×
INTELLIGENT WORK FORUMS
FOR ENGINEERING PROFESSIONALS

Log In

Come Join Us!

Are you an
Engineering professional?
Join Eng-Tips Forums!
  • Talk With Other Members
  • Be Notified Of Responses
    To Your Posts
  • Keyword Search
  • One-Click Access To Your
    Favorite Forums
  • Automated Signatures
    On Your Posts
  • Best Of All, It's Free!
  • Students Click Here

*Eng-Tips's functionality depends on members receiving e-mail. By joining you are opting in to receive e-mail.

Posting Guidelines

Promoting, selling, recruiting, coursework and thesis posting is forbidden.

Students Click Here

Jobs

Why 400 Hz in mainframes?
3

Why 400 Hz in mainframes?

Why 400 Hz in mainframes?

(OP)
Does anyone know why IBM used to require 400 Hz power to their old mainframes?

RE: Why 400 Hz in mainframes?

One is to get high disc drive speed {for faster access.}  Apparently 415Hz is/was used too.  
  

RE: Why 400 Hz in mainframes?

(OP)
The drives are AC, not DC?  Hmmm....

Still seems like a lot of money and space required for those power supplies.  Doesn't seem like a lot of bang for the buck.

RE: Why 400 Hz in mainframes?

Hi, same reason as for aircraft, it keeps the size and weight down.

RE: Why 400 Hz in mainframes?

(OP)
I have trouble buying that, cbarn.  It makes sense on planes to generate at 400 Hz, then, yes, all subsequent conversions use small inductors.  But with data centers, we already have some perfectly good 60 Hz power coming in.  Performing two conversions on that (60-400-DC) just seems like it would require extra space and weight.

RE: Why 400 Hz in mainframes?

Hi peebee, in those old days mainframes were huge enterprises taking up complete floors. They had hundreds of power units and ac motors. years ago when i was in the military all the field portable radar units had 50 to 400 converters and 1 60's anti aircraft radar even used 1600hz power.

RE: Why 400 Hz in mainframes?


One comment that is open for correction/discussion by readers...for higher AC-power frequencies in aviation, isn’t there a savings in iron but not copper for otherwise similar systems?  
  

RE: Why 400 Hz in mainframes?

(OP)
I think you have that right, busbar.  My understanding is that higher frequencies mean smaller inductors as impedance is a function of frequency; higher frequencies mean smaller inductors for the same impedance.  If you want to reduce copper, you need to increase voltage, same as anywhere else.

RE: Why 400 Hz in mainframes?

Hi, a smaller transformer has less copper as well as less iron than a bigger one. Also with power units you need smaller filter caps as well.

RE: Why 400 Hz in mainframes?

About busbar comment..cbarn24050 is right, when you have higher frequency in transformer you have less iron but still large induced voltage. So you can reduce number of winding turns, which means less copper.

RE: Why 400 Hz in mainframes?

Aircraft instruments contain gyroscopes driven by small 3 phase 400 Hz motors. A 4 pole motor will turn at 12,000 RPM, the required speed for the gyros.

The maximum speed from a 60 Hz. supply is 3600 RPM, with 2 poles, not adaquate for a gyro to function.

Speed of drive must be the reason.

RE: Why 400 Hz in mainframes?

(OP)
If you read closely, you'll see benray's answer was right on the mark, with his reference to "speed of drive."

RE: Why 400 Hz in mainframes?

Search for "400 hz" in the advanced search function of this website - there are many threads, although most pertain to use of 400 Hz in aircraft.

RE: Why 400 Hz in mainframes?

Sorry about my lack of clarity. My technical writing skills need improvement. I am accustomed to writing to myself, and sometimes leave a lot of detail out.

I am new to this forum, and don't know the protocol.

400 Hertz is used on warships for gyro stabilized compasses, gyro stabilized operating rooms, gun turrent fire control, and automatic navigation systems, where high constant speed drive motors are required.

Aircraft use is based on the same conditions and need. The automatic flight control, and navigation systems, depend on gyros for stability.

Whenever I see 400 Hertz in use, I think speed plus weight and space, in that order.

RE: Why 400 Hz in mainframes?


Aside from physical size, a 2-pole 400Hz induction motor works out to run at 24,000RPM.  That would be a nice disc-drive speed.  AC…no pesky brushes and commutator maintenance!  {Isn't 10kRPM hot stuff in PCs right now?}

RE: Why 400 Hz in mainframes?

Hi guys, i hate to disagre with you BUT there was nothing on those old IBM mainframes that ran at high speed.

RE: Why 400 Hz in mainframes?

