[ardenum2]

Hi Ardenum,
I use this development board:

https://www.digikey.co.nz/en/product...BoC_DoQAvD_BwE

Lots to recommend it. The micro is optimised for rotating machinery and has all the periphials that you can dream of and more, including really tightly timed complementary PWM with programmable deadzone, sequential PWM
signals for polyphase systems, cycle by cycle current and voltage limits, dual encoder inputs, plus all the usual timers,analogue inputs and all the rest. The board has a galvanaically isolated D100 JTag. If you use the
onboard JTag then you get free unlimited use of CodeComposerStudio, Texas Instruments own IDE including the Insta-Spin software libraries. You can use this development board right up to production. In my case
I just drop it into my drive complete, saves having to write a bootloader. All-in-all Texas Instruments have tried to squeeze everything in that you might ever want, all for free or at least cheap to encourage your use of
Texas Instruments silicon. This particular IC (TMS320F28069M) is about $20USD each (1000 unit rate), so hardly expensive.

Craig

In other words, if you were to work on a linear drive, this board has all that's needed, hardware wise, only needs to be programmed for a specific use? I'm hoping to handle the magnetics part of the linear servo, it's been my long term dream to work on motors. I'm planning to get a base model from SMJ, give it a test with an off the shelf drive and then compare that to the linear motor I'm working on, empirically.

Excitement guaranteed. Results not so much. There's plenty of things to consider in regards to eg. cogging issues like skewing magnets or destructive interference(which sounds awesome) etc. An important part would be a way to measure cogging.

Getting off the shelf drives will cost at least $1500 for 3 axis, I'd be more than interested to divert that to your R&D and end up with something open source that anyone can use, a $80 drive sounds a lot better than a $500. You think you could end up with something functional?

[peteeng]

Hi Ard - You corrected your "funtional" - I like funtional vs functional... fun and function are a good combo. Peter

[joeavaerage]

Hi Ardenum,

In other words, if you were to work on a linear drive, this board has all that's needed, hardware wise, only needs to be programmed for a specific use?

Sorry to burst your bubble but ....NO.

The development board has the micro controller and some support circuitry so you can experiment/design, but it can in no way drive a motor as is. The board is 3.3V and has
20mA outputs, it will never drive a motor on its own. Texas Instruments do some support boards but they are limited to a few amps and 85VDC. You want tens of amps (peak)
at 600VDC, and that will require you design and build your own switching hardware.

My servo drive runs off single phase 240VAC. It has automatic power factor correction and so the DCLink voltage, with the PFC boost, is 400VDC. I have a two stage brake resistor,
the first stage kicks in at 420VDC and the second stage at 440VDC. I have used 600V MOSFETs throughout. I've used two in the PFC boost circuit, two for the two stage brake
and of course the six required for the main inverter, ie ten 600VDC MOSFETs, and these are high value items. That does not account for the snubber circuit components for each and every one of them.
I have used LEM Hall sensors for current feedback, although in the five-six years since I first designed it a line of galvanaiclly isolated resistive shunts with Delta-Sigma modulation have come on the market
and may save some dollars there. I have used ISO5452 galvanaically isolated MOSFET drivers, although they are probably getting pretty dated, but they sure do make it easy to signal high-side MOSFETs.

If you want a DCLink voltage of 600VDC then you are going to need either some real smart high voltage MOSFETs (at a cost) OR 1200VDC IGBTs. IGBTs are great but slow, at least by comparison
to MOSFETs. For all IGBTs advantages they are still saturating minority carrier devices and you must pay attention to second breakdown and be especially careful to account for recombination time
in your switching, just like any BJT design from years ago. All-in-all IGBT designs are very useful but take time to develop, test and prove. Reliability of IGBT designs is good BUT ONLY if you have
been thorough in development.

If I were designing a drive the first working prototype would cost thousands by the time its done. Thereafter you might be able to make them cheaply but would still be woefully under developed software-wise.
You no doubt have seen all the programming parameters of a modern AC servo? It would take thousands of hours of software development to even match that, let alone exceed it.

I would say if you can get off the shelf drives for $500 then take them!. The cost is not about the parts and materials of the drive,or even the profit to the manufacturer, the main cost is the development budget
that has accrued while the manufacturer was designing it....that is what you are paying for. The manufacturer is hoping to sell enough of them to re-coup his development cost...and good luck to them.

