[SOLVED] Wobbly Clock

discuss both original and modern hardware for the bbc micro/electron
Post Reply
iancr77
Posts: 40
Joined: Mon Aug 20, 2018 11:16 am
Location: Swansea, UK
Contact:

[SOLVED] Wobbly Clock

Post by iancr77 » Thu Sep 03, 2020 6:35 pm

Continuing my investigations into my dead beeb... I notice it has a wobbly clock.
IMG_4639a.mov
(1.47 MiB) Downloaded 29 times
The measured frequency moves between 15.9 and 16.1 MHz. (This is from pin 6 of IC43)

Is this normal? Does it matter? If it does matter, what might the problem be? I've replaced IC43...

Thanks in advance... :)
Last edited by iancr77 on Sat Sep 26, 2020 5:17 pm, edited 1 time in total.
Ian
Intermittent tinkerer - BBC Model B, Master 128 + various bits and bobs.

User avatar
MartinB
Posts: 5340
Joined: Mon Mar 31, 2008 10:04 pm
Location: Obscurity
Contact:

Re: Wobbly Clock

Post by MartinB » Thu Sep 03, 2020 6:45 pm

Looks perhaps more like a scope threshold triggering issue? I would have expected a simple varying frequency trace to be stable and fixed at the left but slightly 'stretching' and 'contracting' on the right?

User avatar
1024MAK
Posts: 10292
Joined: Mon Apr 18, 2011 5:46 pm
Location: Looking forward to summer in Somerset, UK...
Contact:

Re: Wobbly Clock

Post by 1024MAK » Thu Sep 03, 2020 7:44 pm

iancr77 wrote:
Thu Sep 03, 2020 6:35 pm
Continuing my investigations into my dead beeb... I notice it has a wobbly clock.

The measured frequency moves between 15.9 and 16.1 MHz. (This is from pin 6 of IC43)

Is this normal? Does it matter? If it does matter, what might the problem be? I've replaced IC43...
Is the TV picture stable? If yes, it’s nothing to worry about for normal use.

If not, renew the 16MHz crystal.

Mark

iancr77
Posts: 40
Joined: Mon Aug 20, 2018 11:16 am
Location: Swansea, UK
Contact:

Re: Wobbly Clock

Post by iancr77 » Thu Sep 03, 2020 8:49 pm

Thanks v much, both.

I'm a novice with the scope, so it's entirely possible it is user error. But I fiddled around with the trigger to no avail, so I think it's probably a real effect.

The machine is dead - no video signal, just a continuous beep (this thread viewtopic.php?f=3&t=20190). I've investigated all kinds of things and replaced various ICs but without success.

I've just got an MiniPro/TL866II+, and some AT28C256s, with the hope that I can try Tricky's test ROM. Will this work? Do need to know anything special about putting an EEPROM in the OS socket?

Thanks again!
Ian
Intermittent tinkerer - BBC Model B, Master 128 + various bits and bobs.

User avatar
1024MAK
Posts: 10292
Joined: Mon Apr 18, 2011 5:46 pm
Location: Looking forward to summer in Somerset, UK...
Contact:

Re: Wobbly Clo

Post by 1024MAK » Thu Sep 03, 2020 10:01 pm

For sideways ROM use: Pin 1 is not connected on most BBC B sideways ROM sockets, on the AT28C256 this is address pin A14. So this needs to be tied to 0V/GND (use addresses 0x0000 to 0x3FFF to program the chip) or tie it to +5V/VCC (use addresses 0x4000 to 0x7FFF to program the chip). Or use a flying lead to connect A14 to pin 12 of IC76 (74LS163) to have two 16K ROM images in one physical chip :wink:

For OS socket use, first use a meter or look on the bottom of the board and look for modifications to the PCB tracks to pins 1 and 27.

Mark

User avatar
1024MAK
Posts: 10292
Joined: Mon Apr 18, 2011 5:46 pm
Location: Looking forward to summer in Somerset, UK...
Contact:

Re: Wobbly Clock

Post by 1024MAK » Thu Sep 03, 2020 10:36 pm

The OS socket on an unmodified board has pin 27 connected to the 6502 CPU address line A14 and pin 1 connected to the 6502 CPU address line A15. Modified boards normally have both pin 27 and pin 1 connected to +5V/VCC.

An AT28C256 needs pin 27 to be connected to +5V/VCC to keep the /WR line high at all times.
Pin 1 can be left connected to A15 in which case program addresses 0x04000 to 0x7FFF.

