[strantor]

My old CNC mill is dying slow electronic death. All the AC servo drives are 20+ years old and two have failed; my functionality is reduced to milling slots in jog mode.

I have $4k worth of 1kW LS L7C servo drives and motors in my AutomationDirect shopping cart and I'm not pulling the trigger for 24 hours. Please tell me if you have bad experiences with these. I have commissioned two of them in the past and had no issues, I thought they were excellent products. But those were not CNC applications. I am not 100% confident these are up to the task of a CNC machine. Only like 90% confident.

In case it matters, these will be in +/-10V analog velocity mode with divided encoder output from the drive to the motion controller.

[joeavaerage]

Hi,
to be honest I do not think those servos are the best choice, but perhaps that is my bias.

As I've posted I use Delta B2 servos and have done for some years. I buy 750W B2s from this company and have been paying $438USD per kit (servo,drive and cables) and have done for nearly five years.
I've bought eight 750W servos and four 400W servos aver five years. Never had a fault or issue with any of them.

https://www.fasttobuy.com/flange-80m...er_p28084.html

This is a 2kW version ($895USD):

https://www.fasttobuy.com/delta-2kw-...nc_p33936.html

While this is a 1kW version and somewhat cheaper ($601USD):

https://www.fasttobuy.com/220v-1kw-d...er_p28099.html

Delta is a Taiwanese brand made in China. Good quality, support, documentation, and most importantly free set-up and tuning software at fair prices.
Your DMM servos, a Canadian brand made in China are also good quality, support, documentation and free setup and tuning software, at fair prices.
The two brands are head to head competitors.

Do the LS servos have setup and tuning software? If they do not then walk away........ Modern servos have hundreds of parameters, my B2's have 278 of them, and the B2 is the entry level servo, the A2 and B3 servos have many
more!. You need software, it is just too error prone to attempt to program them manually by pushing buttons like a microwave.

Craig

[joeavaerage]

Hi,

In case it matters, these will be in +/-10V analog velocity mode with divided encoder output from the drive to the motion controller.

Personally I think having the controller being the feedback device is for the birds....but that is my choice. I have a Step/Direction controller and the individual servo drives
close the loop. I'm of the opinion that a servo drive made by the same manufacturer as the servo is a better matched closed loop solution than anything you or I could program.

Either way, any of the Delta servos that I've listed can be in (analogue -10V-+10V) velocity or torque modes, so if you wish to have LinuxCNC close the loop they will do so.

Craig

[strantor]

Hi,
to be honest I do not think those servos are the best choice, but perhaps that is my bias.

Everyone has a bias, I don't hold it against you. I am curious though if it is a positive bias or a negative bias. Did you use LS and were unimpressed? Or have used a lot of Delta and WERE impressed? It sounds like the latter to me but if I am reading too much into that let me know.

I appreciate the links to the good deals on Delta servos, thank you for taking the time to provide them. The thing is, for reasons frustrating and uninteresting, I must order from AutomationDirect. So it's either LS or SureServo2. I believe SureServo2 is Delta (rebranded), but costs several times more than what was in your links. If I had the freedom to buy from anywhere, I would have polled for "best bang for your buck" servos and drives, but things being as they are I asked "is buying LS a mistake" because if it is, I will get the SureServo2.

Your DMM servos, a Canadian brand made in China are also good quality, support, documentation and free setup and tuning software, at fair prices.
The two brands are head to head competitors.

Personally I think having the controller being the feedback device is for the birds....but that is my choice. I have a Step/Direction controller and the individual servo drives
close the loop. I'm of the opinion that a servo drive made by the same manufacturer as the servo is a better matched closed loop solution than anything you or I could program.

Either way, any of the Delta servos that I've listed can be in (analogue -10V-+10V) velocity or torque modes, so if you wish to have LinuxCNC close the loop they will do so.

Craig

I understand your point and if I were going all the way back to square 1 I would probably do step/dir, but I already have my LinuxCNC hardware and software configured and that was a considerable amount of effort that I invested too long ago for anything useful to remain in the brain. Starting over with new LinuxCNC, different Mesa boards, if I were going to do that, it would probably take me over a week just fiddling with LinuxCNC. If I were to go that far back I probably wouldn't even use LinuxCNC. I would prefer to keep what I got to the extent I can, which means analog velocity (or torque) with encoder feedback to the controller, and the DMM drives have already proven ill equipped to perform in that way.

