Apple II clock circuit question...

19 posts / 0 new
Last post
Offline
Last seen: 22 hours 5 min ago
Joined: May 31 2022 - 18:18
Posts: 302
Apple II clock circuit question...

Looking at the II clock circuit because I'm not seeing a clock on a 79 II motherboard. At first I thought it was a problem with the crystal because I was only seeing a high output from Q2.

 

Looking at the SAMS Compufact Apple II / IIPlus I see the following information on the timing circuit, so I took some measurements:

 

It now looks like Q2 may be bad, and I can pull that and test, but a quick search online shows these parts are long gone. Is there a suggested replacement? I grabbed a datasheet from an archive site so I have the specs, but don't want to search out my own if there's already a good replacment know.

 

Looks to be a fairly simple part of the cirucuit, but I may have missed something.... is there something else I should check before messing with Q2?

 

I am not a fan of the way the author calls out the  transistor pins on page 18 (board picture) , is right, but hard to clearly see and I was wondering why there are two of the same parts on this board.  Even at 300% this was not clear to these old eyes lol:

 

 

CVT
CVT's picture
Offline
Last seen: 10 hours 42 min ago
Joined: Aug 9 2022 - 00:48
Posts: 1071
They are not gone, you can

They are not gone, you can still buy them: https://www.findchips.com/search/2N4258

 

You should not put the oscilloscope probe right at the crystal. Put it where the SAMS book suggests instead.

Offline
Last seen: 5 hours 54 min ago
Joined: Apr 1 2020 - 16:46
Posts: 950
Apple II crystal oscillator issues are legend ...

... a lot has been written here on Applefritter about the challenges with them. It also could be a damaged crystal ... this type of oscillator tends to overload the crystal (most simple crystal oscillator circuits do that) and then the abused crystal gets damaged and refuses to work.

 

Finding the right (series mode) crystal for the Apple II is difficult. You can use a parallel mode crystal with a small (i.e. 47 pF) trimmer capacitor in series but if that thing starts up is yet another question. Some crystals do, others don't.

 

For more information, see the various threads here on Applefritter on the topic.

 

- Uncle Bernie

Offline
Last seen: 22 hours 5 min ago
Joined: May 31 2022 - 18:18
Posts: 302
CVT wrote: You should not put
CVT wrote:

 

You should not put the oscilloscope probe right at the crystal. Put it where the SAMS book suggests instead.

 

Thanks, I was probing the transistors but understand there's nothing between. 

The clock divider outputs were as expected when there is no pulse train. You got me wondering how to check the voltage references they included. 

 

Given what I'm seeing I'm thinking the transistor(s) are the first thing to try, but only because I don't see anything else it could be.

Voltages from the supply measured at the slot connectors are : 5.15,  12.7,  -5.3, -12.5  all within 5%. 

 

This board also came with a bodge wire from UB1p16 to UB2p14 wihch is power but I can't see anything that would indicate any trace problems, they both get power from different sides of the board. The topside trace (B1 74175 FF p16 ) looks good, as does the bottom side (B2 7486 EOR p14) also looks good.  Now I'm second guessing if I should just pull those chips and verify they are good as that's easy enough to do. 

Offline
Last seen: 22 hours 5 min ago
Joined: May 31 2022 - 18:18
Posts: 302
UncleBernie wrote: For more
UncleBernie wrote:

 

For more information, see the various threads here on Applefritter on the topic.

 

 

I will do so, but do you have suggestions on what to search for that will get me there? I had searched for the transistor part number, and to my surprise there was nothing. Searches for other clock oscillator stuffs  mostly returned things for Mega 1 and/or overclocking. I didn't find anything helpful related to the II.

Offline
Last seen: 18 hours 6 min ago
Joined: Jun 25 2020 - 17:00
Posts: 238
Jeff d, does your image

Jeff d, does your image indicate no output signal on the collector of Q2?  e.g "but not this"  No signal here indicates an issue with Q1, Q2 or the crystal if all transistor bias voltages are good.

 

Offline
Last seen: 22 hours 5 min ago
Joined: May 31 2022 - 18:18
Posts: 302
LaserMaster wrote:Jeff d,
LaserMaster wrote:

Jeff d, does your image indicate no output signal on the collector of Q2?  e.g "but not this"  No signal here indicates an issue with Q1, Q2 or the crystal if all transistor bias voltages are good.

