PDA

View Full Version : 12v --> 10v Nowforever D100S2R2B inverter & MKS-V-V02 CDXHCTECH board



CarecaIII
06-01-2018, 10:57 PM
Hello. I bought a cnc 6040, 2.2KW on aliexpress from Chinacnczone.
The cnc works well enough only that I have a problem.
The spindle speed is about 20% greater than the one set as shown in the following graph:
s1000 -> 23.60
s2000 -> 40
s4000 -> 74
s6000 -> 114
s8000 -> 150.80
S10000 -> 187.60
S12000 -> 230.80
S14000 -> 270
S15000 -> 290
S16000 -> 309.6-310
S18000 -> 356
S20000 -> 400

23549

It should go to 400Hz when I set s24000 and not with s20000.
I opened the controller and found that they connected the output of the inverter (Nowforever D100S2R2B) that is 12v to a board input (MKS-V-V02 www.cdxhctech.com) that instead expects10v, as shown in the following scheme.

23550

How can I solve the problem? Where to get 10v?

Thank you.

magicniner
07-01-2018, 02:00 PM
You could build an amplifier circuit which gives a % of input voltage as output, and set it to give 10v Out for 12v In
It's a fairly basic Op Amp project, the plans for which are easily found with a Google search,
Regards,
Nick

EddyCurrent
07-01-2018, 02:43 PM
Maybe you could adjust these parameters, specifically "P0-024 max input corresponding frequency"

But before changing anything, what are they set to now ?

P0-021 AIN min input Setting range 0.00~10.00V
Factory value:0.00V
P0-022 AIN min input corresponding frequency Setting range 0.00~600.00Hz
Factory value:0.00Hz
P0-023 AIN max input Setting range 0.00~10.00V
Factory value:10.00V
P0-024 AIN max input corresponding frequency Setting range:0.00~600.00Hz
Factory value:50.00Hz

routercnc
07-01-2018, 04:53 PM
One for the electronic experts - would a pair of resistors in series across the 12V output with values chosen to give 10V at the centre tap point work? All outputs would be 10/12 ths of the input so it would track all intermediate speeds ? If so just 2 cheap components required

magicniner
07-01-2018, 05:53 PM
One for the electronic experts - would a pair of resistors in series across the 12V output with values chosen to give 10V at the centre tap point work? All outputs would be 10/12 ths of the input so it would track all intermediate speeds ? If so just 2 cheap components required

That's a classic Voltage Divider and Yes, it will work provided your circuit can carry enough current to deliver the signal current required (V=IR)
Do the maths in combination with the input impedance data from the controller manual.
Remember that sourcing any current from a simple resistor based Voltage Divider will skew the output, you need to know the input impedance and available current to work out the values which will safely give the best result.
Hence my suggestion of an Op Amp circuit with fractional gain.

- Nick

Doddy
07-01-2018, 06:49 PM
Okay, had a quick look at the data sheets for both the inverter and the controller. OP is quite correct - the controller expects to be supplied a 10V supply from (e.g.) inverter, and will presumably generate a 0-Vref for the spindle speed. What is needed is to reduce the 12V reference voltage to 10V. A simple way to do this would be to place 3 silicon diodes in series with the 12V output from the spindle.

+12 -------->|------>|------>|------------- +10V In

Diodes 1N4001, or similar would be good enough and cost pennies each.

You could put a LM317 regulator inline - with 2V difference between Vin and Vout, it should manage well enough, but it's probably a sledgehammer to crack a nut.

There is also a AO1 analogue output on the inverter, range 0-10V, but I can't figure out a way to use this to generate a constant 10V output.

Neale
07-01-2018, 06:54 PM
I would go for the diode solution as well. Even if you buy from Maplin, they're still cheap enough and it's a better solution than resistors. Tweak the calibration as also suggested here (if needed - give or take a few per cent, spindle speed isn't that important) and job's done.

magicniner
07-01-2018, 06:59 PM
A fixed 2 Volt drop may be acceptable for the OP's requirements and that's an elegant solution if such is the case, I had Tunnel Vision on an output proportional to the input,

- Nick

Doddy
07-01-2018, 07:04 PM
I'm impressed with the inverter - the manual claims the electrolytic caps within are "wearing parts" and must be replaced after 5 years.

At least they're honest.

Clive S
07-01-2018, 07:12 PM
Great minds think alike

http://www.cnczone.com/forums/chinese-machines/352398-cnc.html

magicniner
07-01-2018, 07:12 PM
I'm impressed with the inverter - the manual claims the electrolytic caps within are "wearing parts" and must be replaced after 5 years.

At least they're honest.

That's interesting, I have a couple of 1hp Siemens VFDs which are well over 10 years old and get daily use, standards must be slipping ;-)

Neale
07-01-2018, 10:35 PM
That's interesting, I have a couple of 1hp Siemens VFDs which are well over 10 years old and get daily use, standards must be slipping ;-)

...or Siemens don't push the limits quite as much as the Chinese? I seem to remember that the HY manual says that you need to re-form the electrolytics if the inverter isn't used for a year, and you should use a variable voltage supply and work up to full voltage slowly. My guess is that daily use is actually better for the caps than letting the thing stand unused for long periods. I try to use 105degC caps for high-power applications rather than the cheaper 70degC versions - maybe Siemens do as well!

