Power supply problems

[Ark1]

Although I am an electrician my conversion has been plauged with problems

Computer device conflicts that took 4 hours to resolve.

Mach3 communication error to Break Out Board that took 16 hours to resolve.

I got to the commisioning stage and the Z axis worked perfectly, connected the Y axis and the motor was hunting. Shut off the power supply to reverse A and B channels on the encoder, turned on the power supply and boom. Y drive had a flash over and the bridge rectifiers on both secondary power supply out puts shorted. 3 circuit boards melted the copper trace and not one fuse blew. That was a week ago. Now waiting for parts and give this another try. I really, really hope you do not have the problems I have been encountering.

All the best.

Ark1

TonyK.

[Ark1]

Here are some pictures of how I laid out my back board for my machine with terminal switches, computer and Power Supply. This could give you some ideas.

The enclosure, and terminal strips I purchased from a used electrical supply warehouse with a bundle of control wire for $100.

I used military style connectors to the control cabinet and the machine incase I have to relocate it.


Ark1

TonyK.

[CNCMAN172]

Ark1,

Just read your post concerning your hardware problems.

"I got to the commisioning stage and the Z axis worked perfectly, connected the Y axis and the motor was hunting."

This normally indicates the encoders are reversed, so your next step seems correct.

"Shut off the power supply to reverse A and B channels on the encoder, turned on the power supply and boom. Y drive had a flash over and the bridge rectifiers on both secondary power supply out puts shorted. 3 circuit boards melted the copper trace and not one fuse blew.

Just reversing the encoder channels should not have caused this issue. More likely you accidently bumped something else as the encoders are normally powered by 5VDC and even if you shorted them out it would not have wipped out your Y drive. The encoders draw next to no current so it they get damaged you would not see anything. My guess is you had a cable loose somewhere and as you went to reverse the encoder wires that cable moved and touched something.... Just my guess


DIYengineer,
I would suggest you test one motor at a time and if you get wire from HomeDepot most likely it will not have shielding at least on the power cables. and most likely not on the ethernet cables either. Anyway it is very important that the encoder cables are well away from the motor power cables as well as any AC lines you have laying around. The AC or motor lines will cause all kinds of issues and MACH will exhibit all kinds of strange behaviour. Another important step is to ensure the motor is not connected to a load initially. You might want to start with the motor that will move your axis that is using the cog gear and just put a block of wood to hold the gear away from the cog rail while doing your initial testing. One important aspect is the steps/unit which can actually be calculated by MACH ones you have the motor moving. Start with your acceleration and speed on the lower ends and then work your way up to higher levels where things continue to work repeatedly.

BEST OF LUCK to both of you. It sometime helps if you have someone close who has already built his cnc and has a little experience. I have helped several in my area and this often goes much faster and going it alone.

Russ

[Ark1]

Just reversing the encoder channels should not have caused this issue. More likely you accidently bumped something else as the encoders are normally powered by 5VDC and even if you shorted them out it would not have wipped out your Y drive. The encoders draw next to no current so it they get damaged you would not see anything. My guess is you had a cable loose somewhere and as you went to reverse the encoder wires that cable moved and touched something.... Just my guess


The 5 Volt DC comes from the Rodgers board. All changes to the encoder channels are done in the de-engergized state. The Power supply was off line when this was done, then when energized both rectifier bridges shorted instantly and a flash over at Y drive occured. Take a look at my layout, each drive has a separate fuse. These did not blow, the fault was ahead of the fuse. The rectifier bridge circuit board trace melted to clear the fault. The Y drive saw AC voltage at the DC terminals.

Both secondary windings from the power supply fed one drive, the X drive motor was not connected. The fault was in the rectifiers. I suspect a bad batch of bridges were obtained by the manufactor.

All power supplies for 5V, 12V and 77 Volt are floating and all component were checked with refernce to ground. Nothing was grounded except the cases of the motors.

Ark1

TonyK.