Hi,
Do the LS servos have setup and tuning software? If they do not then walk away........ Modern servos have hundreds of parameters, my B2's have 278 of them, and the B2 is the entry level servo, the A2 and B3 servos have many
more!. You need software, it is just too error prone to attempt to program them manually by pushing buttons like a microwave.

Craig

I completely agree, software is critical. That's one of the reasons DMM put a bad taste in my mouth. Their software is terrible. Or at least it was when I tried to use it a couple of years ago, maybe it has gotten better, I don't know. Yes, the LS drives have software and it is not bad. It's a little unintuitive, but once you're used to it, it is useful. It has setup wizards that get you 95% of the way to a working parameter set within a few minutes, and the next couple hours are spent figuring out what those other 3 parameters are, that you need to get to 100%. Those last 3 parameters are pain in the butt because the software has no parameter descriptions beyond tagname and no description at all for the variables. For parameter 0x965 should set it to...2? 3? the software doesn't tell you what 2 means or what 3 means; you have to look it up in the manual. That's my only gripe with it.

[joeavaerage]

Hi,

I am curious though if it is a positive bias or a negative bias. Did you use LS and were unimpressed? Or have used a lot of Delta and WERE impressed?

The later, I am impressed with Delta servos. I have used Omron, Yaskawa, Mitsubishi, Allen Bradley and Schiender here in New Zealand....and they are all good, but eye-wateringly expensive.
I sought a well performing and reasonable quality servo at a better price, and have found it in Delta. I've used Delta VFD's for twenty years, and they are damned good too.

I note the LS servos are Korean. I have no reason to doubt that they work as advertised.

I've used several brands of software, and to be honest found every single one of them to have their own quirks, but never have I found one that is unusable either.

The Delta B2 servo requires a programming cable (based on an IEEE1394), and I bought a genuine Delta cable, to be 100% sure that I was doing the right thing. Now that I've go it
programming a new drive is a piece of cake. I have saved files of parameters of all the different drives I've programmed, so creating a clone is a matter of half a dozen keystrokes, and even a
servo in a new application often so closely resembles another servo application that to start with a clone of that previous application and modifying it to suit.

It is certainly a case of familarity with a given brand is likely to lead you to do repeat purchases.

Craig

[joeavaerage]

Hi,
I notice the Sure Servo2 are all listed with 16,777,216 ppr encoders, ie 24bit.

Delta B2 servos have 160,000 count encoders, so the SureServo2 is not a rebranded B2. In fact I don't think its an A2 servo either, as they have 20bit encoders. That leads me to believe that the SureServo2
is a re-branded A3 or B3. All of which is very nice but also goes part way to explain why they are so pricy.

Cant help but think a Delta B2 servo is an economic and performance match for the job at hand, for something like one half the cost of SureServo2.

Even the LS servos fetch a premium over a Delta B2, by a third or so???

My money is too hard earned to let some bugger fleece me of it!!

Craig

[strantor]

Hi,
I notice the Sure Servo2 are all listed with 16,777,216 ppr encoders, ie 24bit.

Delta B2 servos have 160,000 count encoders, so the SureServo2 is not a rebranded B2. In fact I don't think its an A2 servo either, as they have 20bit encoders. That leads me to believe that the SureServo2
is a re-branded A3 or B3. All of which is very nice but also goes part way to explain why they are so pricy.

About these encoders with 6,7, or 8 figure resolution specs, are those realistic numbers?

• Encoder resolution on my existing motors is 2,000PPR
  
  • optical encoders.
  
  
• Encoders on the LS servos are 131,072 ppr
  
  • pretty sure they're magnetic
  • pretty sure they're serial
  • That's a 65x improvement over what I have (theoretically)
  
  
• SureServo 2 encoders are 16,777,216 ppr
  
  • pretty sure they're magnetic.
  • Can't tell if they're serial but that seems to be the new popular thing, so they probably are (even though there's like 30 pins in the encoder plug).
  • That's an 8,388X improvement over what I have and a 128X improvement over LS (theoretically)

• I am not convinced that magnetic encoders are as accurate as their resolution implies.*

• But assuming that they are, if they are serial, then I am not convinced that they can communicate their position to the drive fast enough to matter, unless you're only going like 10 RPM. If it's running 3krpm and serial packet #77 reports a position of 89,587 and serial packet #78 reports 97,998, then it has effectively the same resolution as my 2000ppr optical encoders.