Oh well, back to square one. What is the speed of the drives? Would 60 Hertz present a problem with EMF, that may not appear with 400 Hertz? Does filtering of high cycle  play into this application? How about harmonic frequencies?

What phasing and voltage is this 400 Hertz?

RE: Why 400 Hz in mainframes?

(OP)
I have no idea, benray.  I just know I've seen a bunch of those old power supplies sitting abandoned around old data centers and have ripped a few out.  They are really big boxes, usually at least as big as the 60 Hz UPS systems installed next to them (which are often still in operation), which is why I'm struggling with the proposed idea that they are supposed to be saving space or weight or cost.  Drive speed sounded like a reasonable idea, but I suspect those old 12" disk stacks could not have been spinning very fast (although its not clear to me that 12" stacks are the kind of drives they were running at 400Hz).

I'm kind of surprised this post set up such a heated debate, I really thought there would be one simple answer like there is on planes.  Maybe there never was a good reason for them in the first place other than IBM marketing (let's lock them in to buying a power supply from us), and that's why 400 Hz is gone now.

Not to get too far off the subject, but does anyone have any info on either 1) how to request a data center forum be set up on this site, or 2) any other links for data center and reliability engineering?  This is the best site I've found so far for engineering, but I've found little anywhere on the net that's particularly useful or focused for data center design.

RE: Why 400 Hz in mainframes?

Note that IBM was famous (notorious?) in the old days for building stuff simply to make it difficult for the "7 dwarves," e.g., Honeywell, Burroughs, Sperry, etc., to compete with.  

There may not be a technical reason for the design baseline at all.

TTFN

RE: Why 400 Hz in mainframes?

I don't know what RPM the IBM drives ran at, I do know that the big disk drives built by Texas Instruments for themselves and other brands ran at 3600RPM and from 60Hz.
Hope that info is useful.

RE: Why 400 Hz in mainframes?

Now I am really getting interested. I can not tolerate an unsolved issue, even if it is only trivia.
I worked on the construction of an IBM data center in California for So. Cal Edison.
They moved their computer mainframes from San Onofre Nuke plant. I understand a radiation leak will erase the discs.
I asked them; "if they moved the humans for the same reason?" I understood that a radiation leak would erase them also. My humor was not appreciated.

3600 RPM is the max on 60 Hertz. There has to be a logical reason for 400 Hertz. I will research this with some friends from Sandia Labs.

RE: Why 400 Hz in mainframes?

Hi,

400Hz is easier to create the - 0.85V (Logical 1) and the -1.85V (logical 0) from.
These powersuplies were capable of supplying 100-200 Amps and stayed within a few mV stable.
Try that with 60HZ....

The disk motors run on either 1 phase 50/60Hz or 3 phase 60Hz and when I remember correctly at 3000 RPM. (but I am not sure anymore)
Depending of the type of disk used (E.g. a 3330 had a 1HP 3phase 380V/50Hz motor... here in Europe)
The 400 Hz was not used for the disk motors.... only for the powersuppies.

Best regards, Jan

RE: Why 400 Hz in mainframes?

(OP)
JanH, can you provide further explanation for your statement that "400Hz is easier to create the - 0.85V (Logical 1) and the -1.85V (logical 0) from."  We are talking about DC here, aren't we?  A 60Hz to DC conversion is no more difficult to achieve than a 400Hz to DC conversion, except the capacitors would need to be larger to maintain the same ripple voltage.  It still seems like it would be much smaller, lighter, and cheaper to go from 60Hz directly to DC and provide larger caps than to go through two conversions.

If you register your name with eng-tips and give me just a little more convincing detail I'll vote you for a star. . . .

RE: Why 400 Hz in mainframes?

Hi peebee, it only seems that way because you havent done the calculations.

RE: Why 400 Hz in mainframes?

peebee,

The idea is that at 60Hz you have a 16.6 mS between top-top voltage and at 400 Hz only 2.5 mS.

A DC voltage under load will go down in voltage when the AC voltage is in the process of going through zero (alternate it polarity)
This is simple because the AC voltage is lower that the DC voltage you want.
The longer this takes the more the DC voltage will drop.
Where I live we call it the "rimple" of the DC voltage.

With 400 Hz this time is about 6x shorter then with 60 Hz
That means the this "rimple" is much smaller.

The powersupplies we talk about can handle 100-200 Amps.
This is such a heavy load that with 60 Hz you will have real problems keeping the "rimple" small enough.
When I remember correctly these powersupplies were allowed a max "rimple" of 3 mV... under full load...


