New CNC build - AVID 4848Pro with GCnC Electronics

[Ronald Blue]

Just curious but are there lock washers used? Maybe I should ask are there supposed to be? Socket head bolts use a different lock washer called a hi-collar lock washer. That seems like a place that you would use lock washers or loctite. That could be the difference you need.

[Ben Grefe]

Ronald, good thought and one that I had as well. There are no lock washers anywhere in the kit and Avid's instructions are pretty good elsewhere about exactly what hardware to use.

Turns out I was right, the plate was machined wrong. Avid called me up yesterday and confirmed that they had a batch of them with counterbores about 2mm too deep and their solution is to use 16mm fasteners instead. I used this as an opportunity to upgrade my Z axis to the 12inch ball screw model as well (sticking with the 8 inch extrusion for the moment) for easier expandability in the future. After a quick trip out to Avid I'm back in business. Probably a couple more days and I'll have some build pictures to share.

I figured another topic that might help some others in the future is a full breakdown of what I bought and cost. Seems like lots of people have cost questions and none of the CNC vendors make cost as obvious as it could be. On top of that, CNC's balloon in cost quick. I think I started with an initial 2500 budget 2 or 3 years ago, then went up and up and up. I did ratchet my budget back down eventually when I was looking at machines well over 20k. The other thing is that I know Avid has upped their prices over the years, so these numbers won't be accurate.

So, that being said, here's the breakdown. Stuff in the 'barebones items' is basically the bare minimum for what you need to buy to have a working CNC machine. Sure you can swap a spindle for a router, but either way you're buying something. Stuff in the 'upgrades' section are extras that I'm hoping/expecting to make the system more enjoyable to use.

Barebones items:

PRO4848 CNC Router Kit - $4,375.00 from Avid
3 HP Plug & Play Spindle + VFD - $1,695.00 from Avid
Pepperl & Fuchs Proximity Sensors - $260.00 from Avid.

ReZurrection Control System (base) - $2795.00 from GCnC


Upgrade Items:

3 HP Spindle tramming mount - $105.00 from Avid
Auto Z touch plate - $129.75 from Avid
Leg Kit - $569.00 from Avid
ER20 Collet & Nut Set - $79.00 from Avid
Upgrade to 12 inch ball screw Z axis - $100 from Avid

ReZurrection Closed Loop Upgrade - $1500 from GCnC

Guesstimating - $200 in wiring/dust collection upgrades
Guesstimating - $200 in end mills added to the shop in prep (I had lots of 1/2 and 1/4 end mills, not many tapered/compression/flattening).

Shipping, which was a mix of drop shipping to Avid and shipping from GCnC added another $465 and then there was some sales tax as well. Avid did offer a 1% discount for paying my e-check as well.

So - base/barebones price to get a functional system would have been $9,125 + shipping & tax. Upgrades added ~$2,890 + tax.

I'm not trying to flaunt the price, but more just give people an idea of what getting one of these machines into your shop will cost. Hoping it helps level set some of the other discussions I see in here. I've already found one or two areas that I could have (but didn't) saved money and obviously did a late add on the Z axis.

[ChrisA Edwards]

Much appreciated Ben, that helps a lot.

[Ronald Blue]

Glad you figured it out and glad they realized their mistake. That is great that they took responsibility. Says a lot about Avid as a company.

[Patrick Kane]

Ben, That sounds about right. Like you, i had an original budget around $2000+/-. I ended up finding a cnc shark on craigslist for $700-800 and jumped on it. The machine could vcarve ok, but it was semi-incapable of cutting anything. I used it to cut the templates for a maloof rocker using 3/16" plywood, and it sucked. Gantry flexed like crazy and it took hours to cut a few templates. I then sold that machine for $2500+/-. It was a good lesson for how the lower end machines lack the rigidity to make cuts at any real speed. Im guessing the xcarve and shapeoko are better than the CNC Shark, but i cant imagine they are significantly better. Theres only so much you can do with that budget.

Which brings me to your realization--and i think you are correct--$10,000 is kind of the benchmark for a machine that is worthwhile. Very interested in your take once you begin cutting. CNC routers have wild capabilities, but it is a pretty crazy entry price. $10k would cover my 20" jointer, 20" italian bandsaw, 3hp cyclone, 20" planer, and a unisaw(i later sold this). In essence, a fully capable furniture shop.

[Jim Becker]

There is a reason that so many folks experienced with CNC overwhelmingly say "Buy your second CNC machine first"...

[Ben Grefe]

Alright, it’s about time I post another build update. To summarize, the machine cuts wood! Hooray. It’s not ‘complete’ yet, but it’s dang close.

Picking up from where I left off last time, I got the new fasteners from Avid and continued assembly of the machine. I don’t think I ran into any more hiccups throughout the rest of assembly. I did end up assembling the gantry as one massive unit and then put it on the support arms instead of Avid’s preferred piece by piece method. Using Avid’s directions ‘by the book’, you need almost 6 feet on each side of the machine to thread the linear rails into the machine.

