[Helloagain]

Hi,

Yesterday I just ordered the 4 axes 6040 router from carving-cnc.com.
As I understand it, it's the same as the yoocnc ones and the ones sold on ebay.
Total price (with DHL delivery to Europe) was 2080$.


Do any of you already have some experience with this "carving-cnc" dealer?

Is there any interest from other people in it's quality/etc? For then I would document it.

It's planned to be used mainly for engraving plastics and aluminum. And also to cut some holes in aluminum for panel making. So these should not be strong requests for the mechanical stiffness of the router. This is why we decided to try this cheap chinese one.
I added the 4th axis, as I'm personally interested to maybe do some woodworking with it.

Any comments welcome...

Ah, BTW:
I asked the dealer, what would be about the delivery time, and he explained, that he first does sent it to HongKong and then from there it is sent to me. Am I the only one who thinks, that this is kinda strange???

[aarggh]

These units are all the same, and I call b******t on him sending it to China and then sending it out, they ALL come from China. Although possible, I'm not aware of anyone outside China producing these.

Ok, the good news first, they are really quite good, not great, or extremely accurate, but more than good enough for most purposes. It will be powerful enough for you to do some serious aluminium work. Mine has a 0.8Kw spindle, and I've machined up to 38mm thick aluminium.

The bad, apart from the price seeming higher than I recall almost of them are, the controller that comes with it is garbage. Complete and utter. They all suffer the same crappy design flaws, and earthing problems, assuming your's hasn't had all the earth lugs cut off that is. I can't express just how rubbish they are. I spent weeks trying to get mine working, and scrapped it.

I know you've spent a large amount of money, but if you have any form of serious troubles getting the controller working, I would strongly advise you to immediately lodge a paypal complaint, and try to get a partial refund on the controller, and put that towards a Gecko G540.

These forums are FULL of posts about these controllers, a G540 is the best investment you'll make for your machine and your sanity. If you have problems with it, you could spend months trying to resolve, and in almost all cases, people with the same controller ended up buying a G540, or a Linistepper, or similar.

cheers,
Ian

[whimsical]

These units are all the same, and I call b******t on him sending it to China and then sending it out, they ALL come from China. Although possible, I'm not aware of anyone outside China producing these.

I just received mine from carving yesterday and the plot thickens to say the least.
The label on the box has the originator as
Virtual Volf international
Selangor, 63000
Malaysia

The ups tracking showed the following
Changi, Singapore 08/03/2012 7:20 P.M. Departure Scan
Perth, Australia 08/03/2012 6:32 A.M. Package data processed by brokerage. Waiting for clearance.
Shenzhen, China 08/03/2012 1:39 A.M. Arrival Scan
Sepang, Malaysia 08/02/2012 11:00 P.M. departure Scan
08/02/2012 3:40 P.M. Origin Scan
Malaysia 08/01/2012 6:05 P.M. Order Processed: Ready for UPS

Made in China? From this it doesn't seem to be so.

Mike :cheers:

[Helloagain]

I didn't order it from ebay. I ordered it directly from their website. And from their website they are located in Bejing (in China). So it is IMHO really strange, that he indicated, that he first sends it from Bejing to HongKong and then to me.

Hmm, when I think about it, my personal guess would be, that this "yoocnc"-manufacturer is located in HongKong and that every distributor (even the chinese ones) just tell them, they should ship a machine to a customer.

[Helloagain]

The machine arrived today in 2 packets by DHL.
It really looks, that the heavier the packet the more misused will it be upon transport...

All in all, the packaging was OK.

The probe was missing. At first sight, it looks like everything else is there.

The cables to the steppers look a bit like crap. Don't know, if I will try it with these, or if I should exchange these from the beginning...

The Control box is really something...:

* The Emergency-Stop button is a complete joke. After one usage, it literally fell apart. It would anyway not comply with CE rules for an emergency button.

* The separate 5V cabling for the parallel-port card is also a joke. Just adding a little 5V power supply would have been so easy, and could surely have been bought in china for a very small amount of money...or even integrated into the parallel-card.

* The designer of the parallel-card should get an award for ignorance!
They used 74HC chips on the card. This means the card will not work for modern PC parallel ports (for they all have an output level of 3.3V). These 3.3V are within the parallel port specification, because it was defined as 5V-TTL-levels. But they used pure CMOS Chips, which do not interpret 3.3V as high. They should instead have used 74HCT chips. These are TTL-Level compatible and interpret 3.3V as high. I think I gonna resolder some 74HCT types on the board.

* I think we don't have to discuss the non existing earthing of the control box... ;-) But this I expected from a chinese manufacturer, so no surprise there...

* The genius of assembler managed to exactly mount the cable holder on the fan of the power supply...;-) Sometimes I really wonder, if they do such things on purpose, as I cannot imagine anyone being that ignorant...

Ah, BTW. Did I mention the complete lacking of any documentation?
Especially programming the VFD could be a hard time, without any docs. Will have to look, if I can find any documentation about it online...

All in all, was about like I expected it from a chinese manufacturer. So up 'til now, no big disappointment (yet).

