Hardinge AB 7370

Hello all,
I might have a chance to buy an old Hardinge CHNC-16C-AB lathe in the next few months (Allen Bradley 7370 control). Price should be very reasonable. Machine is in working condition but I am told parts are scarce for the control. Anybody know of possible control retrofits for this machine? Guessing it to be about 20-25 years old.

thanks
Lee

I've got two complete spare 7370s.

Working to get prices on a fanuc retrofit for one of the chncs.

Very dissatisfied with a commercial retrofit on another chnc.

Lee,

The 73 series control was brought out in 1973. That's how old the technology is. We've got several that we're working on replacing. As damonfg says, the "commercial" kits aren't worth a darn. Depending on what your doing, a Fanuc 0T might be a good retrofit.

You had better get it cheap, because it's worth next to nothing with that control. If you do a retrofit, you will have to buy new servo-motors, encoders, and a new drive motor wouldn't hurt. $$$$

Unless you've got a specifice use for it, don't do it. You can buy a new Haas TL-1 for less money than you'd spend on the control.
JR

thanks for the feedback, main reasons I am considering it are;

Machine is in good condition and runs fine.
Would come with alot of tooling and spare parts.
I am familiar with programming and operation.
It's a Hardinge super precision.
Comes with the original documentation.
Would be going into my home shop, 1st cnc.

You guys know if this machine can run on single phase?

thanks for your time
Lee

Lee, the machine can run on single phase for under $100. You'll need to move one wire in the power cabinet and also you'll have to add a small (1/2HP) VFD for the lube pump.

I believe all of the early CHNCs are super precision. They do run very well, just don't use water based lube. Turret drive system is "fun".

If you get it and get in a bind I've spent plenty of time nursing 7370s. Got a full set of schematics and stuff, just holler.

Market value zero. I gave two machines away last month. Kept the controllers as spares for my last 7370... But I'll share my parts stash !!

Go look at either Centroid.com or Ajaxcnc.com

Hire a hit man to SHOOT the AB control in head....
Then rip it off there & retrofit.

dan k

thanks for the feedback, might be worth snagging
if can run on single phase and I can get it for
next to nothing! We'll see what happens.

Definitely not the centroid.

I got a centroid for one CHNC, it is so lousy that it inspired me to keep the AB alive on our other CHNC.

My God!
What could -possibly- be bad about the Centroid
to make anyone perfer the AB instead?????

Please explain in DETAIL!

dan k

Humm...
I wonder if your pile of spare parts that are basically worth "zero" and a guy that might need them
(if he keeps the old control) would have anything to do with the negative slant toward the Centroid?

Other than that, I'm REAL curious what's SO bad about
centroid....

dk

Dan, Damon's Hardinge retrofit sounds like it is several orders of magnitude more complicated than the Centroid stuff for my Tree mill. From a couple of emails I've exchanged with him it sounds like he's not managed to find "the guru" at Centroid and instead has gotten BS'ing sales droids and runarounds.

Centroid puts out a video of the Hardinge retrofit that looks pretty cool, and it seems like they should be able to hand out a set of good instructions with "follow this to the letter and it will work" written at the top. However, it doesn't seem like that is the experience he's had with the project.

On the other hand, my mill stuff looks straightforward enough that I feel comfortable expecting to be able to wire everything up. But the Tree doesn't have all the other tooling and controls that the Hardinge has.

It's sort of like your search for lathe software. Everyone can supply a mill package that works, but lathe stuff seems to be a lot harder to deliver.

cheers,
Michael

This is -not- like my search for Lathe CAM software.
The Centroid CHNC chucker retrofit is an engineered package....or supposed to be. The "other tooling & controls" are the turret and maybe a cutoff slide.
Big deal.
A spindle encoder and threading capabilities are not rocket science.

he's not managed to find "the guru" at Centroid and instead has gotten BS'ing sales droids and runarounds.

Click to expand...

Something sounds fishy here......

If someone purchases the Retro package, and decides to install it themselves, thats a choice. If it doesn't work right, it's time to have an integrator come out & make it work, at additional cost.
I do not believe it's the fault of the hardware or software, or that there are major problems with it and Centroid is selling a package that doesn't
work. Going to a Fanuc because the Centroid "won't work" is going off the deep end.


dan k

Dan, we've never met so I assume you're not really accusing me of giving misinformation so I can clear out junk.

