[Novakon]

To our valued Novakon customers,

We would like to take this time to bring to light the problem we have discovered concerning the current VFD used in our NM-145 & NM-200 mills.
As some are aware, a few VFD’s exhibited low end torque deficiency in a number of machines. Prior to discovering this problem, we did not test for torque, and all machines appeared to work to specifications. In light of reported problems, we have confirmed that some machines did exhibit a problem with the VFD.

In our experience using these machines, we have not exhibited a high failure rate as reported. We are actively working with the motor factory to resolve these issues that appear in the identified machines.

Since we started testing for the lower end torque and reset problems, we have identified in our shop approximately 15% of machines exhibiting this issue. We outlined the problem to the factory and they explained to us that the problem lies in the programming of the on-board processor. They also acknowledge that these symptoms are not exhibited in all VFDs which we have also verified. The problem with the programming created a self-protection shutdown during low speed high load operations. If the VFD experiences an overload condition, the micro-processor improperly shuts down the driver and may no longer be resettable. In this instance, the VFD becomes inoperable. Further testing has been made and was discovered that if the VFD does not initially fail from shutdown then the self protection programming works and the VFD will continue to reset properly.

We have tested in our shop, several overload conditions and the sample VFDs did in fact shutdown. Unfortunately, prior to our realization of the problem through our testing, a number of machines were released to customers that had the issue of the micro-processor overreacting to the fault. We have sufficiently tested and validated that the VFDs in our remaining stock of NM-145 and NM-200 machines are performing properly at present.

Until we understood the magnitude of the problem, we had exhausted our limited inventory of VFDs correcting this issue. The number of spare VFDs was not enough to satisfy all of our customers who experienced the same problem. We have since received additional production VFDs which are available to our remaining customers in solving this matter.

In working closely with the VFD factory on a daily basis, they now indicate that they have isolated the problem and are working on the corrective software programming/hardware. A corrective revision will be released for future VFD production. It is our plan to offer the update to standardize all VFDs to the current production revision which will be provided free of charge in order to bring our customers’ machines up to current configuration standards. The update is significant in reducing our in-house failures that pertain to our warranty with the VFD factory.

It has always been Novakon’s policy and practice to screen and test all machines in our facility against a list of accepted standards within our testing capability before shipment to our customers. As problems are reported via our customer base, we constantly review and update our testing criteria for the most effective screening and acceptance of performance. We believe that this problem should not become an issue in the future.

We would like to assure our customers that since it is our policy to integrate and test our machines in Canada, we can react to these types of problems immediately and effectively. Consequently we do experience longer shipping delays due to the more in-depth testing involved which we believe is warranted since it allows for better testing and control prior to release.

In summary, we wish to express our appreciation for the constructive feedback that has brought this issue to our attention. Rest assured that our efforts are focused in improving the support and technical issues with all of our machines. We are committed to continue and address any problems identified and provide a timely response within our resources. We remain determined to continuously improve the quality and performance of our machines.

Sincerely,

The Novakon Systems Team

[JoeBean_NEW]

To our valued Novakon customers,

We would like to take this time to bring to light the problem we have discovered concerning the current VFD used in our NM-145 & NM-200 mills.
As some are aware, a few VFD’s exhibited low end torque deficiency in a number of machines. Prior to discovering this problem, we did not test for torque, and all machines appeared to work to specifications. In light of reported problems, we have confirmed that some machines did exhibit a problem with the VFD.

In our experience using these machines, we have not exhibited a high failure rate as reported. We are actively working with the motor factory to resolve these issues that appear in the identified machines.

Since we started testing for the lower end torque and reset problems, we have identified in our shop approximately 15% of machines exhibiting this issue. We outlined the problem to the factory and they explained to us that the problem lies in the programming of the on-board processor. They also acknowledge that these symptoms are not exhibited in all VFDs which we have also verified. The problem with the programming created a self-protection shutdown during low speed high load operations. If the VFD experiences an overload condition, the micro-processor improperly shuts down the driver and may no longer be resettable. In this instance, the VFD becomes inoperable. Further testing has been made and was discovered that if the VFD does not initially fail from shutdown then the self protection programming works and the VFD will continue to reset properly.

