Failed Crystal in Heavy Sixer, Anyone?

[+bent_pin]

3 minutes ago, mytek said:

creating the exact same circuit off board and then patching it in place of the original might help to figure out what's going on.

And it's reversable.

 

You can go any route with the new clock, just be sure that no components are left from the old one as you need to reduce inductance. I don't like the original circuit and I think the circuit traces might have degraded a bit, too.

[+batari]

6 hours ago, DrVenkman said:

I appreciate your thoughts, Michael and Fred. Great to hear from folks who may have some insights here. 

 

@mytek - the first post above has the Heavy Sixer clock circuit from the Atari VCS Domestic Field Service Manual and it matches the board I see in front of me. I’ve measured the voltages at pin 10 (DEL) on TIA as recommended by one of the steps in the FSM troubleshooting steps and didn’t uncover any obvious issues, so I think the color pot itself isn’t a problem. Neither end is shorted or failed open, though I don’t recall if I replaced C208 (the filter cap between TIA pin 10 and R211 (the color pot). FWIW, I will switch my scope and probe to 10X to see if it makes any difference in the waveforms measured. 

 

@batari - I’ve already replaced both diodes in the clock circuit with modern 4148s in place of the vintage 1N914’s. Both old diodes tested good out of circuit. I’ve also replaced both transistors with new, though the old ones tested good once removed. I’ve also replaced C202, C203, and C209. 

 

***

Anyway, I’m kind of out of ideas at this point. As I said, I don’t recall right now if I’ve replaced C208 but if I haven’t, I will do so after some coffee. Other than that, there’s nothing else I can think of except to lift each resistor out of circuit to get accurate resistance readings. I know none of them are failed open or shorted at least.

 

Fundamentally, SOMETHING killed that original crystal so it is not out of the question that something else may have been damaged at the same time - I just can’t figure out what it might be. 

I noticed something else - one of the circuits posted above is the same as yours but with a 270 ohm resistor in place of the 1.8k. Perhaps you could try a smaller value there and see if you get a better frequency.

 

Also, notable is that there are two 5v domains on some of the 2600s, separated by an inductor on old H6 models or a ferrite bead on newer models. There was a service bulletin in 1981 that, if I remember right, recommended replacing the inductor with a ferrite bead. Indeed this inductor was the cause of compatibility issues, and I wonder if it might have an effect here. Some have replaced the inductor with a wire and this worked fine as well.

 

If that does not help - did you check or replace the 220 and 100 ohm load resistors on the transistor collectors? Also, I assume you checked the voltages at the points where it claims to be 5v and GND?

 

Anyway, the 4-switch looks like a variant of a Colpitts oscillator which I believe is a far more common setup (and basically what I have used myself and see more commonly today) so if everything else fails and you are going to recreate anything, I would recreate that.

 

 

[+batari]

Sorry for not noticing this earlier, but I assumed the console was video-modded, but from other info in this thread, I am assuming that it is not?

 

If it is not modded, it is worth checking the resistors and caps coming out of the COL and BLK pins on the TIA (pins 6 and 9). Later models don't connect BLK at all. And scoping the signal at pin 9 might also reveal something.

[+DrVenkman]

1 hour ago, batari said:

I noticed something else - one of the circuits posted above is the same as yours but with a 270 ohm resistor in place of the 1.8k. Perhaps you could try a smaller value there and see if you get a better frequency.

 

Also, notable is that there are two 5v domains on some of the 2600s, separated by an inductor on old H6 models or a ferrite bead on newer models. There was a service bulletin in 1981 that, if I remember right, recommended replacing the inductor with a ferrite bead. Indeed this inductor was the cause of compatibility issues, and I wonder if it might have an effect here. Some have replaced the inductor with a wire and this worked fine as well.

 

If that does not help - did you check or replace the 220 and 100 ohm load resistors on the transistor collectors? Also, I assume you checked the voltages at the points where it claims to be 5v and GND?

 

Anyway, the 4-switch looks like a variant of a Colpitts oscillator which I believe is a far more common setup (and basically what I have used myself and see more commonly today) so if everything else fails and you are going to recreate anything, I would recreate that.

 

 

Thanks for your continued thoughts, Fred. I'd have to go back and look at the power section to see how this Heavy's power section is assembled. I do note this Heavy, being made in Taiwan, was probably a production pathfinder for the Light Sixers to come - the switch board is the later Light Sixer design  with the heatsink and regulator mounted to the board instead of being angled away. It looks basically identical to the board in my Taiwan-made Light Sixer from 1981. I would suspect it therefore includes the ferrite bead instead of the inductor but I will confirm.

