New CNC Router - Welded steel

[pippin88]

I'm finally getting around to building a new router.

I've previously built one of my own design which had 900x375mm cutting area (http://viglink.pgpartner.com/rd.php?r=5316&m=829557091&q=n&rdgt=1414239524&it=1414671524&et=1414844324&priceret=194.56&pg=%7E%7E3&k=43a3ca5c6c416eb01c3331885dff7bd0&source=feed&url=http%3A%2F%2Fwww%2Eamazon%2Ecom%2Fdp%2FB000EHL2GK%2Fref%3Dasc%5Fdf%5FB000EHL2GK3099990%3Fsmid%3DA23E2C033ZN7B7%26tag%3Dpgmp%2D1585%2D01%2D20%26linkCode%3Ddf0%26creative%3D395109%26creativeASIN%3DB000EHL2GK&st=feed&mt=%7E%7E%7E%7E%7E%7E%7E%7En%7E%7E%7E). It was based on SBR16 (http://viglink.pgpartner.com/rd.php?r=402&m=809891080&q=n&rdgt=1414239278&it=1414671278&et=1414844078&priceret=280.80&pg=%7E%7E3&k=4d5fac7202abfa485c9fa23a3fa2ef8f&source=feed&url=http%3A%2F%2Fwww%2Eamazon%2Ecom%2Fdp%2FB002HHKHFW%2Fref%3Dasc%5Fdf%5FB002HHKHFW3369780%3Fsmid%3DATVPDKIKX0DER%26tag%3Dpg%2D584%2D01%2D20%26linkCode%3Ddf0%26creative%3D395097%26creativeASIN%3DB002HHKHFW&st=feed&mt=%7E%7E%7E%7E%7E%7E%7E%7En%7E%7E%7E) rails with ballscrews (single screw on X). Made of 25mm plywood and aluminium extrusion. Unacceptable flex and racking, I've never been happy with it.

So after plenty of reading and research, I've come up with my replacement design.

I was initially going to bolt everything together and painstakingly level. I've since bought a welder and will be using that.

Design:
Welded steel construction
Dual drive (http://www.target.com/s?searchTerm=dual+drive) on gantry
Cutting area of greater than 1220x610mm (quarter sheet) - will be ~1250 by ~650mm in reality
Travels ~1250x675x200mm
Self levelling epoxy for rail mounting (probably West System 105 + 209)
Linear profile bearings
- already have 830mm long 20mm rails for the Y axis (short axis)
- 15mm new Hiwins for X (long axis, 1500mm along gantry) and Z axis
15mm ground ballscrews with 20mm lead for Y axis (already have)
16mm rolled ballscrews with 16mm lead for X and Z axis (new, SFE1616)

I have decided to go for a wide machine, with the gantry running the longer axis for two reasons:
1. It works better in the vast majority of spaces as it can be loaded from the front.
2. I already have a very nice pair of ground 20mm (http://viglink.pgpartner.com/rd.php?r=5316&m=1419537551&q=n&rdgt=1414239181&it=1414671181&et=1414843981&priceret=5.42&pg=%7E%7E3&k=4ad4d6903d83bcaa222acc2ad3d57e79&source=feed&url=http%3A%2F%2Fwww%2Eamazon%2Ecom%2Fdp%2FB003ABTH28%2Fref%3Dasc%5Fdf%5FB003ABTH283107783%3Fsmid%3DA21ZNGKUIZH0WR%26tag%3Dpgmp%2D1536%2D01%2D20%26linkCode%3Ddf0%26creative%3D395109%26creativeASIN%3DB003ABTH28&st=feed&mt=%7E%7E%7E%7E%7E%7E%7E%7En%7E%7E%7E) lead ballscrews with great bearing blocks.
The steel gantry will be stiffer than 3060 8020 used in CNCRP machines, and I believe will be adequate.

