I know it has been quiet for a couple of weeks but we took a holiday to get some sunshine, some golf, some snorkeling, some eating and sleeping. Much enjoyed but I am starting to twitch now and need to get back into the workshop. Still it has given me time to think about the next project and I have had time to book for the NYC CNC Open House in Chicago in September.
I got a Dell XPS13 from EuroPC just before we left home which I have been playing with under the guise of sitting ‘relaxing’ on a lounger. I am impressed with it as a machine. Nice screen, decent battery life and quick to boot. It came with Win10 Pro which I had never ventured into and while it has some nice touches I still prefer Win 7. I managed to download all the apps I needed at the airport lounge before we left (including Fusion …) so it has been fun to play but you need a real job to do in order for things to sink in and to learn. Incidentally I did a speed test at the airport and was staggered to see 100Mbps download and upload speeds with a few ms ping time so clearly on fibre.
I also downloaded a load of books onto the Kindle to keep me occupied and was impressed by Los Alamos by Joseph Kanon which is a ‘who done it’ woven around the Manhattan Project. This lead me to Wiki to refresh me on the background on Trinity etc. Incredible engineering and if I have my dates correct, no CNC systems to manufacture the parts.
After a few distractions the Mill Turning Jigs are complete and I have run a test piece that is representative of a clock pillar.
Mill Turning Jigs
The jigs were both designed in Fusion 360. One consists of a large block with space for three 10mm cross section carbide insert tools and a second block with drill and boring related tools. I have fitted three ER16 collet chucks to this to allow flexibility of tooling choice. Both have mountings to fit onto my 25mm hole matrix tooling plate on the Tormach.
The jig manufacture was relatively straightforward with the exception of needing a new 10mm end mill having extended length (35mm) to bottom out the ER16 collet mounting holes. I got this from APT and the edges were lethally sharp.
Trial Clock Pillar
The pillar had simple geometry as below.
I opted to base this on the largest pillar I had come across in any design which was formed on a 5/8″ brass rod. I held the stock in the spindle in a 16mm ER32 collet held in a TTS holder.
I struggled a bit with the CAM for the trial as the tool geometry of the tools I recently received from Banggood were not in the standard tool library. I got some of the settings wrong. That aside the result of the first run is quite pleasing.
My feeds and speeds were a bit coarse and I cringed once or twice at the tortured sound of brass under pressure. I didn’t complete the parting off as I didn’t fancy ducking from a large piece of brass spinning lose at 5000 RPM.
As ever there was quite a bit of learning while making both the jigs and running the trial pillar test piece.
As part of my purchase package of the Tormach PCNC440 I ordered their granite block and height caliper. This meant I could measure and set tool heights directly into the tool table on PathPilot via the caliper USB connection. See the image below which shows the granite block, caliper and the associated dongle box.
This concept works really well and saves manual entry typo errors when measuring tool lengths. I have found one problem however and that is the caliper eats batteries at an alarming rate. These are CR2032 button cells which are not dramatically expensive, but the cost does start to add up. There is the added frustration of the caliper potentially not functioning at a critical moment when a new tool needs to be measured.
It struck me as strange that a device talking via USB should be dependent on a battery when 5V is available from the USB interface. The fact that this did not happen suggested to me that the connection from the interface box to the caliper did not have through continuity of the 5V supply. This wasn’t surprising given that the caliper runs from a 3V cell.
I connected the caliper via a standard USB cable directly to a variable power supply connected onto the 5V power pin on the USB cable. Varying the power supply from 0V to 5V showed that the caliper would work quite reliably over a range of 3V to 3.8V but above this the display would blank or just show 8888 at high intensity so masking the actual reading.
I found a couple of 1N4148 signal diodes in my component stock and put them in series with the 5V feed from the power supply to act as a series voltage drop. This brought the working voltage delivered to the caliper back into the 3V to 3.8V range where it functioned without any problems.
So the question was now as to how to implement this modification elegantly ? ….
Be warned that the modification to be described involves a change to the USB dongle box supplied with the caliper and as such will invalidate any warranty. Mimic what I did at your own risk.
The dongle box has four screws on the bottom cover and removing these reveals the controller pcb. Take care not to loose the three blue switch activator rods in the process. On inspection of the pcb, the USB cable entering the box has all four standard USB connections but the cable exiting to the caliper has the 5V lead (red) disconnected.
I found a pin quite close by to the output lead that was marked 5V. This was a possible feed for my two diodes. On measuring this I found it was at a lower voltage than expected suggesting that there was perhaps some circuitry between this point and the incoming 5V. I therefore chose to ignore this and looked instead to the input connection cable. I found the +5V connection (red) as it connected to the pcb. I connected the two diodes in series to this cable termination and then ran a connecting wire (orange below)across the board to the 5V output (red) cable which was previously not connected.
These modifications are shown below albeit with hot glue over the diodes and connections.
This completed the modification. I checked out all the voltages while the pcb was still outside the box and also checked the caliper was still working. I replaced the pcb back into the box and screwed the lid in place.