We have tested in our shop, several overload conditions and the sample VFDs did in fact shutdown. Unfortunately, prior to our realization of the problem through our testing, a number of machines were released to customers that had the issue of the micro-processor overreacting to the fault. We have sufficiently tested and validated that the VFDs in our remaining stock of NM-145 and NM-200 machines are performing properly at present.

Until we understood the magnitude of the problem, we had exhausted our limited inventory of VFDs correcting this issue. The number of spare VFDs was not enough to satisfy all of our customers who experienced the same problem. We have since received additional production VFDs which are available to our remaining customers in solving this matter.

In working closely with the VFD factory on a daily basis, they now indicate that they have isolated the problem and are working on the corrective software programming/hardware. A corrective revision will be released for future VFD production. It is our plan to offer the update to standardize all VFDs to the current production revision which will be provided free of charge in order to bring our customers’ machines up to current configuration standards. The update is significant in reducing our in-house failures that pertain to our warranty with the VFD factory.

It has always been Novakon’s policy and practice to screen and test all machines in our facility against a list of accepted standards within our testing capability before shipment to our customers. As problems are reported via our customer base, we constantly review and update our testing criteria for the most effective screening and acceptance of performance. We believe that this problem should not become an issue in the future.

We would like to assure our customers that since it is our policy to integrate and test our machines in Canada, we can react to these types of problems immediately and effectively. Consequently we do experience longer shipping delays due to the more in-depth testing involved which we believe is warranted since it allows for better testing and control prior to release.

In summary, we wish to express our appreciation for the constructive feedback that has brought this issue to our attention. Rest assured that our efforts are focused in improving the support and technical issues with all of our machines. We are committed to continue and address any problems identified and provide a timely response within our resources. We remain determined to continuously improve the quality and performance of our machines.

Sincerely,

The Novakon Systems Team

It's commendable that you're communicating about the issue. That's something I would have liked to have seen before, but it's never too late to start.

That said, there are a few things from what I'm reading that don't make sense to me.

For example, you state that according to your testing only about 15% of units have this issue, so the vast majority (85%) are apparently operating correctly. In fact, you state that you have 'sufficiently tested and validated' the stock that you have and they are operating normally. You also explain that the issue is software-related and Sangmutan are working on a fix. But here's the issue:
1. If the problem is simply software-related then should not all units be faulty, as all units presumably have the same software? If the issue was fixed at some point in the past with a software update then why have Sangmutan been unable as of yet to release fixed drives?
2. If the problem is a combination of a software glitch and the use of particular hardware would it not be a simple matter to isolate the problem hardware and avoid it in new units? Why bother with a software solution if the issue can be truly solved with a change of hardware?
3. If it's a matter of some component(s) being out of spec causing the trouble then, besides simple QC steps to solve the matter, if your testing method were truly effective would not a relatively simple solution be to test each unit in an overload condition, verify that it functions normally, and then release it?
3. If 85% of all units are trouble-free, why would you bother replacing units that you know (ie. "have sufficiently tested and validated") are working properly?
4. The controller runs a dsPIC. You state that "the micro-processor improperly shuts down the driver and may no longer be resettable". If that's true, a) how in the world did they manage to lock up a PIC in a motor controller application, and b) would it not be a simple matter to reload the program to the PIC in Toronto and send the controller back out?
5. You state that "[p]rior to discovering this problem, we did not test for torque" and that "[i]t has always been Novakon’s policy and practice to screen and test all machines in our facility against a list of accepted standards within our testing capability before shipment to our customers". It's a massive oversight to have a brand-new drive system and then not bother testing if it's satisfactory. That really raises questions as to what else you've overlooked.
6. You state "We would like to assure our customers that since it is our policy to integrate and test our machines in Canada, we can react to these types of problems immediately and effectively. Consequently we do experience longer shipping delays due to the more in-depth testing ..." I'm not quite sure what you mean by these statements as they seem contradictory at first glance, but obviously the first sentence is not correct in this case. Your 'integrating' the machine in Canada in no way helped you to "react to these types of problems immediately and effectively". You are at the mercy of Sangmutan in China, and many customers who have had this issue have had to wait a very long time for any kind of a solution.

