BP Series 1 - Z axis power transistor died again!?!?!

[bbuonomo]

First, thanks to everyone who helped the first time.

Yesterday i was playing with MACH3 jogging, x,y,z and setting offsets, returning to home, all was good.
The only thing I changed was the motor tuning, but only within the limits that I was told to do. (1000-1500, 30-60, 1)

Today I set everything to the recommended settings I got from Hillbilly to use with the BOB Board (1000, 30, 1).I jogged Z+, up to the limit switch, switch halte dthe machine, like usual, I hit reset and when i tried to Z-, nothing....

Same as other day, this time i knew how to find the bad transistor (instead of replacing all 4) and put in another transistor. Fired it up and it works.

My question is, what else could be wrong that the transistor blew 2 days later? Is there another setting I'm missing? I'ma afraid to try to Z+ and trigger that limit switch again.

I mean, my students are getting good at changing out the transistors, but this is silly.

Any suggestions please let me know.

Tomorrow I'm going to go in and try to make sure it is cutting accurately (aka 1" = 1"). I haven't made any cuts yet and I'd hate to spend the day changing transistors :-)

[Astroboy]

bbuonomo:

Assuming that you are talking about the transisters that are mounted on the back door heat sink. Each axis uses 4. My BP Series 1 CNC has had two axis die due to transistors. What I found and learned from an electronic friend is that the VOLTAGE of the transister is as important as the current it can handle. It turned out that the ones that blew were ones that were replaced some time before I got it. Whoever replaced them used NTE60 (140V 20A) (cheap). The original ones were Motorola 2N6678 (400V 15A). The stepper motor sends a kickback to the transistor which will kill it if it is rated at too low a voltage like the NTE60 (NO GOOD!). I have had good luck with Motorola MJ16012 (450V 15A) or if I could get them MJ16014 (450V 20A). The drive to the steppers is set at 8 amps if I remember so the Motorola transistors have no problem.

Make sure that the is a film of heat sink compound on the transistor and heat sink. Too much can be as bad as too little. It is only to fill in the low spots, not be a good conductor. I have not tried the Artic Silver used on my PC heat sink, but it may be great for this application as well. Make sure the transistors are tightened pretty securely for best heat conduction.

- Astroboy

[keebler303]

First off on the motor tuning, the 1000 steps per inch is not a variable for you to adjust. It is determined by the machine. The reduction on the series 1 is such that 1 step=.001" for a boss 5, 1 step=.0005" for a boss 6. This number is what tells mach how far to make the stepper move to go 1 inch. Changing this number is effectively scaling your part files as you cut them. Not good.

If you only replaced 1 of the transistors then the others may have been really weak too and were the next thing to go anyway. This exact thing happened to me on my x axis after blowing one.

You want to make sure you amperage across FU 12,13,14 is around 8 amps. Too high is not good.

Astroboy is right about the kickback surge to the transistors. When you stop the stepper from full speed to zero (limit switch, E-stop, etc.) there is a pretty big current spike which can toast your transistors.

I am not fully sure how it works but I have an SMS board on my machine which sinks this current in an e-stop situation. It consists of only a few fat resistors and diodes. Machinetek has said that if you don't have this board, you will blow transistors often when you e-stop. The board is located in the bottom right of the control cabinet. If you have it I would check the fuses on it and make sure all the solder joints are good, etc. Machinetek may be able to shed some more light on the SMS board issue.

Good Luck
Matt

[NC Cams]

You might want to look at bypassing the inductive load with an appropriate diode to kill off the inductive feedback. I suspect that this is what the SMS board is supposed to be doing. If you don't have such protection, you probably should figure out how to get it.

Keep in mind that diodes don't live forever - should they start to "zener" as they age in that they will NOT protect like they used to do when new. It might be time to replace any diodes in the protection network for the reason already cited.

1N4001's will NOT work although some folks use them for simple relay bypassing - you need some real fast Schottky's for inductive motor use, especially when you're doing ultrafast switching as during jog's (could this be why jog's cause more problems with blown transistors????)
I'd buy the fastest, highest current rated ones I could find and afford to protect the devices.