 

Yes, the "not this" is hte collector signal. I was trying to use the red arrow to point to the line the graph is pointing to... guess that could have been done better. The Q2C in the scope image is what I see, ie a steady 1.43V with no oscillation. The green voltage values are my readings which belive are "good". 

Offline
Last seen: 5 hours 54 min ago
Joined: Apr 1 2020 - 16:46
Posts: 950
More on crystal oscillators (and their pitfalls)

In post #5, 'jeff d' wrote:

 

" I will do so, but do you have suggestions on what to search for that will get me there? I had searched for the transistor part number, and to my surprise there was nothing. "

 

Uncle Bernie says:

 

There have been numerous threads and posts about Apple-1 and Apple II crystal oscillators and their problems. But I also do have issues finding them again. Here is one thread on the Apple-1 crystal oscillator which is as botched as the oscillator of the Apple II:

 

https://www.applefritter.com/content/wanted-apple-1-owners-frequency-counter

 

The fundamental problems in that thread are discussed and apply the same way to the Apple II: if you get the "wrong" crystal and if it starts oscillating, you still may have the wrong frequency (not the 14.3181818 MHz stamped on the crystal) and so your colors will be gone.

 

The trimmer fixes that as you can 'pull' the crystal to the target frequency.

 

Once you have that trimmer (and a parallel mode crystal, which is easier to obtain) you can also use a lot more transistor types. I think most small signal PNP transistors could be used sucessfully, but I did not try that. One problem with the original series resonant oscillator circuit is that ideally there should be no reactances in the loop, which means the transistors must have very small parasitic capacitances, or, specific ones which fit to the specific crystal. Otherwise the oscillator may not start or refuse to work on the proper frequency. The oscillation criteria for the phase and gain around the loop are not well understood by most hobbyists, and some "professionals", too. There also are many misconceptions as to the differences between series mode and parallel mode crystals. See my post #16 in the above thread for more insight.

 

The bottom line is this: don't try to find a "series resonant" 14.3181818MHz crystal for this oscillator, because they are hard to find and most likely won't work in that circuit unless by sheer luck they are very close in certain parameters (not found in most crystal datasheets) to the ones Apple did use. Instead, go for a "parallel mode" crystal (which are easy to get) and add the trimmer.  But don't expect that everything works instantly ... you might need to tweak the circuit a bit, i.e. adding small load capacitors (10...33pF) to ground.

 

There is a reason why the industry nowadays avoids to design their own crystal oscillators, and uses the ready made "canned" ones in DIL-14, DIL-8 or various forms of SMDs. Proper design of crystal oscillators is a lost art ... and even most books written by "authorities" in the field have misleading recipes or fallacies. For instance, in one such book, their best oscillator (crystal in the emitter to ground path of the sole NPN) sports a large parallel inductor there ... hmmm. The author claims that it won't work without that inductor but also states the reason why is not understood. (Ouch).

 

I laughed and replaced that nasty inductor with something else ... and this book is supposed to be one of the best available on the subject.

 

I could tell you stories ... designers desperately trying to simulate a crystal oscillator with SPICE ... never works, always lies ... oscillators for wristwatches which stop oscillating when the watch gets a small g-shock, and only restart seconds later. (CASIO exploited that common behaviour of some competitor's watches to turn that into a brand name for their watches which had a better oscillator circuit). Never underestimate the pitfalls lurking in crystal oscillators and never build the lousy circuits shown in all TTL and CMOS databooks and cookbooks. All of them are bad despite they pretend to work ... until they kill the crystal by overloading it.

 

- Uncle Bernie

CVT
CVT's picture
Offline
Last seen: 10 hours 42 min ago
Joined: Aug 9 2022 - 00:48
Posts: 1071
jeff d wrote:...You got me
jeff d wrote:
...
You got me wondering how to check the voltage references they included. 
...

 

You can still use the oscilloscope to measure voltages, just ignore the oscillogram. In the SAMS books when they put voltages like that, they are to be measured with a digital meter set to DC relative to the common ground.

