News:

Printed Amstrad Addict magazine announced, check it out here!

Main Menu
avatar_WacKEDmaN

more issues.... clock?

Started by WacKEDmaN, 02:06, 19 November 22

Previous topic - Next topic

0 Members and 1 Guest are viewing this topic.

WacKEDmaN

so couple of nights ago i got usifac working again on my board... played around a bit and it was all looking good...
went to bed that night.. get up the next morning and switch the CPC on... usifac line was showing... but i had to go do somethin for 5mins... when i came back it was locked up with corrupt pixels on the screen...i reboot... and nothing but black screen.....

this is the second time usifac has killed my board (MC0001A replica)

so i changed EVERY SINGLE chip (except the AY)...but the issue still remains the same....

im hearing the tape relay kick in when switching on...
i see weird looking waveform on data lines (one is always low)...
and sometimes after a few seconds of power on the waveform on the data lines changes to a very triangle wave.....

so i started looking at the clock circuit...
i cant really test the 16mhz clock with my scope (but it doesnt look like it normally does)..should be almost 5v peak right?...im seeing it quite low
the 4mhz clock is at 4mhz...but the levels look very low again (800mV peak?)...see attached pic

could the transistor be pulling the clock down?

Bryce

Any chance you've got the 1:1 / 10:1 switch set wrong on the scope probe? Your signal is 525mVpp, about 10% of what it should be.

Bryce.

WacKEDmaN

#2
Quote from: Bryce on 10:38, 19 November 22Any chance you've got the 1:1 / 10:1 switch set wrong on the scope probe? Your signal is 525mVpp, about 10% of what it should be.

Bryce.
no sir!.. 1x on the probe switch....

that image was from pin 39 on the GA... i guess i should check it after the transistor...but out of the GA it should be higher than what it is...right?!

ive got a replacement transistor for the clock line... ill try replacing it and see what happens....

Bryce

Ok, you should definitely be using the 10x position and also setting the software that it knows that you have chosen this. The bandwidth on 1x will be extremely low, maybe only 2 or 3 MHz.

Bryce.

WacKEDmaN

thats interesting...coz ive never needed to set it to 10x before to see the 4mhz clock....
apparently the bandwidth is 80mhz on this probe in both the 1x and 10x positions...
the scope is only 20mhz ..hence why i cant check 16mhz clock!

but ill give it a try none the less!

Bryce

#5
Wow, show me the datasheet (and price) of a scope probe that claims 80MHz on the 1x setting!!?? It must have cost you a fortune.

Here's the specs for a Tektronix P2220 scope probe that will cost you over €210 just for the probe. 10x bandwidth = 200MHz, 1x bandwidth 6Mhz. https://www.tek.com/en/datasheet/passive-voltage-probe-1x-10x

If your probe claims 80Mhz, then it can barely manage 2MHz on the 1x setting.

Bryce.

Edit: Sorry if that came across as sarcastic, it wasn't my intention. But measuring frequencies and proper use of a scope is a lot more complicated than most people realise. The results are effected by so many factors such as attenuation, aliasing, signal reflection, termination, sample rates, Nyquist theorems and so much more. It's not just a matter of sticking a probe on the PCB and getting a nice wave on the screen. If you are really interesting in diving deeper, I can give you a few links and book recommendations to read up on the subject, but be warned, it's a can of worms that can keep you busy for months.

WacKEDmaN

#6
oh i know scopes are a bit voodoo (not as much as antennas!)... and i know im really limited with the cheap scope i have..but its all i can afford...(its better than nothing!)

and i was just quoting what the spec sheet of the probe said... (hantek PP80B)
Attenuation ratio: 1x & 10x
Bandwidth: 80mhz
EDIT: just looked up the probe specs on their site..this isnt on the spec sheet with the probes!
System bandwidth   x1:DC-6MHz; x10: DC-60MHz

i have been through classes on using scopes....
and i watch alot of EEVBlogs scope usage vids and teardowns..
but it seems i forget alot!  :picard:
..and i honestly do appreciate the help!

ok so i just switched the probe to 10x and set the software to 10x

the levels are the same! (maybe slightly higher 800mVpp)

pin 39 of GA..
You cannot view this attachment.

but out of the transistor is looking the right levels...but isnt as square as ive seen it... (right side of R141)
You cannot view this attachment.

WacKEDmaN

#7
ok...so z80 is getting the "good" clock.. so i guess its not the clock circuit....

data in lines on the ram is looking horrible... each chip is showing something different.. only 1 looks close to normal...

also the data seems to change... like soon as switch on..i see the correct levels (not the correct waveform tho!)...and like maybe half a second to a second later, i hear the tape relay kick in and the data changes... like some will go full high..some full low... some with a lot of noise...

its looking like 153s... but i have like 8 spares that ive swapped through and still only get a black screen....and same issues on ram (rams been changed too!)

for example....

pin2 on ram ic120
You cannot view this attachment.

and pin 2 on ic124... i see it as above...until the tape relay switches (automatically a few ms after power on)..then i see this...
You cannot view this attachment.

