View Full Version : Hercus Retrofit Saga Continues...

07-01-2009, 03:17 AM
Since the last thread got somewhat derailed...

Finally made some progress on the lathe. Bought a Tyan S3095 motherboard for the machine. Its a FlexATX form factor board. Supports anything from Pentium M to Core2Duo in a Socket 479 package. I happened to have a 1.66Ghz CoreDuo processor in that package so I went with that. The board itself has built in compact flash, 3 ethernet ports, PCI, MiniPCI, PCI-E 4x, USB, serial, parallel, firewire, and LVDS for a direct connection to a LCD panel. I am currently running the system off a 5GB compact flash micro hard drive.

I got a neat little LCD screen that I am hoping to get working with the LVDS on board the MoBo. Its a little Toshiba 5.6" LCD. 1024x600, 18bit color. Darn well need a magnifying glass to read it! I will have to get an adapter to use it though. The connection on the panel is one of those flex circuit connections. Found one company that makes an adapter to standard LVDS connection.

Ordered a PMDX-122 breakout board. Did some figuring and can run all the major timing critical I/O off of a single parallel port. The board handles charge pump and has an relay to control my spray mist unit. The rest will be done with a PoKeys55 and a ModIO. Peter Homann in Australia is writing a macro for mach3 to control the tool changer with his ModIO board. He really has some well made products.

Last I got a 8 position Opto22 board to control IO to various things.

Yesterday I got the spindle running. Not without issues of course. The darn thing would run faster in one way than the other. Made no sense at all. Checked the encoder timing which was fine. Finally found that the voltage to the drive was different in fwd vs rev. I am using the differential input on the servo drive to control forward and reverse. +10v on one input is full speed CW, +10v on the other is full speed CCW. A relay on the step/dir to analog board switches between the two. I checked the impedance on the inputs of the drive and found one was 19k and the other was 31k. The difference in load was causing a variation in the voltage drop out of the frequency to voltage converter on the board. Added a resistor and fixed that problem. Did some tuning and it seems to work fine. Max spindle speed of 3400 RPM.

Today I got the thing moving. More issues though. Tried yesterday and got no results. Did some research and found a post saying they want 24v inputs on the step and direction inputs. So I made up a little board with a ULN2003 to give the drive 24v pulses from the 5v out of the breakout board. Hooked it up. Smelled something hot! And it was coming from the servo drive!

Opened up the drive to find a burned up surface mount resistor that was in series to the optoisolator. Not good. Luckily I had another drive to compare the circuit to and was able to verify nothing else was bad. Replaced the 91ohm resistor with a non-surface mount variety and tests match a new drive.

So what the heck was wrong? I start looking through the manual on the drive. They are inputting 24v though the on board supply. They show it internally feeding the optos through a 1200 ohm resistor in single ended mode. They also show the resistor, although unmarked, in the differential line driver mode. But the drive and the schematic do not agree. The unmarked resistor in the picture in the line driver schematic is actually the 91 ohm resistor, not the 1.2k resistor in the single ended schematic. Next was the hookup diagram was backwards of what it should be + was marked - and vice-versa. So I ditched my new driver board and swapped the connections on the plug on the drive and wired it up and it moves! Turns out the issue yesterday was because of polarity. When I hooked 24v to the input it dumped way too much current through the reverse polarity protection diode on the board and fried the resistor.

So I got the motors up and calibrated. Runs at about 350IPM max. Everything looks good. The Z axis screw is a bit noisy though. Will have to keep an eye out for a new one. Wired up the limits and estop circuity tonight and everything seems to be working as it should.

Tool changer is next. Need to order one of the ModIO boards. Then I need to get the front panel finished. Need to set up things like soft limits too.

Heres a couple pics of the mess so far.



07-31-2009, 12:53 AM
Made a little more progress. Got spindle feedback working. Had to use a one-shot timer board to lengthen the index pulse coming out of the spindle encoder to something that the computer could see.

But the bad news is the spindle drive sucks. Very little torque under 1000 RPM. I think it may have something to do with the simulated tachometer system the drive uses to work with motors that do not have tachs. So I picked up an Allen Bradley 1398-DDM-009. That ought to do the trick and will allow me to get rid of the step/dir to analog board. I will have to buy the software to configure the drive from AB. What a pain...

Heres a video of it so far. Spindle speed 3000 RPM. Still waiting for the ModIO board from Peter Homann.


Paul in OKC
07-31-2009, 11:24 AM
That's nice. Some day for mine........

08-26-2009, 06:22 PM
Got the machine cutting metal. Here it is threading 6061 at 2k rpm. 1/2"-13 thread.