I'm not trying to sell 7300 stuff here. Ive got a complete set of schematics on the 7370, down to each rev of each board, and I've spent a bunch of time working on these systems. My first two degrees say I'm an EE and I'm really just kinda offering up help or a possible board to swap and try. There isn't really a market for 25 year old electronics... This stuff is junk-obsolete, definitely not priceless-obsolete. I'm headed out on Monday to help out on a CHNC at another shop. Maybe they will take a board off my hands?

The money comes in here when the tools are working. Getting $40 for a board and taking the time to dig up the proper schematics for the board and copy them *and send them out with the board* isn't exactly a get rich quick scheme.


Just making the comment that there are better ways to spend the money than on a centroid.
I'll get into that in a seperate post in detail, I promise. I need a little more free time to compose that post, but please don't call BS on Michael and myself yet. At least let me get the info posted first

Absolutely fair enough.
I sure am curious......

dk

We'll see if this works. Apologies for the hijack.

BTW, it looks like nobody is interested in $40 boards!

A couple of points to make up front –
1) I’d love to hear from anyone running a Centroid on a CHNC because I really need a couple of pointers.
2) This isn’t a personal attack. The first vendor will remain nameless but I can post the docs. The second vendor does post here on PM, is highly respected for his knowledge, would be my first choice for “iron” based repairs, and is completely welcome to keep me honest here. He’s been great and I suspect he’s not getting a great deal of support from the factory.

I can put the issues into a couple of categories.

The centroid ‘experience’:

I'm not sure what it takes to be a centroid distributor. The first vendor I worked with demonstrated the bar was set pretty low.

When I started shopping for retrofits, I had 3 CHNCs. 2 7370s and 1 7370A.

When I was discussing what parts were included in the retrofit package, he said that encoders were not included and I’d be able to reuse the existing encoders. We went around for a little while on this until he decided to holler at me "YOU MUST ALREADY HAVE ENCODERS ON THE MACHINE, HOW ELSE COULD IT WORK?". Okay, well, anyone who’s somewhat familiar with these machines knows they use resolvers, not encoders. You can imagine how unimpressed I was by that.

Centroid sells the base “CHNC” kit and has a couple of add-ons, one being an AC spindle motor and VFD, another being new axis servos, etc.
I was very interested in keeping the DC spindle motor. Remember I’ve got 3 of these, and spare parts coming out my ears. I send him specs on a new DC spindle controller I found (to replace the vintage, expensive, delicate, did I mention expensive GE Hi-Ak drive.) They wanted a couple hundred bucks to change the PLC program to work with this drive. *AFTER* the install I found out that the DC spindle controller I inquired about was a drop in replacement to the AC VFD in their spindle package. More on the spindle drive later.

Same vendor, discussing using a bar feed with the centroid. He said it required a change in the ladder logic of the PLC for an extra couple hundred bucks. (see a theme yet?) We debated if it was worth doing the ladder change at the factory, or later. The conclusion was to do it later, since it was just a single file to be installed on the machine. He made a very strong point of stressing how centroid was looking to help me today and be with me later down the road as my needs change.

The next item we discussed (same conversation) was about a G33 threading cycle. G33 is a variable lead threading cycle. The command you would use to cut a thread that starts at one pitch and ends at another. The AB7300 can do it, so can Fanuc. I have a customer that wants it. Since the centroid includes a spindle encoder and plenty of processor power – (the 2gig Pentium versus a Zilog Z80 in the 7300!) it seemed to be relatively straightforward to implement. I pretty much asked how much would it cost for centroid to write me a G33. I went so far as to send him the G33 documentation from the Hardinge manuals, so they could see just what I was looking for. After a couple of days, he got back to me. He said he discussed it with centroid and they agreed that you’d never get a nut to fit that thread, and they wouldn’t be interested in writing the G code for me. (My application was a special ballscrew!) Well if I can’t make today’s parts on a new controller, I’m sure not going to retrofit all of the 7300s. I guess it was a very short road that they were going to help me down. They didn’t even quote the additional G command, just said ‘nope’.

To top it off, he wrote me a quote that included 4 days of his time for the installation, with me to provide an assistant at his disposal. The crown jewel was he misspelled CHNC about five times in the quote.