Best regards, Jan



RE: Why 400 Hz in mainframes?

That's assuming you're using a brute-force full-wave rectification without any output filtering.  

The only specifications I've ever seen for 400 Hz power is MIL-STD-704, which is nearly unregulated to start with, so any benefits that come from the higher frequency are completely overwhelmed by variations in line voltage and frequency.  

I was going to go on about circuitry and decided to search for the history of ECL and ran across the following link about a computer built at Rice.  Of particular note is the mention of 400 Hz power provided by a Navy SURPLUS generator:

http://www.princeton.edu/~adam/R1/r1rpt.html

So the answer may simply be that there was an abundance of surplus high power generators available after the war

TTFN

RE: Why 400 Hz in mainframes?

(OP)
JanH & IR, stars to each of you.

Cbarn, if I had the background to do a cost estimate to build 1970's vintage power supplies, do you think I would have posted this question in the first place?  I've got a high respect for you & your eng-tips posts, but geez, you seem to have a wild hair going on this question. . . .

Anyway, I still tend to believe that the answer falls somewhere between the marketing thing and IR's surplus power supply suggestion.

RE: Why 400 Hz in mainframes?

Here's another interesting tidbit:

http://w3.uwyo.edu/~jimkirk/cyber_era.html

The jist of which is that reliable, high DC power was a big issue and that 3-phase 400Hz provided a simple solution technically, and there was an abundance of 400 Hz expertise and equipment.

TTFN

RE: Why 400 Hz in mainframes?

(OP)
Now that's what I call a convincing answer.

RE: Why 400 Hz in mainframes?

Hi,
interesting thread....slightly off topic I guess, but why did someone pick 400 instead of 350, 450 or even 500 ??

RE: Why 400 Hz in mainframes?

(OP)
In data centers, they likely copied airplanes.

Why they picked 400 or 415 in planes though seems to be a good question.  You'd think they'd want a nice multiple of 60 Hz since they were using M-G sets, but you couldn't get too high and still fit the windings in easily.  7x60=420.  It's not clear to me, though, how 420 drops to 415 or 400 -- 5% slip?  I guess frequency drift might not have been a big problem, though, so synchronous speed might not have been a concern.

Just a guess.

RE: Why 400 Hz in mainframes?

Hi,

I see that there is an idea that 400 Hz powersupplies were used because there was a surplus of them.
Well, IBM mainframes were using brand new , specially designed powersuplies, so no surplus what so ever.

400 Hz in planes is for the same reason as in Mainframes.
Smaller transformers, smaller powersupplies, smaller electromotors... and for planes very important... therefor less weight

Beswt regards, Jan

RE: Why 400 Hz in mainframes?

If they were starting from scratch, they could have used any frequency they wanted.  Why not 1000 Hz?  The transformers would have been even smaller.

The reason 400 Hz was used by a variety of companies, not just IBM, was the simple fact that the military had already developed the hardware, the design know-how and the engineering talent.  

Don't forget that the electrical engineering talent that IBM and other companies used after the war were mostly busy developing military hardware during WWII.

So, in essence, while IBM hardware wasn't surplus, most of the engineers were.  Additionally, remember that the first computers were almost exclusively used by the military, since they were the only ones that could afford it.

TTFN

RE: Why 400 Hz in mainframes?

400Hz has proved to be best frequency... actually 415Hz when my memory serves me right.
1000 Hz is too high. Transformer coils are actually heavy duty chokes...so to say,...high resistance to high frequency.

I complete agree with you that other companies also made their own supplies.
I mentioned IBM because I knew the had/have their own designs.

One thing good about a war is that technoligy takes a leap forward.
Could very well be that the 400Hz suplies were developed in the war.

Best regards, jan

RE: Why 400 Hz in mainframes?

My quick question is, alot of computer power supplies run at 120VAC/60HZ. By using a Chroma programmable AC power source we were able to run the power supplies at 400hz and up 1000hz. Our intention is to place these into military airplanes. My question is what are the effects on these power supplies due to the higher frequency? Are there any firmulas that rate effieciency of load based in input frequency? ANy help would be greatly appreciated.


joshua.noonan@tag.com

Thanks!

Red Flag This Post

Please let us know here why this post is inappropriate. Reasons such as off-topic, duplicates, flames, illegal, vulgar, or students posting their homework.

Red Flag Submitted

Thank you for helping keep Eng-Tips Forums free from inappropriate posts.
The Eng-Tips staff will check this out and take appropriate action.

Reply To This Thread

Posting in the Eng-Tips forums is a member-only feature.

Click Here to join Eng-Tips and talk with other members!


Resources