Mark

iancr77
Posts: 40
Joined: Mon Aug 20, 2018 11:16 am
Location: Swansea, UK
Contact:

Re: Wobbly Clock

Post by iancr77 » Thu Sep 03, 2020 10:39 pm

Thanks, Mark - that’s brilliant. Very clear.

I’ll try and find a moment to check it tomorrow or on Saturday...
Ian
Intermittent tinkerer - BBC Model B, Master 128 + various bits and bobs.

dp11
Posts: 1199
Joined: Sun Aug 12, 2012 9:47 pm
Contact:

Re: Wobbly Clock

Post by dp11 » Thu Sep 03, 2020 11:17 pm

Set the trigger position to the middle of the screen instead of a 22mS delay will also help. Also make sure the ground lead of the scope is very short ( don't use the ground lead that come with the scope) and is connected to IC43 pin7 if you really want to measure the clock accurately.

iancr77
Posts: 40
Joined: Mon Aug 20, 2018 11:16 am
Location: Swansea, UK
Contact:

Re: Wobbly Clock

Post by iancr77 » Sun Sep 06, 2020 3:02 pm

dp11 wrote:
Thu Sep 03, 2020 11:17 pm
Set the trigger position to the middle of the screen instead of a 22mS delay will also help. Also make sure the ground lead of the scope is very short ( don't use the ground lead that come with the scope) and is connected to IC43 pin7 if you really want to measure the clock accurately.
Thanks for the advice. I since rechecked the clock and it seems fine now, so I guess it must have been a measurement error on my part.

However, the machine still doesn't work. This is the situation:

- Minimal set-up
- Tricky's test ROM burned on an EEPROM and installed in the OS ROM socket as per Mark's instructions (I used a modified socket as intermediary between the EEPROM and the board, to get 5V to pin 27. The EEPROM has the image burned twice, so there's a copy at 0x4000 of the ROM)

Main symptom: no video output. Looking more closely, VSYNC (pin 40 of the CRTC) is held low, and HSYNC (pin 39 of CRTC) pulses briefly at power-on, but then remains low.

I'm satisfied that the CPU is executing the code on the EEPROM. Looking at data and address pins, it's clear there are bursts of activity between periods when the CPU is fetching and executing NOP instructions from the ROM.

The data and address lines are connected properly between the CPU, ROM and CRTC.

I obtained two spare 6845s - installing these in turn gives the same result, so I assume it's not a fault with the CRTC.

Current hypothesis: the chip select logic may be the problem. If this is faulty, then the CRTC may not being programmed correctly. Pin 25 of the CRTC seems to be high all the time. I haven't managed to detect any low-going pulses.

Question: does the test ROM reprogram the CRTC on a regular basis, or does it just do it once at the start? (i.e. should I see pin 25 pulse low every so often? Or is it a blink-and-you'll miss it situation at power-on?)

Thanks for your help!
Last edited by iancr77 on Fri Sep 18, 2020 7:54 pm, edited 1 time in total.
Ian
Intermittent tinkerer - BBC Model B, Master 128 + various bits and bobs.

iancr77
Posts: 40
Joined: Mon Aug 20, 2018 11:16 am
Location: Swansea, UK
Contact:

Re: Wobbly Clock

Post by iancr77 » Thu Sep 17, 2020 9:14 pm

OK.

So I pared the EEPROM down to just programme the CRTC. Then I tried it just programming Port B of the User VIA (having reinstalled it into the socket).

In both cases, it seemed like the chips just weren't being programmed correctly. The CPU seems to be fetching stuff from the ROM and executing it, but writing to the hardware isn't working.

There seem to be the correct chip-select signals, but I will recheck these.

One thing I noticed on the scope was that while the address bus shows quite nice square pulses, the data bus is a lot less pretty
IMG_4713.JPG
Is this normal, and does it matter?

I obtained a replacement CPU, but the results are the same. So, if it is the CPU, then both have the same fault. Alternatively it's something else on the board, but there's almost nothing else on the bus... IC14 is OK as far as I can tell.

I'm completely stumped... :(
Ian
Intermittent tinkerer - BBC Model B, Master 128 + various bits and bobs.

iancr77
Posts: 40
Joined: Mon Aug 20, 2018 11:16 am
Location: Swansea, UK
Contact:

[SOLVED] Wobbly Clock

Post by iancr77 » Sat Sep 26, 2020 5:16 pm

See viewtopic.php?f=3&t=20190 for details...
Ian
Intermittent tinkerer - BBC Model B, Master 128 + various bits and bobs.

Post Reply

Return to “8-bit acorn hardware”