Fujitsu TX1320 M3 - Cheap low power server (barebone)

[da8833]

Well, I am using PETG, too. If the caddies do not sit in the cage overly tight, maybe it does not matter all that much. But still, the lattice would be a concern for me. Effectively leaving just two small spots at the edges that might snap easily under load when pulling out the caddy (sometimes you have to pull in the middle where the pattern is, at least I have to). Curious to hear about your results, though. What works, works.

 

[richard.dzavoronok]

Interesting. Curious to hear how your design is doing regarding overall rigidness, though. I would be really afraid that pulling out such a caddy might be prone to breaking the lattice, particularly if it fits somewhat tight (some of mine do).

The hexagonal holes do print perfectly fine without any supports, by the way (for FDM at least). And the hexagons nicely resemble the design pattern of the original caddies. So actually no need to change this from my experience.

Either way, any design improvements are very welcome.

Just wanted to reduce print time and test it as soon as possible
I was pulling caddy out by the latch so no problems with removing drive. Putting drive in is another thing, you are right on that part.

So my printer decided to just not print few top layers. I had to guide the drive into the socket a little bit because of that. -> I'll print ds8833s design tomorrow and check that also.

Overall dimensions are good fit. Definitely not very tight. (altough I did set extrusion to 105% -> overall dimensions increase from this could be ~2% so no biggie)

Remarks on the design:
1. I did press on the center part of lattice with my thumb a lot when I was putting drive into the slot. -> better support of that part will be good thing to do
2. I would remove manual supports as they are hard to remove in post
3. would be nice to create small holes for LED lights as caddy is now completely blocking them
4. screws being flush with the side is mandatory thing, I tried my flat head screws and was not able to insert caddy, had to use countersunk ones -> if you don't have countersunk screws, it will not fit

Tomorrow I'll print ds8833s design with my remarks included and post findings.

 

[Bjorn Smith]

screws being flush with the side is mandatory thing, I tried my flat head screws and was not able to insert caddy, had to use countersunk ones -> if you don't have countersunk screws, it will not fit

I think thats more or less mandatory with all caddies - at least those that I have ever used. If screws are not flush with the surface you are in trouble.

 

[richard.dzavoronok]

So I did few changes to the design, tested it and unfortunately I'd rather not use it.

Latch is too far to the back. Sides are too short when aligned with original caddy. I was afraid of backplane damage.

I found another design which should be designed specifically for M3/M4. It has solid bottom and is holding hdd just with pressure. I don't like that and I'll change it to properly cool and hold drive.
It takes a little long because I'm able to print only one caddy per day so thanks for patience

 

[da8833]

Thanks for the update.

Well, only thing I can say is that for me the design I had posted previously works absolutely fine. I have mounted several HDDs and SSDs in my machine without any issues, ever. The different latch seems to be a non-issue as far as I can tell and the SATA/power connection is flush.

I wonder how you expect to potentially damage your backplane if the sides are "too short"? As long as the actual drive ends up in the correct position with respect to the connectors (which in my case it does) and you are not pushing overly hard, I can not really imagine any issues.

I also had a go at the other design you mention. Did not really work for me and I really didn't like the solid bottom, either.

Good luck, in any case!

 

[Bjorn Smith]

It takes a little long because I'm able to print only one caddy per day so thanks for patience

Community service is appreciated We will wait with patience - I wish I had a 3d printer myself, but I have looked at many - and its a jungle - I don't want to spend a fortune, but do not want a shitty one either - so I wait with patience on those more fortunate than myself

 

[richard.dzavoronok]

I wonder how you expect to potentially damage your backplane if the sides are "too short"? As long as the actual drive ends up in the correct position with respect to the connectors (which in my case it does) and you are not pushing overly hard, I can not really imagine any issues.

Drive is protruding to the back and when pushed into the cage, frame front is not "flush" with the cage. It just looks like it "should" go a litle bit deeper which can result into pushing too hard.
Maybe I'm just paranoid
You are right you have to push overly hard do damage anything.

 

[da8833]

I see. Given you have the original caddy at hand, you probably have a much better chance at achieving a fully fitting design adaption. Maybe you could even re-create the forked end of the latch? Anyway, thank you for sharing your progress.

 

[richard.dzavoronok]

Okay, I'm printing/testing final design tomorrow (iteration 4). Today I only tried 1 iteration, was hoping for two, but what you can do.
I did merge two designs into the one, to have best of two designs.

What dimensions of screws are you using? Could you guys measure it with calipers? I have only one type and they fit perfectly/tightly into the countersink hole. I could enlarge the hole if widely available screws are larger than mine.

 

[Bjorn Smith]

What dimensions of screws are you using? Could you guys measure it with calipers? I have only one type and they fit perfectly/tightly into the countersink hole. I could enlarge the hole if widely available screws are larger than mine.

To be honest I don't think you should change the size of the screw holes. I think there is a standard size that is used by caddies and this should use that. The one I tested worked perfectly with "normal" screws - but it is certainly possible to get screws with same diameter and thread size with different size heads - but if the holes become too big, the smaller headed screws will potentially fall through the hole and not hold the disk properly in place.

Also I do not have a tool to measure the size of the screws - but now that you mention it - I definetely need one now - its a tool I will rarely use for anything, but its cool to have - so definetely a must have

 

[richard.dzavoronok]

Yep, that tool is called calipers, very useful once on a year
The hole I wanted to enlarge is countersink hole, not hole for thread That place where screw head sink into, to be flush with side.
But it's very easy to make it larger with drill or sandpaper. Or just screw it tighter

 

[richard.dzavoronok]

Shameless edit 3: If anyone is interested in these caddies I can print them for you in PETG and test for fit. (temp resistance 85ºC, plenty room for power hungry drives). Just contact me. Not sure about price, but sure we can come to agreement for sure Shipping from [EU]SK.

