PDP-11/23 weight and floor loading

[retrobits]

I have a PDP-11 almost exactly like this one, in the same type of cabinet:

http://www.cosam.org/computers/dec/pdp11-23/

Except, instead of the blank space in the middle, there is a TU-58 dual tape drive (very light).

I'm trying to figure out if I'm going to overload my floor. The RL02 drives are 75 lbs each. I haven't been able to find weight specs on the cabinet itself, or the BA11 enclosure/power supply/card cage with the PDP-11/23 CPU and cards. (I'm ignoring the weight of the cards themselves for this estimate.) The cabinet is currently resting on a 4' x 2' piece of half-inch plywood, in order to better distribute its weight, and to save the carpet and flooring from permanent dents.

I've heard that typical residential wood floors in the US are made to handle about 40 lbs/square foot, obviously with other factors such as distance from a load bearing wall, etc. being taken into account. I've got to be pretty close to that if not above it - hence my concern.

If anyone has any insight, it would be much appreciated!

Thanks,

- Earl

 

[nige the hippy]

I wouldn't worry over much unless you fill the room with old kit (looks round loft room... now feeling a little more worried than before!)
The difference between acceptable deflection of floor joists and actual breaking strain is VAST, I can't remember the UK equivalents etc, but the 40lb per square foot loading will give you a maximum deflection of something like 1/4" in the centre of the room, and there's always a large margin for error.
A party full of people pogo-ing is much more risky, and that's been allowed for!

 

[Lou - N2MIY]

Earl,

I think you are OK, but I wouldn't put much more on top of that rack. You're probably under 250 lbs. now. That plywood probably distributes the load a bit, but not perfectly. Are you mid span on the floor joists?

Tim may chime in, but I was over his house when he had three pretty full H960s side by side in his living room. They had an 8/I with TC08/TU56 and RK8I/RK02 drives. There must have been near 3/4 ton there, but even that seemed ok.

Lou

 

[tradde]

Earl,

I think you are OK, but I wouldn't put much more on top of that rack. You're probably under 250 lbs. now. That plywood probably distributes the load a bit, but not perfectly. Are you mid span on the floor joists?

Tim may chime in, but I was over his house when he had three pretty full H960s side by side in his living room. They had an 8/I with TC08/TU56 and RK8I/RK02 drives. There must have been near 3/4 ton there, but even that seemed ok.

Lou

I had 4 H960s. But I made sure they were mostly sitting closest to the wall that was over the I-beam in the basement too. I figure I had 800 lbs. Don't forget the
two DF32s. I don't think something like what the OP has is something to worry about. I also had a single rack 8e, 2 RK05s, power controllers in another room never
with any issues. That had to be 400 lbs.

 

[tradde]

View attachment 9848 In fact here is a small picture.

 

[glitch]

Lou is correct, you should be under 250 pounds. I moved one of these in more or less exactly that configuration, plus a spare RL02 and a bunch of terminals. Total weight was under 350.

 

[1ajs]

if u put them on the main floor and load a room up like crazy with this gear just add a temp beam under that room in the basement so the floor does not bow.....

 

[Chuck(G)]

Earl, note that that's 50 lbs. concentrated on one square foot. The obvious answer is to get some framing lumber (e.g. 2×4) and build a grid to spread the weight out over a larger area. Use some 3/4" subfloor-grade plywood on top; finish however you want (I recommend vinyl composition tile--just about impossible to destroy the stuff).

You'll be way under 50 lbs./ft² even if you fill the rack with concrete.

 

[RSX11M+]

Floor Loading of a 19" rack

Floor Loading of a 19" rack

This loading issue is actually something to check, particularly with a wood floor. At some time in the past, probably while building a stone fireplace mantle, workers so overloaded the floor of my living room with materials that one of the 10" floor joists broke. These joists are on 16" centers and cover a 12 ft span. I'd estimate it took more than 2000lbs over a single joist in the center of the span - to do this damage. It was never repaired, and I discovered it when cutting into the sheet rock ceiling of the room below to install some equipment. The living room floors are 3/4 inch tongue and groove hardwood [run across the joists] with a 3/4 inch tongue and groove knotty pine subfloor [run diagonally]. This flooring served to spread the load to the adjacent joists, preventing a collapse.

That 1/2 inch plywood isn't doing anything except keeping the rack's rollers from marring the wood floors. As a spreader, well it isn't.

If your system's weight is below 1000lbs, and you have 10" floor joists, I think you're pretty safe with an unsupported span of 12 feet or less as long as the flooring itself isn't strained too much.

Flooring Load:

The best way to think of it is as if you're rolling the unit across the floor. It's weight is borne by the rollers, which while they are 19" apart, are only a few inches in diameter themselves. A rocking motion, can mean only two rollers are bearing the weight. As these two points of load traverse from being directly over a joist, to the span between - they cross a midpoint of maximum stress about 8" from each joist. This is probably your most critical point.

Imagine half the weight of the rack, supported by a couple of 16" boards themselves supported at each end, however thick your accumulated flooring is. If you're uncomfortable with that picture, then you need a change. Extra flooring layers, and full height bridging between the joists can help support the flooring.

Of course, another way to cope is to de-mobilize the rack [no feet or rollers], and place it directly over two joists and fasten it down. Some racks had this provision built-in. With the rack down on it's frame like this, flooring load becomes irrelevant. If you do this, you'll need to remember not to upset this picture after the rack is loaded [by moving it?]
​

Joist Load:

If your joists are shorter than 10 inch, or the span greater than 12 feet - you'll need to consider how far your rack is from the joist's support at any time. Fortunately, being a 19" rack means two joists will bear the load at all times, if not three. Most people keep their equipment racks near a supporting wall for aesthetic reasons, unknowingly minimizing joist loads. Just remember your decisions when you go rolling that beauty around! [center of the room may be a bad spot]
​

Those of us with multi-bay racks supported over a larger spans with fewer rollers, have a greater concern. These should probably be reserved for concrete surfaces.