Maximum workpiece height is ~100mm
The Z axis travel has been designed to reach the table with 20mm tool stickout (http://viglink.pgpartner.com/rd.php?r=5316&m=1306189769&q=n&rdgt=1414068030&it=1414500030&et=1414672830&priceret=30.10&pg=%7E%7E3&k=7836a2ef10472db2cad711769406c18d&source=feed&url=http%3A%2F%2Fwww%2Eamazon%2Ecom%2Fdp%2FB005H1SGEE%2Fref%3Dasc%5Fdf%5FB005H1SGEE3366756%3Fsmid%3DA2RFSIF56F6W5J%26tag%3Dpgmp%2D1585%2D01%2D20%26linkCode%3Ddf0%26creative%3D395109%26creativeASIN%3DB005H1SGEE&st=feed&mt=%7E%7E%7E%7E%7E%7E%7E%7En%7E%7E%7E), and to clear the top of a 100mm piece of work with 80mm of tool stickout (http://viglink.pgpartner.com/rd.php?r=5316&m=1306189769&q=n&rdgt=1414068030&it=1414500030&et=1414672830&priceret=30.10&pg=%7E%7E3&k=7836a2ef10472db2cad711769406c18d&source=feed&url=http%3A%2F%2Fwww%2Eamazon%2Ecom%2Fdp%2FB005H1SGEE%2Fref%3Dasc%5Fdf%5FB005H1SGEE3366756%3Fsmid%3DA2RFSIF56F6W5J%26tag%3Dpgmp%2D1585%2D01%2D20%26linkCode%3Ddf0%26creative%3D395109%26creativeASIN%3DB005H1SGEE&st=feed&mt=%7E%7E%7E%7E%7E%7E%7E%7En%7E%7E%7E). The spindle will be adjustable up and down if I need to work with very short tools or very long tools.

Overhang over front of machine for vertical machining (65mm thick pieces) and a 4th axis (should be able to machine 150mm diameter by ~1m long)

The table top will be a layer of ~30mm MDF covering full area, with T slots and Tee Nuts, then another ~18mm piece same size as cutting area (http://viglink.pgpartner.com/rd.php?r=5316&m=829557091&q=n&rdgt=1414239524&it=1414671524&et=1414844324&priceret=194.56&pg=%7E%7E3&k=43a3ca5c6c416eb01c3331885dff7bd0&source=feed&url=http%3A%2F%2Fwww%2Eamazon%2Ecom%2Fdp%2FB000EHL2GK%2Fref%3Dasc%5Fdf%5FB000EHL2GK3099990%3Fsmid%3DA23E2C033ZN7B7%26tag%3Dpgmp%2D1585%2D01%2D20%26linkCode%3Ddf0%26creative%3D395109%26creativeASIN%3DB000EHL2GK&st=feed&mt=%7E%7E%7E%7E%7E%7E%7E%7En%7E%7E%7E) as sacrificial board.

The steel:
Large Y axis rails are 200x200mmx6mm RHS.
The Gantry is 200x200mmx6mm RHS with an unsupported span of ~1400mm.
The bearers under the table are 150x50x3mm RHS
I'll likely put a strap joining the bottom of the bearers to box it all up.
Without the stand, it looks like the machine will weight >160kg (352lb)

[pippin88]

Yesterday I got all the steel cut for the router (including a stand, not shown).

Today I squared the ends and then got welding.
The beauty of this design (in theory at this point) is that the epoxy levelling means precision is not important at this stage. Within a couple of mm is fine.

This is my second welding project. The first project I ever welded is the mill stand that is underneath and being used as a temporary welding table.
I'm not much of a practicer, so I just had at it, and my welds are improving fairly quickly. The flat weld on butt joints is going well. Fillet welds are still very hit and miss.

[pippin88]

Thanks for the interest guys.

Progress:

Stand welded and first coat of paint on:
330177

Uprights are 75x75x3mm
Horizontals are 50x50x3mm
Slats are 25x25x2mm
Weight ~65kg
Getting it up on to the router by myself to mark and drill the mounting holes was tricky to say the least. It was fine to move around until I welded the slats in. Before I could stand in the middle, squat and lift from the horizontal in the middle and get a fair bit of height. With the slats I can get in there and lift from the top horizontals and move it around, but can't squat down to lift from the middle so can't get the height.
*Just* managed to get it on top of the router without killing myself, though my neighbour was hanging out their window asking if I needed help.

Paint is $10 a litre ALDI metal paint. Read lots of good things. I'm pretty happy thus far, goes on pretty well. Stand has just been brushed.

Gantry painted:
330176

200x200x6mm 1.6m long. Weighs ~55kg.
I forgot to prepare the new roller (removing loose fibres) so ended up with the first coat covered in fibres :thumbdn:
Scotchbrite pad and a quick rubbing got rid of most of them when the first coat was dry.
Got some more with the second coat, but not as bad. Will likely give it a light sand with fine grit paper.