Move on to the second dealer. He’s a good guy who obviously knows these machines inside and out. I feel lots better. We kick around some numbers. Things are coming together. As I’m about ready to close the deal, centroid puts up a brand new webpage to show their list pricing. They have reduced list price on the CHNC package by about $2k overnight. I get on the phone to order one. Well that was a typo on the website. Centroid won’t honor that price to the dealer. We go back to the old price. He rejects my first PO, we wrestle a little more and he’s accepts the second PO. I end up paying based on the old / correct / revised price. Nonetheless, the dealer does the nice guy thing and throws in a control pendant no charge. It was a good few hundred bucks out of his pocket. …more on the pendant later!


Centroid is pretty slow to deliver the unit. They quoted a 2 week lead time, and somewhere around week 6 I called my dealer to see what was going on. I’m basically bottle feeding one AB7300 throughout the shipping delay! Well he was on his way out the door and he gave me a contact at centroid to call and inquire with. I call centroid and try and figure when I will see the unit. I get a fellow on the phone who hollers at me “YOU SHOULD KNOW BETTER THAN THAT! We can’t give you that information.” Keep in mind Centroid is shipping it directly to my shop, to my attention and they have my name, shop, and number on their caller ID. I get really turned off when people yell at me.


Installation:

Ok, so now the retrofit arrives. Physical installation is very easy. Between the centroid schematics and the hardinge schematics, it’s a breeze. I hadn’t done one before, but I did the physical install in about 15 hours of my own time, with an extra set of hands for about 1 hour (to wrestle in the new spindle motor). Although, after spending so much time under the hood of the 7300, I was pretty familiar with the wiring scheme of the CHNC. There was only one brief moment when I was stumped, and I just had to dig a little to turn up the proper cable. Physical installation was a breeze.

The PLC in the cabinet wasn’t properly assembled and it shed some sheet metal in transit. A couple of emails to resolve which pieces went where. Not a real big deal, but I’m sure glad I didn’t power it up without inspecting it thoroughly.


Ok time to power it up. Centroid sells this as a CHNC package. I was hoping to just run it, not have to reconfigure it. Out of the box, if any of the 200 parameters were set correct it was a total coincidence.
I was very disappointed to try and decipher proper setting for all of the parameters. Since it’s a CHNC package, I had hoped it would be configured darned close to final - maybe needing a tweak for spindle speed range or axis range.

I could not get a list of parameters for a CHNC from centroid, just the advice to not change anything unless absolutely necessary. That really didn’t help me along any.

Part of the reason for the centroid package was one phone call, bolt it on, and run. The more time I spent engineering this the longer I had a machine down. Ajax could have delievered a similar pile of parts for significantly less money.


The spindle motor was an amazing pain. Between myself and another guy we’ve probably got 20 hours into that thing. We used the VFD parameters suggested in the manual and found them to be sorely lacking. The 5HP spindle motor would stall out during toolchanges!

The system to control spindle speed is almost good. The controller outputs an analog voltage. The analog voltage goes to the VFD and the
VFD drives the spindle motor. The spindle has an encoder on it. However, the spindle speed measured isn’t fed back into the spindle speed command.

If you command S2000 and it spins up and happens to run at 2400 RPM, it will run at 2400 RPM and tell you that right on the screen- even though it knows you want to run at 2000.

There was a lot of finagling with the setups. A LOT. We ended up setting the centroid to a valid spindle speed range of 100 to 3000 rpm. Below 100 RPM and it would stall out frequently. 100 RPM isn’t exactly great either. (The original DC motor has lots more torque down low).
We had to play with a lot of the VFD parameters, moving around the max and min velocity points, etc. Using a VFD as the only way to control spindle RPM (no gear ratios in the CHNC) is a little weak. The VFD curve had to be skewed a little bit and we ended up bumping up the high speed frequency and lowering the low end frequency. This was to try and get uniform matching of commanded versus actual speed, since motor slip becomes a much bigger factor (Hz wise) at higher RPM. We were able to get decent matching at low speed (100-1k rpm) and pretty good from 2500-3k. In the middle, the command may vary by 100 RPM from actual.
About 2-3 days of our time to get to the root of this. Why not just send an extra page in the instructions, even as simple as listing the VFD parameters to set.