One final thing I did which is not necessarily essential but felt like a good thing to do, was to put a tantalum capacitor across the former battery contacts in the caliper battery compartment. This would act as decoupling should there be any ripple on the new supply to the caliper. See the image below. Note that the tantalum capacitor is polarised and its + lead goes to the former + battery contact (on the right as shown below).
While this is a potentially useful thing to do, it has the disadvantage that you cannot put a battery in the caliper if you want to use it ‘off line’ when not connected to a USB port. You could however plug it into USB charger via the dongle lead.
The other minor thing I did was to fit a small cable tie to retain the caliper connector in place as I found it easily become disconnected.
On putting the setup all back together, the caliper was working well with a nice contrast to the LCD display. Tool table updates work just as they did before so no issues there.
If you do this modification you might want to experiment with different diodes or the number of diodes needed to the drop the voltage from 5V to within the caliper normal voltage range. Note that you need to use small signal silicon diodes which will have around a 0.6V voltage drop per diode. Don’t use Schottky diodes as these generally have around 0.2V and so you would need at least 3 times as many to achieve the same overall voltage drop. You could try LEDs as they all seem to have different voltage drops but they tend to need a high drive current which if this is the case, makes them unsuitable for this application.
You could of course go really elegant and build a small integrated power supply chip into the dongle box such as the AMS1117. These are available in various fixed output voltages including a 3.3V version (which is popular for Arduino projects and available from Amazon). You can also buy a ready made 5V to 3V module based on the AMS1117 from Amazon. I like the AMS1117 and used the 1.5V version in my power supply modification to the Shumatech DRO systems.
Here is a printer friendly (and slightly updated) version of this article : –
Two weeks in France flew past and all we seemed to do was jobs around the place. Weather pretty mixed but one very hot day which of course was the day before we left. Surprise surprise. It has since been very hot out there but equally the weather here has been excellent.
Background Air Extract System
I have often been slightly concerned about fumes in the workshop. You know the pervasive smell of cutting fluid and welding smoke etc so first job back was to install a background extract system. My colleague Dave arrived this morning and between us we put in a fan and ducting to vent through a custom roof tile (shingle ?). OK the draft won’t rip your clothes off but it will just keep the air moving especially in winter when the hatches are battened down.
Mill Turning Tooling Jig
Having tested mill turning with a Heath Robinson set up I have been accumulating parts to create a proper custom jig to fit on the 440 table. This will take 3 turning tools, 3 ER16 collets for drill bit holding, a centre drill and a boring bar. Currently it is drawn in Fusion 360 but there are one or two issues to sort before committing to CAM. More to follow once it is fully underway.
G53 Tool Change Location Update
Another update (not yet resolved) is my G53 tool change routine which has a slight weakness. As previously mentioned the Tormach post processor does not always put a WCS following a tool change. To overcome this I was hard coding a G54 after the new G53 tool change routine. The problem is that occasionally I might not be using WCS G54 but any one of the eight other available references. So some extra code is going to be needed to make the WCS a variable that mirrors the WCS being used. Head scratching so far reveals that the post processor sees the WCS as a number from 1-9 and then converts this to G54 through G59.3 in a separate sub routine. I need somehow to overcome this. More to follow on this one too.
Anyway good to be back in the UK (and the workshop), great weather for the last three days and it looks we are going to sit out on the terrace tonight with steak and chips washed down with a glass or two of red. Takes some beating.
The Tormach PCNC440 is a lovely machine and is more than big enough for my present needs. The one problem I had encountered was when coming to a tool change on a CNC job sometimes there was not enough Z height to get the TTS collet out of the spindle. This was particularly difficult when using larger diameter drill bits in a chuck style holder.
Once in program there did not seem to be any option to break the run and do a G30 or similar. What I really needed was a move of the spindle upwards and outwards to get it clear of the job and allow TTS access.
Reading up in Peter Smid’s excellent CNC Programming Handbook I could see that care was going to be needed to ensure that any movement was first of all a Z action and then X and Y to avoid the danger of crashing the tool into the job or its fixtures.
I had some discussion with John Saunders at NYC CNC and John was working on a video around this subject. He helped enormously.
The end result is to use G53 machine coordinates to first do a Z and then and X and Y to move the tool up and to the side for tool change access.
This involves edits to the post processor in three places. The first two edits (Lines 44 and 66) are there to give an option for this movement in the drop down selection box. (The line 24 edit is an earlier modification to allow Mill Turning – see separate post).
The third edit gives the instructions for this as a G53 Z move than a X and Y move (Lines 543-538). Note that I later found that I had to add a G54 after the G53 movements as some CAM actions did not include a G54 as part of a tool change.
I later on decided it would be nice to include this G53 movement at program end so this is a fourth edit (Lines 1404 – 1405) and not forgetting the change for Mill Turning edit (Line 25) there are five changes in total.
If you can’t read the edits then drop me an email and I can send you a full listing.
Note that these are changes to the Tormach standard post processor code and if you are tempted to do this you should do a ‘Save As’ on the original code and only edit the newly created and saved file so you have a fall back position. Likewise I accept no responsibility in documenting this and putting you up to potential mischief messing with your machine and causing damage.