[WesM]

I'm revisiting an old project I had set aside due to life events. I have an old cnc knee mill (was originally setup for CNC) that I would like to put new servos/drives and a new control on.

Doing some digging on the AC servos, DMM tech is definitely a strong contender, but at roughly $2k for my XYZ axis, I'm also looking at other options.

I noticed that stepperonline (someone I have used for steppers) has a T6 series servo that spec wise seems very similar to the DMM servos. Has anyone used the stepperonline servos or have any thoughts on them? I could not find much background on the servos (T6 or E6 series), but they seem like a good value at around $1k for a xyz axis setup.

https://www.omc-stepperonline.com/t6...rs750h2a3-m17s


For the control, I'm currently leaning towards linux CNC with a Mesa 7I76e or 7I96s controller.

[joeavaerage]

Hi,
I use Delta servos. They are a Taiwanese brand made in China. They are good quality, have good support and documentation, free setup and tuning software at fair prices.
They are, bang for your buck, a close competitor to DMM.

The same supplier whom sells me these servos have there own brand ToAuto. I do not have any, but experience with this company leads me to believe they are credible,
and of course the price is attractive. If you have had experience with AC servos before then you'll have no trouble programming these, if however you've not had any
exposure to AC servos then you'd be better off paying the extra to get DMM or Delta. The setup and tuning software moderates what can otherwise be a steep
learning curve.

750W, 80mm frame,Delta B2 series with 160,000 count per rev encoder for $438USD:
https://www.fasttobuy.com/flange-80m...er_p28084.html

750W, 80mm frame, ToAuto with 10,000 count per rev encoder for $242USD:
https://www.fasttobuy.com/nema-32-cn...he_p35883.html

Craig

[cncmakers001]

We are suing GSK AC Servo here.

[WesM]

Hi,
I use Delta servos. They are a Taiwanese brand made in China. They are good quality, have good support and documentation, free setup and tuning software at fair prices.
They are, bang for your buck, a close competitor to DMM.

The same supplier whom sells me these servos have there own brand ToAuto. I do not have any, but experience with this company leads me to believe they are credible,
and of course the price is attractive. If you have had experience with AC servos before then you'll have no trouble programming these, if however you've not had any
exposure to AC servos then you'd be better off paying the extra to get DMM or Delta. The setup and tuning software moderates what can otherwise be a steep
learning curve.

750W, 80mm frame,Delta B2 series with 160,000 count per rev encoder for $438USD:
https://www.fasttobuy.com/flange-80m...er_p28084.html

750W, 80mm frame, ToAuto with 10,000 count per rev encoder for $242USD:
https://www.fasttobuy.com/nema-32-cn...he_p35883.html

Craig

Thanks for the advice, I saw a similar post you put out recommending the Deltas and may consider them. They seem a little less expensive than a DMM tech setup (which has gotten a lot more expensive since the last time I looked a couple years ago). I have never setup AC servos before. So the simplicity of having a good autotune software is deffinitly appealing.

The toauto servos come in at around what the stepperonline servos cost (after shipping) but do not have the 17bit encoders. I reached out to stepperonline to see if there is autotune software for their servos or if manual tuning is the only method available.

[pippin88]

Chinese servos are coming down in price further.

Can now get some very cheap on AliExpress with 17bit encoders.

Personally I would go for name brand servos with good quality manual. There are a lot of settings that you may need to change. Programming software is essential.
I would rather pay more $$ than spend 20 hours trying to program and tune a servo without a proper manual and through a few push buttons.

[joeavaerage]

HI,

Personally I would go for name brand servos with good quality manual. There are a lot of settings that you may need to change. Programming software is essential.
I would rather pay more $$ than spend 20 hours trying to program and tune a servo without a proper manual and through a few push buttons.

I agree. I buy 750W Delta B2's for $438 whereas from the same supplier I can get 750W ToAuto for $242....but without the software...and I choose the more expensive Delta's.

Craig

[joeavaerage]

Hi,

The toauto servos come in at around what the stepperonline servos cost (after shipping) but do not have the 17bit encoders

Twenty years ago a 10,000 count per rev encoder was considered state of the art, and not many years before that resolvers were the norm with resolutions of 5 arc min-15 arc min.
17 bit encoders are nice, but not critical, a well set up servo with a 10,000 count encoder will be perfect for CNC in every way.