Frankly, though, I don't think that you and/or Sangmutan truly have a handle on this problem, as evidenced by the still absent solution and the seeming inconsistencies, but I think what is/was far worse is the delay in acknowledging the problem and trying to work out reasonable solutions. The implication has been that the problem was the customer's fault, or at the very best not Novakon's responsibility, and even in the comments above (eg. "In our experience using these machines, we have not exhibited a high failure rate as reported.") the tendency is to downplay the problem. Owning up to issues will win respect, but covering up issues destroys credibility.

Having said all of that, I do appreciate that you've stepped up to the plate and offered the temporary replacement drives for affected customers, as well as the upgrade when/if it becomes available. But in my opinion a lot of damage to Novakon's reputation has been done because of poor communication and a slow initial response. Consequently, I think there's likely to be some loss of good will towards the company that will take time to build back up.

[dbrija]

The lack of a response speaks volumes, doesn't it?

[MichaelHenry]

Hopefully the thread itself doesn't get deleted or moved to some obscure part of CNC Zone - that would speak a lot louder.

Mike

[JoeBean_NEW]

Just for the record, I've noticed some things disappeared a while back on cnczone. But I think there are 2 reasonable explanations:
1. Some things seem to have become archived and thus the beginning of some threads are not available in that thread as found via a search. But they seem to be available through Google.
2. I think there must have been a server crash at somepoint a few weeks ago as several threads in a number of forums appeared, then a few hours later (next morning) were gone.

At any rate, my replacement drive & controller (note: not the fix, just a temporary replacement) should be here soon. Still no word on a permanent fix, though...

[MichaelHenry]

Do threads here get archived somewhere? I'd thiught they just stayed forever.

I vaguely recall at least one thread in one of the mfr forums criticising something about the mill or company and that got deleted or banished to some other forum. If that is necessary, I think that there should at least be a message stating what action was taken and why.

Mike

[newton009]

Do threads here get archived somewhere? I'd thiught they just stayed forever.

I vaguely recall at least one thread in one of the mfr forums criticising something about the mill or company and that got deleted or banished to some other forum. If that is necessary, I think that there should at least be a message stating what action was taken and why.

Mike

Advertising pays more to CNCzone than do our comments.

[JoeBean_NEW]

Advertising pays more to CNCzone than do our comments.

Add to that, though, that the ads are only here because of our traffic/comments. If the admins/moderators do enough to ostracize their audience they will also lose their advertisers.

Note, though, that I don't disagree with the sentiment of your post. It's a difficult line to walk.

[Al_The_Man]

Advertising pays more to CNCzone than do our comments.

FYI, The mods here are generally unbiased, there is no favour given to anyone based on who he is.
For example the ability of advertisers to moderate their own forums was recently taken away because they were deleting posts that did not show their product in a favorable light.
Moderators remove or edit posts based on the CNCzone rules & code of conduct, only.
Al.

[newton009]

FYI, The mods here are generally unbiased, there is no favour given to anyone based on who he is.
For example the ability of advertisers to moderate their own forums was recently taken away because they were deleting posts that did not show their product in a favorable light.
Moderators remove or edit posts based on the CNCzone rules & code of conduct, only.
Al.

Thank you!

I suspect that generally most deleted response are due to re-think of the member. Just as I have done myself.

I strongly support what the community is doing in regards to CNCzone.

A never ending wonder!

Thanks again.

