skeptical... seem too cheap to me... or all the others over priced? :confused:
but... for the price, it may be worth the risk. but IMO, will it be worth it when they fail?
VALUE?

Definitely affordable, one can afford to have more spares & still better off. I had given ups asking why the Chinese can supply at such price... & accept the reliability set back...
My concern is the range, is 4mm sufficient?

4mm is a perfect range. Does that site tell you the diameter and the voltage range? (I couldn't see it at first look)

well my point Ken, was, do we want to spend our time using our machine or working on it...

Gerald, Voltage range 6~36V, 12mm Dia.
They also have some with 2mm range, same V & Dia.

John, from my experience, electronics part will either fail in the first 5min or last the perceived life span lets (just say say 2 years...). Off the shelve price of reputable proxi sensors goes at around US$80~100, at least US$250 saving.... about a days salary of a CEO in a mid size company here.... Changing a sensor won't take more the 2 hours... Under such scenario, I won't mind work on the machine....

He also has a 15mm detection range with a 30MM dia, a 5mm Detection range with 18mm Dia and a 8mm Detection Range with18mm dia. What sesor would be more suited to users who are using the aluminum angle?

Heath,
As Nils and I found out, the Aluminum angle being a non-ferrous material requires a "little" longer sensing distance to work really well. If memory is right, we had the 3-5 mm sense and didn't work as well as the 7-15mm sense.

So the 8mm or 15mm would be a better choice for the aluminum rails. The 8mm seems a little better at 18mm dia versus the 15mm at 30mm dia to work with.

At the price that they are offering, we can experiment with it. ;)
Just place an order for 5 pcs of the 4mm dist dia.12mm sensor, 1 for spare...

Please let us know how long it take to recieve them Ken :) Did you order any relays?

No, I can get relays easily from every hardware shop in town.

They normally arrive within 1 week. Anyway, I'm not in the US ;)

If you have trouble sensing aluminum, just replace the 'hole' in the aluminum with a steel bolt, then instead of sensing the absence of metal, you will sense the presence of metal when you hit the 'target'. My Shopbot is setup that way. When I had the 12mm tubular sensors installed, I set them at 3mm (they were rated to 4mm) and everything worked perfectly. Now I use the APS4-12M-E-D from www.AutomationDirect.com, which costs $16.75 each. That flat style is more easily mounted on my machine.

Mike, that will prevent us from sensing when the gantry lifts off from the rail.

You used the ebay dealer I found? I plan to order from them too, but now that you ordered I'll wait for your review :D

Pablo, it will take some time for me to install the sensors, as I'd just made dust, my priorities list are as such,
1) setting the axis & spindle square,
2) learning CAM
3) proxi switches....
4) .....
5) .....

BUT I'll report when I get there, that's a promise.

Ken, one more question, do these sensors come with a cable with conector or the cable is attached directly to the sensor?

Thanks!

Pablo,
All the Proxy units I have ordered thus far have the cable as an added item. M12 connection. Check with your supplier - a lot of options are available.

Sean

Within the Mechmate community, how often does this happen... just wondering.. My machine has micro-switches , so I cannot detect 'lift off (Houston, we have a problem...) '

just wondering if it ever happened...

When I was doing some Crazy prototype cutting of 3D shapes - I would derail at least once a week.

It's funny how well the Gantry wants to lift up when you drive a 12" long / 1/2" diameter roughing bit into the work. Happens everytime I write the tool file and Rhino would reset the ZERO to the center of the object and not TOP of object. You tell Mach to GO TO ZERO and BANG........300 dollar tool bit goes clunk.

So, for me - the derail feature was very necessary.

Sean

Sometimes we have to decide whether a proximity sensor that can't sense aluminum at standard distances is better than adding a target that can be sensed (at standard distances).

The design of the machine expected that you would use ferrous metal that a proximity sensor could easily sense so that you could sense a hole. If you modify the design and use aluminum, then you have to also modify the design. You have three choices:

1. Buy more expensive sensors that can sense aluminum at standard distances.

2. Use a bolt as a target and loose the ability to know that the gantry has raised.

3. Not use any sensors and hope for the best.

It would be better to use ferrous iron for the rails so that you didn't have to compromise; otherwise, I would at least use a target so that you won't smack the end stops.

We have seldom lifted the big gantry off the rails, but in the early days the y-car was lifted about once a week.
- forget to switch router on
- electrical interference made router go too deep
- bad programming
- bad restart of a program, forgeting that machine moves in a straight line to the next point (from wherever it happens to be)
- run collet/cutter into a clamp
- run collet against a part of the job while cutting a very deep profile
- run dust foot against a clamp, or x-beam
- cutter went blunt within one meter of cutting (HSS instead of carbide)
- rag lying on rail
- scrap jammed inside cut out hole
- try to plunge a big diam cutter that was not designed to drill
- etc. . . . .