Craig

[Karl_T]

I used DMM's largest servos on my Vectrax CNC mill rebuild a few years ago. The machine works great.

I used them with a galil control card in +/- 10V analog mode and had a few learning issues. I think most are using them in step servo mode. Did not like this route as no position feed back to the control.

Ask Jim Dawson for his input. he worked with me on this install. He now suggests another vendor's servos.

[Jim Dawson]

My current prefered vendor of servo systems is Automation Direct. They carry both the SureServo2 (Delta) line, and the slightly less expensive LS Electric servos line.

https://www.automationdirect.com/adc...motion_control

Local support, great documentation, setup software is included. I have used both of the above in projects and have been happy with the result.

[WesM]

HI,


I agree. I buy 750W Delta B2's for $438 whereas from the same supplier I can get 750W ToAuto for $242....but without the software...and I choose the more expensive Delta's.

Craig

I will have to look up the delta software, I saw some tutorials DMMtech had up for tuning their drives and it looked incredibly simple. Is Delta tuning software similarly simple?

Are the Delta B3 servos better than the B2? I noticed the vender you linked has the B3 for almost the same price.

I'm a machinist, not a machine builder (although machine building is something I am interested in), so I don't want to spend all my time trying to inadequately tune a drive if the Delta or DMM are that much more straightforward to tune.

[joeavaerage]

Hi,

Is Delta tuning software similarly simple?

Simple?....when setting up and tuning servos?...I'm not so sure that simple is the right description, but yes overall I would say it is straight forward.

I use B2 series servos, they have been in manufacture for ten years or so and are now Delta's entry level model. They have an IEEE1394 socket to program
the drive. I bought a genuine Delta USB-IEEE1394 adapter ($60USD) to do so, but there are cheaper ways to do the same thing. I understand that later Delta models
have a USB programming socket that circumvents the need for a specialist cable. Of course that one cable can program as many servos as you have.

There are two parts to the software. The first part is the setup software. The are hundreds of parameters that need be set. The software will read the drive and lay
out the parameters in a number of logical pages. You may then edit any or all of the parameters and then program the drive with your selection. The advantage is that
all the parameters are laid out in a table, with things like a description of what each does, the numeric range of acceptable inputs, the default setting. In addition you can save
a set of parameters. For instance if you spend some time setting up and tuning the X axis servo, you can save it. Should you ever have to replace the drive, or maybe
make another machine and require another X axis, grab the copy and your done. You can also use the X axis parameters as the basis for setting up and tuning the Y and Z axes,
often very similar set ups.

The second part of the software is tuning. It is somewhat more complicated to use but is still reasonably straight forward. The feature that most use is Autotune. There you set a number
of movement details and thereafter the software seizes control of the servo and moves it back and forth rapidly so that it may calculate the optimum tune parameters.
My machine has an inertia ratio of 8:1, and is in the 'sweet spot' for auto tuning. If you wish to manually tweak the Autotune solution you can. The software has an oscilloscope built in
to allow you to visualise the dynamic response of the servo and tune individual parameters to your hearts content.

I found that at an inertia ratio of 8:1 that Autotune got me to 90% to 95% of the best possible solution, and while I did try to improve on it I saw little improvement for my efforts.
If the inertia ratio were 15:1 or 20:1 then manual tuning would not only be beneficial but probably mandatory. For this purpose one of the details you are asked to enter is the inertia
ratio, it allows the algorithm to quickly and accurately converge on the optimum solution. If you don't know the inertia ratio the software will try to calculate it from the dynamic response
of the servo, but far better you give it a good estimate.

I personally found the set up software to be the most useful. If you have to program the drive by pushing buttons like a microwave it very tedious and EXTREMELY error prone.
Hitting one button to program the drive in one fell swoop is 'night and day' better.

Are the Delta B3 servos better than the B2? I noticed the vender you linked has the B3 for almost the same price.