 

I have checked that all the resistors are not shorted or failed open, and that they are in fact connected to adjacent components as indicated in the schematics. It's notoriously unreliable to check resistance in circuit - some of these have multiple patches to ground (R203, which I mentioned way back at the beginning of the thread). I can lift them and measure them out of circuit but I haven't done that yet. 

 

***

I did pull out my Light Sixer (which works great) and compare the clock. Here's the Light Sixer:

 

And here is the Heavy - the waveforms look more or less identical except for the frequency being low.

 

I do note that the PtP voltage is a tiny bit higher but the output of the Light Sixer's regulator is a tiny bit higher than the Heavy's (5.05V vs. 4.97). And just to be 100% sure the issue isn't something to do with the switchboard or the modulator, I tried the Light Sixer's switchboard with the Heavy but it made no difference.

 

12 minutes ago, batari said:

If it is not modded, it is worth checking the resistors and caps coming out of the COL and BLK pins on the TIA (pins 6 and 9). Later models don't connect BLK at all. And scoping the signal at pin 9 might also reveal something.

I have at least verified continuity between adjacent component legs all the way out through the resistor ladder and confirmed nothing is shorted. BLK is connected on this board, for what that's worth. I also verified signals to/from the inputs and outputs of the 4050. To verify it wasn't bad, I also removed the 4050 and installed a socket. Of course the old chip tested fine in my MiniPro. I will spend some time this week exploring the output side of the board and see what I can observe.

[zylon]

I'll throw some "simple" into the mix. Many heavy 6'ers had this colored paper dot, stuck just under upper left corner of cart port, Sunnyvale L6ers as well. If I remove it, strange things happen, unless I get that board lined up perfectly, inside the housing. I'd also consider replacing the 4050 hex buffer, as those can fail in multitude of ways, not just in "fire" function, or use of paddles. 

Edit- I just saw that you already explored the 4050

Edited June 11, 2023 by zylon

[+DrVenkman]

I've now verified good +5V at R200, R201 and CR200. Well, each one is right at 4.94V, but pin 3 of the VR is 4.97 so it's what I would expect, and nothing that should affect the clock.

 

I've also lifted and tested R200, R204 and R205. All measure within a couple ohms of nominal value. I've previously removed and tested R203 and it's good, too. Basically the only ones I have not lifted or removed to test are R202 and R201. I'm pretty sure R201 is good because it forms part of a voltage divider which should give just a little under 4V at transistor Q200 and at the positive side of C202, both of which track with my tests on the board with power applied. I guess for the sake of completeness I should lift it and check it for sure though, along with R202.

 

But so far, nothing else I've checked has actually been wrong with this board - it's incredibly frustrating. 

 

But while I was poking around at the other end of the board looking for ANYTHING that might explain what is dragging down the clock circuit, I did notice this example of really poor early-Taiwan manufacturing quality. Note: this cap leg is not broken - it was never soldered at all. Unfortunately it really doesn't matter as to the main issue - this is just a filter cap on a TIA line, specifically the line monitoring the position of the player 2 Difficulty switch. Of course I repaired that while I was at the board. 

 

 

[+batari]

Sounds like you’ve tried most everything. So did you shunt the 1.8k with something smaller? Just curious. 

[+DrVenkman]

9 hours ago, batari said:

Sounds like you’ve tried most everything. So did you shunt the 1.8k with something smaller? Just curious. 

Not yet - it’s on the ever-shortening list of things to try. When I do that, I’ll also verify that R202 is in fact 100 ohms, and that R201 is 470 ohms. 

 

My very last step will be to breadboard out a clock circuit based on the Atari schematic (*) just to verify that, when built exactly as described, the crystal oscillates at the correct frequency when powered off my bench power supply. If it works, I’ll go so far as to isolate the existing clock circuit and wire the breadboard to this Heavy motherboard as a test (as suggested a few days ago up-thread). If it slows down too, then at least I’ll know for sure the problem is NOT the clock and must be something else on the board, unlikely as that seems. 

 

(*) I’ve already verified that if you remove TIA and power on the board, Pin 11 of the socket still shows the slower-than-expected OSC signal. So whatever the root cause is here, it’s NOT due to TIA itself at all. 

[cwilkson]

One sanity note.  "Never" measure a crystal oscillator's output directly.  Others have alluded to potential loading effects.  In many crystal circuits, adding a scope probe's capacitance in the wrong spot can completely kill the oscillations.  Best case, it can slow the circuit down a bit.