Having, partially at least, designed my own AC servo drive, I can tell you the effort absorbed easily 500 hours of my time, and the design is nowhere like finished. Since that time (six years ago)
I've sold my home, relocated closer to the city, changed my job, bought a business, built a new mill with a total investment approaching $30,000 NZD (20,000USD).

As I have said 'I pick my battles' and agreeing to design and build a linear servo drive is not a battle I want to engage in. It would cost a fortune and absorb more time than I could devote ot it.
If you want help, yes I do have some ideas to get you started.......but this is your project, not mine. I have entirely enough on my plate as it stands.

Craig

[joeavaerage]

Hi,
just to give you a feel for where the costs are going to go...... my design required 10 MOSFETs and my preferred device is NTHL110N6553S3F (650V,35A) at $11.65NZD each at 10 off quantities
and to drive them I elected to use an ISO5452DW IC at $11.15NZD each. These prices do not include 15% GST, a local NZ tax.

So just the 10 MOSFETs and their drive ICs will cost $228NZD. ($139USD)

That does not count the snubber components, the PCB, the power supply components, the current feedback devices or anything else.

You'd be more than welcome to search around for cheaper MOSFETs or IGBTs and you could change the signalling so that you only need the ISO5452's on the highside MOSFETS.....
but then you'd still have to transformer signal the low side MOSFETS, so its a wash. And what about the power supply DCLink capacitors? You get good ones or they go BANG.....
and they cost a bomb. You'd get cheap 1000uf Chinese 400V ones for $20 each but if you want good ones you can bank on $50 each or more, a lot more if you want real high quality units.

Craig

[ardenum2]

Well that clears up a lot of things.

[joeavaerage]

Hi Ardenum,
it sure does, designing and building these sorts of drives is far from cheap and easy. If it were cheap and easy everyone would do it and no one would buy off the shelf, and yet
99.999% of drives are off the shelf.

The only reason that I elected to make my own was because my donor servo has a resolver rather than an encoder, and given its power level the only drive I could find was more than $3000USD.
I thought I could do better than that. I certainly have not spent $3000USD but I don't know how many hours I've put in.....many hundreds at the very least. Given that this is/was a hobby project
that sort of time barrier does not put me off, but it would be impossible to justify the time invested on a commercial project, unless there were thousands of sales....not just one!

Craig

[ardenum2]

wait this has me confused, what has the servo's voltage capability got to do with the motors? Isn't a motors power source and the drives a separate line? I only got limited experience with arduino and the board is always isolated from whatever the motor or the hardware part needs, two seperate power lines and the board only outputs pwm to the servo but the power doesn't go through the board to the motor. is that not the case with cnc systems?

[pippin88]

Those tiny servos used with radio control / Arduino contain a driver board in the motor casing. Thus only need power supply and a speed control signal (PWM).
Bigger servos rarely have the driver integrated (there are some E.g. jmc integrated servo - it's just a driver stuck onto the side of the motor).

Technically a "servo motor is defined as an electric motor that allows for precise control of angular or linear position, speed, and torque"

Usually this involves an encoder (or resolver or similar) to feedback position.
Any electric motor can be a servo with a encoder feeding back to a drive / control.

There are specific types of motors that are generally used for what we tend to call servos in CNC use.

Usually system consists of:
Motion controller
Servo drive (take in position command / torque command etc, and spit out needed volts and amps to the servo motor)
Servo motor

[joeavaerage]

Hi,

wait this has me confused, what has the servo's voltage capability got to do with the motors? Isn't a motors power source and the drives a separate line? I only got limited experience with arduino and the board is always isolated from whatever the motor or the hardware part needs, two seperate power lines and the board only outputs pwm to the servo but the power doesn't go through the board to the motor. is that not the case with cnc systems?

A CNC servo, like my 750W Deltas for example, requires at a minimum a power supply that can deliver 750W, and more like 2250W in short bursts.

My CNC motion control board (ESS) is 5V 25mA output, and it in turn is connected to a 24V 5mA output breakout board. Clearly the breakout board has nothing like the energy required to drive
a servo. What it can do is control a servo drive, and the servo drive provides the energy required to the servo motor. So you are correct there are multiple power supplies throughout the system, a 5V 0.5A supply
for the ESS, a 24V 2A supply for the breakout board, an isolated 24V 0.5A flyback inverter within the servo drive for control purposes, and of course the main power supply for the motor, in my case 320VDC 7A (peak).

