PIC 14 Karen

[Slothie]

Originally Posted by Timbucus

Yes sorry I should have warned you that those needed a good scrub - I think they have some kind of coating on them. Sounds like the same place I got those from...

Yes, the last display of this type I used had gold plated circuit board and soldered like a dream. I'm going to get a fibreglass pen I think and clean the next one carefully and possibly even give it a dab with the old flux pen.

[Slothie]

Hmm well I finished the soldering and plugged it in and... nothing. Oh well. I've done a visual inspection and all my joints look OK. There are a few places I damaged the solder mask while soldering so I need to buzz those out to make sure theyre not broke,,,,, So I've either got a problem with the programming of the PIC or maybe the crystal isn't oscillating. I've got to have a think about what diagnostics I can do with what I have. First option is to rig up my Arduino as a frequency meter and test for crystal activity and keyboard scanning. If its not doing that then its probably the PIC.
I did read back the contents of the PIC after programming it so I was resonably assured it was OK. Perhaps I have the wrong firmware? But it is the stuff downloaded from Karens website so should do something even if its not the latest version. I think the display is OK too, because I once saw the digits flash on power up although I've not been able to reproduce that.

Interesting times!

[Phil__G]

Holding a domestic FM radio near the crystal you should pick up the 5th on 100mhz...

[SiriusHardware]

I can still send you a programmed chip or I can loan you a PicKit if you want to be a bit more flexible.

[SiriusHardware]

As regards using the Arduino as a frequency meter, since the Uno? is clocked at 16MHz and the frequency under investigation is 20MHz, won't that be far too high a frequency for the Uno to measure?

I like Phil's dodge, a simple FM receiver will typically make a loud roar/hiss under no-signal conditions and go silent when there is an unmodulated carrier on the frequency it is tuned to. Watch out for FM radios which have auto-mute though. Using a receiver to 'sniff' an oscillator like that gets around one possible drawback with direct measurement - sometimes the very act of connecting an instrument to an oscillator to measure it will actually stop the oscillator.

[Timbucus]

Would it help if I sent up a known good PIC for you - I have a good stock so could send mine and blow another - that removes one unknown?

Tim

Originally Posted by Slothie

Hmm well I finished the soldering and plugged it in and... nothing. Oh well. I've done a visual inspection and all my joints look OK. There are a few places I damaged the solder mask while soldering so I need to buzz those out to make sure theyre not broke,,,,, So I've either got a problem with the programming of the PIC or maybe the crystal isn't oscillating. I've got to have a think about what diagnostics I can do with what I have. First option is to rig up my Arduino as a frequency meter and test for crystal activity and keyboard scanning. If its not doing that then its probably the PIC.
I did read back the contents of the PIC after programming it so I was resonably assured it was OK. Perhaps I have the wrong firmware? But it is the stuff downloaded from Karens website so should do something even if its not the latest version. I think the display is OK too, because I once saw the digits flash on power up although I've not been able to reproduce that.

Interesting times!

[Timbucus]

Even a rough sample should show a predictable sample pattern as long as like 16Mhz it is not an exact frequency that risks only sampling the peak or the trough - e.g. at 16Mhz assuming a sample every 10Mhz could risk seeing only the peak or the trough of the 20Mhz so allowing a variable delay and summing the samples should show a visible pattern - I find my really crap handheld scope good enough to see activity for quick checks even though it is lower than my 20MHz Hameg.

I learned to cope with lesser tools. I still stare at Z80 code until I see the bug rather than using a monitor half the time.

Tim

Originally Posted by SiriusHardware

As regards using the Arduino as a frequency meter, since the Uno? is clocked at 16MHz and the frequency under investigation is 20MHz, won't that be far too high a frequency for the Uno to measure?

I like Phil's dodge, a simple FM receiver will typically make a loud roar/hiss under no-signal conditions and go silent when there is an unmodulated carrier on the frequency it is tuned to. Watch out for FM radios which have auto-mute though. Using a receiver to 'sniff' an oscillator like that gets around one possible drawback with direct measurement - sometimes the very act of connecting an instrument to an oscillator to measure it will actually stop the oscillator.

[SiriusHardware]

You presumably agree though that it would show just about anything but the right frequency? You'd just be looking for a change between what it does when connected to the oscillator, and what it does when it isn't?

It's different with scopes, you can put a scope on a signal which far exceeds its specified maximum working frequency and you will still see 'something'. With an analogue scope - by far my preference - the individual cycles of the signal will just merge together into a 'band' of signal in which individual cycles can no longer be resolved, which is what you would intuitively expect.