• But assuming I'm wrong about that too, I have had no issues related to the 2000ppr encoders on my existing motors. The ballscrews are 5mm pitch (5.08 TPI) so that's 10,160 pulses per inch, for 0.000,098" resolution. I Can't imagine myself ever needing the machine tuned well enough that the comparatively "low resolution" of the existing encoders would be the limiting factor. So 65X improvement or 8,388X improvement, I don't think I will notice.

• Just for fun, linear resolution (if it's to be trusted) would be

• LS: 0.000,001,502"

• SS2: 0.000,000,012"

• WOW. I would have no way to verify accuracy at that level. Would anyone? IS there any earthly technology able to measure 12 billionths?

(*) I think they are just an analog magnetic displacement sensor and the "PPR" comes from the input resolution of the ADC that the sensor is connected to. I could connect the potentiometer out of an old tape deck to the 32 bit input of a AD7177 chip and call it a 4,294,967,296ppr single-turn encoder. I'll give them credit for using something better than an old tape deck pot, but still, it "technically could be anything." Hypothetically SureServo2 and LS could use the same sensor, but sureServo2 has a higher resolution ADC. Hypothetically, that analog sensor could be less accurate than my 2000ppr optical encoders. Again, unless I'm wrong. I am actually talking out of my butt. I don't know how they actually work, just making educated guesses. And I know nothing about the serial protocol they use.

Cant help but think a Delta B2 servo is an economic and performance match for the job at hand, for something like one half the cost of SureServo2.

Even the LS servos fetch a premium over a Delta B2, by a third or so???

My money is too hard earned to let some bugger fleece me of it!!

Craig

I hear ya, and thanks for looking out for my pocketbook. Just so you know, I don't have anything against Delta, and if my hands weren't tied, I would follow those links you provided.

[joeavaerage]

Hi,

I am not convinced that magnetic encoders are as accurate as their resolution implies.*

I tend to agree, but I'm not sure that resolution is the most important thing beyond a certain level. What a higher encoder count does provide is a less noisy (numerically) calculation.
The greater the number of bits in your calculation the finer the result. My B2's have 160,000 ppr encoders, and they are serial, and yes they work well up to and past the max speed of the servo, 5000rpm.
So I disagree that serial encoders are not quick enough. Think about Ethernet for instance, 1 Gbit/sec, so there is no question that serial can do the business.

As I say my B2's have 160,000 ppr encoders but with electronic gearing I have them at 5000ppr effective, or 1um linear resolution. So I don't 'need' 160,000ppr but it does allow me a wide choice of electronic gearing
to satisfy my application......and that is important to me.

I mention the encoder resolution, not so much as an advantage, but rather trying to identify what Delta servos have been rebranded, because the cost of SureSerov2 seems overly high. I don't give a 'tinkers cuss' about
encoder resolution over and above a practical level, but I do care about paying for features and performance that I'll never use....and can think of better uses for that extra money.

Craig

[strantor]

Good points all around, thanks for the perspective and the explanation.

[joeavaerage]

Hi,
if there is an unavoidable business reason you are locked into Automation Direct, then there are two choices, LS or SureServo2.

Neither are rubbish or bargain basement Chinese. Either would do the job on your machine no trouble.

I'm not 100% sure what Delta model the SureServo2's are, but its beyond the Delta B2, which is in itself Delta's entry level model.
That it should be more expensive is not therefore surprising, but I rather think that they are 'better' than is actually required for the retrofit you have in mind.

My feeling is that B2';s would be entirely adequate, they have certainly exceeded my every expectation. While 'better' models like A2's would be nice....would I get any
advantage....beyond bragging rights? I don't think so.

I would guess that Delta B2's are very competitive with the LS servos, both performance and price. While Automation Direct price is somewhat higher than my preferred Hong Kong Delta supplier, neither are they outrageously priced.
I have no experience with Automation Direct, but if you have reason to believe they are a credible and reliable supplier then have at it!

Craig