WacKEDmaN

#8
very very strange.....

looks like it was a combo of a bad 153...and possibly a badly seated ram chip....

pulled all the 153s and the ram... ran the diag from the lower rom board... got all errors (of course!)
then put one 153 chip in at a time testing the diag with each one... till i noticed the pattern change all over the place.. thats the bad chip!.. replaced it with a good one... then moved to puttin one ram chip in at a time..testing each one along the way... all good now!...

im thinking this is probably the same issue as the other MC0002C board im trying to fix aswell... ive still got to socket 2 153s on it

and btw... the clock outta the GA.. pin 39...its STILL at the same low level (800mVpp)...
its boosted at the transistor to just under 5Vpp b4 going into the z80 :)

thanks for ya help again Bryce!

EDIT: if i had a better scope... id screenshot every major pin.... just for reference!

Bryce

Your scope is fine for the troubleshooting that you're doing. The "roundness" of the clock signal doesn't matter, a clock signal doesn't need to be square, it just needs to pass through a certain voltage level at the correct rate. The roundness is due to the switching time of the transistor plus the resistance of the entire circuit (and some other factors).

Even though it's now repaired, it would be interesting to know why it failed in that way and whether the expansion played any part in it.

Top marks for actually learning about the usage of a scope. Many don't seem to realise the importance. But you have fellow Australian (Dave) to go to, who has really great videos on scope usage.

Bryce.

WacKEDmaN

Quote from: Bryce on 11:48, 20 November 22Even though it's now repaired, it would be interesting to know why it failed in that way and whether the expansion played any part in it.
id really like to know that definitively aswell!...
i was having all sorts of issues with the usifac expansion... and then it just came good...but it didnt last and seemed to take out a chip... and now it seems fully rocksolid again (apart from some firmware issues introduced in the last update)
im fairly sure it killed a 153 once before aswell....

thanks for the encouragement Bryce! as i mentioned.. i tend to forget alot.. comes with not using a scope regularly i guess...
i think im finally getting my head around the whole system (only took what?.. 2 years!)

Bryce

It would be very unusual for an expansion to damage one of the 153's as they are only connected to the expansion via address lines and these aren't controlled or driven by the expansion (or at least they shouldn't be!). Do you have issues with static charge in the room that you have the CPC?

Bryce.

WacKEDmaN

#12
Quote from: Bryce on 15:22, 20 November 22It would be very unusual for an expansion to damage one of the 153's as they are only connected to the expansion via address lines and these aren't controlled or driven by the expansion (or at least they shouldn't be!). Do you have issues with static charge in the room that you have the CPC?

Bryce.
not that i know of... i play with a fair bit of electronics in this area and have never had an issue... (arduinos, esp8266s, esp32s, rpis, z80-MBC2.. PC and now the CPC)

i havent seen anything like sparks jumping...but i really should be using a ground strap.. just hard to setup being on 2nd floor level..i dont really wanna hook up to the mains earth!..

maybe it was due to a bad connection on the extension?.. like bridging a couple of fingers on the edge connector from trying to get the usifac to boot correctly...its just strange that i got usifac working... switched it all off..went to bed.. got up next morning and switched it on...without touching the usifac.. and within 5mins the CPC locked up and failed..
i never have had any connection issues with your lower rom board...or the DDI-5 on this board...(i even swapped the connector from the lower rom board to the usifac and had the exact same issues)

i killed the PIO on my CPC back in the day from soldering up a joystick while machine was up and running... seen some sparks jumping from the tip of the soldering iron to the connection i was soldering..i was like thats cool!..... then keyboard went all funky!.. had to put it in for a repair!.. that was one hard lesson to learn! (im pritty sure it was just the PIO that was changed, it was a new chip and socketed when i got it back.. dunno if the AY-3 and 145 where changed tho)

EDIT.. just comparing... the edge connector 'fingers' on this replica board look slightly wider than on the MC0002C board i have here.. and coz they are wider, they are also closer together and easier to bridge if the connector isnt bang on straight

Bryce

What's wrong with using the mains earth? That's what I use. There's even plugs designed specifically for this: https://cpc.farnell.com/vermason/j6303/earth-bonding-plug-3-x-10mm-stud/dp/SA00898?st=earth%20plugs

(Link to Farnell, because they use CPC in their URL :) )

Bryce.

WacKEDmaN

#14
Quote from: Bryce on 19:34, 20 November 22What's wrong with using the mains earth? That's what I use. There's even plugs designed specifically for this: https://cpc.farnell.com/vermason/j6303/earth-bonding-plug-3-x-10mm-stud/dp/SA00898?st=earth%20plugs

(Link to Farnell, because they use CPC in their URL :) )

Bryce.
...here...we're told specifically NOT to do it!
might be ok in UK...but not good enough for 'straya! (find me one of them^ with an AU plug! good luck!)..UK plugs are fused too IIRC

if theres any leakage.. or a device shorts...the earth wire becomes hot...(sinking the current to earth..and kicks in the RCDs)

ill have to go find Daves vid on setting it up... see what he did in his lab...(iirc strapped to a cold water pipe)

found a coupla good ones...
old school Dave!

