PDA

View Full Version : IMSAI 8080 Chassis warping up under transformer - why?



Crawford
April 21st, 2010, 02:56 PM
Folks,

I noticed that my IMSAI 8080 chassis is warped under the power transformer. The catch is that it is warped UP (yes up) about half an inch. At the recent MARCH CP/M workshop I noticed that Rich Cini's IMSAI is warped up in exactly the same place, but less deflection. There were 2 other IMSAI's and they were not warped.

Anyone else notice this, or have an explanation? I think that it may have been a) heat from continuous running or b) extended storage upside-down.

My IMSAI is clamped to my workbench with spacers trying to convince it to 'go straight'.

Just curious if anyone else sees this on their IMSAI 8080...

Crawford (taking a short break from 12-bit to hang with the 8-bit crowd)

Chuck(G)
April 21st, 2010, 03:07 PM
Got a photo? I've never seen this before.

glitch
April 21st, 2010, 05:13 PM
Is the warping under the mounts for the power transformer? If so, there may have been a lot of current draw from the power supply at some point that could have heated the transformer, and the chassis under it. Since it was essentially clamped by the transformer mounts, it'd have to expand up or down. I've actually seen this before, but in tube equipment, which generally gets a lot hotter than computer stuff!

paul
April 22nd, 2010, 01:40 PM
Just speculating, if there are pressed-in studs installed to support the power transformer the local stress may be affecting the straightness of the nearby bend in the chassis.

Chuck(G)
April 22nd, 2010, 03:28 PM
Just speculating, if there are pressed-in studs installed to support the power transformer the local stress may be affecting the straightness of the nearby bend in the chassis.

No, bolts, nuts and threaded spacers. Here's a photo of the bottom (http://maben.homeip.net/static/S100/IMSAI/photos/bottom.jpg) and you can see where the power transformer attaches. Paul may be right--it might well be heat, although you'd think it'd be a 50-50 chance of bulging up or down.

RichCini
April 22nd, 2010, 07:02 PM
Assuming that heat plays a role *and* the machine is sitting in its normal upright position, wouldn't the base plate deflect downwards because of the weight of the transformer? If the base made out of steel, could the transformer act as an electromagnet and possibly deflect the base plate upwards? I can't imagine the heat getting so high as to soften the steel (no idea as to the alloy, but it's somewhere around 1000 degrees C). I might pull out my IMSAI and put a thermocouple under the transformer and take some measurements.

glitch
April 22nd, 2010, 07:36 PM
It's not softening the steel (or aluminum, et c.), it's just heating it. The heat causes expansion, and since the metal is secured by the bolts of the transformer mount, it can't expand uniformly. It's not deflecting because of the weight of the transformer, just because it's expanding from heat. You can see this in cheap aluminum pots and pans that have been taken from the stovetop and had cold water ran into them -- the bottoms are never flat.

You wouldn't get any magnetic pull from the transformer since it's passing AC and not DC. If anything, the AC would degauss the metal.

Dwight Elvey
April 23rd, 2010, 06:52 AM
Hi
If it was heat, it makes sense that it would bend towards the heat.
The cold side wouldn't expand more.
Another possibility is that the feet on the transformer are
not flat. This might also have the same effect.
Dwight

Fredbug
April 25th, 2010, 10:44 PM
It's not softening the steel (or aluminum, et c.), it's just heating it. The heat causes expansion, and since the metal is secured by the bolts of the transformer mount, it can't expand uniformly. It's not deflecting because of the weight of the transformer, just because it's expanding from heat. You can see this in cheap aluminum pots and pans that have been taken from the stovetop and had cold water ran into them -- the bottoms are never flat.

You wouldn't get any magnetic pull from the transformer since it's passing AC and not DC. If anything, the AC would degauss the metal.

If the bottom is aluminum, the magnetic field may be acting similar to a electric meter.

Bruce

Crawford
April 26th, 2010, 01:38 PM
Folks,

Well, some interesting observations for sure. I regret not taking a 'before' picture of the phenomena. The check I did last night showed the deflection was down quite a bit so the clamps are working - I am cranking it a bit below 'normal/flat', so the end-result will be straight.

I do know that my rubber bumpers on the bottom were mashed flat, even under the transformer warp. So, I'm leaning towards the heat-based warp.