[aarggh]

The Control box is really something...:

* The Emergency-Stop button is a complete joke. After one usage, it literally fell apart. It would anyway not comply with CE rules for an emergency button.

* The separate 5V cabling for the parallel-port card is also a joke. Just adding a little 5V power supply would have been so easy, and could surely have been bought in china for a very small amount of money...or even integrated into the parallel-card.

* The designer of the parallel-card should get an award for ignorance!
They used 74HC chips on the card. This means the card will not work for modern PC parallel ports (for they all have an output level of 3.3V). These 3.3V are within the parallel port specification, because it was defined as 5V-TTL-levels. But they used pure CMOS Chips, which do not interpret 3.3V as high. They should instead have used 74HCT chips. These are TTL-Level compatible and interpret 3.3V as high. I think I gonna resolder some 74HCT types on the board.

The 5v supply is really bad news as it completes an alternate earth path to the parallel cable that can help introduce noise, especially when the spindle is plugged in. Even worse if you fix up the earth on the mains input, from memory that would potentially blow the chip depending on how it's tab is connected. I wouldn't spend any time resoldering IC's, as I'm fairly sure you'll still need to spend ages trying to track down the cause of the lack of control and wild movements, and lock ups, and end up buying a Gecko or Linistepper, or similar. These controllers really are garbage. But, the machines themselves are good value and work quite well once the controller is sorted out.

cheers,
Ian

[Helloagain]

Lol, actually I'm a professional industrial automation developer. So I have some ambition, to try the original solution (with some changes...).
I will sure not spend too much time on it. If it doesn't work well, I will buy another board.

In the mean spare time I had a short look at the 2 pcbs (stepper driver and parallel interface).
Here is my analysis of them, if anyone is interested:


Parallel interface board:
Actually this board doesn't get you much. Actually it makes things worse.
As I already said, the used 74HC chips are not meant for TTL-level interfacing.
Due to the PullUps it could work, if your computer really still has TTL outputs, but no newer PC has that. All modern PC chips I've seen so far, drive the parallel port with 5V tolerant 3.3V cmos drivers. And the HC chips need at least 3.5V levels for a high...
As I said, you could use HCT chips, these wouldn't have these problems.

The second problem is the driving strength. Neither the used HC, nor the HCT version are strong drivers. So you actually make things worse by using these chips in between your port and the optocouplers, for most parallel port drivers are stronger drivers.

The third problem. The chips don't have schmitt-trigger inputs.This is why I see absolutely no reason, why to use such chips in between the port and the optocouplers, or even for the inputs (where Schmitt Triggers would have been nice).

The funniest thing is the third used IC. It's a 74HC14 (hex Schmitt inverter). "Well here are your Schmitt Triggers", some of you might say now. Lol.
Well, only one inverter of this IC is used. And this one goes to the "PWM IN" pads on the pcb. A short analysis shows, that this port is no PWM IN port, it's an output port. And the stange thing is, the signal goes from the parallel port PIN 17 through a 74HC244 on the 72HC14. So all it does is invert the signal. The Schmitt Trigger feature itself is completely useless, as it is directly been driven by the 74HC244...
My personal guess is, that they at first really wanted to make a PWN input. For this the Schmitt-Trigger would have made really sense. But to me it seems like later on someone else finished the project and now made it the other way around, which doesn't make sense anymore...

Really a very strange design. I'm yet still undecided, if I should just drop the card completely, and wire all the signals directly to the optocouplers, or if I should just exchange the two 74HC244 with 74LS244 types.
For the LS types don't have any problem with the TTL-levels, they additionally have Schmitt-Trigger inputs and the can drive a lot more...
I think I will go for the 74LS chips, as then I don't have to recable everything...

I personally think it's bad designing from the beginning. If they would really have wanna make a new design, they should have taken the 74ACT1284 chips, as they are especially designed for interfacing parallel ports, and they are also quite cheap and strong drivers...

BTW, as there's no documentation and if anyone else is having such a card:
For the main in and outputs (Dirs, Steps, Probe, PWM and ESTOP) the used parallel pins are indicated on the board. For the unused connector on the bottom left, the wiring is as following (all inputs are PU):
Pin 13 (together with Probe)
Pin 12
Pin 11
Pin 10 (together with ESTOP)
Pin 15 (this pin is directly wired to the parallel port, no chip in between!)


The Stepper Driver:

I had a look at the datasheet and I discovered the following problems:

1) The logic power (5V) should be present before the driving power (here 24V). But in this design, it's actually the other way around, as the 5V is generated out of the 24V. This could really be a problem, for if the internal registers don't have settled yet, they can have any state. And if you then put some power on the output stage, the FETs could have any state. In the worst case, this would lead to a shoot through and therefore an immediate death of the chip.
I think I will just add a separate little 5V power supply which is just always on. Then disconnect the Vin Pin on the LM317(U3) and instead attach this supply directly on the VCC of the drivers. This should solve this problem.