I really think that it’s an incredible weakness to not actively control spindle speed. There’s an encoder on the spindle. You can set the control up to view either command spindle speed or actual speed measured from the encoder. The controller doesn’t really adjust spindle speed.

The control code implementation doesn’t disable constant surface speed during a toolchange, so every time the turret retracts the spindle slows down. It’s a bit annoying. When the tool offsets are cleared upon indexing to a new tool, the Txx00 offset is very large and makes the controller think the part is now lots bigger in diameter, so the spindle drops down to a crawl.

I sent a writeup of everything my guy did and all of the vfd settings back to the distributor, and hoped to get some insight on the low end torque. He hasn’t got anything from centroid yet. But lets face it, an AC motor at 5Hz isn’t exactly going to be a powerhouse. I haven’t tried tapping yet. I am partial to doing that at lower speeds but in order to get the torque going, this may be have to be a faster ordeal.

The spindle deceleration is through the VFD to a braking resistor mounted in the rear cabinet of the lathe cabinet. The cabinet becomes available after removing the large Hardinge power cabinet from the back of the machine. That cabinet contains about 150 wire connections using screw-type terminal strips. Since this panel was originally covered by the Hardinge power panel, there isn’t a Hardinge supplied cover for it. Centroid didn’t supply a cover either. What’s a $20 sheet metal cover with a $10k plus package? Well it was a couple hours extra work, but it would have been a nice thing to include.


On control software :

On the initial runs we were getting all sorts of crazy things happening. The default configuration for this controller (still the CHNC package here!) was set up to assume the turret would be behind the centerline of the part. From the tailstock, M03 (spindle fwd) would turn clockwise. M04 (spin rev) would be CCW. Moving the turret to X+ would move the turret away from the operator, and X- would be towards the operator. Those settings are all completely backwards for the CHNC. It took a bit of digging to get those set properly, and they are still not resolved. Using MDI, a G-code program, or the pc keyboard, they now all function properly. M3 is CCW, X+ is towards the operator.
However, if you use the hand held control pendant, the spindle and X axis respond completely backwards.
This is a truly amazing feature. Here’s a little scenario that happens every day. You put a new tool on the turret, and touch it off on a piece of stock to set the tool offset. Get it all dialed in, then you want to back the tool off the stock by jogging in the X+ direction. Well you push the X+ button on the pendant, which is located below the X-, so you’d think that the bottom button would move the turret towards you anyway. When you touch the X+ button, the turret goes X- and rams the tool right into the work.
Centroid says they think that’s a software issue. I heard that 3 months ago and haven’t seen an update yet.

The other great one was the M codes to open and close the collet. When running either of those commands, the g-code program would crash. That was a known issue from centroid, and they guided my programmer in getting them to run properly. The specific cause of that was a software port from dos to the linux platform. There were some EOL characters that were improperly converted and the centroid didn’t know what to do with them. That really showed me that they tested the software thoroughly. Since the CHNC package goes on a chucker you would think collet operation would be worth checking out?

There was another issue with the machine able to jog. In manual mode, if the spindle was running, the controller wouldn’t let me do a fast jog, it would force it to slow speed. It reported an error message saying that it won’t let me jog fast while a probe is connected. I don’t have a probe in the lathe. They told me what parameter to change, and after changing it to their specification, I couldn’t fast jog at all. Well I went back to the old setting, that way it will fast jog if the spindle is off.

To this day I don’t know what is up with the turret. It’s got a dandy little encoder on it (Hardinge’s factory encoder). Sometimes on power up it doesn’t know what station the turret is at. It’s an 8-position turret and 8-position encoder. Rocket science this isn’t. You can
turn on the machine, go into MDI mode, enter Txx00 and the turret wont move, but it thinks it did. Tyy00, nothing again. Hit the ‘turret index’ button on the pendant a couple of times and then the T-commands seem to work. I have booted the machine, gone to run programs, and it
approaches the work with the wrong tool selected. That is incredibly not cool.


So after getting the machine running, carefully, a couple of days go by and the machine shuts down and locks me out. It says my demo software had expired. This was on day 11 of having the machine in the shop. The machine sat for over 2 days before I was given an “extended demo code” to enter in. I didn’t need to actually RUN the lathe or anything. At this shop we make a profit just by just sitting around and having coffee, until it’s time to have beer. The CHNC made a fine “thing to lean on” but the centroid didn’t. It had wheels so it kept rolling away when Vern would lean on it. Maybe it is time to close the shop and just open a museum.