EddyCurrent
08-01-2018, 11:52 AM
The reason I suggested changing the parameters is this;

The manual for a Nowforever E100 shows that it uses 12v for the 0-10 input.
The graph shown in the first post of this thread shows that speed error gets worse as the frequency increases.
It suggests to me that the upper frequency setting should be adjusted so that at maximum input voltage to Analogue Input 1, the frequency was such that spindle speed was correct.

Regarding diodes, we used to use this method on reversing roller tables to give a central dead band for the operators, but on some of the newer drives this did not work as planned.

Doddy
08-01-2018, 01:27 PM
The reason I suggested changing the parameters is this;

The manual for a Nowforever E100 shows that it uses 12v for the 0-10 input.
The graph shown in the first post of this thread shows that speed error gets worse as the frequency increases.
It suggests to me that the upper frequency setting should be adjusted so that at maximum input voltage to Analogue Input 1, the frequency was such that spindle speed was correct.

Regarding diodes, we used to use this method on reversing roller tables to give a central dead band for the operators, but on some of the newer drives this did not work as planned.

My interpretation of the E100 manual is that - yes - it provides a 12V supply to the speed controller but that the AI1 input still remains expecting a value from 0-10V from 0-100% RPM. That's pretty much backed off with the graph that the OP provided (full RPM at 20000dmd/24000full x 12V(ref) = 10V). Reading the manual the value for the AI1 input is programmable from 0..10V, but not beyond. There is a gain setting that can be applied also, but I'd expect, again, that requires that the AI1 input is limited to 10V. My concern is that the response to the AI1 input will be limited at 10V, and so no amount of scaling applied by the controller in the VFD would counter this.

Of course, remote interpretation of a Chinglish manual is never an exact science.

Ref the diodes / roller table - I guess you're using the forward bias of the diode to generate a dead-band around the pot's central position?, nice idea (strange that modern machines didn't work with that - be interesting to look at the input characteristics of the modern machines - maybe super hi-z could cause some unexpected results. In the OP's case though - driving the interface board down through 3 silicon diodes should generate a reasonably steady voltage drop somewhere around the 2V region.

EddyCurrent
08-01-2018, 02:07 PM
My interpretation of the E100 manual is that - yes - it provides a 12V supply to the speed controller but that the AI1 input still remains expecting a value from 0-10V from 0-100% RPM. That's pretty much backed off with the graph that the OP provided (full RPM at 20000dmd/24000full x 12V(ref) = 10V). Reading the manual the value for the AI1 input is programmable from 0..10V, but not beyond. There is a gain setting that can be applied also, but I'd expect, again, that requires that the AI1 input is limited to 10V. My concern is that the response to the AI1 input will be limited at 10V, and so no amount of scaling applied by the controller in the VFD would counter this.


What you say is correct but;
1. Regardless, the vfd diagram does show the input connected to 12v
2. I never intended that the parameter should go beyond 10v because that is obvioulsy it's limit set in the firmware, my intention was to set it so that whatever was output from Mach3 e.g. S20000 actually produced 20k RPM at the spindle.

Neale
08-01-2018, 11:00 PM
What you say is correct but;
1. Regardless, the vfd diagram does show the input connected to 12v


Again, we are interpreting a Chinglish manual (or diagram, in this case) so there is a bit of guesswork here! The given diagram shows a pot supplied with 12V. If the idea is that the pot is uncalibrated and you use it in conjunction with the VFD display, then this is fine - in effect, it just reduces the rotation of the pot spindle to go from zero to full. Ditto, I guess, if the pot has a calibrated scale associated but again you won't use the full pot travel.

Slightly different situation where the analogue input is driven from an external source, although it's a pity that the BOB or whatever doesn't regulate the DC input to ensure that the output range is correct.

magicniner
09-01-2018, 12:09 AM
Guys,
With the right pot the source voltage is close to irrelevant so long as you can't blow the VFD input with the output.
Instead of arguing semantics why not posit a functional solution to the problem as it stands.
Offsetting voltage with diodes chops off the bottom of the range BTW ;-)

- Nick

Doddy
09-01-2018, 12:13 AM
No it doesn't. Read my post - it's reducing the Vref voltage from 12V to 10V. The controller's charge-pump will then generate a voltage proportional to the M-S ratio of the demand vs Vref - from 0v to 10v.

It is a functional solution.

I didn't think the OP had a "pot" - if he did, then the solution would be to introduce a positive-side static resistance, e.g. 2k on top of a 10k pot, which would give a sufficient offset from the 12V reference voltage. But that's not the design case here.

Re. the chop-off. Nope, what you have interpreted is the insertion of the diode chain into the Ain1 input, that's not what I posted.

magicniner
09-01-2018, 12:29 AM
So the BOB will function properly with 10V input rather than 12V?

Bonus, you've fixed it!

Nice one!

Doddy
09-01-2018, 12:35 AM
The BoB requires a 10V ref input. The implementation is flawed by driving that with a 12V reference.

23566

magicniner
09-01-2018, 12:47 AM
I know it would be a bit more faff but if it really needs 10V in from a 12V supply I'd put a regulator in ;-)