2
\$\begingroup\$

I have a few projects restoring old arcade video game PCBs. Some of the boards need a slightly higher voltage (~5V DC) to power all the components (esp. the ROMs furthest away from the power connector (e.g. JAMMA input). This requirement is often mentioned by other restorers.

Is this necessity to increase the voltage normal or is it related to components and/or age related resistance? I.e. if you got a perfectly working PCB from the factory years ago, would it have needed the same increased voltage requirement?

\$\endgroup\$
4
  • 1
    \$\begingroup\$ Nothing age related, just normal voltage drop due to power currents through wire and trace impedances. In old arcade games these both could be significant. TTL chips are only spec’ed to work from 4.75V to 5.25V so goosing the supply so that it enters the board at 5.1V or 5.2V may be a good way to ensure voltages at far side of the board remain above 4.75V. \$\endgroup\$
    – td127
    Commented Mar 2 at 23:52
  • \$\begingroup\$ I understand, but sometimes still have to remind myself that increasing the voltage at the power supply is not the only way to get enough voltage at all the chips. \$\endgroup\$ Commented Mar 2 at 23:59
  • \$\begingroup\$ I will be helpful if you post a good picture of your boards and power supply, and which points do you use to make voltage measurements. \$\endgroup\$ Commented Mar 3 at 2:46
  • \$\begingroup\$ I think there might be confusion here. I'm using a Namco Tekken 2 PCB with a switching PSU. If the PSU measures 5.05V, the JAMMA pin measures 4.75V and one of the chips measures 4.74V. If I up the PSU to 5.35V, the JAMMA pins measure 5.08V and one of the chips measures 5.08V. It is the drop from 5.05V to 4.75V that I am referring to in my question. \$\endgroup\$ Commented Mar 3 at 13:29

2 Answers 2

7
\$\begingroup\$

Is this necessity to increase the voltage normal

It depends what we consider normal :)

It has nothing to do with aging, but everything to do with the age in which these systems were designed, and with market pressures involved.

Back then, getting a PCB prototype manufactured was an expensive proposition and it usually took weeks to get the boards back, unless you wanted to pay rush fees. The PCB design software was rudimentary, and some layout was done by hand and thus was very time consuming. And multi-layer PCBs cost significantly more than two-layer, so for most people 2-layer was "the PCB" and they couldn't afford more layers than that.

Laying out a board, getting the prototype, realizing that "oops, the voltages are low, we have to beef up some traces" is not a big deal today most of the time, but back then it was just not done unless you sat on an excessively large pile of money. Getting heavier copper load on the board was an option that would save on the cost of redoing the films and the NC drill tape, but that increased unit costs, so was avoided. If there was a way to get something to work cheaply, even if not very elegantly, it was the way chosen.

A designer without a good mentor would do a 2-layer design and not draw out and subsequently calculate the power distribution tree. The power and ground traces compete for space with signal traces, and usually leave something to be desired. And unless you calculate voltage drops, you are blissfully unaware of what's going on.

Thus, the voltages at various points in the circuit had values lower than they otherwise would, if the designer had time and wherewithal to design the board to work with 5.0V±10% at the power input terminals.

Today, you can set up a finite element voltage drop model of a PCB layout in an hour or less, and just see exactly what the voltages will be. It is not particularly complicated if you've done it once, and in expensive EDA software this functionality is just a few clicks away. Open source tools also support it but it's a pain to initially set up.

Back then, you could do these sorts of simulations in big engineering departments, or if you were an engineer who "graduated" such a department and had good contacts for computing resources for rent by the hour to run the simulations. I had personally seen a Gerber to SPICE DC model extraction workflow in the 80s, done on a low budget, but that took a rather capable engineer, who had to know that such workflows even existed. There was no Internet to look things up - a trivial task today, but back then such experience was not googleable, you had to have either had done it, or been active in continuing engineering education and met people who did it, or had good mentors.

A lot of the arcade and similar "prosumer" systems made by media companies were marginal engineering done under extreme time constraints, with budgets that people in larger engineering firms would laugh at.

If you got a perfectly working PCB from the factory years ago, would it have needed the same increased voltage requirement?

Sure. It was designed-in, whether intentionally or not. But most of the time, it was just a "shrug, crank up the power supply and ship it".

Sometimes I still have to remind myself that increasing the voltage at the power supply is not the only way to get enough voltage at all the chips

It was common for the boards were under-designed in terms of power distribution (too much resistance), but there was no time and/or money to fix that. In arcade business, if it was good enough to ship, it got shipped, and sometimes each unit had to be tweaked individually. Design engineers were asked to make it work, and as soon as they said "I got it to work but..." they were cut off and the production got their grubby hands on it :) If you read the stories, some of the products were released with no money left. They shipped or project was canceled, and if it was The Project that was to keep the company alive, then the company went bankrupt.

\$\endgroup\$
1
  • \$\begingroup\$ Thank you for the explanation. That makes me feel a bit better about always having to fiddle with the PSU before plugging the next board in. \$\endgroup\$ Commented Mar 3 at 0:22
2
\$\begingroup\$

It they are TTL logic chips such as the old 74xx series then they should be run at 5 V with quite a low tolerance for reliable operation and can be damaged at < 6 V if I remember correctly. (The datasheets will provide specifications. Remember the Absolute Maximum ratings are just that and not the recommended operating point.)

You could measure the voltage on the chips at various points on the board to see if there are voltage drops due to PCB resistance. If so, you could consider adding jumper wires from one side to the other to provide a parallel current path.

If the power supply is some distance from the PCB then check the voltage at the PSU terminals and at the board to see if you have excessive voltage drop. You could then consider increasing the conductor size for those.

\$\endgroup\$
7
  • \$\begingroup\$ In other words, if there is a lower voltage reading on chips further away from the power source, it is considered normal for that PCB? \$\endgroup\$ Commented Mar 2 at 23:52
  • \$\begingroup\$ It's the way it worked when shipped unless something somewhere is sinking current between one side of the board and the other. A bad capacitor, for example. \$\endgroup\$
    – Transistor
    Commented Mar 3 at 0:45
  • 1
    \$\begingroup\$ Does the PSU, by any chance, have voltage sense wires coming back from the input connector on this PCB? (This would cause the PSU to output a higher voltage so that there was 5 V at the load.) And are other TTL boards powered by the PSU? \$\endgroup\$
    – Transistor
    Commented Mar 3 at 0:47
  • \$\begingroup\$ I am currently working/switching between boards depending on availability of replacement parts. I only connect one PCB at a time. Some of the PCBs work perfect with exactly 5V while others need, for example, 5.1V to prevent glitches. I am worried about faulty capacitors, too (hence the question). I don't understand what you mean with "voltage sense wires coming back from the input connector" \$\endgroup\$ Commented Mar 3 at 1:02
  • 1
    \$\begingroup\$ @RenierDelport For "voltage sense wires coming back from the input connector", see the What is the use of SENSE + and - in a PSU?. \$\endgroup\$ Commented Mar 3 at 8:43

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.