[CNCMAN172]

Tony,


"The 5 Volt DC comes from the Rodgers board. All changes to the encoder channels are done in the de-engergized state. The Power supply was off line when this was done, then when energized both rectifier bridges shorted instantly and a flash over at Y drive occured. Take a look at my layout, each drive has a separate fuse. These did not blow, the fault was ahead of the fuse. The rectifier bridge circuit board trace melted to clear the fault. The Y drive saw AC voltage at the DC terminals.

Both secondary windings from the power supply fed one drive, the X drive motor was not connected. The fault was in the rectifiers. I suspect a bad batch of bridges were obtained by the manufactor.

All power supplies for 5V, 12V and 77 Volt are floating and all component were checked with refernce to ground. Nothing was grounded except the cases of the motors"

*******

DIYengineer will have a similar layout as I recall seeing his clear enclosure so this is probably a useful discussion for his thread. I am not familar with the Rogers board so all I can comment on is based on the photographs. It appears that the rectifier boards with large capacitors have three fuses each. I am assuming there is one fuse for each supply 5,12,77 right? If this is the case do you use one larger transformers and your feeding two sets of rectifier boards to get enought current to supply each of the larger drive?

Rectifiers can go bad, BUT since you had power running before you reversed the encoder leads it seems unlikely that was the root cause of your failure. Normally encoder cables have four leads: 5VDC, GND, A, B for standard encoders and if you have differential encoders you also have A/ and B/. Since you switched everything powered off and it fried instantly when you turned it back on it sounds like you had a direct short. The most likely cause would have been a short between the GROUND WIRE and the 5VDC supply as there are not many possibilities if you only worked with two of the four wires. If you have not disconnected the encoder wires and they are exactly as they were when this occured you can measure between the two leads and see if you had a direct short. Even if this was the case it is interesting you did not blow the 5V fuse. How much current does your 5V supply? What is the fuse rated at? I think I would check very closely before I installed a new rectifier board and I would also test my encoder maybe with a separate supply. You can just supply five volts across the encoder power leads and spin the motor by hand and you should see the channels change from something close to ground to something close to five volts as you slow rotate the motor if it is still working.

Best of luck

Russ

[CNCMAN172]

"My encoder wires are 8 wire, and my homes are 3 wire. I was going to use these small C27 RJ45 breakout boards to break out the wires and use shielded flex cat 5 from igus to make the runs. this would give me a nice plug and play situation, while allowing me to use the proper shielded flex cable."

DIYengineer,

If you have eight wires on your encoders I am guessing you have differential encoders with the following outputs A, A/, B, B/ ,Z ,Z/ , +5V, GND

Not all drivers require all these signals but I can not recall which amplifiers you are using. I use Panasonic amplifiers and in my case I also have all these signals actually in my case my incremental encoders are 11 wires.

Russ

[Ark1]

Originally Posted by CNCMAN172 Tony, "The 5 Volt DC comes from the Rodgers board. All changes to the encoder channels are done in the de-engergized state. The Power supply was off line when this was done, then when energized both rectifier bridges shorted instantly and a flash over at Y drive occured. Take a look at my layout, each drive has a separate fuse. These did not blow, the fault was ahead of the fuse. The rectifier bridge circuit board trace melted to clear the fault. The Y drive saw AC voltage at the DC terminals. Both secondary windings from the power supply fed one drive, the X drive motor was not connected. The fault was in the rectifiers. I suspect a bad batch of bridges were obtained by the manufactor. All power supplies for 5V, 12V and 77 Volt are floating and all component were checked with refernce to ground. Nothing was grounded except the cases of the motors" ******* DIYengineer will have a similar layout as I recall seeing his clear enclosure so this is probably a useful discussion for his thread. I am not familar with the Rogers board so all I can comment on is based on the photographs. It appears that the rectifier boards with large capacitors have three fuses each. I am assuming there is one fuse for each supply 5,12,77 right? If this is the case do you use one larger transformers and your feeding two sets of rectifier boards to get enought current to supply each of the larger drive? Rectifiers can go bad, BUT since you had power running before you reversed the encoder leads it seems unlikely that was the root cause of your failure. Normally encoder cables have four leads: 5VDC, GND, A, B for standard encoders and if you have differential encoders you also have A/ and B/. Since you switched everything powered off and it fried instantly when you turned it back on it sounds like you had a direct short. The most likely cause would have been a short between the GROUND WIRE and the 5VDC supply as there are not many possibilities if you only worked with two of the four wires. If you have not disconnected the encoder wires and they are exactly as they were when this occured you can measure between the two leads and see if you had a direct short. Even if this was the case it is interesting you did not blow the 5V fuse. How much current does your 5V supply? What is the fuse rated at? I think I would check very closely before I installed a new rectifier board and I would also test my encoder maybe with a separate supply. You can just supply five volts across the encoder power leads and spin the motor by hand and you should see the channels change from something close to ground to something close to five volts as you slow rotate the motor if it is still working. Best of luck Russ