To be honest I don't know. One would presume that the latest model would be better than the B2 which has been around for a while. The only difference that I know of is the encoder.
The B3 has a choice of 24 bit incremental, 24 bit absolute and 24 bit multiturn absolute. While the basic (24bit incremental) might be the same price...the other higher grade options are certainly
more than the B2. The B2 has a 160,000 count, also called 17 bit encoder. I have never had any cause to think the the encoder has limited the ability of the servo. I rather suspect that I'd notice
no difference if I went to 20 bit (like the A2 and A3) or the 24bit like the B3 series. If you want or need 24bit multiturn absolute then presumably you don't mind paying for it.

If I were going to improve my servos I would choose the A2 series. The A2 series has a secondary encoder input, maybe a glass scale for instance, that you can use for closing the position
loop. This is called 'load sensing' and is the last word in machine accuracy. They cost about another $100 to $150 over the same B2 servo. I don't even come close to taxing the B2
performance....and I have nothing like the budget to get good (not cheap s***t Chinese made ones either) linear scales to make the A2's be worthwhile.

My machine with B2 servos and C5 ground ballscrews gets me parts to around 0.02mm to 0.01mm accuracy. If I wanted even better I would need not only A2 servos and expensive
linear scales but need to redesign my machine to be twice as stiff, ie five times the cost plus all the servos etc. Chasing um level accuracy is a hugely expensive business, way more
than my means. My budget means that I'll just have to be satisfied with 10um to 20um.

Craig

[WesM]

Yeah I wont see any value from the kind of investment dual input would give me on a knee mill. My expectations for this machine are at best +/- .001. If I can repeat on that I will be happy.

How would I go about calculating the inertia ratio of my machine? Is this engineering level calculus, or is it something a layman can understand?

[ardenum2]

have a look at lichuans A5 23bit

https://lichuanmotor.en.alibaba.com/...motor_kit.html

[joeavaerage]

Hi,

Yeah I wont see any value from the kind of investment dual input would give me on a knee mill. My expectations for this machine are at best +/- .001. If I can repeat on that I will be happy.

Quite so, that's the same conclusion I came to. There are many things I can spend money on to improve my machine, but load sensing servos and linear scales would be an improvement without doubt,
but the cost is just too high and would leave other more important features unattended by virtue of having sucked up all the available funds.

How would I go about calculating the inertia ratio of my machine? Is this engineering level calculus, or is it something a layman can understand?

Simplified calculations are perfectly doable, but if you start including frictional terms they become quite complicated quite quickly. Normally friction accounts for as little as 5%
of the calculation, so I tend to ignore it, favoring simplicity and clarity over technical rigour.

The inertia ratio is the ratio of the load, which includes the ballscrew and the linear axis mass divided by the inertia of the armature of the servo. Just as an example I'll lay out the inertia caluation for
my new built mill (18 months or so):

I use 750W Delta B2 servos which has a first moment of inertia of 1.16 x 10^-4 kg.m². This is published by Delta. Theses servos are called 'low inertia'. You can get 750W
servos which are larger in diameter but shorter in length, typically developing more torque at lower speeds, but most importantly have a higher first moment, say 2.5X10^-4 kg.m².
These might be called medium inertia servos. Medium inertia servos might be a better bet for large and particularly heavy machines whereas low inertia servos would suit lighter faster
moving machines.

For the purpose of this discussion I'll use the low inertia (1.16 x10^-4kg.m²) data, but be aware that there is some considerable choice about the inertia of different models
of servos.

I use a 32mm diameter ballscrew 700mm in length. The first moment of inertia can be calculated easily, it is a well known engineering calculation:

J_screw=8.23 x10^-4 kg.m²

The linear axis of my machine is 150kg, and the pitch of the ballscrew is 5mm. There is a calculation that equates linear momentum to a first moment in terms of angular velocity:

J_axis= mass x pitch² /( 2 x pi)²
J_axis=0.95 x 10^-4kg.m²

The first thing to note it that the linear axis momentum is actually a small fraction of the total, about 9%, and the momentum of the ballscrew dominates, 80%.
This surprises many people but is best explained that while the linear axis mass is very heavy it actually only travels very slowly, a few mm/sec, whereas a ballscrew
while being much lighter is spinning very fast, typically 25 to 50 revs/sec.