08-26-2009, 06:39 PM
Threading at 2000 rpm ? Cool ! ! ! That would be a very useful little lathe around my shop.

I admire you for working through this project. I wish I had time to do projects like that.

08-27-2009, 04:44 AM
Very nice, very very nice...

08-27-2009, 05:21 AM
Yes, these are great little machine for screwcutting, that's mainly what mine gets used for.

08-27-2009, 06:00 AM
That is a truly great project topped only by the OP's (macona)'s tenacity and ability/skill-set.

I can appreciate the threading at high speed and the pretty well limitless range of threads/pitches that can be cut etc.

The ability to turn tapered threads and an almost unlimited range of shapes ands profiles to the time, finish and costs that it does is what gets my attention.

I am really intrigued with that tool turret.

Question anwered when I saw the video.
[End idit]

I think I recall you saying that it was made by Hercus in OZ. I don't doubt that its true. I'm amazed that I've never seen or heard of that level of manufacturing and design.

Here is a "run-down on "Hercus":


I doubt there would be too many here in OZ who did any Fitting and Turning in a Technical School etc. who did not use one of these:

08-27-2009, 10:18 AM
I think I recall you saying that it was made by Hercus in OZ. I don't doubt that its true. I'm amazed that I've never seen or heard of that level of manufacturing and design.

HI Tiffie, We have made and still do make a bit of CNC gear here in Oz...Hercus also made the compumill, which was a small CNC mill...

Then there is ANCA http://www.anca.com/en/home.aspx which among other things makes CNC tool and cutter grinders...Zenford Zeigler which is not around any more but they made CNC mills and large CNC gantry mills

There is also another company in Melbourne making CNC machining centres http://www.h-h.com.au/products_overview.htm, and there is Farley making CNC flame cutters http://www.farleylaserlab.com.au

08-27-2009, 10:30 AM
Thanks Ringer.

I feel better already.

They must be pretty cost-effective if they can survive and thrive in what I'd have thought was a pretty competitive market.

Paul in OKC
08-27-2009, 11:12 AM
That is sweet. I am still twiddling my thumbs on where to go with mine, but you have re-inspired me!

08-27-2009, 01:14 PM
ANCA made the original control for this guy. Apparently it was a control the same as a big machine. I can believe it. There were spots for components for 6 axis' I believe.

I dug through the manual and found some of the original specs. 80w servo motors with resolver feedback. 39ipm cutting, 150 ipm rapids. Now it does 380ipm. Could go faster but my computer gets a tad unhappy past 65khz.

08-27-2009, 01:19 PM
Oops. Double post.

08-28-2009, 07:09 AM
Hi that is one sweet little lathe, I wish I had seen this thread earlier :rolleyes: I did a search on this lathe and found very little except an action 3 weeks ago here in OZ that listed not 1 but 2 hercus compulathes and a hercus v300 cnc mill

09-14-2009, 11:25 PM
Made quite a bit of progress since the last post. Got spindle indexing working. I am currently using the index pulse out of the 2500 line encoder through a 555 one shot timer to get a pulse length long enough to trigger the opto22 input module. A EE friend of mine came up with a little circuit that will divide the output of the encoder by 625 and give me 4 pulses, one longer than the rest to allow me to get better speed indication at low speeds. I need to make this board and install it.

I got the ModIO board from Peter Homann and after a few firmware revisions I now have a fully working tool turret on the lathe. The ModIO also displays the current tool position on the back of the machine with a Omron LED display.

I have the auto oiler installed and working. It was kind of a pain since Mach2 has no way to trigger a pulse on home. The bed of the machine has oil ports drilled in the v-ways where the carriage ways are resting at home position. The cross slide has ports to lube its dovetails as well when it is in position. I had a Robertshaw Oiler I picked up from Boeing when the surplus store was still in operation. It is adjustable to run in a couple different modes. Timed run with time interval. It also has a mode that allows it to count pulses. You set with a dip switch how many pulsed to activate at and when you hit that on the pulse it runs for a set time based on the other dip switch in the oiler. It puts out over 100psi and have ruined a couple gauges before I figured that one out!

The Bijur mister is working as well. It is controlled through a SSR in one of the Opto22 module slots. Amazingly enough the pinouts were right on the SIP style SSR and I just had to bend the leads to fit in the sockets. I am using Koolmist 77 as my coolant, the same as on my CNC mill. I picked up the Bijur unit NOS from Plaza Machinery for $50. He also threw in a Noga misting unit which I was able to modify to connect to the Bijur Unit.

To connect the machine to the outside world I made a back panel out of acrylic on our Epilog laser cutter. Made out of 1/4" plexi and laser etched the graphics on the back side. Filled in the markings with dry erase marker and then painted it Sun Yellow with a rattle can. It is mounted in a square hole plasma cut into the back panel.

