Why is R74 missing on the 502 board?

[Jeff]

According to the schematics, this connects +5 through a 1K resistor to pin 41 of the backplane and calls the signal VMA. The 540B feeds this signal into two chips on the 540 video board that decode the addressing to the color ram and the video ram.

Without this line pulled high, access to the color video should not work. Also, the signal floats, and so, sometimes access to video character memory fails.

Anyone know why it’s not populated on any 502 boards I see on the internet?

Anyone know why it is on the CPU boards (505 and 502) and not on the video board itself?

/Jeff

[dave]

Interesting. I never noticed that before.

The VMA signal was there to support the 6800 processor. In a 6502 system, it should be tied high. I have no idea why this would be omitted. For LS TTL parts, internal biasing should weakly pull a floating input high, so the 540 board would typically work, but the input circuit properties may vary between manufacturers, the circuit would be noise-sensitive, and it's not a guaranteed behavior. HC parts would not behave properly without pulling up the input.

The signal would be tied high on the CPU board because it's supposed to be generated on the CPU board. Tying it high on the video board as well would sink a couple of mA extra current through the 6800 output, if a 6800 is installed.

Dave

[Jeff]

Thanks, Dave for your usual well explained answers.

I came across this issue while testing my newly built 540B and 502 boards. I was using my stock C4P as a test bed, and interchanging boards. Neither of the 502's (stock or replica) would drive my replica 540B (no writing to the video character memory... essentially no clear screen on RESET), while the stock 502 would! Yet neither of the 502's had R74. I also checked many internet photos of 502 boards and they were missing R74 as well. So I initially dismissed this as the cause of why my 540B wasn't working and I moved on to more circuit checking. After about two hours, of painstaking circuit probing, I came back to it and decide to add it to the circuit.

So basically, with a combination of hardware, aging chips, new chips, different manufacturers, etc. the VMA line was floating and simply a crap shoot if it worked. As it stood, my stock C4P video COLOR memory wasn't working.

Adding R7 to both boards made both work properly.

People love pictures, so I thought I would include one...

Schematics.png (1.6 MiB) Viewed 6306 times

...and ask one more question: Do you know why R74/R93 a different value on the two boards?

/Jeff

[dave]

Thanks, Dave for your usual well explained answers.
. . . Neither of the 502's (stock or replica) would drive my replica 540B (no writing to the video character memory... essentially no clear screen on RESET), while the stock 502 would! Yet neither of the 502's had R74. I also checked many internet photos of 502 boards and they were missing R74 as well. . .

Is it the stock 505 that worked?

...and ask one more question: Do you know why R74/R93 a different value on the two boards?

The function is not particularly sensitive to resistor value. Since the VMA line is only connected to high impedance inputs on a 6502 machine, very little current will ever be flowing through those resistors. So there's little difference betwen the 470 ohm and 1K values.

A very high value, such as 100K, may produce voltage drops from input biasing and protection circuits and may also be more susceptible to noise. So usually the value is chosen as a compromise, high enough to keep current reasonably low when the line is pulled low, but not so high to provide a stable logic level and abstract out the realities of noise and various input circuits. Usually pullups will be in the 2.2k to 10k range for a 5V logic design.

In this case, the line should never be pulled low (the 502 and 505 are 6502 only), so it seems the compromise apparently went more toward the low range.

Dave

[Jeff]

The Stock 502 worked with the Stock 540, but no access to color video ram.
/Jeff