Now we can complete the inertia ratio of my machine:
Ratio= (8.23 +0.95)/1.16
=7.9

To do this calculation for your machine you need to know the diameter and length of the ballscrew, the pitch of the ballscrew, and the total mass of the moving axis.

Craig

[WesM]

Interesting, since the mass on my XYZ axis is quite different, I guess that means I will have different tuning on my XY and Z servos? I'm only driving the quill for Z, so that should have significantly less mass and a shorter ballscrew then my X or Y axis.

I will try and make a cad model in solidworks over the weekend of my XY carriage so I can get a decent weight estimate.

I am leaning towards getting 3 of the 750w Delta B3 servos/drives at this point. I like the DMM tech, but they end up being about $2k for the kit (3 servos/drives and cabling) vs $1500 for the Deltas.

The only thing that keeps me coming back to the DMM tech servos is they probably have warranty/support, while I assume since I'm getting the Deltas from China they will come with no support or warranty from Delta.

[joeavaerage]

Hi,
getting a mass estimate is useful, but as my calculation showed the axis mass is 'small beer' compared to the ballscrew.

Metal cutting mills tend to have large diameter ballscrews, 25mm, 32mm, 36mm and 40mm, with fine pitches, say 5mm. This results in high
thrust at slow speed for a given amount of torque, appropriate for metal cutting.

CNCRouters and woodworking machine tend to have smaller ballscrews 16mm, 20mm and 25mm, with coarser pitches, say 10mm,20mm etc.
This results in lower thrust but higher speeds for a given applied torque, just right for high speed low(ish) mass woodworking machines.

This is all intuitive and straight forward but there is another concomitant feature of the same argument, namely that the momentum, acceleration and inertia ratio
calculations vary in character also.

Large diameter fine pitch screws have high mechanical advantage and therefore the momentum calculation is less sensitive to axis mass but very sensitive to
ballscrew momentum. Small diameter medium pitch screws have low mechanical advantage and so it may well be that the axis mass dominates to momentum calculation.

As an example, if I recalculate the inertia ratio of my machine but with double the axis mass (300kg cf 150kg):
Inertia ratio=8.7
If I now calculate the inertia ratio with half the axis mass (75kg cf 150kg)
Inertia ratio=7.5

So the inertia ratio of my machine, a metal cutting mill, is very insensitive to axis mass.

May I suggest that you pay particular attention to the rotating components, the ballscrew, any large of heavy couplers and the servo armature itself. If your machine
is a metal cutting mill it is very likely that these components will dominate the calculation and the axis mass be 'small beer'.

Craig

[joeavaerage]

Hi,

The only thing that keeps me coming back to the DMM tech servos is they probably have warranty/support, while I assume since I'm getting the Deltas from China they will come with no support or warranty from Delta.

That is a good question. When I get these servos they have a warranty card issued by Delta. I've never had to make a claim, but I presume I would return it to the supplier and they would
secure a replacement. I have had many purchases (over a dozen) from this company and believe they would do so....and I do know that they have direct contact with Delta....but without trying the
system with a claim I don't know for sure.

DMM have a good reputation for support and warranty service in Canada and the US. Its somewhat dubious that the same service would extend to me in New Zealand, but who knows?.

My own experience of the Delta products that I have purchased (20+ servos and a dozen or so VFD's) is that I've never had a claim nor expect to do so. I suppose there is always a first time.

Craig

[WesM]

Got out and measured the ballscrews tonight. I will try and model up the X and Y axis tomorrow to get weight estimates on those.

X Axis is 1000mm x 32mm diameter x 5mm pitch. 6.0kg
Y axis is 760mm x 32mm diameter x 5mm pitch. 4.5 kg
Z axis is 320mmx 25mmd diameter, 5mm pitch. 1.13 kg

[WesM]

Maybe I am missing something, but it seems like I cant edit posts I have made? odd.

Anyhow

X Axis is 1000mm x 32mm diameter x 5mm pitch. 6.0kg
Y axis is 760mm x 32mm diameter x 5mm pitch. 4.5 kg
Z axis is 320mmx 25mmd diameter, 5mm pitch. 1.13 kg

X axis weight is right at 300 lbs
Y axis (which carries the X) weight is right at 600lbs
Z axis only weights around 50lbs best as I can tell.