D7EF7FAE-05D0-4FA1-A62E-58614178C3B3.jpg

At this point it was time to setup the electronics package. Part of Avid’s instructions have you install the steppers into the gear assembly, so that’s done early on, but then I hooked them up 1 at a time to test operation. Here’s where I ran into one of my first problems (actually 2). First, the Z axis assembly provided by Avid must expect steppers with slightly shorter axels. Specifically, about ~2mm shorter. This means that my Z axis stepper wouldn’t fully seat into the Z axis assembly. Oddly, NEMA34 is a industry specification, so I’m not sure whose violating the spec. I decided my temporary solution was to use grade 8 ‘thick’ washers to space the Z axis stepper. In the future I might built an interface plate out of aluminum to make the connection even more solid. I have noticed that the stepper ‘rocks’ the entire top of the Z axis (nothing to do with the washers, part of the axis itself). I might need to build a stiffener plate in the future as well.


Anyways, after my highly complex workaround of adding washers, I started turning the steppers and ran into the second hiccup. The Z axis would only turn in a single direct and my slowness on the keyboard caused a couple soft crashes. This was a quick fix with some help from Matt - one of the control wires was swizzled in the control box. Hooking up the X and Y axis’s worked without a problem. This left me with a fully moving machine, but without proximity switches and no spindle.

The spindle unfortunately was the next challenge. I put the whole thing together (Avid’s trimming plate although pricy for what it is, is a wonderful example of ‘keep it simple’ engineering that works) and hit the green button in CNC12. .... Nothing.... I tried and tried and couldn’t get anything to turn. I started poking into the settings on the VFD, but that’s really not my area of expertise. My initial thought was a problem in the control box, so that’s who I asked first for help.

After about 45 minutes of debugging on the phone, Matt helped me figure out that the 14 pin cable going from the control box to the VFD had a broken wire in it. Centroid was correctly outputting voltage on the 0-10V spindle speed adjustment pin and the relay in the control box was shorting the pins to turn on the VFD, but in the VFD control box we weren’t getting that ‘on’ signal. I ran a new jumper cable from the control box to the VFD (by-passing the 14pin cable) and the VFD/spindle worked great.

Avid is actually the supplier of that 14 pin cable, so I reached out to them and to their credit they, without much proof (just my word), sent out a new cable which arrived the next day.

So that leaves me to where I am now. The machine moves, the spindle spins and I’ve done some test cuts on scrap wood (just using the keyboard, haven’t loaded any g-code yet). I still need to hook up the proximity sensors and tram/square/flatten the machine with a spoil board. That’s my goal for this week with the intention of doing some actual ‘CNC’ cutting by next week.

I will say, I’ve had a couple realizations during the build process that might help others. First, my ‘wisdom’ for today, is to think of the CNC as a table saw in the shop. What I mean by this is that most people position their table saws in the middle of the shop and have the ability to walk around most sides of it. That’s needed with the CNC as well, especially during setup. In my mind I was planning on shoving the CNC as far into the corner as possible and that’s just not realistic. Right away I decided to put another foot between the CNC and the wall (more than I was planning) and frankly I wish I had space for 2 or 3 more feet. I can already see that attaching stuff to one side of the table is going to be a challenge.

My second realization (and I sorta knew this, but really want to emphasize it) is that cable management is incredibly important. There are dozens of cables in this thing, all of them long enough for a machine that’s 3x the size and every one is critical in application. I’ve ended up needing to isolate a few of them for debug and that means keeping everything organized. These things turn into a rats nest the moment you turn away.

My final ‘realization’ for today is that the system I bought (and I’m guessing some others, but I can only speak to what I’ve experienced) is really built by integrators and there are dozens of suppliers. Because I did final assembly that means that problems are going to arise. Just like any complex machine, some tweaking is required and I’m confident anyone can do it, but don’t go into this assuming it’s going to work perfectly right off the bat and make flawless cuts off complex designs. Thankfully the tech support from both Avid and GCnC has been excellent and I’ve absolutely used it. I’m sure I could have figured out my problems on a ‘no name vendor’ machine (think import where the moment you pay the conversation is over), but it would have taken 10x as long. I bet on the flip side there are vendors who will sell a whole fully packaged machine that’s been 100% tested and delivered with an onsite technician. I bet that costs 100k+ (think Fanuc).

[Ben Grefe]

I need to start thinking about spoilboard ‘designs’ on the machine. I know this isn’t make or break, but I’m curious to hear what others have done and found what works vs what doesn’t. I’m thinking about the following options:

Single layer of 3/4 MDF - no holes, just screw directly into the MDF, hot glue to the MDF and expect to resurface it a lot.


Single layer of 3/4 MDF and cut a Festool-like grid of 20mm holes - probably my top option right now, I have a bunch of accessories for my MFT already and this would allow both side clamping as well as pull down clamping.

Double layer of 3/4 MDF, second layer broken up with T-tracks - I see this a lot of ‘nicer’ machines. Probably one of the more costly designs.

Single layer of 3/4 MDF with big holes cutout for F-clamps. Think Frank Howarth’s CNC table design.