2) The Reset is also just tied PU to 5V (R8). But this probably isn't that a bad problem as the one before. Maybe I just add a capacitor, for a RC-delay to help things a little.

3) The heatsink is directly on the chip. This isn't so good, as if the heatsink should get direct electrical contact this could lead to unwanted currents through the substrate. The datasheet says, either connect the heatsink to GND, or isolate it. I will simply put a heat pad in between.

4) The MO output pin is directly connected to a LED, with a resistor, so that about 4-5mA are flowing. This is bad, as the outputs max rating is 1mA! I will simply disable the LED (remove the Resistor R9 to VCC )

5) The Clk capacitor is 330pF. This is OK. But if you wanna go to the speed limit of the driver (15KHz), then this could become a problem. So I will simply go to 100pF.

6) The thing with the Enable signal is really funny. It is intended on the PCB to be driven by an optocoupler. But it hasn't been mounted. And I can also clearly see why. For as it is designed, the Enable signal is on, until the optocoupler is driven. Which means, you actually disable the driver, by driving the optocoupler, and for an enable signal, this surely is rubbish.

7) The Dir signal optocoupler is a much slower version, than the Step signal one. So one has to make sure, to configure the SW so, that the Dir signal has a profound lead before the Step signal, otherwise some steps will go in the wrong direction.IMHO also the driving Resistor (R13) is quite big with 750Ohms.
I will give it a try (maybe with a smaller resistor). If it doesn't work, the Dir optocoupler can easily be replace with a 6N137 (just some small fiddling needed).

But on the positive side, I have to say, that I was quite astonished to see a quite well made current reduction for stepper idling. For this, an extra IC U2 was used (CD14538), which toggles the current to 50% if some delay time since the last step has passed.

So I will keep you informed...

[Helloagain]

As I was just at resoldering the Stepper Drivers, I recognized, that the Load Resistor (R16) of the Dir optocoupler is very big (4k7). This will result in astronomically long off times. This resistor should be at least 10x smaller.
If you then configure your SW, to have a delay of 10us between dir and step, it should work.

[tascheri]

welll. that is solved because I don´t have a water pump.

anyways... routed a new shielded wire on the X axis and now it works well. ( shield grounded to the connector on the box side ). looks like the original was already broken in less than 2 hours of running.

the Z axis driver seems to be dead for good. it will only stutter... did the resistor switch to no avail. switching driver boards around points me to a defective driver really. is there a way to check the driver chip? dosn´t look smoked at all...

I´ve put a separate 5v supply ( from a regular cell phone wall USB charger ) using the supplied USB power cable. Q: Will it support driving 5 volts to the boards as suggested by removing the regulators as suggested by @Helloagain?? I´t´s rated around 500mA 5V

Finally I´m cutting stuff with manual Z ( LOL ).

cheers.

[mateusz1945]

I just ordered same model of that machine. I really like you writeup about how to improve it. My question is where can i order all that electronics parts?

[Helloagain]

The few parts that are needed, you can get from every better store, which sells electronic components.
Except the 74LS244 chips. Not all do have them in the narrow soic package (which is needed here). (So you have to look out for the 74LS244NSR chips)
E.g. I order my stuff always from digikey. But as said, I'm a professional anyway, so I don't know the best places to buy for hobby people...
Maybe someone can help here.

[jaru-eri]

They offer now a new model of the 6040 router.

CNC 6040Z-S80 ROUTER ENGRAVER DRILLING AND MILLING MACHINE - carving-cnc.com

Do you think this will be a better choice than the older one?

[Helloagain]

Why a new model?
That's exactly the model I bought.

[aarggh]

The only difference I can see is this model has black panels on the controller case whereas the older one has blue. And this one they give a probe.

Construction appears identical.

cheers,
Ian

[Helloagain]

The basic difference is the stronger spindle motor (1.5kw) and that the controller box is already equipped for a 4th axis.
The disadvantage is, that due to the bigger spindle, the total work volume is a bit smaller.
BTW: On my table, the max x/y values are 379.7/579

[MarcusWolschon]

I find the Z-travel to be quite limited with 60mm.
Thus my question
6040Z+S80 - using the 4th axis - CNCzone.com-The Largest Machinist Community on the net!
if the T-slot plate can be removed to mount the 4th axis at Z<0 and mill cylindrical objects up to n+120mm diameter.

[jaru-eri]

I think the sellers on ebay cooperate or have to sell out their old models before they can list a newer and improved model which will make the older models almost impossible to sell.. I can´t see any other reasons for this..

[MarcusWolschon]

So, what is the inductance on the steppers (needed to select a power supply for the advised Gecko G540)

[aarggh]

A great deal of the individual sellers on ebay are actually one massive company. I got an email from a seller who was changing companies, and the new company had almost the top 10 Chinese cnc suppliers on ebay as store fronts for this one company. That's whay the postage and product costs vary seller to seller, but the total price including shipping amounts to around the same!

cheers,
Ian

[port_huon]

I just noticed that carving-cnc now offer a hardware only model, no controller, spindle or VFD. Perhaps they're starting to realise that the crap controllers are affecting their business.