A frequency meter fed with a signal which exceeds its maximum input frequency will usually just go haywire.

A compromise: Take the oscillator signal to the input of a CMOS IC configured as a divide-by-ten circuit and use the Arduino-Meter to read the output frequency of that. If it says 2MHz, great.

Phil's simple trick of using an FM radio to listen on 100MHz is still a lot easier though. Most smartphones (and a lot of dumb phones) have an FM radio hidden away on them somewhere, although it usually needs wired earphones to function because it uses the lead as the aerial.

[Timbucus]

Of course. You could never determine any useful frequency information only the presence of a regular signal that is more than random noise or a fixed level. A sample from a lower frequency is a bit like us today peering into the subatomic space... I suppose even a meter on AC (perhaps with a negative DC bias) could show something you could view and it would be a clue. I was merely thinking of the difference between oscillation happening (at any speed the PIC would do something) and not starting - which was what happened to me on the SCRUMPI with the 'not' SC/MP 8050 and the flat line on the hand scope showed there was no oscillation.

I was thinking of these things after watching curious Marc last night during a period of insomnia:

https://www.youtube.com/watch?v=X_cywNLu7VE

I wonder if only the one sliver had been missing and the contact was still present that it would have worked with unforeseen circumstance - I have since wrapped all my crystals in bubble wrap...

In a disclaimer this week has been hard work and by 20:01 Friday I have had a few beers and a fair number of glasses of wine with dinner while watching an episode or two of the Sky 1 adaption of Brave New World; so logic and thinking may be more based on gut feeling than any mathematical method.

Tim

PS: Slothie the brute force method of the offer of a tested 'in Phil's board' PIC stands... if not lost in the SNR on this thread I am generating.

PPS: shocked by the current bay value of a genuine MK14 manual I was watching...

Originally Posted by SiriusHardware

You presumably agree though that it would show just about anything but the right frequency? You'd just be looking for a change between what it does when connected to the oscillator, and what it does when it isn't?

It's different with scopes, you can put a scope on a signal which far exceeds its specified maximum working frequency and you will still see 'something'. With an analogue scope - by far my preference - the individual cycles of the signal will just merge together into a 'band' of signal in which individual cycles can no longer be resolved, which is what you would intuitively expect.

A frequency meter fed with a signal which exceeds its maximum input frequency will usually just go haywire.

A compromise: Take the oscillator signal to the input of a CMOS IC configured as a divide-by-ten circuit and use the Arduino-Meter to read the output frequency of that. If it says 2MHz, great.

Phil's simple trick of using an FM radio to listen on 100MHz is still a lot easier though. Most smartphones (and a lot of dumb phones) have an FM radio hidden away on them somewhere, although it usually needs wired earphones to function because it uses the lead as the aerial.

[Timbucus]

Originally Posted by SiriusHardware

Phil's simple trick of using an FM radio to listen on 100MHz is still a lot easier though. Most smartphones (and a lot of dumb phones) have an FM radio hidden away on them somewhere, although it usually needs wired earphones to function because it uses the lead as the aerial.

...and of course Phil's solution to use the best proven and authentic method (cf Apple first music demo at the Home Computer Club) is the best and deserves its own thread. The earliest testing I did on the Spectrum Next was with my radio receiver They may even speak to us from the rest of the forum if we talk about radio...

[SiriusHardware]

Originally Posted by Timbucus

PPS: shocked by the value of a genuine MK14 manual I was watching...

I didn't follow how much it went for although I was notified about it. I would have wanted to know which version it was (V1 or V2). I don't think the source specified that.

The spare clean (V1) manual I have, I bought instantly as soon as it appeared for a B.I.N. of £7 just before the various replicas started to stoke up new interest in the machines. I think I've mentioned before that my original manual is in such dire condition that I would probably have to pay the council to tow it away. That's why I sought out a replacement, obviously.

Some time after that I chased after a copy of "Understanding Microprocessors With The Science Of Cambridge MK14" and had to bale out when it shot past the £20+ mark.

Even replica machines go for surprisingly high prices but that is understandable to a degree, given what it costs to populate one now.

[SiriusHardware]

...Which is what makes Karen's PIC14 such a good alternative. (There. Back on topic.)