Then I faced the problem of how to move a 55kg 1.6m long beam that has paint that is just tack dry - I can't just pick it up and move it like I did to get it there.
So I walked it:
You'll notice it's on a 2x4 between the two saw horses (BTW these are from Bunnings and are , do not recommend, very warped and uneven). There is ~a foot of length spare.
1. Advance the back saw horse so it is close to the end of the gantry.
2. Advance the front saw horse so that it is at the front end of the 2x4
3. Slide the 2x4 forwards (making sure you don't drop of the back saw horse)
4. Rinse and repeat.
Took a little while but worked.

Next up is a few final bits of welding and grinding on the main router body and then it gets painted.

[neksmerj]

Hi Pippin88, my god, what an ambitious project. My piddly little camera slider project fades into insignificance by comparison.

I don't spose by any chance you will be using an Ardinuo uno R3 board and Adafruit stepper controller in your project?

If yes, may I call on you for some programming pointers in the near future?

Ken

[pippin88]

Ken,

Thanks for the reply.

This project requires much higher power steppers and drivers. I haven't done any motor stuff with arduino, but it should be fairly straightforward. I'd suggest start by googling, as I'm sure there should be some guides / other people's experiences.

Happy to try to help out if you get stuck, though I'm not an expert.

Cheers,
Nick

[pippin88]

Got some more done today.

Main thing was preparing the gantry for epoxy pour, which included building the dams and doing an initial pour to seal any small gaps.


330784
Gantry dams built. The form in the middle is to reduce epoxy needed and create space for the ballnut to travel. It's going to be quite tight fitting the ballnut in.

330783
As expected the area around this end was a problem. The wooden block in the middle is sitting in a cut out for the stepper motor. I had some leaks around here.

So I quickly grabbed the blutack (only putty / mouldable I had on hand) and had at it:
330782

In retrospect I think it would be easier to do the initial leak sealing pour with fast setting epoxy instead of slow (I'm using West System 209 extra slow hardener)

Hopefully do main epoxy pour tomorrow. Have to go buy a heat gun for the bubbles.

[pippin88]

Epoxy pour seems to have gone ok from a quick peek. I had some problems with an oily residue on the surface, likely because I didn't thoroughly clean before the pour. It's visible, but I hope hasn't affected the surface flatness. Going to leave it for a full week before I pull apart the form work.

Today I got the aluminium I need for various parts, and firmed up the CAD model.
330956

I also moved the router body. It weighs about 120kg at present and is not an easy shape to move around. I'd been scratching my head on how best to move this without hassling 2-3 other people to just carry it (I'm going to need to move it a few times). Then I realised I can very easily make a cart:

330958
Couple of timber beams with a few screws, and some casters I had lying around.

330962
I had the router upright from painting. I moved it to the front edge of the table, and put two saw horses, lined up with the big beams. Some rags to protect the paint.

330960
A controlled topple / slide forwards on to the saw horses.

330957
Moved the table back and rolled the cart in to position.

330959
Controlled descent on to a couple of cushions. Then I pulled them out (Note, you've got to be comfortable lifting a third to half the weight to get the cushions out, if you have a bad back I'd suggest other techniques).

330961
Tip it upright fully.

330955
And we are rolling.

Somehow I don't think it's going to be as easy working against gravity...

[ttjarrett]

Hi Pippin88,

I noticed in your CAD image that you're only supporting your ball screws at the motor end. Is there a reason for this?

Thanks,

Toby

[pippin88]

Toby,

The CAD model isn't complete. The X and Y will be supported at both ends. Z only at motor end as it doesn't need support at both.

[ttjarrett]

I thought that might be the case given that your rotational speeds would be greatly reduced without the additional support. Looking at the Hiwin ball screw catalogue, a fixed-free ball screw mounting configuration only allows for 16% of the RPM of a fixed-fixed system. Fixed-supported is around 69%. Good choice using only one support on the Z axis, many people use two which adds cost but no real benefit given the length of the screw.

Great work on your machine! I would like to build a welded machine one day. Are you worried about residual stress from the welding? It's something I've heard mentioned before as a limitation with welded frames. A bit of past research on the subject quickly suggested that stress relief in welding is a black art!

[pippin88]

Yep, appropriate support certainly makes a large difference. I use a calculator from Nook which provides the critical speed (whip rpm) for input diameter, length and end fixation.