The servo motor is a '200V Class motor'. That is to say that it requires about 200VAC to have it develop its full power.This would be a natural choice for a machine with a single phase 230VAC input.
There are many servo motors out there of the '400V Class'. They require something like 400VAC to drive them to full output. They are a natural choice for a machine with a 400VAC three phase input.

There are many servos with much lower voltage requirements as well. DMM have a line of servos that operate at around 70V peak or about 50VAC. The servo drives for these servos require a separate
high current power supply of around 60VDC.

The same principle applies though....the servo drive provides the energy to the servo motor, while the motion control/breakout board provides only low energy signalling to control it all. It for this
reason that they are sometimes called 'servo amps'. In the early days they were indeed just transconductance amplifiers, but in the modern era with AC servos they are not strictly amplifiers at all,
nonetheless are still called 'amps' sometimes.

I seem to recall, whether it was in this thread, or your own thread, that you claimed the linear servos that you were interested in buying require a 600V drive? It is for that reason that I adopted 600VDC
as the target DCLink voltage. If you are designing your own magnetics then that 600V is hardly sacrosanct. Having said that I suspect that for practical purposes you will require high inductance
motor windings (large number of turns of a fine wire) which will in turn require a very high voltage drive. A DCLink voltage of 600VDC is a natural fit for a 400V (line to line) three phase input.

Certain Miller inverter welders had a boost circuit such that the DCLink voltage was 800VDC, and that was obtained with virtually any power input. For instance if you had 208VAC single
phase input the boost circuit would take that to 800VDC. A three phase 230VAC (line to line) input would likewise be boosted to 800VDC.A 400VAC (line to line) input would also be boosted to 800VDC.
They were great machines, and top quality US made, when US made meant high quality, but man did they go BANG when they let go! Additionally 800VDC is right in the 'almost certainly
fatal zone' with electrocution. The danger band for human fatality is in the region of 800VDC to 1500VDC. These voltages are seldom used in domestic and even commercial equipment but are used
widely in traction systems, trains, trams, cranes, lifts and in this case an inverter welder! You want to be absolutely certain before you delve into this area typically called 'medium voltage', you will
not survive a mistake.

Craig

[DavidHe1359]

Hi All and Sundry - I've pinched parts from P2P so its config is complete. So now I think I'll look at the lifting gantry again and see if there are any inspirational moments. Peter

this machine is a good machine ,suitable for machine small parts.

[Lesteraction]

You’d be in relaxing mode while using ScreenPad Plus. Its 14-inch additional matte touchscreen adds extra efficiency for fast users. Now, you can have a new and amazing Blender experience. You can also count it as the best laptop for 3D projects.

[DavidHe1359]

Fullmida Automation Equipments Co. LTD.

David He
http://www.aliexpress.com/store/908764
Mobilephone: +86 13902952931 Tel/Fax: +86 760 88853349
wechat: +86 13902952931
[email protected]
whatsapp: +8613902952931

[ardenum2]

You no doubt have seen all the programming parameters of a modern AC servo? It would take thousands of hours of software development to even match that, let alone exceed it

The way I understand it, commercial drives are like windows, designed to work with a variety of hardware variations. In this case, I thought we would provide a set of fixed 'profiles' that are preconfigured to work with specific forcer sizes that people can choose from. And only for these forcer/motor parameters, instead of something that can be used with any other motor. You are right though, someone would have to configure all the parameters initially in the firmware, someone as in you probably. I tried my hand at programming and let me tell you you'd have a better chance teaching a chimp how to code than teaching me.

These voltages are seldom used in domestic and even commercial equipment but are used
widely in traction systems, trains, trams, cranes, lifts and in this case an inverter welder! You want to be absolutely certain before you delve into this area typically called 'medium voltage', you will not survive a mistake

I'm hoping to put all the electronic schematics into an ecad and have it analyzed before doing anything in the physical world. The electrical part I have to pay a certified professional to handle.

[peteeng]

Hi All - still waiting for the 3350. 3 days delivery they said, now its 3 weeks.... Hmppp... So I have been playing with the Veronoi generator in fusion. The free one, Its a bit clunky and unpredictable. It does not place the pattern in the profile selected which is poor and it can make the pattern the wrong scale..... I suppose its free... I have made the moulds for my grout laps and will seal them today and pour soon.... may stick to isogrids and diagrids... Peter

[peteeng]

Hi All you lot - The 3350 arrived today so I can get onto casting the coupons. I have waxed the mould so onward and upward. The pain is waiting 28 days to send it to the lab!! Peter

[ardenum2]

Hi Pete. was looking forward for a long time to see 3350 in action. How much do your lab guys charge for the testing?

