Spindle speed control for SX3...

[neilw20]

Measure/test all the other IGBTs.

Look on google to see how to test them with low voltage.
http://www.fairchildsemi.com/an/AN/AN-9020.pdf

Testing with a meter:
[nomedia="http://www.youtube.com/watch?v=AfV7gYChILc&feature=related"]YouTube - ‪How To Test An IGBT Transistor With A Digital Multimeter DMM‬‏[/nomedia]

Replace the IGBT(s) and the fuse.
If there is a blown up IGBT is quite likely the IR driver is bad too.

Test it and hope the smoke doesn't get out again.
Use my circuits to do use a guide to where to test.

If in doubt ask. The worst answer is nothing, nada, zip.

[devers6]

Neil, I'd be interested in your troubleshooting help.

When I blew my board, it took out 2 IGBT's which were dead shorts and would blow the fuse every time the board was powered. I replaced both of them, and looked over the board to see if there were any obvious burned traces or components. The board no longer blows fuses, and when power is turned on I get one beep from the front panel, and the tach display shows 0000. This leads me to believe the ATMEGA is OK and communicating with the front panel, as I understand the beep would continue if the program in the ATMEGA wasn't running.

However, if I try to start the motor using the Start button, nothing happens; and after about 5 seconds the display shows "Err". Hitting Stop clears the "Err", but hitting Start again just repeats the process. So I am assuming that the ATMEGA is trying to start the motor, but is seeing no change from the Hall Effects and is timing out.

My first thought was the IR2103 drivers, but I have replaced all 3 and there is no change. About the only other thing in the path from the ATMEGA to the driver circuitry are the 6N137 optos, I have replacements, but they are not socketed so it will be a bit of a hassle to change. Any ideas?

[neilw20]

Unplug the hall sensors from the motor and use a 9v battery to power them. Each output should go up and down independently.
CN101 - pin 1=5v (9v+) pin 2 0v. Pin 3,4,5 sensor lines. Use a 470ohm in series with each LED to pin 1.
The motor I repaired had blown hall sensors after a blowup of similar nature (wire to chassis at stupid door switch.)

I made a battery operated test socket with leds to test the hall sensors. I have replaced some before. If you need to replace one, replace all three with the same type.
Ensure NO WHISKERS on the hall sensor magnet in the motor by using sticky tape to make it clean. One whisker can cause all sorts of weird running problems.

If that all works, suspect the current sense circuit indicating over current.
You will get err signal if start is pushed, and the motor does not rotate very shortly after.

I have had success testing, using a variac with a 250W lamp in series to current limit to a safe value supply. Only try this if you really know what you are doing. All trial and NO ERRORS!!

You need to refer to the circuit to make sure other supplies are satisfied.
The display will generate err if the Atmega is not working, if I remember correctly, but I may be off with the fairies too.

This was when testing the early model with the front panel PCB that had the Atmega, again running of a bench supply.

Stay well insulated, set it all up on a wooden bench, and only use one hand.

In all likelihood you will need to take out the motor to repair the sensors.

The not socketed optos: Individually cut the pins off the optos near the package to remove it. Heat each pin to remove from the PCB then use a solder sucker to remove the solder. Solder wick will work but it will damage the PCB with the heat required.

At a pinch, a stainless steel spike can be poked through hole while solder is molten. Use a tip that is not too small so the heat is only applied for the minimum time required.

[JimsCNCZone]

Quazar, any progress on this project? I know a lot of us would really like to be able to control the SX3's Spindle via Mach-3, even if I could turn it on/off via relays I'd be a happy camper, but your project sounds like the Greatest. Thnx

[compunerdy]

I agree, people are dishing out over $500 to convert the motor and driver. Someone could be making a fortune if they make a driver that could retain the stock motor or a controller for the current driver that interfaces with Mach.

[Al_The_Man]

I am still tinkering with the BLDC motor and drive I inherited, has anyone got info on the LCD display?
It has SMS0324CG but nothing shows up on Google, it appears to be serial data in.
Or any present timing details from the drive?
Al.

[neilw20]

Not even serial data.
Just put in faster pulses and it reads higher.

Can't remember what the frequency/indication is.

Because it is so crude, if you add the high speed pulley mod there is no way, short of a frequency multiplier, to get it to read the correct speed.

If I remember from testing it maxes out 4000 indicated.

That is just another thing that need some smarts.

[Al_The_Man]

Not even serial data.
Just put in faster pulses and it reads higher.