[ihavenofish]

FYI, The mods here are generally unbiased, there is no favour given to anyone based on who he is.
For example the ability of advertisers to moderate their own forums was recently taken away because they were deleting posts that did not show their product in a favorable light.
Moderators remove or edit posts based on the CNCzone rules & code of conduct, only.
Al.

they do like to remove links to certain companies websites though.

anyhow, we wait patiently for new drives for our spindles. while i agree that novakon should have been keeping everyone updated about any issue and solutions much earlier, i also respect that its difficult when you havent really got a source of the problem, let alone a viable solution.

screwups by manufacturers like this one - which affect many machine sellers - take patience and communication on all sides i think.

in the meantime i have been searching for possible alternative drives and motors for the nm145. its a very bothersome situation though, as almost every affordable drive is made by samutang.

[MichaelHenry]

FYI, The mods here are generally unbiased, there is no favour given to anyone based on who he is.
For example the ability of advertisers to moderate their own forums was recently taken away because they were deleting posts that did not show their product in a favorable light.
Moderators remove or edit posts based on the CNCzone rules & code of conduct, only.
Al.

It's good to hear that non-sponsor moderators are the only ones that can delete or move posts now. Do they leave a message in the thread to state that one or more posts were deleted/moved and if moved do they link to the new location? That would seem like a fair thing to do.

Mike

[Al_The_Man]

. Do they leave a message in the thread to state that one or more posts were deleted/moved and if moved do they link to the new location? That would seem like a fair thing to do.
Mike

In some case it is done, but due to the time involved to daily monitor all the spammers, policy violators, reported postings, and general miss postings, it is not always possible.
Al.

[JoeBean_NEW]

anyhow, we wait patiently for new drives for our spindles. while i agree that novakon should have been keeping everyone updated about any issue and solutions much earlier, i also respect that its difficult when you havent really got a source of the problem, let alone a viable solution.
in the meantime i have been searching for possible alternative drives and motors for the nm145. its a very bothersome situation though, as almost every affordable drive is made by samutang.

1. They shouldn't claim they know the problem when they don't (or, at least, the guy explaining doesn't). They said "If the VFD experiences an overload condition, the micro-processor improperly shuts down the driver and may no longer be resettable. " That's not exactly accurate. The uPIC on the contoller is fine. The problem in on the driver, a blown power module by the looks of things. It's a Fairchild FSBB30CH60F. Good luck finding one, though, as I ordered the only 3 that I could find in North America last week. Mouser lists them but they're b/o. The Fairchild is primarily what that oversized heatsink is cooling. And in this case, oversized is not necessarily good, as you need the proper ratio of surface area to volume to allow heat dissipation and I believe this combo is inefficient. If we had a massive surface that was putting out heat this heatsink might be more appropriate, but we're dealing with an area about the size of the cross-section of a AA battery. Perhaps adding a fan, or changing the cooling system, will solve these dying drive issues. I'm thinking of installing a more suitable cooling system for a PC CPU, maybe liquid cooled.

2. Sangmutan are not the only suppliers of cheap BLDC motors & drives. Adlee Powertronic are importable for a decent price, and were recommended once or twice on here. The problem is, I contacted them with the info about the application and they basically said it might work, but they won't guarantee it, and they're not comfortable with this application because of (*SHOCK*) the high-torque requirements on the low-speed end. FWIW Adlee has been in the business for a LOT longer than Sangmutan. So if they're not comfortable I'd be a bit leery.
So the next best option is a 3-phase motor & VFD. A quality vectorless VFD should give a decent torque and a decent speed range, but we'd likely still need a couple of pulleys for full speed range.
Or, we could try adapting a Fanuc spindle motor. Not sure what they require as far as a drive, though, and cost would not be cheap unless you could get a deal. I believe Al_The_Man has commented on a few threads on the topic of using Fanuc spindle motors on machines so he'd likely know what we'd need to do.

