Problems with replica Apple II Rev. 0

Have you got video?  Show us the screen.

Have youy got a beep?

Have you got a 14mHz at B2?

Have you got the 7 mHz video sync signal there as well?

It would be great if you could insert an F8 autostart ROM and see if it works that way.

Also, I'm curious - what is the blue tape on some of the chips?

What is the make and model of your oscilloscope? I want to see if it has a built-in frequency counter. If it doesn't, you can still measure the frequency of 7M on the Apple II bus (pin 36) and multiply it by 2, but it's not going to be very accurate. The idea is to see if the oscillator is running.

No frequency counter, but you should still be able to measure the 7M frequency to within a couple of kHz.

I would please ask you to write here. Every question and answer will be written here and other users could help them. It has helped me a lot to read old posts, I have learned and I still learn from you. Thank you so much.

Macintosh_nik wrote:

I'm like a monkey with a grenade, but if I'm given the settings I can definitely handle the changes. 

You are not the only one, I also include myself. I liked that expression, I will add it to my favorite phrases.

Macintosh_nik wrote:

Thanks for the help! Nothing on the screen, no sound either, I connected the speaker from my Apple //e. I don't think it's related to my monitor, 

 

Can you tell me how I should measure 14 MHz on B2 and whatever else, I'm not really good with an oscilloscope but I have one. 

No video at all means there probably no clock signal to the video generation circuitry.  I asked if you had any video at all because it's important to know where the problems lie.  Often troubleshooting stalled CPU issues you still get video signal onto the screen - like garbage characters or a screen of all white, or SOMETHING.

Having zero video is important because it means either there is a cataclismic failure of the video circuitry or there is not clock signal at the video circuitry, which is more likely.

No beep means the CPU did not start reading RAM and ROM and could not execute any code.  Reasons could be many, but if there is no 1 mHz clock signal, nothing happens.

You need to start at the beginning, which is a healthy clock signal.  Do nothing until you are certain you have the correct clock signals.  14 Mhz, 7 Mhz, 3.58 Mhz, and 1 Mhz, both phases (one is inverted at the CPU) 

Then you can start poking around with ROM and RAM addressing.

That ROM card you show contains an F8 ROM.  With it inserted in the F8 position on the motherboard you can basically go back to the "beginning" and possibly eliminate any problems the ROM Replacement Board might be causing.  Also, the EPROM on the ROM Replacement Board has an exposed window.  Is it programmed?  If so, put a piece of opaque tape over the window to prevent intrusion of UV light which may partially erase the EPROM.

The clock timing circuit starts with the crystal and the chip at motherboard position "B2", and is distributed as shown in the diagram.  You should refer to the Rev-0 schematic to see what parts of each chip do what, and what pins contain each signal.  This diagram shows which chip locations to look at:  B1, B13, C1, C2, A2.

Chip designations are as follows:  

86=74S86 (Note this is NOT a 74LS86 chip, it is a 74S86 chip.  They're similar, but the 'S86 has a much higher current output than the 'LS86.

175=74LS175

195=74LS195

02=74LS02

00=74LS00

153=74LS153

clocktiming.jpg

Folks, please don't overlook the evidence Macintosh_nik documented in the OP!

• It's a replica Rev. 0 board, without a Disk II controller nor keyboard, so it will not beep at power on because there's no RESET source.  So Macintosh_nik needs a method to RESET the computer, if he hasn't already done so.
• He posted an oscilligram of the 1.02MHz Phase-0 clock at pin 37 of the 6502.  That indirectly confirms that the 14MHz crystal oscillator is functioning because it is derived from the crystal oscillator by dividing its signal by 14.

About RESET in the Rev. 0 board:

When you turn on a Rev. 0 Apple ][ (without any peripherals) it will be frozen until the user manually generates a RESET signal. 

Here are some ways to generate a RESET signal:

• If the motherboard is installed in a case with a keyboard, then press the RESET key and release it again.  (On later keyboards you would need to hold CTRL while pressing RESET.)
• If you have a known-good PS/2 keyboard, plug it in and press CTRL+ALT+DEL.  (Note: a USB keyboard won't necessarily work, even through an adapter)
• Attach a pushbutton switch between the 0-volt supply (GND) and RESET.  Your PS/2 adapter exposes a convenient location to solder a pigtail & pushbutton, see diagram.
• Install a Disk II controller in any slot.  The Disk II was designed to automatically generate a RESET signal for Rev. 0 motherboards.

Keyboard RESET connections.png

About those timing signal measurements...

Here's the oscilloscope measurement of the phase-1 clock from the 6502's pin 37.  In addition to showing the shape of the signal, this oscilliscope automatically displays a lot of measurements at the bottom of the screen -- including the measured frequency of the signal.