I'd originally planned to use bolted construction, but it would have been a lot of drilling and tapping and aligning etc. In the end it worked out a neater design and easier to weld. I'm then using an epoxy pour to level and provide a flat surface.
I have thought about residual stress, but my understanding is that this us really a problem when you machine part of the welded structure. If you don't machine then the stress is relatively trapped. Over time I believe that some level of stress relieving / change occurs spontaneously. I think this heavy steel design will move much less than many routers constructed of wood or bolted aluminium extrusions. However, we are talking about a wood router and if it can hold less than 0.1mm tolerance that should be fine.

[pippin88]

Work is progressing. I had a set back while milling Z axis parts: the brass motor gear in my mill shredded itself leaving only flakes and dust where the teeth had been. I'd started a belt drive conversion ages ago but abandoned it as my lathe blew a gear and had some other issues. So after much struggling and kludge fixes on the lathe I now have the mill belt drive done.

In the meantime I've got the main router epoxy pour done. It went ok and looks good now. I had a couple of stuff ups:
1. I tried to cover it with a suspended cloth to keep dust and chips out - the cloth dumped a load of fibres all over the epoxy... It didn't do that when I used it for the gantry epoxy cover.
2. I managed to droop the cloth on to the epoxy. It'd been curing for a few hours and so was starting to gel and a few humps were created by the contact, but still seemed to start to level / smooth.
3. Watch the heat - I'm using a pocket butane blow touch to get rid of bubbles. It also temporarily makes the epoxy flow easier (viscosity falls with rising temperature). However I went a bit overboard trying to fix the spots the cloth had touched and managed to cook a few small areas of epoxy. This meant I had to fish these cured discs / thin bits of epoxy out and the epoxy had to self level again.
4. I switched to fast epoxy for the dam sealing step on this last pour. This works much better, but pay attention to the pot time. I was a bit slow and had a pot start smoking then suddenly go off. Had to drop the cup on the concrete and the disposable cup melted.
Despite those problems it looks like it's gone well overall.

Tips for self leveling epoxy
You need the right epoxy - slow, low viscosity for self levelling. I used West System 105 + 209 hardener. Certainly not cheap, but others have used with good results.
Get the object as level as possible as you'll need less epoxy
Clean the object then let it dry
There are two main stages:
A) Building a dam and sealing it
B) Pouring the main epoxy

A) The dam
You need something the epoxy doesn't stick too. It won't stick to the non sticky side of most tapes. It will stick to metal and wood.
I tried gaffer tape, duct tape, and blue painters masking tape for various parts.
Definitely use the blue tape (also called 14 day painters tape).
Advantages:
1. It's stiff and therefore forms a nice straight dam
2. It's low tack so doesn't pull your paint off, and it's possible to reposition if you don't apply it perfectly
Gaffer tape works but is too sticky and pulls off paint, and is also hard to apply as well as the blue tape.
Duct tape stretches and curls, so doesn't give a nice flat edge.
I found a single layer of blue tape carefully applied worked well, with particular attention to the corners.

You must then seal the dam. This is done by applying a small amount of epoxy around the join between the dam and the object. You only want a small amount - this is done by applying with a spoon or knife, not pouring. If you stick to a small amount it will bridge small gaps and hold in place with surface tension - if you pour too much gravity will overcome this and it will flow through the gaps. Use fast curing epoxy for this. I used West System 105+205 hardener. The fast curing epoxy tends to be more viscous (an advantage at this stage) and works quickly to seal gaps. I used very slow epoxy for my first couple of dam sealing and it was a pain - it's runnier and it has time to all flow out before it cures.
Be ready for leaks. Tape doesn't really stick where the epoxy is running, but if you whack enough on, the tape sticking to either side of the area will be enough to hold it and you can create an effective barrier. Blutack works also and is of course mouldable, but try to avoid it. It won't adhere where the epoxy already is, but does get incorporated in to the epoxy edge. However I did use it effectively to physically cover small leaks that would have been unfixable otherwise. You'll need big lumps of it.
Now you have to wait till the sealing epoxy is solid enough. It doesn't have to be fully cured / full strength. If you used fast curing epoxy (seriously don't even consider using slow) you can proceed to the main pour a couple of hours later.