I don’t understand why this is running on a linux machine. The software looks to be a dead on port from their dos version with a couple of ugly leftovers.

Eight-dot-three filenames. Why on earth have a linux application that will only support 8.3 ?? Of course this linux box isn’t set up with anything like FTP or telnet. It will only network with a win NT/2k box. Personal preference but the directory layout is just plain queer. Like a /cnc10/cncroot/a directory that’s supposed to be the A drive, and a /cnc10/cncroot/c that’s supposed to mimic the C drive. And an /e that’s the network drive. Of course when you want to use the network drive, you can’t use the ‘network drive’ button. You have to use the ‘hard drive’ button, then navigate up the tree and down again to the networked drive. Don’t go too far up that directory tree or you’ll be rebooting!!


Really, if you are opening a file, either a .NC or conversational file, don’t navigate up to the top directory. If you try to go to the top directory the centroid controller will crash. That’s a pretty nice feature.

Tool offsets were a real bugger. We had nothing but trouble setting the offsets from the centroid graphical tool definition screen. I never got to the root of what was the cause there. There’s a text based screen that works a lot better for us.

The documentation is totally lacking any clue as to what the tool wear offsets change. Yes, I know what a wear offset should do, but are the X & Z units in radius or diameter? Which way are + and - ?

And if you happen to open the lathe’s window a little too quick, sometimes (like on the last line of a program) the control software will crash. Other times it just stops the program like one would expect.

When thread cutting, the centroid generates cryptic warnings that are not described in the documentation. When I inquired what the “vector too small for federate” warning meant, I was just told to ignore it.
I still don’t know what it is, just that I should ignore it.

The conversational software is pretty good for the basic stuff. It’s great for generating roughing and finishing passes. It’s very easy to define complex contours if you are going from a drawing.

I have but one objection to the conversational programming. It lets you key in various feature start points, end points, taper amount, taper angle, arc radius, arc center point, arc length, etc. If you want to program a 5degree taper from X1 Z0 to Xwhatever Z-2 you are not able to do it. The control wont recognize which points you want to specify and which ones you want to solve for. You can go over to their trigonometry tutor and work it out yourself, then plug that result back into the conversational front end.


The centroid also has a couple of software routines to tune the motors. It’s basically a stripchart recorder to look at drag on an axis and safe maximum speeds. This is actually pretty darned neat.

You can specify overall screw backlash in the control and it will handle proper positioning. To those who haven’t been that deep into the CHNC, they use a single ballscrew on each axis. They are not the anti-backlash screws. To get more precise, you can edit the lead error tables on the screws. The software appears to be set up to interface with a laser for calibration, but after I purchased the controller I was told the laser cal was not available.

The 7300 also has support for part scaling and taper compensation, those are not present in the centroid.

One thing definitely worth mentioning is the resolution of the new encoders with the kit. The centroid kit has lower resolution than the 7370. I do not have my notes in front of me to quote exactly the differences, but the centroid ‘steps’ are significantly larger than the 7370s with it’s use of resolvers and tachometers.

The 7370 does thermal compensation of the screws. The centroid only estimates temperature of the motors, and will warn and disable accordingly.


After the sale:

Well, after getting running with the extended demo mode unlock keys, I got the unlock keys for many of the extra features. Not all of the features I ordered though. After I got on him about those extra features the distributor says he didn’t read the whole purchase order.
He read it enough to reject the first purchase order and work price some more with me. He found some of the options I wanted though.
But as it sits today, he wants me to cough up another $3k so he can ship *some* of the features I listed on the PO. One of the features isn’t even available.


Controller crashes on directory navigation or opening the guard.
‘my first linux’ that is done so poorly it may as well be dos.
Lower step resolution than the 7370.
Back-asswards control pendant.
Ignored items in the purchase order, will cost me $3000 more.
Complete lack of low end spindle torque. No factory help. $2600 wasted.
Can’t fast jog with spindle running.
Ocasionally loses position of turret on power up.
Conversational program impossible to set what value to solve for.
Will not accept the Hardinge tool probe to set offsets.