Osc 6502 pin 37.png

From that single measurement, we can infer that the timing generator is emitting the correct signal from pin 7 of the LS175 at motherboard location B1.  That suggests the 14 MHz clock, oscillator, and crystal are all working correctly too, so I've highlighted them in green.  To verify the rest of the signals, connect that oscilliscope to each of the identified pins and read the "Freq:" measurement.

• Pin 9 = "Freq: 14.3MHz"
• Pin 2 = "Freq: 3.58MHz"
• Pin 6 = "Freq: 1.02Mhz"
• Pin 7 = "Freq: 1.02Mhz"
• Pin 15 = "Freq: 7.16Mhz"

Timing generator.png

S.Elliott wrote:

Folks, please don't overlook the evidence Macintosh_nik documented in the OP!

• It's a replica Rev. 0 board, without a Disk II controller nor keyboard, so it will not beep at power on because there's no RESET source.  So 

Good point about the oscillator already running, but I don't think the issue is lack of reset at this point, since his video is not working, which should start before the reset.

CVT wrote:

Good point about the oscillator already running, but I don't think the issue is lack of reset at this point, since his video is not working, which should start before the reset.

 

 Yes, my point exactly...

And, @S.Elliott: good point about the reset.

CVT wrote:

Good point about the oscillator already running, but I don't think the issue is lack of reset at this point, since his video is not working, which should start before the reset.

Then let's inquire more precisely what Macintosh_nik meant when he wrote "Nothing on the screen" in comment #3.  That might mean that it's not generating any video signal.  Or it might mean that there's a good video signal that just happens to be empty, which is exactly what would happen if the LS194 at A10 is a fake or counterfeit part.  (The OP raised the issue of "junk sold on aliexpress" and marked unverified/untested parts with blue tape . . . including the LS194 at A10.)

To Macintosh_nik some clarifying questions:

• When you tested the video output, could you tell whether there was any sort of signal present?
• For example, does your monitor show a message like "no signal" or is it just blank? 
• As baldrick suggested, can you show us a picture of it?  (even if it's just a blank monitor)
• Can you try attaching the video output to your tablet oscilloscope, and attach a screencap like the one you posted earlier?

Yes.  Those are the frequencies you should be seeing out of the 74S86's pins.

You should follow them out to where they belong at the video circuitry and to the CPU.  And verify that you have 1MHz at pin 37 of the CPU along with pin 39 and pin 3.

You should also follow @S.Elliott's advice about causing a momentary RESET condition after power up to start up the CPU.

"pins 9,2,6,7 and 15" refered to above are on the B1 chip as you discovered. It is possible the XOR gate at B2C is bad - in that case there would be no signal to probe anywhere else down the clock chain. You could try looking at B2 pin 10, but I don't know how badly the oscilloscope probe capacitance will affect the oscillator. Testing B2 off the board should be fairly simple if you haven't done that already.

If you have time while waiting for new crystals to arrive, you could compare your oscillator circuit to the one on the Apple III schematic. It is very similar, but a few resistor values are changed, the base of Q1 is biased differently and there is more feedback between the base of Q1 and Q2. (If I was doing it, I would build the oscillator on a bit of protoboard so I could swap crystals and maybe part values around easily.)

Full disclosure - I don't have an Apple II, I'm just going by the kicad schematic. I do have and Apple III, but the only schematics available for that are the hand-drawn ones from 1980 that may not be too accurate. I have been following your progress in this thread with great interest. Best of luck!

A lot of threads and posts have been written about the infamously problematic two transistor crystal oscillator of the early Apple II.

APPLE (the company) had the luxury to be able to order custom made crystals - just the right type of crystal - to work in their terrible circuit.

See posts #3, #4 and below for this thread:

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

Don't get confused it's for the Apple-1, which has an even worse TTL gate based crystal oscillator circuit. The issue with both types of oscillators is the same. They are meant for series resonant crystals while most crystals sold are parallel resonant. These will not work on their specified frequency if put into an oscillator meant for a series resonant crystal. And being off frequency commonly leads to startup problems. You can try to add a small trim capacitor of 47pF max in series with the crystal. This would allow you to tweak a parallel mode crystal to the right frequency even when it's in a series resonant oscillator. It's not guaranteed to work but worth a try.

If you decide to order more crystals you can order one of these DIL-8 or DIL-14 metal can crystal oscillators like this one:

https://www.digikey.com/en/products/detail/ecs-inc/ECS-100A-143/20509

These are not much more expensive than crystals and are guaranteed to oscillate at the frequency that is printed on them ! No surprises !

