Hello Guest it is March 28, 2024, 09:43:21 AM

Show Posts

This section allows you to view all posts made by this member. Note that you can only see posts made in areas you currently have access to.


Messages - Pandinus

Pages: 1
1
It is DEFINITELY Mach 3 ignoring acceleration values when the THC UP/DOWN are used...
I swapped the pins to my y axis that is rack and pinion, about 2080 steps per inch on a Gecko 540(10 microsteps) and a 387OzIn motor, capable of over 1700IPM and accelerating well over 100IPS/S I can't run it that hard or it will shake the machine to pieces... but it still managed to stall a couple of times.

I agree when I make a more permanent Z-axis at a lower step count it "might" work better, but the fact of the mater is the acceleration rules are not being applied... if that's the case then why do we bother tuning the motors for top performance in the first place???



As for EMC... (the PC I'm using can't run faster than 25KHz In Mach3 = 800IPM Max) yes the pulse train is way faster / more stable... but its running in Linux and isn't exactly easy to configure past a very basic setup(they are improving the configuration utility but it's not quite there yet). It can be customized, but you have to spend some time reading up on how the hardware abstraction layers and .INI files are configured. I think I read, but forgot to bookmark that there is a routine someone wrote to slave and properly home a two motor axis... In Mach you click a radio button and a checkbox... DONE...
EMC though...lol

I really like the look and functionality of Mach3,  but if I figure out how to Modify one of their sample THC setups to my machine before I can get Mach3 to work for me, I'll be using EMC instead. Maybe Mach4 will work properly??

2
General Mach Discussion / Re: Anomalies in the LPT signalling
« on: February 21, 2010, 03:42:04 AM »
Here's a thought in another direction, Do you know how much current your port can sink????

If a pull up resistor is supplied from the the PC PSU +5V or a wall wart etc.... to Pin 1 of the coupler, and the port connected to pin 2 the opto-coupler would protect the output from reverse voltage and can likely sink more current than it can source...

A quick look at "a" tlp521 data sheet says
INPUT DIODE
Forward Current 50mA
Reverse Voltage 6V

so 35mA isn't enough for the one I looked up

1.15 typical forward voltage drop from 5 volts = 3.85V/.050 = 77Ohms so thy a 82Ohm 10% or a 75 ohm 5%
at a port high forward V of 5-3.5-1.15 =0.35V/75=4.6mA so you should be good to go.


Same "MIGHT" work with a couple of diodes and a resistor, but no guarantees on this one.... .7v for a regular diode... shotkeys are lower .3ish

supply +5V------>|--------^^^^-----------X---------->|--------+3.5V port
                                 1Kohm   

Take your signal at "X"   it will boost it .7V to 4.2Vhigh and .7v low... which "should" satisfy CMOS and maybe TTL Levels and only 5mA to sink but the i think the inputs to the drives should be happy with the voltages. (The reason I suggest two diodes is that 1.4V+3.5V=4.9V and 1KOkms on .1v is 100uA so it shouldn't affect the port)
If someone is willing to try this maybe on a PCI card that puts out 3.3 - 3.5 volts Id love to hear the results, because I just thought of this.


My answer for one of your questions is that if, as I believe, an active high means that the port is low until it sends a high step pulse, of a few micro seconds, the low will not supply current so the transistor is OFF and the LED is forward biased with voltage=ON. If the active low puts the port high, the transistor is now biased and is sinking the current, so the LED is OFF. If you go +5V to resistor to LED to transistor the LED will be on when the port is high not low.

I think I spent way to long on this, but as long as this helps you or someone down the road.... all is not lost :p

3
The Powermax45  has a built in voltage divider 50:1 (its actually a 51:1 according to their schematics 100k over 2k) So it has a theoretical max of 7 volts... I'm reading up to about 5  before the torch will extinguish.
I have one input to ground via 10kokm input and a 100uF cap. The other input is the 0 to -7 volts 10k to the positive of a 1458 OpAmp as a voltage follower so the input impedance is MegOhms. Then the other half of the 1458 is a 1:1 Voltage Inverter. A voltage divider with a pot tunes the hysteresis of the output. Each side of the pot goes to a comparator, and a second voltage divider pot sets the reference voltage. the comparators go through an and gate to disable the outputs, via an input from the PC, and/or a switch. Then on to two transistors to drive LEDs and the outputs to a Gecko G540.

The Torch Lit signal from the PM45 goes straight to the G540.

The table is in a metal shed, bare ground, standing on 4 12"concrete pads, on the soil. That is it for grounding. Work clamp is to the cutting slats (surplus catwalk grid)

Maybe I'm just lucky, but there was about 1V of ac noise on the "ground" side of the input and the 100uF dropped it to 10mV. There is built in filtering inside the PM45 at the 300Volt level.

