ZX81 Power Regulator

The 7805 power regulator, with heatsink attached, alongside the three solder holes (ground at the corner, then something which seems unconnected and finally 5v) which Mark says are the, "best place to pick up a +5V supply on a ZX81 or TS1000"

The voltage regulator is responsible for taking the 9V input and delivering a 5V output. You can read more about the 7805 here.

The main problem with this older component is that it generates a chunk of heat and requires a heatsink to disipate it. So... we replace it.

Perhaps the main replacement is the muRata I50115, model number OKI-78SR-5/1.5-W36-C (or you can also go for the OKI-78SR-5/1.5-W36H-C, as the extra H just means it's got horizontal pins, which you'll need to change anyway) which is listed as a 5V DC power regulator with max draw of 1.5A.

One of the advantages... in theory... of going for the H variant is that you can remove the plastic pin spacer, then heat the solder and push the pins through to the other side of the board, then replacing the spacer when you're done. But check whether the board where you're going to install it, allows enough room for this, because you might have to replace with right angle pins to get a distance advantage. Preferably check this before you buy anything :-)

Some people sell them in a package which they've already altered to be a direct drop in replacement for ZX81 and Spectrum computers like these. (search for "murata")

One of the obvious benefits of this, many decades newer part, is that it runs a lot cooler and more efficiently. It doesn't require a heat sink.

This is how the muRata codes work...

So, how much power does the ZX81 pull? Well, I hooked it up to my bench power supply and with the 7805 fitted it draws 330mA.

NOTE - This is after the RAM has been replaced with a modern RAM chip and there are no external packs attached to the system.

Looking to power everything in one case, I investigated the Raspberry PiZero and that draws 190mA, although I'll note that it is has a max of 220mA, so I'll use the higher figure to build in a little safety.

The last thing I want to power is the ZX-Wespi. On a USB connector it came in at an amazing 90mA, occasionally flashing up to 110mA. For safety, I rose that to 150mA in my calculations.

In fact, I was in for a little bit of a shock. With the new power regulator fitted and both the Pi Zero and the ZX-Wespi running off the 5v internal points, the whole thing came in at less than 400mA on my bench power supply. Wow.

This didn't make much sense to me as my calculations had this pulling nearly 800mA. Mark explained why, but it's still a fair chunk to wrap my head around. I also started a conversation on the EEV Forum about it.

CaptDon said...

I calculated the individual powers and added them. 5x.19=.95w, 5x.09=.45w and 9x.33=2.97w Something is lying in your test setup because you state 9 volts at .385a total which only comes to 3.465 watts but your individually calculated loads come to 4.4 watts. Something is wrong in your current measurement test setup.

The first thing I noticed is that CaptDon converted the draw into watts in order to do the combined calculation whereas I was trying to use ohms law. Mark appeared to do the same but added by saying this...

With a 7805, the input current will be almost exactly the same as the output current (the 7805 uses a small amount of current itself, typically less than 8mA).

That highlighted the problems with my measurement. The zeddy was using the 7805 at the time, supplied by the desktop power supply, whereas the other two units were being measured by USB which was "after" the transformer.

The final measurement was taken from the zeddy after the I50115 had been fitted, and Mark said this...

DC/DC converters / Switching regulators (like the I50115) work differently. Now rather than thinking about just current, you also have to think in terms of power. Power in Watts (W) for electricity is Voltage X Current.

...and this...

Remember, energy cannot be made or destroyed, it can only be converted from one form to another. The 7805 converts the unwanted electrical power into heat (which is why they get hot). A DC/DC converter / Switching regulator produces much less heat, because its much more efficient. Hence as the input voltage goes higher, the input current reduces.

The point of efficiency curve was detailed by Jwillis on the eevblog forum...

You also need to take in account the efficiency of the switching regulator. The 5V regulator has an efficiency of between 89 - 90.5% at nominal 12V input.At Minimal input of 7 volts efficiency rises to 91.4 to 92.5%. That means there is anywhere between a 11 to 7.5% loss between Nominal input voltage of 12V and Minimum input voltage of 7V.So for a 9V input that would equate to approximately an average of 90.2 to 91.5% efficiency with a loss between 9.8 to 8.5%. That loss is transferred into heat and dissipated.

If you look at the Performance Data (Efficiency vs. Line Voltage and Load Current @ +25˚C. (Vout = Vnom.) chart you can see that the 12V nominal is the "sweet spot" for operation. 9V will fall somewhere between the 7V min and 12V nominal. You you calculate your current in and current out based on efficiency percentage.

For a giggle, I fitted a board with a modern Z80 processor, the MyRetroStore RAM, the uLA81 and a modern EPROM chip on which I'd blown the Big Bang ROM.

61mA. How's that for a low power consumption computer!!!

By comparison, my mobile phone operates between 100 to 200mA, but that's obviously got the screen, etc. running. It's a little eye opening, though!