question about microcontroler oscillator

saga continues, my repair project device uses a phillips MAB 8051 AH microcontroller.

while trying to isolate trouble with a badly shaped clock signal i happened across microcontroller oscillator.

the oscillator CSA 12.00 MT has a slightly warped sine wave out of one side and a very bad dent in the crest of the sine wave on the other side like... _ / \\__ / \\

-/------------\\

is the output of the oscillator suppose to be clean sine wave ? and same on both sides ?

oscillator is connected exactly as 8051 datasheet recommends, with the two pins from the oscillator each to small caps which connect to 0v ref

thanks for any help, robb

What value are the caps ? I think you mean you have a ceramic resonator btw, not an oscillator.

Graham

Hi,

Make sure your scope probe is set to 10x, and has a low capacitance. The probe may have more capacitance tha the small capacitors near the crystal (10 pF?) and may load the signal. So you may be measuring what it does when you load it with a capacitor.

Regards, Pieter

signal

out

on

?

which

capacitance.

near the

measuring what

Thanks Pieter, I was planning to post pics of the wave forms on "alt.binaries.schematics.electronic" in case i am not describing correctly or if it helps someone, help me, figure out if i am seing the correct thing.

i just read about intel's specs on the "oscillator for microcontrollers" and my head is spinning { loading, gain, drive, tolerance, reactance, optimal loading ...} they described using a

*MOSFET buffer* to view the signals on o-scope and setting to vertical AC measurement which seems to address the problem you were alluding to.

so i will try to verify i am not measuiring my loading if that majkes sense. my probes are the tek 10x 6109B (~12 pF)

thanks for help, robb

I'd expect them to be the same value too (and about 30pF typically - check the resonator data sheet). See any 8051 application note too. Is the 8051 version you're using CMOS btw ?

Graham

signal

out

on

?

which

resonator btw, not

thanks for help Graham,

the values for caps (grey w/black top) are 15p and 33p

after reading the intel oscillator spec i expected them to be the same value but then maybe i missed something (well, alot)

thanks for help/reply, robb

ceramic

the

typically - check the

the 8051 version

the (phillips 8051ah) datasheet says +5v depletion-load,N-chanel,silicon-gate,N500 MOS

so my slight comprehension leads me to "yes ?"

i am having difficulty finding the resonator datasheet which is why i am troubling "s.e.b" my reliable finds come from alldatasheet.com but not for the resonator

thanks for help Graham, robb

"robb" wrote in news: snipped-for-privacy@corp.supernews.com:

I was thinking of that before I saw it. Look at the thread above this one called "Search OpAmp chip as a voltage follower" for a description of using a FET to buffer a high impedance source. That same idea will let you get a look at that waveform without disturbing it much. The FET and a single resistor to a stable DC rail should be all you need, then scope the FET output signal.

FET input op-amp as non-inverting unity gain buffer might also be enough, and more likely to be close at hand.

If you are seeing 3-4 volts swing on one of the pins, then I would say the resonator is working ok. Don't worry too much about the wiggle in the waveform.

No, that's not a CMOS part and I'm puzzled why you're using it. Is this some old piece of equipment ?

Where do you buy it from ? The "CSA 12.00 MT" marking is the same as ones I've sourced from Farnell (Murata ?) and works fine with 2 x 33pF.

Graham

A FET input probe is NOT required to look at the 8051oscillator for routine purposes. A 10x probe is fine.

Graham

Yes, it'll be around 3-4 V pk-pk on one pin (Osc 2 ?) and about a couple of hundred mV on the other one.

Graham

From memory, yes they are a bit 'warped' as you put it. They're certainly not clean square waves. I wouldn't worry about it.

Graham

I doubt it. It expect it's simply different to the usual 80C51 CMOS version that I'm familiar with. That NMOS version must be VERY old.

It IS being used. That's what the XTAL1 and XTAL2 pins are for. See any 8051 datasheet.

Graham

application note

MOS

Is this some old

Yes, you are correct Graham. It is the main control board for a vintage "computer controlled" sewing machine (~25 years old)

i am trying to restore it to functioning ... as a "fun / learning" electronic hobby project it has just enough complexity not to completely overwhelm (hobby-ist) me ; micro-controller, stepper motors and controls, memory, VFD user interface and programming buttons etc.....

which is

same as ones I've

it was already on the board.

i was trying to backtrace the CLK signal going to the VFD driver IC which had some big spikey ring in the leading edges. i was led back to the 8051 and was checking various pins and noticed warped ? sine waves at the oscillator pins XTAL1 / XTAL2

thanks for your help, robb

above this one

description of

let you get a

a single

scope the FET

also be enough,

would say the

in the

Thanks for reply, that is what i see about 4 volts height on wave forms on both pins . One is slightly bent sine wave the other has a big dent in the wave

thanks for help robb

would say the

wiggle in the

a couple of

well i am seeing about 4 volts on one side (the 33 pF side) and about 3.8 volts on the (15pF) side ?

so does that mean i have a problem ?

this 8051 chip supposedly has and internal oscillator ? why would that not be used ? i mean is there a reason one would not use internal oscillator if there were one.

thanks for all the help everyone, it helps me learn, robb

Eeyore wrote in news: snipped-for-privacy@hotmail.com:

Didn't say it was necessary. So long as the wave crosses the zero point regularly enough to make the timer work who cares what shape it is? Nothing much routine about this anyway, the guy's new to it and wants to be sure of what he sees.

I was just pointing out a useful similarity with another post. If someone wants to learn, there's no doubt that adding a FET and a resistor and a DC cource is a neat way of making sure a signal involving small capacitances might be less disturbed than by loading it directly with a cable connection. It's not a question of whether it's necessary, but whether or not it's useful. Robb wants to learn, so that's a good cheap way to do it. It would confirm that the waveform really is as he sees it, and that it's not being changed by his connection to it.

Don, are you using the ground clip at the probe? long ground leads distort the picture. rw