[ihavenofish]

1. They shouldn't claim they know the problem when they don't (or, at least, the guy explaining doesn't). They said "If the VFD experiences an overload condition, the micro-processor improperly shuts down the driver and may no longer be resettable. " That's not exactly accurate. The uPIC on the contoller is fine. The problem in on the driver, a blown power module by the looks of things. It's a Fairchild FSBB30CH60F. Good luck finding one, though, as I ordered the only 3 that I could find in North America last week. Mouser lists them but they're b/o. The Fairchild is primarily what that oversized heatsink is cooling. And in this case, oversized is not necessarily good, as you need the proper ratio of surface area to volume to allow heat dissipation and I believe this combo is inefficient. If we had a massive surface that was putting out heat this heatsink might be more appropriate, but we're dealing with an area about the size of the cross-section of a AA battery. Perhaps adding a fan, or changing the cooling system, will solve these dying drive issues. I'm thinking of installing a more suitable cooling system for a PC CPU, maybe liquid cooled.

2. Sangmutan are not the only suppliers of cheap BLDC motors & drives. Adlee Powertronic are importable for a decent price, and were recommended once or twice on here. The problem is, I contacted them with the info about the application and they basically said it might work, but they won't guarantee it, and they're not comfortable with this application because of (*SHOCK*) the high-torque requirements on the low-speed end. FWIW Adlee has been in the business for a LOT longer than Sangmutan. So if they're not comfortable I'd be a bit leery.
So the next best option is a 3-phase motor & VFD. A quality vectorless VFD should give a decent torque and a decent speed range, but we'd likely still need a couple of pulleys for full speed range.
Or, we could try adapting a Fanuc spindle motor. Not sure what they require as far as a drive, though, and cost would not be cheap unless you could get a deal. I believe Al_The_Man has commented on a few threads on the topic of using Fanuc spindle motors on machines so he'd likely know what we'd need to do.

on the first, i think hes only quoting what samutang has told them, which is why i say they dont really know what the issue is - asside from it being the drive programming. this makes it difficult to predict when anything can be fixed permanently. my understanding is that the nm200 motor has been fixed at the factory, but the 145 is still problematic. i dont fully understand what the difference between the 2 is besides size.


on the second, i *think* fanuc motrors have proprietary control and encoder signals (although that might just be on their servos). i think most surplus models will be 380 or 560v as well. this may make them more challenging to implement than a more common high speed induction motor and inverter.

the aldee stuff looks interesting. i honestly dont think torque is a problem though for bldc when they are working right. the motor in my nm070 (sieg kx1) is 500w, and has LOTS of low end torque, although the controller is slow to respond to load.

there is an italian company i think i showed you before that makes a 12000rpm 3hp induction motor. its basically a router spindle with a keyed shaft. the price is decent (well under a grand with inverter), but mounting would be tricky as it has no nema style flange.

[JoeBean_NEW]

on the first, i think hes only quoting what samutang has told them, which is why i say they dont really know what the issue is - asside from it being the drive programming. this makes it difficult to predict when anything can be fixed permanently. my understanding is that the nm200 motor has been fixed at the factory, but the 145 is still problematic. i dont fully understand what the difference between the 2 is besides size.


on the second, i *think* fanuc motrors have proprietary control and encoder signals (although that might just be on their servos). i think most surplus models will be 380 or 560v as well. this may make them more challenging to implement than a more common high speed induction motor and inverter.

the aldee stuff looks interesting. i honestly dont think torque is a problem though for bldc when they are working right. the motor in my nm070 (sieg kx1) is 500w, and has LOTS of low end torque, although the controller is slow to respond to load.

there is an italian company i think i showed you before that makes a 12000rpm 3hp induction motor. its basically a router spindle with a keyed shaft. the price is decent (well under a grand with inverter), but mounting would be tricky as it has no nema style flange.

I still don't think they should be making authoritative claims about understanding the problem when they (or at least the poster) apparently doesn't, but oh well...

Sangmutan claim to have a fix but AFAIK nobody,including Novakon, actually have the fixed drives in their hands. Keling has the 1100W back in stock, so I guess they must be working, but I believe from another thread that the new motors are 8 pole, so that's part of the "fix". AFAIK none of those who had the troubled earlier models with the 4-pole motors are fixed, but I may simply have not seen any posts on the topic.