So your saying it is a simple up/down pulse counter?
What about the fwd rev stop indication?
Also it is rather strange for the controller only to switch direction when a start has been issued in the opposite mode?
Al.

[neilw20]

Front panel control.
This is for the EARLY version with the CPU on the front panel pcb and no piggy back board on the MOTOR PCB.
______________________

The later version uses the same display pcb, so I assume signals are the same.
______________________

Pin 9 (front panel) receives tacho signals from the motor PCB.

1 pulses per revolution from hall signal A on motor PCB via CN12 to opto U12 and out pin 9 to the front panel control pcb.

The tacho signal is processed by the front panel CPU and feeds an analogue signal back to the motor PCB via P2 (SV1- pin 6).

The up down buttons operate a digital pot on the front panel PCB creating the P2 signal.

The tapping function puts 5v on the F/R line (SV1-pin 8), as does selecting forward or reverse.

The tapping led is controlled by the front panel PCB.

And that is what happens in manual mode. CNC mode is another stupid story.

[Al_The_Man]

Thanks, I am exploring some possibilities right now during a bit of a lull.
I have never seen it actually on a machine so I am not sure what the Tapping function is supposed to achieve?
This one has the piggy back board with the Atmel CPU.
Al.

[neilw20]

Tapping mode enables buttons (3) on the quill handle to toggle forward/reverse.

Useless due to the unreliable nature of the brush pickup.
OK if quality external button used.

Then when you stop it in reverse, as you do, when you hit start you have to wait for up to speed then reverse it. grrrrrr!!!! Dumb sh1t.
Same with reverse button.

The stupid designer has never used a machine.

[Ozcad]

Hi all just about to embark on a cnc conversion of an SX3 using a Promica kit and Mach3 I would be very interested in any spindle control product

[helitp]

hello, thank you very much for the schematics.
I can provide the program of the front panel PIC16C72:

[neilw20]

I don't have circuit of front panel.
Thanks for the PIC hex file. I must disassemble it to see how it works.

Is there any source code for it? It needs a big fix.
At least it would be good for the stop button to work instantly, before up to speed.

[helitp]

The schematic of my SX3 front panel in this thread: http://www.cnczone.com/forums/x3-sx3...rol-sx3-3.html
The schematic: http://www.cnczone.com/forums/attach...ntrolboard-pdf

Partial dissasemble, with some comments:

[helitp]

Can somebody give me the source (Eagle?) files the schematics?
http://www.cnczone.com/forums/attach...ntrolboard-pdf and
http://www.cnczone.com/forums/attach...-sheet2-a2-pdf
I'm working on modifying the program of this machine and have found errors in the schematics. I would like to correct in the original and not have to draw new ones from scratch.

Current disassembly:

[neilw20]

Can you point out the errors and I will try to patch the PDFs. They are just as I found them.
I decided it was not worth the trouble to redraw. It is not same version as my machine.
Mine has the Atmel piggy back and simple front panel (push button board)
The piggy back needs the big fix, and until I figure out the communications to front panel, it is not worth pursuing (for me).
1. Even when accelerating, the stop button should override any thing - and same with CNC signal.
2. The stupid remote CNC input just fakes button pushing on up and down, in effect, searching for correct speed, and never finds it properly!!
3. Forward reverse should be instant. How are you supposed to tap a hole with that? !!! It is actually capable of servo operation with available signals.
4. If done properly, the speed regulation and hunting/overshooting would not be evident.
5. It should send correct spindle speed to display by measurement, as it already knows actual motor speed, then regardless of pulley it would show correctly.
6. It would then be capable of displaying power as a percentage or actual current

[oldraven]

I did those a long time ago.

This is the only complete file I have left in my , sloppy, archives.

Feel free to alter it.
I know that the drawing might not be accurate, it was a time consuming job to retrace the schematic.

Have a look at this;
http://tinyurl.com/nkygwsa

Scroll down and there is a picture of what looks to be Sieg machines.

Jos

[neilw20]

Just rattling the cage to hopefully get this thread back to life.
SX3 is still a good gadget, that just needs some attention.

[Quazar]

I'm actually back working on this again after a long absence. The original approach did not work because I was trying to run closed-loop control on the spindle with very course inputs (±10RPM button clicks), which was unstable when it interacted with the motor driver's own control loop. I've got a slightly simpler approach I'm trying now, and early tests show decent success. Progress will be spotty though, as I'm in the middle of moving and my shop is going to be packed up for a few weeks...