Enjoy guys. I hope it will be hassle free for you.

It has good latching (drives are not going anywhere), good fit.
I did my test prints in PLA, so keep that in mind. Print in ABS or PETG can yield different results (I think latching could be a little too stiff).
If you have any control over print settings, use 50% infill (at least this is what I used ). I used Cubic Subdivision type but that should not matter that much. No supports needed as @da8833 already said.
Note that you should handle empty caddy with care. If you bend the sides too much, it will snap in the part I highlighted in picture below. I tried adding reinforcements and fill that part of the grid, but then it could end up being too stiff hence unable to release latch. And simply snap in the different part.
After drive is installed, it's very rigid.



Results:

Edit: sorry for dusty machine, I received it today
Edit2: Licensing: I'm sharing my design with you to 3d print it for free. But only I as owner can sell design. And only I can sell 3d printed caddies printed from my design attached in this post.

Design:

 

[joeribl]

Looks good, thanks for sharing your work with us

 

[Bjorn Smith]

Enjoy guys. I hope it will be hassle free for you.

Yes - thanks a lot for your effort - I will get a test print made in PETG just to see how it turns out.

Infill - I am not sure that that controls - both of my test prints in PETG only had 20% and seem sturdy enough - does it control how much material that is in the "solids" areas?

I wonder if it makes more sense if you are printing a solid figure, where it would be a waste to make it solid on the inside?

 

[richard.dzavoronok]

I will get a test print made in PETG just to see how it turns out.

Probably the best approach before ordering 20 of them Please share your feedback so we can do better design.

does it control how much material that is in the "solids" areas?

In 3D printing, the object is printed multiple ways. Walls and infill. Walls are outer shell, other is infill. Infill is sort of like scaffolding but on the inside. When slicing (process of converting 3D model to movements of print head in form of gcode) you can specify how many walls you want and how thick they should be, everything between walls is infill. You basically can't see infill after object is printed. i.e.: if you print simple cube and specify two 0.4mm walls and 20% infill, you get 0.8mm thick shell and 20% of volume inside that shell will be filled with infill.
If you set infill to 100%, you get solid cube. That's base concept and I can't really answer your question yes/no.

I wonder if it makes more sense if you are printing a solid figure, where it would be a waste to make it solid on the inside?

Infill was introduced to save printing material on large objects. Printing solid (100% infill) will consume more material, will be more rigid and it will print faster. With my printer, the solid print is 40minutes faster than 50% infill. As printer has to move a lot (depends on infill pattern).
I would say just compare prices if you can. Really depends on what is more expensive. Print time or material.
Making it solid is not waste in this case I think. This model has thin parts and very few areas will benefit from smaller infill.

 

[Bjorn Smith]

So I finally got my print delivered of the model in post: https://forums.servethehome.com/ind...p-low-power-server-barebone.37804/post-355184

And it works perfect.

Screw holes are on the right side and locking mechanism locks.

So for me this is the perfect model.

I paid 3.52 EUR to get this printed+shipping - and I will now order a batch to get my servers filled up.

In case of any confusion as to what model, its the "TX1320 M3 caddy.stl" file

And when ordering 12 I can get them all printed for 18.60 EUR total (1.55 a piece)+ shipping - and then it becomes very cheap , 30 pieces=34 EUR=(1.13 EUR a piece)

If anyone is interested I can order more so the price probably go down a bit - and then I can ship them out when I receive them. I do have to warn you though - shipping is expensive from Denmark - so no matter how many you would order expect shipping to be around 20 EUR. ( up to a couple of kilos)

 

[Bjorn Smith]

I guess nobody is interested in getting cheap caddies I will order a batch for myself and @Rand__.

I will wait until 14:00 CET and then I will order, so anyone interested have until then to get cheap caddies with expensive shipping from Denmark

 

[phariseo]

Enjoy guys. I hope it will be hassle free for you.

It has good latching (drives are not going anywhere), good fit.
I did my test prints in PLA, so keep that in mind. Print in ABS or PETG can yield different results (I think latching could be a little too stiff).
If you have any control over print settings, use 50% infill (at least this is what I used ). I used Cubic Subdivision type but that should not matter that much. No supports needed as @da8833 already said.
Note that you should handle empty caddy with care. If you bend the sides too much, it will snap in the part I highlighted in picture below. I tried adding reinforcements and fill that part of the grid, but then it could end up being too stiff hence unable to release latch. And simply snap in the different part.
After drive is installed, it's very rigid.

View attachment 26127

Results:
View attachment 26125View attachment 26126
Edit: sorry for dusty machine, I received it today

Design:

Thank you for this design. What kind of screws should I get for this one? Is it countersunk M3?

 

[richard.dzavoronok]

Thank you for this design. What kind of screws should I get for this one? Is it countersunk M3?

Yes, should be it. I'm not 100% sure, because I used screws that came with my other caddies. But according to the Intel page for 2.5" SSD these are the specs:

Metric M3
Total length: 3mm
Thread length: 2mm
Head diameter: 4mm

 

[phariseo]

Yes, should be it. I'm not 100% sure, because I used screws that came with my other caddies. But according to the Intel page for 2.5" SSD these are the specs:

Metric M3
Total length: 3mm
Thread length: 2mm
Head diameter: 4mm

Thank you, I printed it in PETG and it works perfectly now with countersunk M3x4 screws I found in my workshop.