VAX 780 on eBay

[Paul Koning]

> On Jan 1, 2022, at 3:08 PM, Grant Taylor via cctalk <cctalk at classiccmp.org> wrote:
> 
> On 1/1/22 12:44 PM, Paul Koning via cctalk wrote:
>> ...
>> And of course 400 Hz gave them an additional advantage in reduced transformer size and reduced ripple (for a given filter capacitor).
> 
> I was going to ask how the frequency altered the size of the ripple, but "for a given filter capacitor" makes sense.  You are messing with the ratio and having a fixed value "for a given filter capacitor".  --  If I'm even remotely understanding correctly.

The impedance of a capacitor is inversely proportional to the frequency.  Given a desired ripple for a given load, you get the needed capacitor impedance, so if you increase the freqency you'd end up with smaller capacitors.  You see this, big time, in modern switching supplies, where the frequency may into the many kHz or sometimes even a MHz or so, resulting in tiny filter capacitors.

>> They'd do this even in devices that don't draw large amounts of power, for example the DD60 console display gets its power (other than fans which are single phase mains power) from the 400 Hz 3-phase supply feeding the mainframe.  I'd be surprised if it uses even a kW, so this was probably a case of "why not since it's there to be used".
> 
> My understanding is that some of the special power requirements had more to do with isolation and cleanliness / quality of the power.  Meaning don't try to power the equipment off of the same circuit(s) that motors and the likes are on which will induce lots of noise on the line, some of which might make it into the computer and adversely effect things.

True.  And I forgot about the fact that you'd want to run devices interconnected by logic level signals (which is the case here) from a common power source.  A lot of CDC peripherals are transformer coupled (the coax signals used in I/O channels are built that way) but the DD60 connections are DC-coupled.

> The motor / generator provides a quite strong isolation for things like that.  Probably more so than just a normal transformer.

Way more because of inertia.  Not quite a UPS, but short power glitches enough to blink the computer room lights are not going to be seen by the mainframe.

	paul