This is why even APPLE (the company) did use those in the later Apple II models (i.e. IIc) --- seems they were fed up with all the problems caused by their primitive two transistor oscillator. Which, if it works, won't work for all too long because it overloads the crystal. Which greatly reduces its lifetime. This is the root cause why so many people compain about crystal troubles with their Apple II - they found the crystal is dead and then tried to replace it. With the wrong type, of course. The right type (series resonant) must be custom made ... unless you are super lucky and find the random bag of series resonant 14.3181818 MHz crystals in a bag in some electronics surplus store. I never had that luck.

- Uncle Bernie

Let's not sidetrack the problem with the crystal discussion at this point, since the oscillator is clearly working at a frequency that should produce at least a monochrome video signal.

Having the correct crystal and adjusting its frequency is only critical when trying to get an NTSC color image and that comes much later in the game, when everything else is working. The crystal frequency may be way off and you would still get a monochrome composite image. For example the Apple II Europlus uses a 14.250 MHz crystal.

In post #20, CVT wrote:

" ... since the oscillator is clearly working at a frequency that should produce ... "

Uncle Bernie comments:

Ooops ! I was misled by the two photos in post #14 showing a flatlined clock input on the 6502 and a bunch of crystals. This deceived me in thinking the oscillator doesn't work.

Agree that any clock near 14 MHz should produce a video signal. Unless the trim pot is turned down to zero or something else upstream does not work.

- Uncle Bernie

In order to know where to measure, you need to follow the schematic for the computer.

All the information is there, the chip locations and the pin numbers.

Start with this - print it out and paste it together with tape.

And the clock starts here at the lower left at XT0001.

clock2.jpg

You'll notice the crystal oscillator enters B2 at pin 10 and exits at pin 8.  This should measure the nominal 14 MHz.

It then enters B1 at pin 9 and exits at 14, 15, 12, 3, 10, 7 and 7 in varioius divisions of 14 MHz : 7MHz and 1MHz.  You should be able to detect all those at B1.

If you look up at the top left you can see lines for 14M, 7M, ø1, ø0 and what pins they connect to.

The schematic is here: https://downloads.reactivemicro.com/Apple%20II%20Items/Hardware/II_&_II+/Schematic/Apple%20II%20Schematics.pdf

If you are sure your crystal is nonfunctional and want a known-good crystal I have one.  Message me.

S.Elliott wrote:

Folks, please don't overlook the evidence Macintosh_nik documented in the OP!

• It's a replica Rev. 0 board, without a Disk II controller nor keyboard, so it will not beep at power on because there's no RESET source. 

Exact.As soon as I turn on, in my replica I see on the junk screen.There is no whistle.

Basura en pantalla , antes del reset

When I connect Pin 3 and 8 ... then it starts.

pin 3 y 8

I recorded this video the first time it was put into operation.

Primer encendido Apple II Replica

I see you use an Ikegami. good choice.

I also use a similar monitor, from another brand.

Are called .. PVM?

Here we call them "Surveillance Monitor" or "Security Monitor"

In the rest of the monitors and TV I have not been able to see anything. I know without seeing anything.

I can see Apple1 Replica in almost all, even on the new tv 46" MI

UncleBernie wrote:

A lot of threads and posts have been written about the infamously problematic two transistor crystal oscillator of the early Apple II.

 

APPLE (the company) had the luxury to be able to order custom made crystals - just the right type of crystal - to work in their terrible circuit.

 

 

I have been fortunate to make both the ECS 143-S-1X and  CTS MP143 work with reasonably matched PN4258/PN4209.  They are available from several distributors. Transistors are another story. Many of the A2 clone and replica repairs i've performed for others found random 2n3906's in use.

On several clone motherboards where the oscillator layout on the PCB was less than optimal, the ECS 100AX-143 DIP-14 oscillator ended up being the best solution. It just always works. I use them just to inject 14M into boards during diagnostics.

Macintosh_nik wrote:

one guy advised me to do a simple test. I had to connect a multimeter to pin 40 of 6502 and GND, and when I pressed Reset +5v as if reset and then reappeared.

The first things to check when diagnosing any vintage computer are always, voltages, clock, and reset.  You don't need to look any farther until you have those.

The Nymph NMP143 crystal in your photo is series resonant and should work if not damaged. The Raultron crystal is the incorrect type (18pf parallel) and will likely not work, or may be off frequency if it does work. The smaller HC-49/U crystal looks like it *could* be a CTS 14.31818MA, which is also incorrect as it is 20pf parallel.

If you have a working oscillator frequency close to 14.31818MHz the  machine *should* at least execute code (but don't expect frequency dependant functions like NTSC color to always work).

