issues with SC720 new build

[shadow 338 (shadow338)]

I'm seeing this strange behavior on a SC720 new build where "nothing" seems to work not even the clock signal seems right. (attached logic capture).

I ordered new PCBs, new components from Digikey (minus the SIO chip) and when I turn ON the computer with the SCM rom selected, I don't get anything on the serial nor the self test LED scrolling.

I then hooked up a logic analyzer and I was surprised how weird things look like.

Since then, I checked all solder joins, re-applied heat on all of them, cleaned the PCB and confirmed that all chips are getting ~4.8V.

I then re-checked the values of all used components (resistors/capacitors/crystal/etc) and the all seem correct.

I built a few RC2014 family computers already and never had major issues. Also I'm using the same batch of components so I don't suspect of anything there.

Can you point me in the right direction on what I should be doing next?

Much appreciated!

[Tadeusz Pycio]

Digikey sends out fake CPUs?! Zilog stopped labelling its processors in this way around 2004, and the picture shown shows the date 2012.

[shadow 338 (shadow338)]

sorry, the SIO and the Z80 were not sourced from Digikey, however that Z80 works on another RC2014 system (same cpu clock etc). I also tried with an original Z80 (CMOS) but the same happens.

[Tadeusz Pycio]

Ok. The waveform from the analyser is not authoritative, as it must be sampled at least twice as fast as the timing of the circuit under test. Maybe try increasing the sampling rate or decreasing the CPU clock then maybe a problem will be spotted.

[shadow 338 (shadow338)]

I sampled at 16MHz and I'm getting the same. by the way, I used the same settings from the initial capture and tested a working RC2014 unit and all the waveforms come up correctly including the clock signal.

[Tadeusz Pycio]

In that case, I would start with the clock and its components, because it is so unstable that all sorts of strange things can happen.

[Steve Cousins]

You need to fit a jumper shunt to JP2 or fit switch SW2 (but not both). Without a shunt or a switch neither ROM is selected.

[Steve Cousins]

Oh, I've just noticed what looks like a small switch (SW2) half hidden behind one of the chips, so forget the above suggestion!

If the clock signal is really as erratic as shown in your picture then I agree with Tadeusz, look at that first.

[shadow 338 (shadow338)]

Thanks Steve and Tadeusz. I will look into the clock circuit and report.

[Fernando Carolo]

This might be a silly question, but are you sure about the power supply you are using? The SC720 has its own voltage regulator and requires 8 to 15 volts to operate, not 5V like other Z80 based boards.

--
You received this message because you are subscribed to the Google Groups "retro-comp" group.
To unsubscribe from this group and stop receiving emails from it, send an email to retro-comp+...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/retro-comp/bb45fd03-1a29-463a-bfd2-0732f5c26002n%40googlegroups.com.

[shadow 338 (shadow338)]

I'm using a 9V / 1A power supply and I'm getting around ~4.9V on every chip. (just re-measured). I also tested with other power supplies (12V)

[Sławomir Zegarliński]

Hi Henrique.

I think you should check out the Z80 CPU first. I have several with the same markings, but the inscription does not match the contents of the package.

Z80CPU-1240GZ

Regards. Zegar.

[shadow 338 (shadow338)]

As I mentioned before, I also tested with a genuine Z80 (Toshiba CMOS) and I get the same result. I'm aware that the Z80 in the picture is a chinese fake one and for sure it does not work at 20MHz as it says in the marking, however a lot of these fake Z80's, despite the fact that they only work at lower frequencies and probable are not CMOS  (most likely NMOS), they indeed work. The one in the board for example is being working on a RC2014-Mini for around 4 years and never had issues as the RC2014-Mini uses the exact same clock frequency. I also don't see how a fake Z80 would mess up with the clock signal and I wish the problem was because of the CPU, unfortunately does not seem to be the case here.

[Steve Cousins]

Hi Henrique.

Here's a few things to try: (sorry if they sound rather basic ideas but it is always worth going right back to assuming nothing)

1/ Double check the supply voltage reaching the ICs (as you have already done) but also check with an oscilloscope (if possible) that it is a good clean DC voltage. 

2/ Check the reset LED comes on for about half a second at switch on then goes off. Similarly, it should come on when the reset button is pressed and stay on for about half a second after the button is released. Ideally, check the reset signal gets to the CPU.

3/ If you have an oscilloscope look at the clock signal on U16 pin 1. It should be 7.3728MHz. It won't look perfectly square but it should be a clearly identifiable 7.3728MHz frequency that reaches very near ground and 5 volt.

4/ If the clock looks wrong try removing the shunts on J3 and J4 to isolate the clock from the SIO. Look at the clock again.

5/ If the clock looks wrong try removing U12 and U13 to isolate the clock from the IDE interface. Look at the clock again.

6/ If the clock looks wrong try removing U16 pin 1 leg from the socket to isolate the clock signal from the PCB. Look at the clock again (on the pulled out leg).

7/ If the clock still looks wrong then double check component values around the oscillator circuit and look for bad joints and shorts.

8/ If the clock still looks wrong then replace U16.

9/ If the clock still looks wrong check all the pins on U16 are at 0 volts, 5 volts or have an active signal on them. I've seen odd things on clock signals if the oscillator IC has any floating inputs.

If you don't have a scope then select the SCM ROM and watch for activity on the out LEDs. Even without RAM and a serial interface SCM should get as far as outputting the LED cycle pattern. As you have a logic analyser you could just use that to monitor the clock signal, set to highest sample rate possible.

Carry out these tests with no accessories connected (no USB adapters, CF card or expansion modules). 

Once you have a good power, reset and clock report back and we'll see where to go next.

Steve

[shadow 338 (shadow338)]

Thanks Steve for the comprehensive troubleshoot guide! very helpful. voltage seemed fine but the erratic clock was still puzzling me. After step 5 or 6, in had the idea of removing the SIO from the socket and bingo. When U4 is selected the LEDs scroll one time and then stops which is what I was expecting.

I got the SIO from Aliexpress so maybe I got a lemon? I have no good way of testing as I don't have another SBC that uses one. Well, I guess the next step is to order another SIO from a trusted source and try again.

Btw, I did a new logic capture and things look much better now. One curious thing is that the clock is not consistent when I compare to other clocks from other Z80 computers I have. It seems that there is a pattern where one square sometimes is shorter than the others.

I will report back once I get the new SIO!

cheers.

H.

[shadow 338 (shadow338)]

well, if I sample at 12, 16 or 24MHz I see the pattern in the clock signal, but I sample at 8MHz OR 48MHz the square is always perfect. Go figure. maybe something erratic on my cheap logic analyzer.

cheers.

[Sławomir Zegarliński]

Hi Henrique.

This is results from the measurement method. If the sampling frequency is not a multiple of the sampled signal frequency, there may be a different number of samples per period. When the sampling rate is much higher than the measured rate, the problem is negligible.

Regards.

[shadow 338 (shadow338)]

SC720 is now working! Confirmed, I got a SIO lemon from Aliexpress.

Thank you all for the tips!

H.

[Tadeusz Pycio]

You've probably received a faulty SIO/2, as its markings look authentic unlike the CPU (although anything is to be expected these days). Such a marking and in this colour usually came from the Indonesian factory, the number 4 is also characteristic in this case. It is good that your problem was solved, although it is hard to explain the behaviour of the clock with a damaged SIO.

[shadow 338 (shadow338)]

Tadeusz,

Indeed it's difficult to explain why the faulty/fake SIO caused all sort of issues in the circuit. My initial limitation is that I don't have any other circuit that uses the SIO otherwise I could rule that out (or point the root cause in this case) sooner.