Most of those, and more were part of the learning curve and the number of jumps off the rails are now very much reduced, but they still do happen

Let's talk about the proxy switch and do we need NO or NC. My understanding is we need NO, (as Mike used) in mid air it's open, when it senses the rail it closes, then when it gets to the hole or jumps rail or even if it loses connection it opens again. Please read post 252, 253 in this thread and reply. Thanks

Proxy sensors do sense aluminium, albeit at a reduced distance (the gap between proxy and alu rail can be set to less than 1mm). The NO (normally open) type should be used. "Normal" means in free air, without metal nearby.

....and for some reason I now feel that Gerald had a security camera installed in my shop to monitor my progress with that very clear list of things I have done in the past few years! :eek:

A/C proxies
 

My thoughts on the proxies are to use the 12mm AC units SI12-AE4 NO with a 4mm range from FactoryMation. I chose to use a low voltage (24v AC) system for the E-Stops, so I have that power on hand. These proxies are two wire and I would wire them in series to one 24v AC relay. When the machine is on the Normaly Open proxies would be closed when in close contact with the rails which would provide the 24v to the relay coil to close the relay switch so that the signal from the G540 would be grounded. If one of the proxies were to open then the 24v would be broken to the relay coil causing the relay switch to open also and the signal would go high to the G540. Does anyone see a problem with this setup?

Steve,
The only noted challenge might be the noise potential on the ground. I tried some Allen Bradley prox units early on in Beta testing and found that the ac freq on the machine ground (from the z axis) caused the steppers to erratically pick up the noise. Even with shielded cable. Thus, I tested the DC units and no noise. I can't say with certainty that the noise was from the ac prox circuit, but it went away when I changed units.

.....ground loop and noise is REALLY hard to iron out. The bugger in my MM's thus far on all of the machines I have built.

Good luck.

Steve,

AC proximity sensors are slow. Those that you specified are rated at 12Hz, which is 83 milliseconds. The DC proximity sensors are rated at 1000Hz, which is one millisecond. The DC proximity sensors operate 83 times faster than the AC sensors.

At the slower speeds, precise positioning might be a problem. At 2-ips (the speed that I use to 'sneak up' on my zero point), an axis would move 0.166 inches during that 83 milliseconds. The DC sensors would see the target within 0.002 inches at that same speed.

Thanks, Sean and Mike! This is great stuff to know. I thought that 'noise' may be a problem, but never even considered hertz speed! I will need to rethink this.

Hi all
I have just finished my plan for the proxies (Parts ordered but not tested yet)
I note that my sensor choice for detecting the Al rails has a sense distance Sn of 8mm (For standard steel sensor) this gets derated by X 0.55 or a sense distance for detecting Al at 4-5mm.
So if the sensor is 2-3 mm above the Al rail (6mm thick) it has the ability to detect the steel below the Aluminium!!!!
This could be the difficulty some have had setting up sensors on Al rail (Plus the decrease sensing distance.)

My plan is with holes in the Al and an underlying hole in the steel to ensure prox works by sensing the Al.

I have on order three PNP NO 12mm diam 8mm sense distance for Al rail detection and one 8mm dia NPN NO with 3mm sense distance for the Z axis.

I will follow up on the testing when they arrive.

Cheers
David

I don't know how I left this question unanswered... Better late then never ..
Yes, it came with the cable attached directly to the sensor, no connector in between.

I really don't understand why problems arose when the first guys used the common 12mm diameter proxy on alu rails. These proxies typically work at 2 to 4mm sense range for steel and 1 to 2mm for alu, and this should have been okay for either material. As you point out, those going for much extended ranges (18mm dia. and/or 8mm sense on steel) are going to make a new set of problems for themselves.

Gerald,
The 2-4mm sense - 12mm targets worked just fine, but I found that the clearance of 1mm was little too close. (little things liked to get jammed in the clearance area)

Electrically, they worked fine. Mechanically, they were a cow pusher for dust.

Thus, went with the extended distance units for the aluminum for peace of mind.

Sean

Hi Gerald
The problems might be as the switch off distance is what being used, it is obviously longer than the switch on distance.
European standards suggest a plus or minus ten percent of sense distance in manufacturing tollerance is ok to meet the standard.

It is suggested that the off state is clear for three times the sense distance.

So if we start with a 4mm sense distance the "Ideal" target is either a square the size of the diameter of the sensor or three times the sense distance, whichever is the largest.

Also the recommended off state is clearance for three times the sense distance or 12mm. This is the guideline that the installation of sensing the hole in the Aluminium supported by steel breaks. Hence my earlier suggestion of a hole in the steel under the aluminium as well.

Regards
David

Here are my 2 cents to this thread. After reading the options available to wire the sensors and buttons I liked Mike (Richards) idea. Then Gerald added suggestions to this schematic and based on the design an ideas I did this in expresspcb:

http://img830.imageshack.us/img830/1...ultiplexer.jpg

changed 2x4 TTL AND gates by 4x2 CMOS AND gates
Added LEDs to all proximity sensors
Added an 8pin connector to the BoB
Added 3 connectors to pushbutton stations

Then I continued reading the PMDX125 manual and I found the inputs section:

http://img202.imageshack.us/img202/3562/pmdx125.jpg

They suggest using the sensors in parallel without extra interfasing as shown in fig.4. PMDX122 users will need to provide 12V to the sensors

The similar section for PMDX 122 is not this detailed:

http://img837.imageshack.us/img837/3576/pmdx122.jpg

But I guess it can be used the same way, I mean without relays and the design by Mike.