[peteeng]

Hi Ard - This lab does not do commercial testing (I'll explain soon) The lab I use for commercial testing was the University of Southern Qld at Toowoomba. It would be about $800AUD to do the test at USQ. But they closed their lab a couple of years ago and that has created many issues with various projects I have heard. I stay out of that sort of loop these days. I have been dealing with a company for over 30 years and they have an internal lab. Over the years I have sent them considerable business and I usually have interesting projects. So they are happy to get one of their young engineers to do some testing for their experience, if I can wait until a suitable gap shows up in the workload. So it's free to me... Peter

[ardenum2]

Hi Ard - This lab does not do commercial testing (I'll explain soon) The lab I use for commercial testing was the University of Southern Qld at Toowoomba. It would be about $800AUD to do the test at USQ. But they closed their lab a couple of years ago and that has created many issues with various projects I have heard. I stay out of that sort of loop these days. I have been dealing with a company for over 30 years and they have an internal lab. Over the years I have sent them considerable business and I usually have interesting projects. So they are happy to get one of their young engineers to do some testing for their experience, if I can wait until a suitable gap shows up in the workload. So it's free to me... Peter

In the coming months I'll be building a test rig, a linear stage and getting basic components to make sure I can even handle it, among that I'll get 50kg epument from rampf and 50kgs of ducretes uhpc. If you're interested I can mail some samples to you, I'd need to know what shape the lab requires them to be. Would be interesting to know how they compare to 3350 and lanko.

[joeavaerage]

Hi Ardenum,

In this case, I thought we would provide a set of fixed 'profiles' that are preconfigured to work with specific forcer sizes that people can choose from

Yes, I could see that working. I rather suspect that customers would demand this feature or that feature all of which are broadly similar amongst all servo drives. I would think
if we made our own linear drive we would always be playing catch up to existing designs, which in turn means continuing software development over years.

My own design servo drive is a case in point. I have in mind a specific set of functions that it needs to accomplish. I also have the advantage of being able to code using the C source code
whereas any commercial product can only be programmed using parameters. If I decide I need another feature or modify a behaviour I can program the micro controller directly from my PC
using the built-in JTAG.......which is not an option for commercial equipment.

I'm hoping to put all the electronic schematics into an ecad and have it analyzed before doing anything in the physical world.

It is certainly advisable, essential really, to run any and all designs through some sort of software....but that does not in itself prove a design. High voltage, high frequency switching
designs are still as much 'art' as they are science. Things like snubbers for the IGBTs or MOSFETs are designed using a fairly loose set of principles and rules but actually have to be built
and stress tested before you could say they are acceptable. It is very much true that you can do much of the design in software, but there is still the costly reality of building and testing it.
I can't think of how many times I've designed something, built it, only to have it blow up in smoke after a few minutes due to things like second breakdown, or hot spotting or a snubber
dissipating much more energy that you imagined it would. Existing servo drives are usually the result of years of development and is far from easy to get a working and reliable design out-of-the-box.

I might for instance be able to make a bare bones linear drive to suit specific hardware within short order, a month or so. Trying to convert that bare bones design to a commercial offering, and even then
still in the 'experimental hobby class' would take more like a year.

I don't want to be discouraging but I do want to be clear eyed about any design challenge I might take on. Its not impossible to design and build your own servo drive, but personal experience is that
its very time consuming, especially if you are designing for a commercial product. The development time is huge. As consumers we tend not to notice or even be aware of all the development that
has gone into a product.....until or unless you try to manufacture that product yourself.

Craig

[ardenum2]

I'm fully aware that $500 drive I found probably cost millions in R&D to get to where it is. That is probably why drive prices are high, same with axial flux motors, drives go for $5000. The more advanced the tech the harder to control it. My biggest misconception is also probably that there is a major difference of making the linear motor work and making it work good enough for machining.

by the way I wasn't thinking of selling them. I wanted to open source everything so that anyone can make them, both the hardware and the gerber files. forcers and stators made with sheet metal cutting primarily(fiber/laser/waterjet) and ideally a board that you can just upload the gerber to pcbway and have them fully assemble it. At the very least I'll try to make the first step with the forcer/stator and maybe someone smarter will pick it up in the future to further improve it.