The main pour
As mentioned above, you need slow curing low viscosity epoxy.
Mixing in a large shallow dish gives you more pot time.
Mix thoroughly but avoid vigorous mixing as you want to avoid introducing bubbles.
I mixed in batches based on my mixing container size.
When each batch is mixed, just pour it on. Spread out where you pour fairly evenly so the epoxy doesn't have to travel as far.

Be aware that surface tension is a significant factor and there is a minimum thickness to self level. This will vary with the epoxy, and I'm not sure what it is, but I guess and good 2mm at least. You won't get self levelling with 0.5mm thickness. I used to spoon to spread the epoxy around a bit also.
Fingers crossed you have no leaks.
Now you need to get rid of bubbles. This is done with heat, either a small blow torch / gas soldering iron, or a heat gun. Do not use a hair dryer as the air flow is too much and will blow dust in. I opted for a small blow torch as I can use it for other things. Wait a little for the bubbles to rise then gently waft the torch over the surface. Don't stay in one area too long - you'll cook the epoxy and I'm sure burning epoxy fumes aren't good you for either. You'll probably need a few passes, it's surprising how much bubbles you'll have. Bigger bubbles cause significant surface irregularities. Often you'll only see that bubbles are present if you use look at the epoxy with a bright light, such that the epoxy is a mirror and you are looking at the light bulb on the surface of the epoxy.

I left each pour a full week to achieve good strength before disturbing.

Heating the epoxy components reduces viscosity. This is done by putting the bottles in a bucket of warm water. However, this reduces pot life / speeds up curing. Apparently heating the object to be poured on still gives the benefit of better flow but doesn't speed up curing, which would be ideal.

Once your epoxy is cured you'll need to remove the fairly large meniscus. Be careful, the edges will be sharp. My hands are covered in cuts from my clean up.
Use a file to break the edges then take down the meniscus. Epoxy files easily and doesn't clog a file like it will sand paper.

[pippin88]

Progress slowed a bit as I was away for a couple of weeks.

I've done the epoxy for the 4th (rotary) axis on the front of the machine.

Major work has been the Z axis, which is largely done.
Fairly happy with how things have gone. The large backplate is bigger than the working area of my mill, so required a fair few setups to do the edges. Fortunately the holes on the face are all just within the area so could be done in one go.

It's loosely assembled below.

[pippin88]

Work has progressed fairly significantly in a number of areas, but I have found a major problem. I finally got around to checking the epoxy for level and it's not good.

I poured 3 separate areas.
1) Two beams 200mm wide, 800mm long (the beams the Y rails mount to, this is the most critical one). The final epoxy thickness is 3-4mm and these seem to have worked fairly well.
2) One beam 150mm wide, 1300mm long (the front rotary axis area). The epoxy thickness varies from ~1.5mm to ~5mm in a fairly linear fashion from one end of the beam to the other. There is a significant depression in the middle (lengthwise) of the beam of about 0.75mm.
3) A grid with members 50mm wide, 1300mm long with cross beams at 1/3 and 2/3 along length (the table area). Again minimum epoxy thickness of ~2mm. Significant depressions / not level across the grid. The grid before pouring would have had areas 1-2mm low / high.
(I also did the gantry but do have a good way to check that)

I mixed carefully and thoroughly based on weight as per the directions.
I suspect that the problems relate to surface tension / a minimum required epoxy thickness. In 2) the area with the thinnest epoxy (~1.5mm) is proud of the depression, where the epoxy is ~3mm thick and I wonder if the epoxy clung to the area rather than flowing.
With each pour I poured the epoxy over a broad area. I used a pocket blow torch to remove bubbles. The epoxy appeared uniform.

Unfortunately I've now got a few litres of not very level cured epoxy.
I've sent an email to West System to see if they can give any advice / shed some more light.

[foobillious]

sorry to hear about the lack-of-level problem pippin. how hard is it to chip off so you can try again? or would you just go straight over the top?

[pippin88]

Not too hard to chip off, but if I do any more epoxy I'll just go over the top, as it's flatter than it was before the epoxy.

The problem is now I've moved the machine position and the table needs to be on the same plane as the large beams (which worked ok). I don't have a precision level or long straight edge either.
At present I'm thinking of getting the router going and using it to skim the epoxy.

[Lucky Phil]

Yours is one of the machines I have book marked in my browser to follow as parts of my machine are almost copied from yours. I have decided to change the channel in my gantry for an "H" beam to provide more stiffness without distortion when welding it together. What size steppers are you using to move your gantry.
philip

[pippin88]

Phil,

It's a pretty simple design but quite rigid.