Look at this power supply.

On the left is a 5 volt power supply not fused, then a 12 volt power supply not fused then 2, 77 volt power supplies that are fused. For, X, Y, and Z axis then for A, B, and C Axis. Each group of 3 fuses are fed from a secondary coil. Since the Z carries more load to lift the table than the X and the Y I put the X and Y axis on one secondary coil and the Z on it's own secondary coil. When testing the Gecko 320X drives the manual requires that all drives be tested with 5 volt power supplied from the Sound Lodgic board and turning the motors by hand to check for error and fault this is done with the servo motors disconnected. All checked okay. Next test is to connect 1 motor at a time, Z was first and it performed flaulessly. This was on the #2 secondary coil of 77 Volts. I connected the Y axis and turned on the power supply and imediatly the motor drove in one direction then the other until it settled down. The drive went into flault when energized. The Gecko 320X manual states that should this happen to reverse the encoder channels. The power supply was removed from Grid power and the terminals were reversed for A and B channels. The power supply was put back on Grid power and a flash over occured at Y drive and both Secondary rectifiers faulted. I have not stated that the reversal of A an B channels caused this to occur, this is a sequence of events that do not relate to the event. The Sound Logic board did not suffer any damage. I suspect that the rectifier bridges were defective in the 77 volt power supply. Length of time energized was less than 5 minutes total time of all tests performed. I have contacted the manufactor and with saying very little they are sending me 2 new 77 volt regulator blocks no charge for the parts or shipping.
If this is my fault why are they being so nice at their expense?????

My comments to Diyengineer were to explain that sometimes even though thought and process are well planned the out come is unexpected.


Ark1.

TonyK.

Ark1

TonyK.

[CNCMAN172]

DIYengineer,

My encoders have 11-wires and they are consumed by the following functions.

A, A/, B, B/, Z, Z/, RX, RX/, 5V, GND, SHIELD

The Z channel is really just the index pulse that fires once each revolution of the motor.

The RX channel is used to receive communtation data as AL the Man suggested.

The final wire is just the shield over the encoder wires for noise immunity.


You asked about encoders, so here are just a few points as there is a great deal of information on encoders which is beyond of the scope of your thread.

There are two primary types of encoders: Incremental and Absolute

Incremental encoders just provide a given number of pulses per revolution however they always need to be counted by the driver electronics.

Absolute encoders on the other had have a battery normally about 3.6V that is used to hold the encoder data and these type encoders know exactly where they are at all times. When you turn off a commercial type machine and turn it back on later the machine reads the encoder and it knows its exact position immediately. There are several types of absolute encoders but multiple turn encoders are very common. Google them and you can get a very in depth description.

The key element of encoders is their resolution. They come in various precisions normally measured in pulses per revolution. The encoders on my Panasonic motors are 2500 Pulses/Revolution. Now most electronic drivers and even MACH3 do what is called Quadrature encoder counts. What does that mean? The two phases A and B both look like square waves and they are 90 degrees out of phase from each other. If you draw a simple picture of these two square waves just slightly offset (90 degrees) and just show the second wave form slightly above the first you will see there are four vertical lines. Two for each wave form, When the wave goes positive and when it comes back down to zero. The electronics typical monitor the level of these waveform edges and that is where you get the term quadrature. If you count the edges of the two wave forms you get a resolution of 4X the number of pulses per revolution. So, an encoder that has say 500 P/R * 4 = 2000 Counts/Revolution. The higher the number of counts the greater the precision, however on it also takes many more pulses to get where your going so that can limit you speeds. This is why you need a faster pulse engine for high precision encoders. MACH3 is limited but your using the smooth stepper your machine will be able to run much faster than MACH3 alone.