Put it this way - if you can "rock" your system's rack, and you perceive the floor flexing - you may want to re-think.

Disks with 12" platters - don't like being rocked on a flexing floor by people walking round.

 

[nige the hippy]

Disks with 12" platters - don't like being rocked on a flexing floor by people walking round.

that's an interesting & worthwhile point!

 

[Qbus]

Interesting thread makes me wonder what’s the heaviest DEC PDP processor is? I have an 11/23 system but that’s light weight compared to the 11/34 I just started with. I will say that I was in no way prepared for the fact that the 11/34 would weigh 110 pounds by itself. Cannot imagine what the big Unibus systems like the 11/45 or 70 weigh? Wonder if those large systems had separate power supplies from the card frames to cut down the weight? Would also be interested in how others are doing things like lifting and placing things on rack rails without killing themselves? I bought a bunch of the old roll around A/V carts at a school auction and find them real useful for moving things or rolling an item right up to a rack and then sliding it to where it would be installed but trying to just pick up some of this stuff has turned into a real issue. Think the RK-05 may have the title for heaviest item for its size.

 

[tradde]

Interesting thread makes me wonder what’s the heaviest DEC PDP processor is? I have an 11/23 system but that’s light weight compared to the 11/34 I just started with. I will say that I was in no way prepared for the fact that the 11/34 would weigh 110 pounds by itself. Cannot imagine what the big Unibus systems like the 11/45 or 70 weigh? Wonder if those large systems had separate power supplies from the card frames to cut down the weight? Would also be interested in how others are doing things like lifting and placing things on rack rails without killing themselves? I bought a bunch of the old roll around A/V carts at a school auction and find them real useful for moving things or rolling an item right up to a rack and then sliding it to where it would be installed but trying to just pick up some of this stuff has turned into a real issue. Think the RK-05 may have the title for heaviest item for its size.

You have never lifted an RA-80 or RA-81 then. Those things are very heavy. I could barely move the one I had. An RK05 appears light compared. I'd guess
the heaviest PDP would have to be something in the PDP-10 family. I know the KA-10 processor was in an extra wide cabinet (might have been two). I think
the CPU consisted of 1500? flip chips from the discrete logic era. It was huge, but was a beauty too. The front panel was ever nicer than the 8i front panel.

 

[RSX11M+]

Heaviest PDP CPU? Hmmm... I have to include Memory and Power Supplies in my thinking...

I like the nomination of the PDP-10. A DUAL KA10 taken to the extreme of filling it with Core Memory cabinets would rate highly, although a 2 CPU KL10 (1099?) filled even with Fast MOS memory (16 MG10 cabinets) and Cache might take the prize away.

A PDP-6 system, may well have a heaver "CPU", but when Memory is included I doubt it would retain leadership simply by virtue of the KL10's larger physical address space.

The "King" of the PDP-11's would have to be the 4-CPU PDP-11/74, filled with MOS memory. I think it would come up to a DUAL KA10 (1055) with one memory cabinet.

They're all pretty big machines.

Here's a couple views of The University of Queensland KL10 System, taken around 1992.

Even discounting the weight of all those RP06's, it's a big beast. Just imagine it with 10 more memory cabinets.

This system would definitely have had to consider floor loading. I think it was on the 5th floor?

 

[Chuck(G)]

Yup, but with those raised floors, you had a steel frame that distributed the load very nicely.

I wonder what happened to those things? I would have thought that they'd still be available on the surplus market--a great place to run cables and raise rats.

 

[RSX11M+]

Yup, but with those raised floors, you had a steel frame that distributed the load very nicely.

I wonder what happened to those things? I would have thought that they'd still be available on the surplus market--a great place to run cables and raise rats.

Curiously, the structure of the raised floors, didn't spread the load much beyond the footprint of the system above - perhaps a foot in total. Remember too, the space below was often a part of the cooling / ventilation scheme [see 11/74 pic].

The 11/74s are said not have made it out of DEC Engineering, but I know different. I'm not sure if any survives however.

The UQ KL10 system is more of a mystery. Even their KA10, which belonged to the UQ Museum of I.T. may not have made it to safety when the Museum was closed. I see no reason for secrecy surrounding these disbursements, so I am suspicious that it ever took place.

 

[Doug G]

I wonder what happened to those things? I would have thought that they'd still be available on the surplus market--a great place to run cables and raise rats.

I was one of the "rats" once. In the 70's when I worked at CDC I volunteered to provide unskilled labor to a New Years Day "let's move a Fortune 50 datacenter 20 miles across town" project. They had some 20,000 sq. ft. of raised flooring for their 2 or 3 CDC Cyber whatevers, and the new flooring was high enough that I'd dive in with a handful of cables and crawl under the floor for 20-50 ft, and pop up and hand the cables to someone who knew where to hook them up. It was quicker than opening every floor tile and fishing the cable under the frame for each square. This worked because the facility was brand new and there wasn't any clutter under the floor already.

Their systems had to be up and running the by the first work day of the new year. CDC had nearly 100 employees working on the project, all of the Cyber techs and all of us they could get from other groups that didn't work on mainframes. It was a successful move, everything was up with quite a few hours to spare.