I'm using 381oz-in steppers (Homann designs or AusCNC / AusXmods both carry them) and they are fine. I'm using 20mm pitch ballscrews to drive the gantry, and 16mm pitch screws across the gantry and on the Z.
If you don't have ballscrews already, I recommend the 16mm pitch screws from BST Automation on Aliexpress. A bit more expensive than the standard 1605 (5mm pitch) chinese screws, but quality seems higher and they will make for a much faster machine.

I've made a lot of progress and the machine moves and cuts. I've cut some aluminium faceplates for buttons, and a couple of circuit boards. It's helping to complete itself.

Unfortunately I'm now working much longer hours and as such things have slowed down a lot.

[pippin88]

I've been slowly chipping away at this.

Everything takes a little longer than I expect...

I've now got a working basic machine.
Things I've knocked over:
Linear rails: Drilled, tapped and aligned the linear rails. This took a fair while. For alignment I eventually used the tight wire method, with thin steel wire over some pulleys with weights hanging on the end. I searched for ages for piano wire, but couldn't find anywhere with thin enough stuff. In the end I found that fishing shops have thin stainless wire with rated breaking strain (fishing leader wire) and it worked great. I used 27lb stuff. It's pretty cheap too. I then used a simple circuit with an LED ring and a razor blade on a cart on one of the linear bearings. When the razor contacts the wire, the LED ring lights up. I was able to get it so along the whole length it just skimmed and only just lit up the ring. I think the rails are straight within 0.05mm per metre or so. This was a painstaking process. I had to turn the heads of the bolts down for the gantry, as I'd hand drilled using the rails as a guide, and the rails are not straight when not bolted down.

Ballscrews mounted and aligned.

Homing (and limit) switches.

Wiring / electronics: Managed to kill a G540 driver (just one of the G250 drivers in my G540) due to making an incorrectly wired connector. I've since moved to a Mesanet 7i43 + 7i76 for step generation, and Leadshine DM856 digital drivers. I'm very happy with the combination. They are very quiet at low speeds, with the G540 the machine was loud at low speeds. I'm happily rapiding at 20m/min. I fit the PC and the electronics in an old aluminium mid tower PC case. I've got some buttons on the front: ESTOP, feed start, pause and stop, feed rate increase, decrease, spindle speed increase and decrease.

Table Two layers of 18mm MDF, will have a spoilboard on top. Involved using a very long bit to skim the epoxy flat. Then, because the back table support beam is outside of my cutting area, I had to level it to the others. I used an angle grinder and a straight edge, and just kept grinding until it was flat compared to the other beams. Then drilled and tapped mounting holes. Then drilled the mounting holes in the MDF. Then I drilled a grid of holes, on 6cm centres, and milled pockets for tee nuts to sit in. I have 242 tee nuts in the table. They take a while to install - use a screw to pull them in to the MDF, don't use a hammer (I've seen some people recommend using a hammer).

344536
Holes drilled and pockets routed for tee nuts

344535
242 tee nuts installed - took a while!

344534
Very bad photo of the table installed

[pippin88]

Did a bit of cutting today.

6mm carbide single flute spiral, 6mm DOC, 24000rpm, 5metres/min (~200ipm) - cut beautifully.

On my old machine I was cutting at 3mm DOC, 2metres/min with poor cut quality.

Next step is definitely dust collection. I'm going to build a floating dust shoe.

[foobillious]

Just a suggestion for next time you make a spoil board - I found routing hex shaped pockets about the same size as a 1/4" nut (very slightly oversize so you have a tight fit) and then press fitting 1/4" nuts into them works just as well as T nuts and was substantially cheaper. From memory a box of 50 was $2.50. Flip the board over obviously, so the nuts are on the bottom.

I can then use standard 1/4" bolts and screws (which are easier to get and cheaper than metric ones where I am) with my hold down blocks.

[pippin88]

That's a great idea and hadn't occurred to me for this. I've used that technique for other projects in the past.

The tee nuts didn't too much - I found a huge variation when I shopped around.

[Lucky Phil]

Not sure whether you can still get them or not, but in the mild steel range you used to be able to get 1/4 & 5/16 nuts with an over sized hex which would make them even better (not able to pull through the MDF if over tightened).
Philip