Supercomputers, fishing for information

Guy Sotomayor Jr ggs at shiresoft.com
Tue Nov 8 11:33:37 CST 2016


> On Nov 8, 2016, at 9:22 AM, Paul Koning <paulkoning at comcast.net> wrote:
> 
> 
>> On Nov 8, 2016, at 12:08 PM, Guy Sotomayor Jr <ggs at shiresoft.com> wrote:
>> 
>> 
>>> On Nov 8, 2016, at 8:47 AM, Jon Elson <elson at pico-systems.com> wrote:
>>> 
>>> On 11/07/2016 10:31 PM, Jon Elson wrote:
>>>> On 11/07/2016 07:59 PM, Mark Linimon wrote:
>>>>> On Mon, Nov 07, 2016 at 11:23:58AM -0800, Chuck Guzis wrote:
>>>>>> But if you're a suburban resident living on Mulberry Street, anything
>>>>>> but single-phase is pretty much out of the question.
>>>>> Oh, you can get it -- but be prepared for a large hassle.
>>>>> 
>>>>> A former neighbor had a 440V 3-phase Italian lathe in his backyard shop,
>>>>> among other toys.  After he was laid off from his aerospace job doing
>>>>> machining it was how he made his living.  He was a very handy person
>>>>> to know :-)
>>>>> 
>>>>> mcl
>>>>> 
>>>> I have two 3-phase machines in my shop (Bridgeport mill and Sheldon lathe) and run them each off a properly-sized VFD. 2-phase in, 3-phase out, plus variable speed and dynamic braking.
>>>> 
>>>> Jon
>>>> 
>>> And, of course, that is really SINGLE-PHASE power on 2 wires, just to save anybody the trouble of correcting my error.
>>> 
>> 
>> I’m looking to have to do something to get 3-phase for the IBM 4331 gear.  I haven’t quite added up the power requirements yet but I’m guessing its going to be in the 10-15kVA range.  Since the power to all of the gear is really split between 3 loads (string of 4 3340 drives, 3803 control unit + 2 3420 tape drives and 2821 control uint + 1403 printer + 2540 card reader/punch) I need to figure out if it’s best to have one big converter or 3 smaller ones.  It’s unlikely that I’d be running all of the peripherals at once.  The 4331 itself runs off of single phase 220v.
> 
> A VFD is a good option and may be quite economical if you get one of the low cost simple ones.  I have one (3 hp model for my lathe) that cost only a bit over $100, though the price has gone up since.  (Westinghouse TECO brand.)  VFDs specified for single phase input tend to stop around 3 hp, as far as I have seen.  Rumor has it that higher power units will also work (possibly with some derating) even though they claim to be 3 phase input, when you feed them just one phase on 2 of the 3 wires.  I haven't tried that (but it matches how my VFD is connected).
> 
> The other option is a "rotary converter".  Basically that's a 3 phase motor connected to one phase power (with a start and run capacitor); it generates the missing phase roughtly in dynamotor fashion.  Those can be built (articles on the web) or bought from machinery supply companies such as Enco; they show models up to 20 hp, i.e., about 15 kW.  When I was looking into converters, I found VFDs to be the less expensive option.  The instant reverse and variable frequency features were also attractive for lathe use; for powering computers that would not apply.  Well, not unless you need 400 Hz for your Cyber 6600 -- in which case you'd need to check the VFD will go that high, not all do.
> 
> Given that you have a number of smaller devices and that not all might need to run, several smaller converters sounds like a good option, especially if that gets you into the "economy VFD" range.

Yea, that’s what I’m struggling with.  The issue is that the control units power the devices that are connected to them (from what I can tell), so I have to power the entire string as one unit.  The same goes for the 3340’s - the entire string is powered as a unit.  The string of 3340’s need ~5kVA (I don’t know how that translates to HP).  I’m still trying to figure out the requirements for the other strings.

TTFN - Guy



More information about the cctalk mailing list