I finished wiring up the servos and spindle drive. Now if any of them fault it will trigger the machines E-stop circuit bringing everything to a rapid stop. I tested it, it works. When I was parting a piece of aluminum the piece flew off and jammed between the turret and the chuck jaws jamming the spindle. The drive faulted and stopped everything. I have also tied all the reset inputs to one oushbutton switch on the back panel. Also when any drive or the turret faults there is a corresponding light on the back panel that illuminates "MALF" when it happens. Had to build a small board to control the lamps. I wanted to isolate the drives from each other so I used opto couplers to control the lamps and the input for the oiler. Of course the lamps pulled more current than the optos so I had to add a 2N3904 to drive the lamps.

Also on the back is firewire and USB as well as the connectors and fuse for the oiler and mister. The power for these is provided from a little 220 to 110 v transformer since the machine runs off 240. Right now the whole thing is running on a pretty massive 120 to 230v isolation transformer I pulled out of a LASIK machine. Power to the everything but the computer is triggered by a Omron SSR that is driven off the 5v rail on the computers power supply.

I had some issues with the spindle drive tripping overvoltage warnings on me during deceleration. I talked with a friend or mine who designs servos drives and decided that adding input capacitance to the main buss should be enough to absorb that regenerative surge on deceleration. It did and now I can stop about 3 times faster without faulting. This also allowed me to tune the drive much more aggressively.

I made a chip guard from acrylic based on what was left of the original. Used whatever I had laying around which was a bunch of fluorescent plastic. Laser Etched a "Crush Zone" waring on the front of it.

I am starting to get the front panel wired up got the power switch and reset for the computer as well as the E-stop button on the front panel. The lamps on the panel are telephone slide style lamps, not the common T-2 style but the more expensive T1-3.4 style. For the lamps on the power and reset I removed the glass bulb from the slide and replaced it with a LED and resistor. I ground the lens off the front of the LED to give a better spread. The power button lights up on power and the reset is connected to the mainboard's HDD indicator output. For most of the panel I will be able to use the existing lamps which will be ran off 24v from the main power supply. I have brought up power to the panel with a molex as well as the wires to trigger a manual tool change. The rest of the board will be controlled by a PoKeys 55 USB interface.

Tooling has been a pain. The tool holders are clamped from the bottom of the tool holder with a taper lock. I have found that the top hard stop of the tool slot is ~.070 lower than the center line of the lathe so I need to mill or surface grind the top of the tool holder off.

Heres the back panel in the laser cutter:


Pic of the mostly finished electronics:


Front of the machine:


Back of the machine:


Paul in OKC
10-01-2009, 05:42 PM
Well, I took mine over to the place that does our shops repair work, and they say it's not worth messing with. At least not having them do it. I guess I will take it home to the garage and decide to be brave and jump in and learn(with help, of course).

10-02-2009, 09:14 PM
Mr Macona when I see nice stuff like what you built and all those wires and electrical stuff I just realize how dumb I am. Man I would love to be able to do stuiff like that. i can do most machining mechanical but to put it all together and actually wire it make it work wow. Im impressed big time. My hats off my Blad head to you sir.

10-02-2009, 10:48 PM
Thanks. Getting closer to finishing it. I ordered the PoKeys board so I hope to have the front panel finished soon. I have ditched the small LCD display idea for now. A friend of mine gave me a 12" LCD montior with touchscreen. Native 1024x768 resolution means I dont need to make a custom screenset to fit the other screen I was going to use.

This the one I got. Front panel is rated IP65 so I dont have to worry about wet fingers. Intended for use in police cars. It has 6 programmable buttons I can use to do things like turn on coolant.


10-23-2009, 08:45 PM
Almost done! The front panel is finished. The pokeys works really good. The keyboard matrix function works great. I pretty much used up every IO on that board. The MPG does not work great. They are working on that. I was going to relamp the panel with LEDs but I decided it would be easier to just make a little board with transistor to drive the oddball mini telephone slide lamps at 24v.

The spindle speed override is still not working, same with the feed hold light. Something in mach is causing both of these problems.

Have the display mounted to the side of the machine. Used a Anthro flat panel mount and made a support for it out of acrylic on the laser cutter.

Now I just need to mount an external antenna connector on the back. The mini-pci wifi card has a tiny little rf connector so I need to get an adapter for that.

One thing I did change on the lathe is the cooling. They had a fan on each side, one blows in, one out. I changed it to both pulling in to the cabinet. I made spacers out of some scrap alder that hold the control panel 3/8" from the front so air exits under the bottom edge of the front panel.

Couple pics as usual, Back of the panel and the machine in general.