[Slothie]

Originally Posted by SiriusHardware

As regards using the Arduino as a frequency meter, since the Uno? is clocked at 16MHz and the frequency under investigation is 20MHz, won't that be far too high a frequency for the Uno to measure?

I like Phil's dodge, a simple FM receiver will typically make a loud roar/hiss under no-signal conditions and go silent when there is an unmodulated carrier on the frequency it is tuned to. Watch out for FM radios which have auto-mute though. Using a receiver to 'sniff' an oscillator like that gets around one possible drawback with direct measurement - sometimes the very act of connecting an instrument to an oscillator to measure it will actually stop the oscillator.

Using the timer1 as a counter and timer2 to gate it, you can measure up to 6-8 Mhz reliably. Although the clock is 20MHz, all the other pins will be much lower in frequency, I just need to check that anything is going on. Other than that I have a £5 multimeter with no frequency range, so detecting high frequency pulses is challenging. I suppose I could try a capacitor and diode detector to see if theres any AC anywhere, but as you point out mucking around on the crystal pins might cause the oscillator to stop. If the PIC looks DOA I might take you up on that programing offer. I'll have time this weekend to decide what the next step is....
I don't have an FM reciever to hand sadly although I could ask around, I'm sure theres one stashed away somewhere here.

[Slothie]

Originally Posted by Timbucus

...and of course Phil's solution to use the best proven and authentic method (cf Apple first music demo at the Home Computer Club) is the best and deserves its own thread. The earliest testing I did on the Spectrum Next was with my radio receiver They may even speak to us from the rest of the forum if we talk about radio...

When I got my hands on my first computer (A commodore PET) I wedged a small transistor radio inside and spent several days experimenting with BASIC and assembler programs to see what efffect they had on the inteference issued by the CPU. Since the PETs clock was 1MHz there were plenty of harmonics to choose from...! I didnt make any noises you could call music in anything but the most abstract of terms, but I had a lot of fun trying. I then moved on to connecting the CB2 pin of the user port to the radios volume control and had fun programming much more tuneful results hampered only by complete lack of musical talent.

[SiriusHardware]

Originally Posted by Slothie

I don't have an FM receiver to hand sadly although I could ask around, I'm sure there's one stashed away somewhere here.

Not even on your phone, did you scroll through all the Apps? Even my little Nokia 100 'dumb phone' (~£11 when bought) has FM radio on it, but it needs earphones plugged in to work.

Have you any kind of audio amplifier, stereo with an AUX input? A high impedance audio transducer like a passive piezo sounder? A crystal earpiece? The display / key scanning waveforms are probably at a low enough frequency to sound like a buzz or tone, if they are active.

[Mark1960]

Originally Posted by Timbucus

They may even speak to us from the rest of the forum if we talk about radio...

20MHz square wave, so probably doesn’t matter what frequency you tune to you’ll probably detect something. They won’t be speaking to you ever again

[Timbucus]

Originally Posted by Mark1960

20MHz square wave, so probably doesn’t matter what frequency you tune to you’ll probably detect something. They won’t be speaking to you ever again

We should create recordings of the various computers ready for the loss of analogue receivers - Elektor are doing a good SDR kit at the moment - then it will be real... I wonder what a close coupled Pentium sounds like...

Trying to ignore Sirius's a attempt to get back on topic.

[Timbucus]

Can we create a Neural Net that knows what the fault on machine is from a Fast Fourier Transform of known code running...

[Slothie]

Originally Posted by SiriusHardware

Not even on your phone, did you scroll through all the Apps? Even my little Nokia 100 'dumb phone' (~£11 when bought) has FM radio on it, but it needs earphones plugged in to work.

Have you any kind of audio amplifier, stereo with an AUX input? A high impedance audio transducer like a passive piezo sounder? A crystal earpiece? The display / key scanning waveforms are probably at a low enough frequency to sound like a buzz or tone, if they are active.

I thhink the last time I tried the radio on my phone it had some kind of auto tune thing that would be a pain, but I'll see if I can find my headphones and give it a go.

I do have a piezo sounder, NPN transistors and some random components so I may be able to rig up a mini-amplifier and/or a pulse detecting logic pobe. Hmmm...

[SiriusHardware]

As I mentioned, the oscillator will sound like dead silence if you can tune to F x 5 =100MHz.

If you are using phones, try strumming on the PCB with your fingers, it should be sufficiently microphonic, ie, frequency will vary in sympathy with the physical vibration of the PCB, to transmit the sound to your receiver so you'll know it's definitely your oscillator you are listening to.