WacKEDmaN

#15
...hmm seems Daves setup is mains earth referenced...
seems i can just buy a standard socket... remove the active and neutral pins and just wire directly to the earth pin...
..but i still dont like the thought of any leakage onto the earth rail....

oh geeze..how olds this one?!

pelrun

Mains earth is probably fine; any appliance you have with an exposed metal chassis and your plumbing is going to be directly bonded to it, so you should already be painfully aware if there was a fault on the neutral. And a big resistor in series won't hurt, either!

Still a good idea to second guess anything relating to a mains socket though...

Bryce

#17
Quote from: WacKEDmaN on 03:05, 21 November 22
Quote from: Bryce on 19:34, 20 November 22What's wrong with using the mains earth? That's what I use. There's even plugs designed specifically for this: https://cpc.farnell.com/vermason/j6303/earth-bonding-plug-3-x-10mm-stud/dp/SA00898?st=earth%20plugs

(Link to Farnell, because they use CPC in their URL :) )

Bryce.
...here...we're told specifically NOT to do it!
might be ok in UK...but not good enough for 'straya! (find me one of them^ with an AU plug! good luck!)..UK plugs are fused too IIRC

if theres any leakage.. or a device shorts...the earth wire becomes hot...(sinking the current to earth..and kicks in the RCDs)

ill have to go find Daves vid on setting it up... see what he did in his lab...(iirc strapped to a cold water pipe)

found a coupla good ones...
old school Dave!

That's a different issue he's speaking about, ie: testing devices where the negative of the device is not ground referenced such as switched mode PSU's and USB lines etc. These problems occur because your scope is (or should) always be ground referenced and you could cause a short through your scope.

When it come to anti static, you can only use earth (via a 10M resistor) to be certain that you are properly ground referenced.
Yes, you could use any socket, or you can even use the 4mm ground socket on front of your scope which is also directly connected to the mains ground.

If you are working on something that is NOT ground referenced such as a switched mode PSU you need to use differential probe (rather expensive) on your scope or use a battery powered portable scope.

Bryce.

WacKEDmaN

yeah Bryce... i should have mentioned that vid wasnt entirely on topic!..still..was a good refresher none the less!
good idea on using the ground point on the scope...i would have never thought of that!..thats probably what ill end up doing when i end up getting an antistatic strap and lead...

anything above about 12v@2amp i wont touch ...im too scared! ???


copped a coupla small shocks from mains in the past.. luckily just across my hand.....never again!...

this is probably gonna sound weird... but i can "feel" the field around a 230v step-down transformer!..not sure if its the power im feeling or the magnetic component!...when i "feel" it thats when i back away! LOL

Bryce

I'm working on a device at the moment with 400V DC on the PCB. I "feel" the voltage and back away when my elbow jerks and a sharp pain shoots up my arm! :D

Bryce.

WacKEDmaN

#20
LOL...DC aint that bad comparied to AC tho... DC might make ya punch yaself in the head....but it wont grab hold of ya like AC!
id rather be hit with 400VDC than 230VAC!

..i should have said.. i can feel it from a few cm away....dont need to actually touch anything...


...damn..this thread has gone through some topics already!...

Bryce

It's your thread. It can be any topic that helps you with your diagnosis and understanding of electronics.

Bryce.

WacKEDmaN

...thanks for the help and encouragement again Bryce!

i should be right for a while now!.. if this board doesnt fall over again!

WacKEDmaN

well... i knew that last comment i made would come back to bite me.....

all day yesterday i was using the usifac.. switched off cpc lastnight..went to bed... got up this morning..turned it on... usifac showed up as it should...

but....keyboard is dead!...(joystick too)

boot up the diag rom.. no beeps for good/bad memory... 
and it locks up at the CRTC line... (swapped out for spare CRTC..still locks up)

this is absolutely driving me crazy... it was workin fine..switched off and back on a few hours later n dead....

this HAS to be something up with the usifac expansion.... ive only ever had issues when using it.. never with the lower rom board and never with the DDI-5...

i dunno where to go from here..im really not having much fun... i havent got a spare AY-3... and i know the 28 pin jobbies are getting hard to find...

WacKEDmaN

#24
...ok so its not the AY-3...

i got given a new PIO the other day (after i broke leg 20 on my original PIO).... thats the dead chip...
replaced it with my broken one and it now keyboaord and sound works....

SMH... WTF! how?!

EDIT: was thinking about the usage of the scopes ground point for anti static control... my scope is USB ..and afaik usb is isolated (or is that just the data lines)
...i could very simply use the exposed chassis of my LCD monitor i have setup in my box... i know thats directly connected to earth

and that brings up another possible issue.. ...that LCD monitor earth is all ive got grounding this whole setup....
like the cpcs being powered by a isolated transformer, and through a buck converter.. so no real earth point on its power...  the cpc gets its earth through the gbs8200, and the vga cable to the lcd monitor which has the earth point and is powered separately....
i dont know but maybe could be part of the problems...

im eventually going to setup an old ATX power supply to run it all off the same psu

Powered by SMFPacks Menu Editor Mod