shifter but pin 2 is stuck high still. I also have no display on the monitor, but if I take a voltage measurement at the left side of Y2 it shows a white screen when it shorts it out. The burned inductors have been renewed, but it didn't make any difference.
One thing I have noticed is that Q7 and Q9 are shown on the schematic as 2N3906 transistors, but on the board they are 2N3904 transistors. They look original.
I seem to remember in the Atari service manual it says replace the crystal when there is no 32MHz. I believe these crystals are no longer available?

[ijor]

here is no clock to pin 2 of the Shifter. The clock crystal is stuck high ...
I have only a logic probe and multimeter. I have some pulsing on the output of the shifter. The shifter is confirmed functioning.

Please don't take it wrong, but I think that not using proper instruments is a bad idea. You don't need a high end scope. There are very affordable scopes, and even cheaper USB based logic analyzers that you could use to reliable measure a clock, especially when you don't care too much about the accuracy.

If indeed the Shifter output clock is working, that means that the 32 MHz input clock is working. In turn that would mean that you wasted a lot of time trying to diagnose the crystal circuit, just because you didn't probe the 32 MHz clock with a suitable instrument.

Again, sorry if I misunderstood the issue.

[Smonson]

Ijor is of course totally correct. If you are seeing output pulses from the shifter (that being, the 9 RGB pins plus MONO), then the system clock is functioning and quite a lot of other stuff is also working. You are probably even running some ROM code from the beginning of TOS, because the palette registers would need to have been loaded with values for anything other than 1s to come out of the RGB lines.

Maybe try pin 43 of the GLUE chip, that's a 0.5MHz clock generated by dividing the 32MHz original clock repeatedly. I think that's far more likely to be recognised by your logic probe due to the slower rate.