Am I missing something?

Pablo,

I like your design. It's clean and efficient. So far, the only issue I have is using a CMOS chip to drive (I assume) a TTL chip. Traditional CMOS might have a problem with TTL. To get around that limitation, the HCT and VHCT chips were designed, i.e. 74HCTxxxx or 74VHCTxxxx, where 'xxxx' is the part number. Some of my older design books show the use of a 1k resistor connected between the CMOS output and ground. Other designs used a transistor per output gate as a buffer.

I haven't found data on the chips used on the PMDX-125 board, but the manual states:

Inputs (including E-Stop and Fault):
Low must be able to sink 6 mA at less than 0.8V
High must be open circuit or greater than 3.0 volts (and no greater than 25VDC)
Response time is 20 µs

That's a huge load for most CMOS.

Its not my design, its yours!! :)

Changing back again to TTL is no problem at all, just need to change to 74LS21 or 74HCT21 as you suggest. Im drawing the part since its not available.
I will try to make the PCB single layer, so it can be machined with any router

I still wonder if the pmdx125 wiring suggestion have any problem or drawback. [edit] I mean, Im no expert in the differences between TTL and CMOS, switching the design back to TTL solves what you posted in blue color?
Attached is the new expresspcb drawing

Pablo,

I'm comfortable with the 74HCT logic family as the replacement for the CMOS part. The HCT series has the best of both the CMOS and the TTL worlds. It is used to bridge logic families so that different types of logic can be used without unforeseen problems.

Fairchild Semiconductor has an application note, AN-368, that explains when and why to use HCT. It's the kind of document that I like to read instead of reading the latest fiction best-seller.

If you Google "the differences between LS, ACT, HCT logic families", you'll find a lot of other documentation that helps to explain why there are so many logic families and when to use each type of logic. In a perfect world, all of the logic families would have chips with the same functions, but we live in an electronically imperfect world, so knowing what happens when you need to mix logic families helps to keep the glitches out of the design.

When I'm designing circuits for use with other products and when I don't have a schematic and parts list for that other product, I usually try to use 74ACTxxxx logic. The ACT family sinks OR sources 24mA per pin (usually with a 100mA total current draw per device). The problem is that the 'ACT' devices are more expensive and the logic family contains fewer parts, so it's not always possible to find a perfect match.

Things were easier when I started out. If an output had to drive more than one input, I used standard TTL. If the output only had to drive one input, or if the input used 'LS logic, I used 74LSxxxx chips. It was a given that all logic was Active Low, meaning that the output was a current SINK. Now, designers tend to mix and match logic families. Some outputs SOURCE current and some outputs SINK current. The Geckodrive G201 and G202 have inputs that are driven by current sinking outputs. The Geckodrive G203v has inputs that are driven by current sourcing outputs. The Geckodrive G201X can be driven by either sourcing or sinking outputs.

Because I don't like to repair circuit boards, I normally use simple, off-the-shelf parts. I always use chip sockets. (One of my pet peeves is the manufacturer that solders chips directly onto the board. When chips are soldered in, a 30-second chip replacement requires careful unsoldering with a high likelihood that a trace will be damaged by the heat.) I leave the surface mounted chips, the programmable chips and the microprocessors to the more exotic designs for customers who are willing to buy spare boards, which means that the customer's down-time is only long enough to swap out a circuit board.

Although I use the PMDX-122 break-out-board (and I really like that board), the PMDX-122 is much more than is required to do the job. The purpose of the board is to protect the Parallel Port and to amplify the signals. A 3-chip board could do that. One chip for the Step/Direction signals. One chip for the auxiliary input signals. One chip for the auxiliary output signals. Mach3 can easily handle the polarity of the signals. By adding opto-isolators to the auxiliary inputs and outputs, noise would be kept to a minimum. Every time I get ambitious and decide that it's time to design a new break-out-board, I take a reality check and ask myself why I would want to spend a week designing a board and $500 to get two prototypes made when I can buy a very nice, very reliable PMDX-122 for less than $100. So far, logic has won out. As senility progresses, I might someday abandon logic and design a board that is not needed in the market place just so that I can say that I designed and manufactured my own break-out-board.

Hi Pablo,
I was reading a document from PMDX earlier and it had suggestions on wiring e switches for J10 and J13 to use. This is a bit more detailed than in the manual. The document is AN125-1_MachinePower_02.pdf

Doesn't really help with J12, but I hope it's useful.

I'm looking forward to hearing how you go.

Received mine today, I was happy to see they have a LED where the cable meets the sensor, did not checked them yet.

Maibe just a dummy question .... but
 

I assume that from the box situated on the gantry to the pdmx the connection is with a 7 core cable this means that proximity and pushbutton has a common ground. This is possible only if the power of the pdmx is +12 vdc from the same supply of the +12vdc of the proximity isn't it ?
if the pdmx is powered with 9 vac I can't use the same ground, right ? in that case I need 8 core cable right ?
maybe just a question by a dummy .... but ...
thanks