All axis are on a 2:1 reduction L type timing belt right now.

With that info does the 750w medium inertia B2/3 servos seem reasonable? I'm also wondering if I would be better off with the 400w B2/3 on the Z axis, since it is much lower (.277 vs 1.3) inertia than the 750w and is pushing less mass. But then I don't know if the Z ends up needing more torque/power to hole its position when experiencing cutting forces because its so much less weight. What is the conventional wisdom on this one, just spec all axis 750w?

I'm also talking to teknics to see if they have a servo solution in the same price range. Interested to see what they recommend.

[joeavaerage]

Hi,
the most problematic of the axes is the Y axis, by virtue of its mass.

The Y axis inertia ratio:

J_screw= m .r²/2 where m is the screw mass and r is the radius of the screw
J_screw=4.5 (0.016)²/2
=5.76 x 10^-4 kg.m²

J_axis= m .p²/(2.PI)² where m is the linear axis mass and p is the screw pitch
J_axis=273 .(0.005)²/39.48
=1.73 x 10^-4 kg.m²

Lets assume for a moment a servo with a first moment of 1x10^-4kg.m then the inertia ratio is:
Inertia Ratio=(5.76 +1.73)/1
=7.49

This is a good number, any well performing servo of approximately 1x10^-4 kg.m² inertia will handle this ratio with aplomb. If you chose a 400w servo
of only 0.2 x 10^-4kg.m² the ratio would be 37.5!!! That would be near impossible to tune, even if it had enough torque to satisfy other requirements.
Alternately a servo, still say 750 W but slower rated speed with 2x10^-4kg.m² inertia would have a ratio of 3.75, and so would be very easy.

Note here that I have assumed direct coupling, ie made no allowance for the 2:1 reduction. All these figures get better with that reduction, ie the effective inertia ratio improves.
I do not think that a 400W servo would be adequate even with with the reduction, at least for the Y axis. I'd have to recommend a 750W servo. If a 750W servo is right for the Y axis then
its probably also correct for the X axis. The X axis is lighter, but as you have seen the axis mass is not the biggest determinant of inertia ratio, that being dominated by ballscrew, and
that is what leads to my recommendation that the X axis servo be 750W also.

The Z axis inertia ratio:
J_screw=1.13 x (0.0125)²/2
=0.9x10^-4 kg.m²

J_axis=22.7 (0.005)²/39.48
=0.144x10^-4kg.m²

Lets assume a 400w servo of 0.2x10^-4kg.m²

Inertia Ratio (Z axis)= (0.9 + 0.144)/0.2
=5.22

This is easily in the range of a 400W servo, again this assumes no reduction, and the situation only improves with reduction.

The mechanical advantage of a 5mm pitch 25mm diameter screw is (25 x pi)/5=15.7
For 1Nm applied torque the thrust is 1000/12.5 x 15.7=1256N of 125kg force.

Thus even a 400W servo would easily hold the axis aloft using only a small proportion of it's available torque to do so. It would be a good idea to have a servo with a brake though,
so when the power is cut off the weight of the Z axis does not back drive the screw and the Z axis fall towards the table. So 400W should be ample, but if buying two 75W servos for
X and Y then one more for the Z is no trouble. Often there is little price difference between a 400W and 750W servo, so I would go bigger. If at some stage you increase the spindle mass
say then you will still have the Z axis authority to accommodate it.

I'm also talking to teknics to see if they have a servo solution in the same price range. Interested to see what they recommend.

I do not like Clearpath servos, they are too expensive for what they are. Over $500 for a 400w servo. They have only one digital output, a sub par encoder and just too bloody expensive.
Compare that to a 750W Delta B2 for $438, with a 160,000 count/rev encoder, 6 digital outputs, 8 digital inputs,2 analog outputs, auxiliary encoder output, 230VAC line input, ie no
power supply required.......twice the power and vastly better IO for less money. Tecknics do have an excellent reputation for quality and support, of that there is no question.
They also follow the forums and hound me every time I post this opinion.

Craig