 

You should measure the frequency of PHI_0 on TIA-pin4 instead.  The ideal frequency here is (105/88)*1MHz.  But you should divide your crystal's marked frequency by 3 to get your "correct" value for the CPU clock.

 

To get the best resolution from most scopes, put a single period of your waveform on the screen.  And turn on averaging if your scope supports it.  Use a big number for the averaging.

And measure both frequency and period.  Period may be more accurate on your scope.  I really don't know...I've never used Siglent gear.

 

Compare your PHI_0 frequency from the problem machine with another one instead of the using OSC frequency.

 

 

[+DrVenkman]

7 hours ago, cwilkson said:

One sanity note….

Thanks for your thoughts.

 

To be clear, however, I’m not working in a complete vacuum here. Every location from which I measure frequency on this troublesome board is low compared to other 2600 or 7800 machines (running 2600 carts). I’ve compared OSC, Phi0, and Phi2 on this board with a Light Sixer, my daily-driver 4-switch and my daily-driver 7800. This board has some weird electrical quicksand happening that drags down the clock frequency compared to my other machines and I cannot for the life of me figure it out.

 

I’ve set it aside for now until or unless a sudden blast of inspiration strikes. 

[cwilkson]

16 hours ago, DrVenkman said:

Thanks for your thoughts.

 

To be clear, however, I’m not working in a complete vacuum here. Every location from which I measure frequency on this troublesome board is low compared to other 2600 or 7800 machines (running 2600 carts). I’ve compared OSC, Phi0, and Phi2 on this board with a Light Sixer, my daily-driver 4-switch and my daily-driver 7800. This board has some weird electrical quicksand happening that drags down the clock frequency compared to my other machines and I cannot for the life of me figure it out.

 

I’ve set it aside for now until or unless a sudden blast of inspiration strikes. 

Ok.  I know you've been to hell and back on this thing. 

But I didn't see where you had confirmed measuring a buffered clock (PHI_x).  It's mainly a check box/sanity check kind off thing and I also wanted to document for posterity.  I'm glad you did that that check.

 

One other thing I thought of if you ever come back to it.  Maybe the board is dirty?  Not standard dirt, but something special.  Like carbon scoring or chemical contamination.  (even soda could do it).  You mentioned the original TIA was damaged.  Or at least is was suspect.  If there was an "event" it could have left tiny conductive paths that are stealing energy from the oscillator.

Edited June 27, 2023 by cwilkson

[+DrVenkman]

8 hours ago, cwilkson said:

Ok.  I know you've been to hell and back on this thing. 

But I didn't see where you had confirmed measuring a buffered clock (PHI_x).  It's mainly a check box/sanity check kind off thing and I also wanted to document for posterity.  I'm glad you did that that check.

 

One other thing I thought of if you ever come back to it.  Maybe the board is dirty?  Not standard dirt, but something special.  Like carbon scoring or chemical contamination.  (even soda could do it).  You mentioned the original TIA was damaged.  Or at least is was suspect.  If there was an "event" it could have left tiny conductive paths that are stealing energy from the oscillator.

For clarity, all the original chips are fine and work in other systems. The only thing from this board that I am certain was damaged was the original crystal. I was getting a solid low-amplitude high signal on OSC and absolutely no other response from it. Replacing the crystal got the system to work but the crystal is simply not oscillating at its natural frequency (I’ve tried two modern replacements and one known-good crystal borrowed from another Atari). The rest is documented up-thread. 

 

When I come back to this machine again someday, I will start by washing the board in iso  and then rinsing with distilled water and air-drying to see if makes a difference. 

 

****

 

As a coda to this saga: I picked up a second Heavy Sixer yesterday, a US-made model with a somewhat beat up top case and a damaged RF cable. After disassembly and testing with a different cable, that machine works perfectly. So I removed the RF cable from the Taiwan case and installed it into the case for the Sunnyvale unit. Unfortunately the rubber case grommet was pretty hard and brittle and it didn’t survive removal. I then swapped the top of the case from this problematic Taiwan unit to the Sunnyvale machine and now have a lovely, working Heavy Sixer. 

 

If I do ever get this board working properly, I’ll need to source a replacement thin coax cable, probably from one of my spare 4-switch machines. I’ll also need to find a replacement case grommet, and do some work to clean up and restore the beat up case top I swapped over to it. 

 

But all of that is a project for another day. In the meantime, it will serve as a potential parts donor for my two working Heavy systems.