Transistors: After perusing several threads involving this subject, it seems that semiconductor development has transpired since the machines inception resulted in various parts being used in various machines. Also, newer, more robust devices are available for service

In order of introduction/evolution, the transistors were as follows:

2N6547 (400v, 15 amps. 20 amp peak)

then the

BUX48A (450V, 15 amps, 30 amp peak, 60 amp overload, 2nd generation of the transistor series)

then the

MJ16012 (450v, 15 amp, 20 amp peak, 3rd generation of the transistor series)

then the

MJH16012 (same rating as MJ16012 but more gain at same base current).

MJ16014 (450v, 20 amps, 30 amp peak, gen 3)

MJ16016 (same rating as MJ16014 but more gain at the same base current)

The "high gain" versions transistors should "turn on harder" than the "regular" models at the same drive. This should reduce heating effects caused by less internal resisitance while conducting.

At this point, the MJ16016 is probably the most robust part you could possibly use in the application.

It can not be emphasized enough that the use of 'brand name' part numbers as opposed to "aftermarket equivalent/consolidation" p/n's will offer the best possible chance of achieving the rated performance of the high performance transistors.

[bbuonomo]

Originally Posted by keebler303 First off on the motor tuning, the 1000 steps per inch is not a variable for you to adjust. It is determined by the machine. The reduction on the series 1 is such that 1 step=.001" for a boss 5, 1 step=.0005" for a boss 6. This number is what tells mach how far to make the stepper move to go 1 inch. Changing this number is effectively scaling your part files as you cut them. Not good. If you only replaced 1 of the transistors then the others may have been really weak too and were the next thing to go anyway. This exact thing happened to me on my x axis after blowing one. You want to make sure you amperage across FU 12,13,14 is around 8 amps. Too high is not good. Astroboy is right about the kickback surge to the transistors. When you stop the stepper from full speed to zero (limit switch, E-stop, etc.) there is a pretty big current spike which can toast your transistors. I am not fully sure how it works but I have an SMS board on my machine which sinks this current in an e-stop situation. It consists of only a few fat resistors and diodes. Machinetek has said that if you don't have this board, you will blow transistors often when you e-stop. The board is located in the bottom right of the control cabinet. If you have it I would check the fuses on it and make sure all the solder joints are good, etc. Machinetek may be able to shed some more light on the SMS board issue. Good Luck Matt

Yeah, i figured out that I shouldn't change the 1000 in the motor tuning. I noticed that things were off.

Initially I replaced all 4 transistors, but then blew one of the new ones 2 days later, last move was a rapid z to the limit switch.

I do have an SMS board, i'll check all the connections and fuses and see if anything comes up. Thanks.

As for adding/replacing diodes, i'll keep that in mind as well. Thanks.

[machintek]

SMS board function test:
When adjusting current, as I said never attach or detach anything with the drives ON. When the meter is in place of the fuse, press e-stop to drop the dvives out. If the current drops to zero immediately, the SMS is not doing its job. If it drops to about .5 amps and decays slowly from there, then the SMS is doing its job.

George

[creep_pea]

1N4001's will NOT work although some folks use them for simple relay bypassing - you need some real fast Schottky's for inductive motor use, especially when you're doing ultrafast switching as during jog's (could this be why jog's cause more problems with blown transistors????) I'd buy the fastest, highest current rated ones I could find and afford to protect the devices.

Hi

I believe the orginal diodes were 1N4003, in your opinion will these work properly, they have a Average Rectified Forward Current of 1 Amp, Peak Repetitive Reverse Voltage of 200 Volts, Forward Voltage of 1.1v and Maximum Full Load Reverse Current of 30 microAmps.

You say go for higher current rating, Schottky's type of diode but I can't find any with a Peak Repetitive Reverse Voltage of 200 Volts or does this not matter? there is a few at 60volts and a couple at 100 Volts.

The other option I found was a ultrafast one MUR420 with a Average Rectified Forward Current of 4 Amps, Peak Repetitive Reverse Voltage of 200 Volts, Forward Voltage of 1.28v and Maximum Full Load Reverse Current of 125 microAmps. This works in 35 nano seconds.

Also if I put a high current diode in will it cause something else to blow instead of the transitor, I'm just a little worried of of burning something more expensive or harder to replace than the cheapish transitors.

Cheers

Chris

[NC Cams]

Unitrode/TI recommends the use of a UC3610 integrated diode bridge (or "equivalent") for use to protect various steppers from induced voltages.

This protects the pass element (the transistor) from the spurious currents/voltages that get induced during operation - something that rapid feeds at high rates could negatively affect an insufficiently protected drive transistor.