 

To eliminate the possibility of anything other than the generator, you can simply remove the 74S86 chip in position B2 from its socket and then take a look at the oscillogram on pin 10 of the empty socket. If you still get nothing, then the problem is definitely in one of the elements making up the generator.

Offline
Last seen: 22 hours 5 min ago
Joined: May 31 2022 - 18:18
Posts: 302
Thanks Uncle Bernie. I am one

Thanks Uncle Bernie. I am one that understands what I don't know and once I saw clock high I though oh nooooo not the clock...

I had a talk with the board's current (not orignial) owner tonigh and had said many of the same things, including this is generally done with a single part these day. Your description of trying to work in a differetn crystal sounds full of problesm or a lots of effoet to get dialed in. Sounds like a fun challenge, or hell!

 

I think a good plan is to work back to this, and hope it's something else in that path. I'll start with the ICs and work back to the oscillator.Without an obvious reason for the bodge wire, I can't help but wonder if the original owner was attempting to fix something with either the XOR and FF, so maybe this is related... ? I don't know.  Just power bu without obvious trace damage so curious to it seems strange.  

Offline
Last seen: 22 hours 5 min ago
Joined: May 31 2022 - 18:18
Posts: 302
CVT wrote:To eliminate the
CVT wrote:
To eliminate the possibility of anything other than the generator, you can simply remove the 74S86 chip in position B2 from its socket and then take a look at the oscillogram on pin 10 of the empty socket. If you still get nothing, then the problem is definitely in one of the elements making up the generator.

 

 

Thanks for the suggestions, that is similar to what I was planning to do. I was going to pull B1 and B2 and test them. The reason for B1 is because of the wire. 

I had been wondering how to isolate the generator from other circuit stuff, and hadn't yet got that figured out. You suggestion to deadend the singal at the XOR will do the trick, thanks!

Offline
Last seen: 22 hours 5 min ago
Joined: May 31 2022 - 18:18
Posts: 302
Couple of updates, tried CVT

Couple of updates, tried CVT's suggestion of removing the EOR and found voltage on B2p10 but it was steady at roughly the same voltage as in the earlier picture.  Also noticed is pin8 (output) was sitting at 1.5V which I didn't trace through the schematic, but expect that's fine  (unless someone knows otherwise). I may look later, but this point I I'll let that slide.

 

I removed the bodge and was done for a good reason, B2's power is lost without it. I wonder what's going on behind that socket after looking at the pad and pin can't see any problems. The original work was well done, pro work. I don't think Apple would have done that, but who knows maybe it was a production rework... anything is possible in production.

 

I reflowed the transistor joints and took a look at the X1's joints.... someone else had touched these. They weren't horrible, but also not well done and better now. 

 

No change but during a discussion with the owner last night I did learn they were folowing the advice of others and had tried reseating chips and removing the CPU to check the board revision. Fingers crossed that all is fine, but I should check things I didn't think to check before noticing no clock on the CPU. I think he said he may have inserted the CPU wrong, but I now wonder if that would have been discoverd after applying power or what. I'll be tracking down transistors today....

CVT
CVT's picture
Offline
Last seen: 10 hours 42 min ago
Joined: Aug 9 2022 - 00:48
Posts: 1071
jeff d wrote:Couple of
jeff d wrote:

Couple of updates, tried CVT's suggestion of removing the EOR and found voltage on B2p10 but it was steady at roughly the same voltage as in the earlier picture.  Also noticed is pin8 (output) was sitting at 1.5V...

 

With the 74S86 removed, this is what you should have seen on pin 10 if the generator was working:

 

Offline
Last seen: 22 hours 5 min ago
Joined: May 31 2022 - 18:18
Posts: 302
That is definately not what I

That is definately not what I'm seeing.

I was wondering about rather than harvesting parts from a II, this looks like the same core parts as is found in the //e  but I don't see any specifics on the Y1 in the //e other than it's frequency. Could a replacement for the oscillator and transistors come from a donor //e? 