In post #28 "Macintosh_nik" wrote:

Then I decided to trust my intuition. Since there are two 2n3906 transistors in this circuit besides the crystal, I decided to change them to other ones. And it worked 100%!

Uncle Bernie comments:

See, after all it was the crystal oscillator refusing to work, so my post #19 was right, and "CVT"'s rebuttal in post #20 was wrong (it still eludes me how he came to the conclusion that Macintosh_nik's was "clearly working". But it could be that the oscillator did work ... sometimes. And other times it didn't work (Startup problem). Which would explain the confusion very well.

The bottom line is that if an Apple II appears to be "dead" despite all power supply voltages are in spec, the next step should be to see if the crystal oscillator does oscillate or not. For that, the poor fellows with no money to afford an oscilloscope must use / build a "logic probe pen" which not only has a LED for "Low/High" logic level, but also a LED for pulses. These can be built for less than $1 in materials if a discarded / empty marker is used as an enclosure. Use the "fat" markers (Such as "Sharpie Accent Highlighters") because their inner diameter is large enough to shoehorn a DIL-14 IC into it. With the smaller diameter pens you would need to use small SMD parts.

- Uncle Bernie

P.S.: the measurement of 978ns indicates that the oscillator runs at  14.315 MHz ... if we calculate for 977ns and 979ns boundary cases, this is 14.32958 MHz and 14.3003 MHz (take the reciprocal of the time period and multiply with 14 --- the number of M14 clock cycles per CPU cycle). The reason for the  +/-1 ns cases is the (unknown) actual time resolution of the oscilloscope, and the unknown accuracy of its internal timebase.

The wanted oscillation frequency is 14.3181818 MHz.

So you might need to add a small trim capacitor (5-47pF) in series with the crystal to be able to tweak the frequency to the correct value, at which the NTSC colors should appear. (See the link for the frequency counter thread in my post #19 above, to see numbers at which the colors appear and where they disappear. Your monitor may be slightly different in which frequency band for the color subcarrier it accepts, but the message is that if the frequency is off by too much, there is no color).

UncleBernie wrote:

...it still eludes me how he came to the conclusion that Macintosh_nik's was "clearly working...

The conclusion came from the incorrect screenshot posted in the very first post of this thread showing a working oscillator and a 1.02 MHz square wave.

Macintosh_nik wrote:

Then I decided to trust my intuition. Since there are two 2n3906 transistors in this circuit besides the crystal, I decided to change them to other ones. And it worked 100%! 

I'm very happy to hear it.

What crystal have you used?

Can you upload a photo?

Now for the color it would be easier to use a parallel crystal + a trimmer capacitor, instead of trying to find the correct serial crystal. This is the one I used in my Apple II+, but any parallel crystal of that frequency will work. The capacitor is from an old Sokol transistor radio and it is set to about 25 pF:

Crystal.jpg

In post #32, CVT wrote:

"... but any parallel crystal of that frequency will work  ..."

Uncle Bernie comments:

Correct. But to avoid any misunderstanding, I meant to put that trimmer in series with the SAME crystal which Macintosh_nik already uses. It seems to run too slow (as far as we can guess from the time period his oscilloscope tells us) and adding the trim capacitor will make it faster. Running too slow in a series resonant oscillator circuit is typical for a "parallel resonant" crystal which by definition are always ground to have their series resonant frequency a little bit lower than what is stamped on them. This allows for additional reactances in the circuit and still get the right frequency which is stamped on the crystal. In this case, the added reactance is the trim capacitor. Sounds confusing ? Well, this is the reason why nobody in the industry dares to design a crystal oscillator anymore, especially if long life and low drift / high precision is wanted. The exception from the rule are the crystal manufacturers who make these little DIL-8 and DIL-14 oscillators (nowadays also much smaller, as SMDs) because they know what they are doing and they can tweak the crystal that goes into their oscillators for all the desired properties in the particular circuit they use. So disappointments and fiddling on the production floor until the damn things work is prevented. Back in the 1950s to the 1970s it was common practice to have adjustment steps in the production process. Nowadays a part must drop right in without any adjustments at all, and do its job and be in spec. This is why everybody uses ready made crystal oscillators nowadays and why discrete crystals are a slowly dying species.

- Uncle Bernie

UncleBernie wrote:

So you might need to add a small trim capacitor (5-47pF) in series with the crystal to be able to tweak the frequency to the correct value, at which the NTSC colors should appear. (See the link for the frequency counter thread in my post #19 above, to see numbers at which the colors appear and where they disappear. Your monitor may be slightly different in which frequency band for the color subcarrier it accepts, but the message is that if the frequency is off by too much, there is no color).