When the diode is NOT there or is too slow in conducting, the induced voltage tries to be conducted improperly thru the transistor - sort of like hooking it up backwards which they don't like at all.

I'm not an expert on Shottky's but when I ran into problems with fets crapping out, I used the logic outlined in my prior post and replace the 1N4001's that were there(4003's merely have a higher voltage rating but not the speed of a Schottky).

From regularly replacing "cheapish transistors", I had no transistors fail since the use of "fastest, highest current rated ones I could find".

Frankly, I should probably should have looked at the voltage rating but didn't. However, they work and seem to work well.

In all the stepper drive application notes I can find, they specifially spec out the use of Schottky's across the stepper coil to protect the driver transistor - they specifically do NOT list the use of 1N400x's.

Perhaps someone more literate than I with respect to electronic circuitry would make a more specific recommendation. But, in my experience, 1N4003's would not be what I'd use to "protect" a stepper drive circuit.

[creep_pea]

I found this on a audio forum

The reverse rating of the Schottky is not forgiving at all. A few Volts "over the line", for few nanoseconds, and you're out. One good glitch from the power line will do this. Hell, you can blow one up walking across a carpet!

It's got me woried.

What do you think of the ultrafast ones?

Also I found a 150v Schottky's diode STPS10150CT, it's a rectifier not a Dual Schottky Diode Bridge as you suggested, are these just different names for the same thing?

Sorry for all the questions, I'm trying to learn a bit.

Thanks

Chris

[NC Cams]

I'll share my understanding of diodes and Schottky's in particular.

Diodes allow current flow in 1 direction. The "arrow" in the insignia points to the direction the current flow is allowed. As current flows thru the diode, there is a voltage drop that is a function of internal resistance and some other "magic" that goes on.

The voltage rating is the amount of potential difference that can exist before the diode breaks down and allows current to flow in the "wrong" direction.

The purpose of diodes in a driver circuit are to bypass the induced current around the pass element, ergo the transistor. If you don't use them, current tries to flow "backwards" thru the emitter/collector stage which the thing doesn't like.

Schottky's offer lower resistance and faster reaction to the current flow phenomenon that takes place during an induced feedback. Thus, if you get an induced voltage, the Schottky bypasses more current faster around the pass element than a conventional silicon diode.

The "carpet zap" might be more of a voltage potential causing high current flow potential issue. If you have a 200 volt differential (sort of easily handleable) as oppose to a static charge that can reach multiples of that, the mometary current flow can burn holes in darn near anything.

Again, I'm a shade tree expert with respect to the selection and use of Schottky's in inductive load switching circuits. I can't necessarily spec out the right ones but I"ve lucked into the use of robust/adequate ones via the method previously cited.

I also know that all the application notes I've seen clearly show that Schottky's should be used to bypass the pass elements in stepper drive circuits, not conventional 1N400x series diodes.

Keep this in mind: in the early days of CNC, Schottky diodes were quite expensive. Moreover, machine speeds (especially with steppers) were no where's near what can be achieved with today's technologies. At that time, 1N4003's may have be quite adequate.

However, today, when you go to faster jog speeds followed by dead fast stops, you generate inductive problems that the original designs did not have to deal with. It would seem only logical that if you speed things up and ask for higher performance from a circuit, more appropriately selected components will be required.

EDIT: the Schottky "bridges" are essentially pre-packaged individual devices that take up less board space. They are effectively the same thing as the discrete parts only packaged more conveniently. Why 4 or 6 to a package? Check out some of the stepper driver IC's and those for brushless motors and the reason for the packaging becomes self evident.
END EDIT

[creep_pea]

Thank you for taking the time to explain. Much appricated.

I'll order some Schottky's and some of the transitors you recomended ready for the next time I blow one and see how they go.

I think I'll try these Schottky's diodes STPS10150CT, Average Rectified Forward Current of 5 Amps, Peak Repetitive Reverse Voltage of 150 Volts, Forward Voltage of 0.75v and Maximum Full Load Reverse Current of 120Amps. The only problem wih them is they are in T0-220 style not axial. I presume these should be a high enough rating with the voltage as the machine drives run on 70 volts.

Cheers

Chris

[bbuonomo]

Just wanted to say thanks to everyone who helped!

Machine is up and working great!

Now I'm just trying to clean up the BOB Controller board and wiring, I'm hoping to get the entire PC inside the Tape reader enclosure.

Brian