If I leave the THC speed below 28IPM It seems to work . that's probably fine for flat stuff, I'm more worried about cutting thinner materials at high speed... Is there lag in the THC UP/DOWN inputs or not... will a Low Z rate work on thin mats. with the fast federate required. I know the BRAINS cause a small delay in the output to the enable portion of the circuit, because as I posted before and no response yet, the THC enable button In Mach3 DOES NOT APPEAR TO WORK..... Well not when I install the Parallel port driver... don't install it and it functions...just not in the real world... lol   so If your using a Smoothstepper or something like it I guess your good to go. So as an alternative, I have a brain calculate x and y federates to enable the Torch Up/Down . ...Basic triangle square root of a squared plus b squared. I have yet to work the bugs out exactly how I want it to function, add a min THC speed DRO too the screen ETC. but for now it does move the torch up and down :p

There is also a weird hiccup once and a while.. If the THC had moved the axis, sometimes the jog up and down buttons loose their acceleration parameter so the motors stall and I have to reset.... So that makes me think the THC is turning acceleration off to move the axis then turning it on again for regular jogging... Or is there a separate acceleration value for the THC buried somewhere I have to set ???? That would almost make sense to have a diff acceleration. for tiny moves than regular moves/jogging.

(as a note the current Z axis is a 5/16 18TPI screw with 36000 steps per inch... might have something to do with the acceleration ???) I think Ill go try swapping pins to the y axis to try it out ... see what happens.)

Well no luck... definitely ignoring the acceleration rules though... my x and y are rack and pinion up to 1700IPM with moderate acceleration..in EMC2 or acceleration in the 300s+ if i stay under 800IPM... but that's just way too hard on the machine lol. and I cant hit those speeds with Mach and my current PC. Well anyways, I have room to reinforce things and add a little weight :p


Short answer to a simple question??? LoL .. have a good day.

4
It appears to me so far that the THC UP / THC DOWN inputs trigger the z axis to move up and down at her THC RATE, without acceleration.... if I put the THC RATE anywhere close the the max Z axis rate the motor stalls. Is there a way to correct this??? (If I emulate the THC inputs you can tell there is no acceleration.)



On a positive note I do have a working THC circuit for my Powermax 45, It just needs some dialing in on resistor values and i need to stop both outputs being able yo go  high at the same time. Might be a noise thing.

It is one OP-AMP for input impedance and signal inversion, a quad comparator , quad and-gate, and transistor output... resistors and a tun of jumpers all over my breadboard....LOL.

5
General Mach Discussion / Re: Mach 3 Slave Axis problems.
« on: February 19, 2010, 02:10:50 AM »
I think you have a cable problem.... you said the charge pump was switched off to get it to come out of fault but looking at the xml I think the charge pump should work properly... the charge pump on in e stop was checked and the pins looked correct.
Check the cable for continuity on all the pins.
I dislike the female connector on the g540 simply because a male to female cable will usually be a straight through / extension cable, but a male to male cable is more likely to be a serial or crossover or custom configuration...
but I think thats the only thing about them I don't like , I have several male to male adapters from days gone by  ;)

6
Hi all, Im in the process of building a cnc plasma table, and having played a little with a little cnc mill and mach I had noticed that there is a plasma version
When looking into the plasma screen set I noticed there ar some THC commands on it. and figured out to use the inputs (emulated on my laptop) to drive the Z-axis.
After a liitle bit of design work and troubleshooting/re-design, I have the start of my THC circuit. So off to the shed I go , but I cant for the life of me get this thing to go...
the main problem is I wanted to use the THC enable output (#5 on the config screen) to enable / disable my circuit output into Mach3s torch up/down command inputs...
When I click the button, the led and output wont come on... It took some time before I figured out - get this - If I install Mach3 WITHOUT the parallel port driver, the button works... led comes on and off. If I reinstall mach3 WITH the driver.... nothing ???????? Am I missing something ???
I have verified this behavior on 3 different PCs with all the versions available for download.... including R3.042.037

I suppose I can find a way around this... edit my own screen... write a brain... but I just thought It's weird and would like to see it work, because If I can get it working I will post my results as There seems to be lots of questions on a DIY THC but few solutions ... so far I have used a quad comparator , an and gate and a bunch of resistors... might be able to lose the and gate but we will see....

7
Brains Development / Re: Brain timer not behaving as expected.
« on: February 10, 2010, 10:41:41 AM »
Try adding another identical input, second timer 1 second initial delay to put them 180 degrees out of phase, then "or" the timers.
Even add the direct input and "or" all 3 for redundancy if the two timers ever overlap.

Pages: 1