-Crawford

paul
April 26th, 2010, 05:18 PM
If you remove the middle rubber foot under the transformer, place the computer normally on the table and then press down on the top of the transformer you may just as easily straighten it out. The casing design should have another screw from the side panel to the chassis floor near the middle.

Crawford
April 28th, 2010, 06:33 AM
Paul,

The IMSAI chassis is 0.075" (13 gauge) aluminum, and formed into a boxed shape. This is serious metal (with an equally serous warp in it). I have multiple bar clamps on it, I would guess at least a couple of hundred pounds of pressure on it. Unfortunately, there is no center screw (at least on my IMSAI), only two on the sides, each 1/4 of the way from the front or back. All that just to say that just leaning on it wouldn't work. I have another heavy transformer (about 20 lbs) that I balanced on top of the installed transformer, and there was no deflection downwards.. that's why I tried the clamps.

Dwight Elvey
April 28th, 2010, 04:57 PM
Paul,

The IMSAI chassis is 0.075" (13 gauge) aluminum, and formed into a boxed shape. This is serious metal (with an equally serous warp in it). I have multiple bar clamps on it, I would guess at least a couple of hundred pounds of pressure on it. Unfortunately, there is no center screw (at least on my IMSAI), only two on the sides, each 1/4 of the way from the front or back. All that just to say that just leaning on it wouldn't work. I have another heavy transformer (about 20 lbs) that I balanced on top of the installed transformer, and there was no deflection downwards.. that's why I tried the clamps.

That pretty much sinces it. It is surely caused by heat difference.
Dwight

Chuck(G)
April 28th, 2010, 06:10 PM
Agreed, Dwight. This is where I'd probably go at the aluminum with a rawhide hammer and a dolly. I think if you want to keep it from happening again, I would recommend that you cut and bore a piece of fairly heavy sheet steel to sandwhich between the PS PCB and transformer and the bottom of the case.

Crawford
April 30th, 2010, 06:01 AM
All,

Thanks for the inputs. For now I think I'll see if the clamp-straightening works - last check showed only a 2MM deflection. I thought of something similar to what Chuck suggested, but heat-treating the chassis in an oven whilst clamped to a plate, and perhaps quenching it. (Honey? what are you cooking? Ummm ... nothing dear!). I'll keep the truss idea in mind for if it recurs.

-Crawford

Dwight Elvey
April 30th, 2010, 07:49 AM
All,

Thanks for the inputs. For now I think I'll see if the clamp-straightening works - last check showed only a 2MM deflection. I thought of something similar to what Chuck suggested, but heat-treating the chassis in an oven whilst clamped to a plate, and perhaps quenching it. (Honey? what are you cooking? Ummm ... nothing dear!). I'll keep the truss idea in mind for if it recurs.

-Crawford

Hi
An alternative to bracing is the opposite. Drill a bunch of holes
under the transformer. This will give the metal a place to releave
the tension from the heat and also improve the cooling of the
transformer.
Another attack would be to place an insulating sheet under the
transformer. I don't like this one as it would decrease cooling
for the transformer.
Dwight

glitch
April 30th, 2010, 08:37 AM
What I've done with tube equipment is to drill out the transformer mounting holes, and place a rubber or plastic grommet in there. You then secure the transformer to the grommet with washers on either side. In my experience, this allows for enough movement of the chassis to prevent deflection. It also decouples the transformer (mechanically) from the chassis, which can reduce microphonic hum in tube stuff, since a transformer under load vibrates.

paul
April 30th, 2010, 02:23 PM
..heat-treating the chassis in an oven...Hmm, you might be very sorry if you did that. The chassis will be full of residual stresses from manufacture of the sheet metal, fabricating the bends, and aging - which might result in a warped mess if given the chance to re-distribute. Best to leave it as-is.

Chuck(G)
April 30th, 2010, 03:40 PM
Got a friend with an English wheel and a flat anvil (lower roller)? He'll make short work of smoothing your chassis out. Otherwise, a rawhide hammer and a dolly will do the trick, but be a bit more difficult to get exactly right. Someone who knows how to wield a planishing hammer can also do the job.

(disclaimer: I have an English wheel)