Of the points I’m critical on, they are all either present in the 25 year old Allen Bradley, or common on Fanuc over the past 25 years.


I was extremely disappointed in needing to configure everything myself, and having stupid stuff lingering like the backwards spindle and X on the pendant is very disappointing. I have found the experience to be quite far from the plug and play they advertise. A couple thousand more dollars for options they agreed to deliver. It won’t let me use a tap without the unlock key. Setting a tool requires a special backwards train of thought to not crash the tool. The controller itself crashes at the drop of a hat. (Even worse is I now have folks getting setup crashes on the 7370 CHNC out of habit of doing the centroid CHNC backwards.) The AC spindle motor idea without a gearbox just doesn’t cut it. Complete unwillingness to even quote me for writing the G33 threading cycle. No realtime spindle control is a drag. We wasted 2 days getting the spindle to behave decent. Not great, just decent.

When you have a program running it runs well, that’s not much of an issue. I am using it in production on parts without tapped holes or variable pitch threads. It’s getting plenty of use and when in a simple ‘run parts’ mode it does a fine job. Unless you get a little eager and open the hatch to change a part too soon, then the controller crashes.
But at least it reboots with ‘complete part’ count intact ! But the whole thing is so unpolished as a system, it is little more than a box full of parts. I certainly did not find this to be the the case. While it is not a total dog, it’s got a lot of issues. If you are entering this arena with a 7370, then odds are you already have some issues. The large amounts of time spent setting parameters, motor functions, and stuff was really a downer. Especially with all the time that machine was down for. Ugh. The ‘demo’ software expiring was really a nice touch. Brand new controller, 2 weeks old, and we can’t run it. To quote from their website “Simply hook it up and it runs!”. Yup, in my case it tried to run the turret right off the saddle.

It kills me that I can’t even get a stinking parameter listing from centroid on the normal setup params for the CHNC. Or the ‘software issues’ that just seem to linger. It’s not a path I’d go down again. I’ve got a lot of time into this personally, and I’d be more than willing to lend a hand to anyone setting up one of these machines. Just before you write the check, ask about these issues. Or maybe they won’t matter to you. That’s up for you to decide. But in my case I’m bummed with the whole deal. I am glad I kept the good iron. The centroid is paying the bills in it’s own way, but its not running the volume I have on the 7370. We tap a lot of parts here, and it will kill me to write another check.

It’s a reasonable machine. Utopia it isn’t. I’ll be the first to scream from the rooftops when mine is quirk free. Looks like it may be a little wait.

Wow....
Some WOW! and some not-so-wow.

Here's my short? version response....


I've heard of people dissatisfied with integrators (vendors) and I know one guy that tossed a "CNC services" (Centroids approved vendors-retrofitters) guy out of his shop. Nothing new with retrofitter types really. Not good either. If the guy dosen't know the machine uses resolvers......that's just dumb and a BAD sign.

I've NEVER done a Centroid Ajax retro.
Yet.....

I have done several Fagor lathe and Mill retrofits.
They (and I suspect -most- controls) have some similar issues as you describe.
Read: You are NOT alone here, these are the part of the initiation process of learning how to retrofit a machine tool. Knowing the difference between an encoder, resolver, and a pot, -already- puts you in the Junior-advanced catagory!

I have some real crappy hour and WEEK wasting war stories of my own...that I'll spare us all from right now....well almost....

Spindle motors 101.
Welcome to the club.
20 hours?..ha!
That's 40 less than it took me to learn on my first one. See below.

DC motors are WAY better for spindles (without a gear box) than AC VFD's.
Hey, BETA was supposed to be WAY better than VHS too remember!??
Lucky for me, before I replaced a spdl motor on my small 5C spindle CNC lathe, I did one on a Bridgeport for a "paying project". Customer specified 2HP VFD for 2HP motor. I bought a 3HP
VFD. Real eye opener.
Words like "mush" and "rubber band putty" come to mind in describing the non-existent torque below 500-800 rpm.
Then it was my fault when they couldn't drill 3/4" holes in steel w/o a pilot hole (not in back gear, 200RPM). haha.
So, when the 8HP DC spdl drive on my lathe popped, I did some research on torque curves, threw the notes away, and installed a 15HP AC motor & 15HP VFD. And it -still- doesn't seem to have -as much-
torque as the DC below 500 RPM.
The 5HP VFD for the 16C CHNC just makes me chuckle. WAY way under powered. That HB headstock, bed & collet size should handle the torque from a 20HP VFD -NO- problem.
Then it wouldn't thread. Because (after 4 days of head-scratching) the VFD was SO noisy, the cnc couldn't see the spdl encoder marker pulse.
Days & dollars$ later, with reactors & supressors & chokes installed, she threads just so fine!
But that's another war story....