Many electronic servo drivers allow what is called electronic gearing where you can adjust the number of pulses per revolution electronically so you can balance precision with speed required. No idea if your drivers have this feature but many Yaskawa, Panasonic, and others have this feature.

Hope this helps. We are all still waiting to see the first movements of your machine. Keep up the great work.

Al the Man... You right the other discussion on the power supply problem with Tony ARK1 should be moved so we don't hijack DIYengineer's thread.


Russ

[Al_The_Man]

Originally Posted by CNCMAN172 When the wave goes positive and when it comes back down to zero. The electronics typical monitor the level of these waveform edges and that is where you get the term quadrature. If you count the edges of the two wave forms you get a resolution of 4X the number of pulses per revolution.

Just to clarify, many use the term Quadrature to indicate the x4 factor, whereas in fact the differential encoder gets its quadrature description from the fact the two outputs are at 90° relation to each other, not because the controller may extract 4x the basic resolution, the Quadrature description applies whether the electronics multiply x1, x2 or x4.
Al.

[CNCMAN172]

AL,

"Just to clarify, many use the term Quadrature to indicate the x4 factor, whereas in fact the differential encoder gets its quadrature description from the fact the two outputs are at 90° relation to each other, not because the controller may extract 4x the basic resolution, the Quadrature description applies whether the electronics multiply x1, x2 or x4.
Al. "


Actually differential encoders output each Phase 180 degree out of sync from each other at the same time. Meaning when you have the A phase you also have A/ phase, one being high while the other is low. The same is true for the B phase.

The PHASE A signal and the PHASE B signal are always 90 degrees out of sync with each other so you can determine the direction of rotation.

All I was saying is the electronics can count actual hi/low transitions of the entire waveform or it can count edges which results in 4X as many transitions. There are many variations depending on your encoder and your electronics, in fact some do not use quadrature or even square waves. Some you SIN COS, etc, lots of varability when it comes to techniques that have been designed and implemented by various companies. In fact some companies have multiple techniques depending on the precision required. Henderhain is a good example of a company which encoders with various implementations and various resolutions.

Russ

[Al_The_Man]

I was just pointing out that there is a misnomer spread through the CNC community itself that often states that a differential encoder is called quadrature because of the x4 factor that many controllers produce from the basic two pulses.
Quadrature comes from the fact the two signals are 90° apart.
The Sin/Cos encoders are still quadrature because there is a 90° difference between the two signals.
In fact this is how a incremental encoder signal starts out in all optical encoders and then it is amplified and squared up in a TTL or other format.
Manuf. such as Heidenhain that implement the sin/cos output versions obtain a much higher resolution than the basic count by extracting a varying angular difference between the two are able to obtain a very high increase in resolution.
The reason for the 180° complementary signals are for noise suppression using RS422 or similar transmission standard.
Then there is the Moiré effect used in order to read the fine grating, but that is another story
Al.

[CNCMAN172]

AL,

YOU are 100% CORRECT. A GOOD DEAL OF MISINFORMATION ON THE ZONE, MOST OF THIS IS PEOPLE ATTEMPTING TO HELP OTHERS AND JUST DON'T REALLY UNDERSTAND THE DIFFERENCES THEMSELVES.

"Manuf. such as Heidenhain that implement the sin/cos output versions obtain a much higher resolution than the basic count by extracting a varying angular difference between the two are able to obtain a very high increase in resolution.
The reason for the 180° complementary signals are for noise suppression using RS422 or similar transmission standard."

YUP, SIN/COS allow you to get much finer by going after small changes in the angles involved.

YUP, the noise suppression is really important on commerical machines because they typically use much larger spindle motors, and servo or stepper motors than the hobbiest. They also tend to be installed in industrial environments where you have lots of very high voltages running all over the place not to mention other electrical interference from radio systems, lighting systems, etc.

Russ