Anybody have any other suggestions on designs to try or things to avoid? I’d love to build a full vac hold down table but it’s not in the cards yet.

[Wes Grass]

Stay away from screws, if at all possible. They're really hard on carbide cutters. Same would go for clamps. Assume whatever you use WILL get hit sooner or later.

'We' (past tense) started with small aluminum nails to hold 1/8" thick panels of carbon fiber laminated on PU foam. Even 1/8" endmills went through them without any apparent damage. That got old quick. I would have loved to have tried one of the composite nailers that Avid sells, but there were 'budget issues'.

We went with a vacuum table, running off of shop vacs. I plumbed it with 2" PVC (I think, the holes were about 2.2'ish?), and zoned each 4x8 section (4x16 machines) into 4 with ball valves so I could isolate it at my will. There was no connection between the 4x8's, so 2 vac's required for our 14' panels. You won't have that issue.

I think I had to bore 2" couplers out so the vac hose would plug in. So there's something to plan for. Maybe there's a rubber coupling that would work here. If you go straight to the MDF, no plumbing for zones, that isn't an issue.

There are plenty of designs online. Best is a solid surface base plate, but MDF will work ok if you seal it up well. 2 coats of primer was barely adequate IMO. You can do this with 1 sheet (2 layers) of material and get by ok. But you've got plenty of 'Z' so maybe a 3 layer stack would help. Solid on the bottom, grid in the 2nd, spoil on top.

If I couldn't commandeer both shop vac's I'd throw a drywall screw in each corner of a sheet for 'safety'. Mold segments, 3/4 ultra light, would be pre-drilled for alignment dowels anyway, so a couple drywall screws were added there as well. Make sure those holes don't get included in the contour program as well. DAMHIKT. The 16 foot table gave me time to remove them and the parts while another sheet ran at the other end. But ... 'No rest for the wicked', as they say.

Small things can be done on a hard vacuum setup pretty well. I built a box of MDF, with a hole in the bottom that mated with a hole thru the table so the hose could run underneath. You have to plan ahead a bit so you leave a void in your grid pattern for that. It's basically a torsion box design, with voids around the internal web for flow. It was the only way to hold blocks of PU foam for 3d contouring. I put some supervisor's tape on the bottom, and blocked any open areas with scraps of whatever was laying around.

You could port it off the end, but then it needs to be tall enough for the hose to pierce the side.

No matter how lightly you 'cut thru', the surface eventually crumbles way and has to be surfaced regardless of whether you need it for vacuum seal or not. So I look at the t-track option with a leery eye.

[Scott Buehler]

20210113_095346.jpg
2 layers of mdf with t track and dog holes work great!

[Jim Becker]

I have tee tracks and do use them with clamps and so forth from time to time, But honestly, most of the time I'm using screws for hold-down...I favor the good quality pocket screws for this purpose and yes, there is always risk when there's steel on top of the table.

Because using screws also leaves behind "volcanoes" that have to be knocked down after cutting a job, I'm considering adopting a setup that Roger MacMunn (a signmaker friend and 'Creeker) uses with his machine. He has a grid cut in the spoilboard so there is a regular pattern of areas lower than the surface and any screws are placed so they enter the spoilboard within those channels. That eliminates the "volcano" problem without extra work while cutting.

[Prashun Patel]

I am considering and evaluating a similar build. I am having trouble locating GCnC to find out more about the Rezurrection control system.

Any queries lead me to the Centroid website, which is difficult to navigate. Any help appreciated.

[Bill George]

I am considering and evaluating a similar build. I am having trouble locating GCnC to find out more about the Rezurrection control system.

Any queries lead me to the Centroid website, which is difficult to navigate. Any help appreciated.

I am going to guess its Acorn control system, re-named by a vendor.

[Jim Becker]

I am considering and evaluating a similar build. I am having trouble locating GCnC to find out more about the Rezurrection control system.

Any queries lead me to the Centroid website, which is difficult to navigate. Any help appreciated.

Centroid Acorn, AFAIK. I believe that "Rezurrection" is the vendor name for their integration that includes Acorn as the controller.

[Gary Campbell]

Guys...
REZurrection Control was the name given to the control systems that I developed in 2013, primarily for ShopBot machines. In 2017 it became too difficult to justify using SB hardware and software due to connection and other support related issues and a replacement controller was sought out. At first this was UCCNC. Luckily, Centroid Acorn arrived in the market and we have been building systems using them along with WinCNC.

The "REZurrection" name has remained. No longer just a replacement or upgrade controller, it has become a system of choice for users building new, especially recently with the AVID crowd. In the last couple years the "AVID Closed Loop NEMA34" system has become our largest seller.

With all of that being said, I personally no longer build retail stepper systems. I still do a limited number of servo controls using WinCNC and Centroid controllers, and usually with ATC, but most of my work is product development for CNC OEM's. And fishing. Here is a link to the Centroid web page with my contact information: ShopBot Acorn CNC retrofit upgrades by GCnC (centroidcnc.com)

I pass all the stepper based inquiries to Matt. (mattdevincentis at gmail) He is a member here.