 

Two other notes on the hunt for replacement transistors... the PN4258 is easy to find today, but not the N and finding having to double check what's available is not really the PN. Also those line item and order minimums along with shipping >$20 for two tiny parts is >$40 for two transistors which I find crazy! I could have an emu running on a pi for about the same! LOL

CVT
CVT's picture
Offline
Last seen: 10 hours 42 min ago
Joined: Aug 9 2022 - 00:48
Posts: 1071
jeff d wrote:That is
jeff d wrote:

That is definately not what I'm seeing.

I was wondering about rather than harvesting parts from a II, this looks like the same core parts as is found in the //e  but I don't see any specifics on the Y1 in the //e other than it's frequency. Could a replacement for the oscillator and transistors come from a donor //e? 

...

 

Yes, you can take the crystal and transistors from a IIe, if you have one lying around for parts.

Offline
Last seen: 22 hours 5 min ago
Joined: May 31 2022 - 18:18
Posts: 302
Thanks, the replacement

Thanks, the replacement search is feeling more hopeful!

 

Bring out yer dead II's! 

Offline
Last seen: 1 day 3 hours ago
Joined: Jan 31 2024 - 06:40
Posts: 158
All Bulgarian Apple2 clones

All Bulgarian Apple2 clones (except Pravetz 8C which have an oscillator) use 2N3906 in the very same ciruit instead of 2N4258. 

CVT
CVT's picture
Offline
Last seen: 10 hours 42 min ago
Joined: Aug 9 2022 - 00:48
Posts: 1071
Also there are some on eBay:

Also there are some 2N4258 on eBay: https://www.ebay.com/itm/404088565243

Offline
Last seen: 5 hours 54 min ago
Joined: Apr 1 2020 - 16:46
Posts: 950
Why cannibalize Apple II for oscillator parts ?

In post #16, 'jeff d' wrote:

 

" Bring out yer dead II's ! "

 

Uncle Bernie complains:

 

IMHO it's a sin to cannibalize another Apple II just for the oscillator components.

 

It would be better if somebody made a small kit based on a 'parallel mode' crystal, a trimmer, and the two transistors to allow replacing the original oscillator circuit. Such a simple solution of course would also overload the crystal, so it will die sooner or later (I see this as the primary cause of all the oscillator trouble with the Apple II) but these 'parallel mode' crystals for 14.3181818 MHz are easy to find and cheap to replace.

 

Anyone who tries to preserve / repair the 'serial mode' crystal oscillator will have huge difficulty sourcing the right components. Even back in the day, Taiwanese clone manufacturers faced the same crystal oscillator woes and so their cloned motherboards often had added pads and traces on the PCB to put in extra components around the basic oscillator circuit they copied 1:1 from Apple. These 'options' would allow them to make the oscillator work even with 'parallel mode' crystals and they could then tweak the frequency to the required accuracy (otherwise, no colors !).

 

It eludes me why Apple II users facing a failed oscillator always choose to go the hard way full of pain, suffering, and additional costs. Maybe the choice of that problematic 'serial mode' oscillator by Apple was done intentionally to make cloning of the Apple II more difficult ... it's quite hopeless to produce such clones unless ordering the right 'serial mode' crystals from the crystal manufacturer, which must be tailored for that specific oscillator. Otherwise the oscillator won't start, or the frequency would be off target. This, BTW, is the reason why 'parallel mode' crystals are preferred and dominate the marketplace. They can be 'tweaked' to target frequency and oscillators based on them start better and run more robustly than series mode oscillators (which theoretically are more accurate and more stable). The reason is the different region on the phase over frequency chart of the crystal on which these oscillators operate. But explaining these details to hobbyists is hopeless. It involves higher math far beyond what typical hobbyists do master (in terms of math). Remember the old adage that half a year experimental work in the lab can replace half an hour of invoking evil, higher math, on a piece of paper. Magic spells come to mind ... the modern DIE way (or is it called DEI ?) is to work experimentally - no evil math - even for pedestrian bridges or aircraft hangars which then may collapse during construction or shortly after the opening ceremony. The collapsing pedestrian bridge killed a few pedestrians who happened to use it at the time of its collapse - oh the great virtues of DIE and 'diverse' design and construction teams ! No need for real structional engineers anymore ! No need for evil math ! No need for the ability to read and write ! It's called 'progress' as in 'progressive' !

 

- Uncle Bernie

Log in or register to post comments