Many II+ clones, including the excellent Unitron variants had a colour trim "trimpot" near the crystal oscillartor circuitry to adjust the crystal frequency so that the computer rendered colour properly.

And another reason why I hate that the preferred build seems to always be a Rev-0.  Rev-0s are challenged in many ways, one of whic being they can only render four colours.

Home replica builds should be a Rev-7 at minimum, and better yet, a 100% clone build like a Unitron clone that could accept EPROMS and had nice features like a trimpot for the oscillator.

First you need 50pF max variable capacitor in series with the crystal to get a color image. Only then you can use the COLOR TRIM variable capacitor on the right side of the motherboard to adjust the tint of the color image. Looks like you are jumping to Step 2 without having done Step 1.

Adjusting the trim capacitor that you put in series with the crystal is a bit tricky. You have a very narrow position that you need to hit. You need to load up some color program and adjust it while the machine is on. If you go too far in either direction the machine freezes and doesn't boot.

Macintosh_nik wrote:

I soldered a 5-50pF trim capacitor on the back of the board, there is a photo above. It's just the same as the COLOR TRIM. I twisted it left and right but nothing changed. I tried this with all 3 crystals. So is this how I stabilize monochrome video?

If it's on the back side, how is it in series with the crystal?

Sure, here is before and after:

AddTrimCap.png

This is why it needs to be on the front side, if you don't want to interrupt traces on the motherboard, which I assume you don't.

Just have the computer load something with color and look at the screen while you adjust the trimmer. Keep in mind that you might have to reset the machine many times while you are adjusting it, so if you can automate the color image appearing without having to type too much that would be nice.

The way you have connected it looks correct, but obviously the picture doesn’t look right. 

I cannot tell you if the image you are looking at is supposed to be in color or not. But you need to be doing this with an image that you are 100% sure it's in color.

What is the model of the Ikegami monitor and does it support NTSC? Also there is this waviness – did it appear when you added the trim capacitor?

Without the trim capacitor does it still look like in this video: https://www.youtube.com/watch?v=0C-J3i-18fo ?

Also I don't know if it's supposed to work with a serial crystal like the one you have right now. The one I have in my machine is parallel, like your Realtron 14.311818-18.

You see parallel crystals resonate at slightly lower frequency, so using the trimmer cap we can raise it to what it needs to be. However if the serial crystal is at or slightly above it, then we cannot lower it using a trimmer capacitor in series. We are only raising it and overshooting even further.

The composite video signal appears to be missing its horizontal synchronization. This can also happen when the signal level is too low, much lower than 1Vpp. Try to increase it using the potentiometer in the top right of the motherboard.

If you are looking at the composite video signal with your oscilloscope, make sure the monitor is plugged in. If the monitor has a switch in back for "Hi Impedance" or "75 Ohms", set it to "75 Ohms".

Remember how I mentioned Rev 0 boards having issues with NTSC compatibility including sync issues?

Even back in the days of analog CRT monitors the early Apple IIs didn't always work right with every monitor.  Later revisions generated better qualty video although no ][+ really was very close to "standard" NTSC.  The //e was a little better and the IIgs better still but no Apple II was really ever more than "close enough".  If you ever tried to record onto a VCR or even worse tried to hook into a TV broadcast you'll know what I mean.  We used to have to use a box made by a company called "Blonder Tongue" (believe it or not, always seemed like a weird name) called a "Broadcast Stadardizer" to filter computer video enough for it to look right on TV.  That box cost about what an Apple II did.

You might want to look for the fixes for that.  If I remember right there were some posts about it either here or on the Facebook Apple II groups.

Macintosh_nik wrote:

Okay, I changed the crystal. Now I get almost the correct picture, but as soon as I press Reset it disappears. CVT's words that the signal to the monitor should be greater than 1 v made me think to see what's going on in the oscilloscope.

By standard it should be 1 volt peak-to-peak, but CRT monitors like yours can accept much wider ranges. It is impossible to tell from your video what the composite signal looks like on the oscilloscope, but in general it should look like this when displaying color bars:

BarPattern.JPG

Vertical bars are super-convenient for oscilloscopes, because every horizontal line is the same, so it doesn’t matter which one you trigger on.

Some time ago I wrote a very short program in Basic that displays verical color bars on the Apple II: PATTERN.zip

This program shows this on the screen:

Pattern.png

Also in your pictures it looks like the solder joint beween the crystal and trim capacitor is really close to the case of the crystal. Make sure it's not touching it.

Very nice! Since it's a Rev.0, are you planning to add a color killer transistor or the extra color mod?