But I digress......

Seems like any +/- 10VDC analog (or is it 0-10V?) spdl drive should work withthe Cent control.

PLC changes for extra charge sounds like nickle/dime padding to build profit into services
that should be included. Another bad sign of apparent not-so-good integrator.
Shame.

G33 VARIABLE lead?
Interesting.....
Never heard of it.
Have to check the Fagor & see if it allows Var. lead on the G33.
Not suprized Cent steered clear of it.
Probably not much call for it for what it would take to write the G code.....and DEBUG it.


$2000 Price typo.....
What can I say?

Things (like the AC servo drives) they were "selling" appear and disappear on their web site all the time...

Centroid being distant with your direct inquiry:
Probably acting "in the best interest" of their "approved vendor".......??????
A good impression for you? NO!
Smart on their part?...nope.

Patameter mismatch:
Unfortunate, but goes with the retrofit territory.
Don't bet on or believe any parameter until you check them .....2 or 3 times.
A bitch when it's supposed to be "bolt on", but the vendor (a sharp one) normally would deal with it -fairly- easily....having "been there done that".

But now YOU know so much more about retrofits!

Fagor dosen't "govern" the spindle speed either.
Don't think many (OK a $20K FANUC might) controls do. Don't really need it. In G33, if the spdl slows down, the z axis feed slows proportionally.
In G1/IPR feed, same thing.
Same with programmed vs actual speed.
Usually have to play with min/max speed parameter settings and "fake out" the control to get the programmed speed to sorta match the actual RPM.
LOTS more hunting and less accurate with a VFD than a DC drive! I've had DC spindles hold within 1 rpm over the 25-4000 range. DC drives are just plain tighter.


CSS on tool change:
Fagor does that too. Big deal.
Switch back to RPM before the T/C in the program, then CSS after T/C. I usually rapid close to the part -then- go into CSS, then start feeding.
Sometimes I don't. And the spdl motor is ramping up/down alot. So what?

Braking resistors:
You can sometimes save big bux by going to Home Despot and buying a $8 electric stove element (or 2) instead. The ohms are usually close enough to work. Haas is so cheap THEY use them on their
machines.

This is getting sorta depressing......

Anyway.....


Program crashing with M codes or door opening.
Let me guess.....there's some sort of switch on the door??
It "sounds" like some relay or solonoid coils are spiking the PC. You might need some Arc-killers or suppressors or diodes across those coils....??

Pendant:
I really WANNA believe you can change some parameter to swap those X keys.
Please BUG them till you solve that (saftey) issue.
Hint: Can't hurt to mention it's a SAFTEY issue with their hardware!
.......and call 2 dozen times on the problem in 3 days if you must.
(Before you or an operator gets injured).

Turret encoder:
The potted one?
What a rube goldburg pile of crap.
Maybe the encoder itself is faulty, magnet weak,
mis adjusted???
Wish it was a "real" encoder or prox switch.

Linux:
Agreed, what for?
DOS works fine.

Have you calculated the resolution of the system?
Is it .0002 or .0001 with 2000 line encoders?
PLC Direct sells Koyo encoders for around $75 that I think go up to 2500 line, maybe higher. They are very relaible.

Getting a written quote with itemized list of whats included helps elminate the "didn't see it" excuse from vendors.

I hope the issues can and WILL be resolved on your machine. It's all up to you to make "squeaky wheel" noises....LOUD if need be.
How far are you from Howard,PA?
I'd arrive on their door step some day and camp out
till I got results.


>Dan, we've never met so I assume you're not really >accusing me of giving misinformation so I can >clear out junk.

Ok, but I've exposed a potential opportunity
for you to cash in and get rich with an inflated
pricing scheme.
..........JUST KIDDING........

dan k