Arcade game board voltages

[dog]

Trying to settle an argument. Based on over fifty years experience as a coin-op tech and consultations with other professionals over the years, I have always set the game logic boards at 5.17 volts to eliminates pesky reset problems. (and provide a cushion for less than ideal line voltages at various locations). An esteemed colleague has now said this is too high. Anyone have any arguments for or against? Thank you...

[Andy Chee]

I feel both of you are “wrong”.

If 5.17 versus 5.00 is enough to fix a reset issue, I would argue that you haven’t fixed the issue, and have just performed a workaround without investigating the actual problem.

As for being too high, I would argue it’s still within tolerance. Anything that is damaged by 5.17 versus 5.00 is likely on the edge of failure anyway.

Rather than fixate on DC values, you should really get an oscilloscope onto the voltage rail to look for ripple.

[fzabkar]

In the 1980s I was employed by an American minicomputer manufacturer which advised its field engineers to set the 5V supply somewhat higher than nominal, but I can't remember the exact figure.

[pcprogrammer]

I agree with Andy Chee.

The top limit depends on the IC's and other components used on the board, but it should be still fine with 5.17V. Although the datasheet of Fairchild 74LS00 shows a recommended max working limit of 5.25V, it's absolute maximum is 7V.

But using raising the supply voltage to fix some issue, is bad practice in my opinion. When it is due to a degrading component, the issue might return over some time and then you can't raise the voltage again to fix it, because there is a risk of destroying things.

Best to properly exam the issue and fix it for real.

[tycz]

Dog,

It depends on where you make the measurement. If you are measuring right at the power supply terminals only, you don't really get enough information. The best way is to measure the voltage at the game board, preferably at multiple places if the game board is very large. This way you can exactly compensate for the resistive losses in the power wiring (and also know if these losses are excessive). There can also be losses in the PCB itself, especially if it's a multi-board stack. Go over 5.2V and risk damage, go under 4.8V and risk going into under voltage reset. Given that you are able to measure the voltage right at the power supply output, but perhaps not at every point on the PCB, I'd say it's safer to higher than lower. A game that is going into reset is still 'broken'.

[JoeyG]

These days with low cost Oscilloscopes you can check for "valid"   Vdd  supply  ripple and reliability,  Vdd signals at the IC's and also the logic signal integrity on some of the logic and memory ics.

If you have two of the same boards   you will be able to signal comparison

[Siwastaja]

Seems irrelevant enough to have a fight over. If the original spec says supply voltage is 5V, it will work with 5V or 5.17V or 4.9V unless the design is total broken shit. And if it fails at 5.00V and seems to work at 5.17V you are not having basically any margin at all so expect it to work only slightly better but still not reliably.

If some of the modules draw significant current, then it's well possible the internal power distribution (wire gauge, card edge connectors, backplanes, PCB trace widths, whatever...) is not up to this and you sure can compensate for the voltage drop by increasing the voltage at the source. This is exactly what many USB power supplies do, outputting 5.2V so that you have at least 4.8V or so left when you combine a crappy cheap cable and a device which charges at significant current.

With increasing voltage, comes risk of overvoltage. 5.17V isn't too much for any sensible 5.00V rated system, but what about accuracy of that regulator? You set it at 5.17V, how much will it drift e.g. due to temperature? Probably not too much but it's worth thinking about. Realistically, you could start seeing problems when beyond 5.5V, so there is still good margin left, but sure, you did eat part of that margin.

If your practical experience is that setting voltage to 5.17V instead of 5.00V decreases weird problems it is well possible you are right, but that tells a lot about the design quality of the products you are working with.

[amyk]

ECL used a nominal -5.2V supply. Any relation or just a coincidence?

[alpher]

Hi, just happen to have the same experience.
Working in the 80's with various boards, you allways set the voltage at the psu at minimum 5.2v.
Sure enough, TTL chips of the day needded 5V, but you must not forget that beside main edge connector to the board there was usually a custom adapter that "standarized" boards for the operator. Voltage drop accros these was sometimes more than .5V.