I know all the Fanuc motors use special drives but there are a few threads around that suggest people have modified them. I'm not sure how though. The higher voltage isn't too big of a deal, we'd just need a transformer. It's the drive that would worry me....

A simple 3ph motor and VFD like Tormach uses might bethe simplest issue if we have to go that way.

[MichaelHenry]

In some case it is done, but due to the time involved to daily monitor all the spammers, policy violators, reported postings, and general miss postings, it is not always possible.
Al.

Thanks for the reply Al. Just speaking for myself here, but I don't really care if a spammer's post gets deleted with no record. I'm more concerned with posts that discuss problems with products when deleting them could be construed as censureship to the benefit of the manufacturer. If it is necessary, I would like to see some public record of what was done.

Mike

[JoeBean_NEW]

I just noticed Novakon has removed the 2200W (3HP) BLDC drive from the new NM-200s:
NM-200 Pro Series 2
http://www.novakon.net/NM-200%20DATA%20SHEET.pdf
Pertinent info from that second link:
"Spindle
Motor HP Rating 2HP AC Spindle 1500W"

So it looks like hope for a true fix is, indeed, not very good. *sigh* My guess now would be the temporary replacement controller that was supposed to be a hold-over until a real fix came might be all we'll get.

The NM-145 still seems to come with the BLDC unit, though with a note saying to expect up to ±10% variance in actual output speed.

[ihavenofish]

I just noticed Novakon has removed the 2200W (3HP) BLDC drive from the new NM-200s:
NM-200 Pro Series 2
http://www.novakon.net/NM-200%20DATA%20SHEET.pdf
Pertinent info from that second link:
"Spindle
Motor HP Rating 2HP AC Spindle 1500W"

So it looks like hope for a true fix is, indeed, not very good. *sigh* My guess now would be the temporary replacement controller that was supposed to be a hold-over until a real fix came might be all we'll get.

The NM-145 still seems to come with the BLDC unit, though with a note saying to expect up to ±10% variance in actual output speed.

ive been wondering. does anyone think there is there a reason a granite devices servo drive wouldnt work? its supposed to control brushless dc motors, and can run at 160V 10A or 14A continuous (which should work with a simple transformerless rectified 220v supply).

there must be some cheaper options than the granite as well. ($280/$340cdn).

the only thing i could think of making this not work is if the motor is fundamentally messed up.

hmm.

[JoeBean_NEW]

ive been wondering. does anyone think there is there a reason a granite devices servo drive wouldnt work? its supposed to control brushless dc motors, and can run at 160V 10A or 14A continuous (which should work with a simple transformerless rectified 220v supply).

there must be some cheaper options than the granite as well. ($280/$340cdn).

the only thing i could think of making this not work is if the motor is fundamentally messed up.

hmm.

If the software is the main issue then new software should solve the problem. The controller's dsPIC30f based, which is often used on BLDC controls. In fact, the specific model used is this one.
From what I could see the Granite Drives unit requires an encoder.

Note that they even have sample code available there. I'm guessing the actual code they're using has very little modification. And since this chip and it's ilk are used a fair bit within the RC community I'm guessing there's room to improve and make this functional. Programming should be simple enough with a picKit 3 as the header is right there on the board - just plug in and go. I have one on order which I'll be using to experiment with.

I still think cooling is a big problem, though. I was looking at a few cheap (<$100CAD) CPU coolers on newegg and I think that might bw the way to go. Then put an opening on the top of the cabinet with a filter to vent the hot air up and out.

If I decide to give up on the controller I think I'll retrofit a 3ph motor & vectorless VFD. At least it has a proven track record in this application.

On a totally off-topic note, I got a great deal yesterday on a Haff Schneier 3D Taster on eBay. I've been wanting one to play with for a while now but they don't come up often. When I was first thinking about it (last fall) there were 3 units on eBay. I figured I'd wait, no rush as they're a dime a dozen. WRONG. Sine then I've only seen 1 unit, and it sold for more than new. I almost had decided to buy the KBC knockoff when this one came up... I guess that's how it goes.

Watch now, KBC will go out of business and sell them off for $25 each.