[Xerxes]

Greetings everyone!

I'm starting to design the next drive that should be available in some point 2008.

Target specs in short:
-Supply voltage range 12-80 Vdc
-Output current 10A cont, 20A peak
-Discrete MOSFET power stage
-Fourth power output pin for 2-phase stepping motors
-Card-like construction that can be installed in control box wall without intermediate wiring from drive to back panel
-Sub-D connectors for motor windings and encoder
-Short circuit & overtemp protections
-99% based on VSD-A firmware but adds couple of new features

It is not replacing VSD-A but is merely an addition to drive selection (a low voltage version with premium stepper support).

I'm asking you folks whether you have some ideas to be added/changed in this drive compared to VSD-A :-) It would be very helpful if you described your "dream drive" for your particular needs.

Thanks,
Tero

[Eson]

How about a FPGA drive with open source software? You could support both 2- and 3-phase servos/steppers and all kinds of encoders.

[Xerxes]

Eson, thank you for sharing the thoughs. FPGA is an interesting approach but is not required to meet these requirements. The drive will support AC/BLDC/DC/Stepping motors and differential & single ended encoders.

Anyone interested about 3-phase stepper support? It could be added but 3ph steppers are quite rare so I don't know if its worth it.

[Episs]

how about modular design by using separate power stage from the board then thous who have small motors could use cheaper power stage than for higher power motors + could support more motor types by changing power stage and firmware. same thing could be with encoder feed back, you could support also sinusoidal encoders .
+ your main board could be multiaxis.

[Eson]

I belive support for more encoder types are one of the most wanted.
Time will tell.

[Xerxes]

Modularized encoder receiver actually has been in my plans for some distant future drive. It could be a small PCB that is connected to main board.

I have been thinking modular drive a lot. However, some compromises would need to be made compared to a monolithic design. Alternatively we could make a drive with only logic section so desired kind of power stage could be supplied by customer. However, that could generate large amounts of support request if inexperienced people would start building power stages themselves. There is nothing wrong in support requests if there is enought time to handle them

[Al_The_Man]

-Sub-D connectors for motor windings and encoder

Is that wise to spec in a sub D connector for motor windings? especially for 10a+.
I wouldn't trust a D connector at that current.
Al.

[Xerxes]

Is that wise to spec in a sub D connector for motor windings? especially for 10a+.
I wouldn't trust a D connector at that current.
Al.

I'm planning to use D25 connector so one phase gets 6 pins in parallel. Typical connectors have specs of 300V and 3A per pin, so in parallel it would equal 18A cont per phase.

[Al_The_Man]

I am just questioning the logic as to what you are saving over say a single screw terminal connection, this would mean terminating 18 wires into the D connector for each motor?
At some point I assume the motor itself will have 3 wires or connections, so each 6 conductors will have to terminate to a single connection, or am I missing something?
Al.

[Xerxes]

I like D connector because it has mechanism to secure cable position by screws and connector nuts can be used to mount card on enclosure wall (just like PC cards are connected to L-shaped sheet metal part).

You could still use the same low pole count cable and just solder it to multiple pins at connector end.

Also hybrid D-sub connectors exist for high currents but those are harder to find and are quite expensive, too.

[PaulH]

Why not use something like an Anderson Power Products connector? The PowerMod series is rated for 30A per pin, can only be plugged in one way, and can be circuit-board mounted.

http://www.andersonpower.com/product...onnectors.html

[Xerxes]

Paul, that kind of connectors seem to be quite good alternative. However, these too are non-standard and harder to find for user. I have put that kind of connectors in my CNC drive box and they're quite good indeed. But I still missing the possibility to mount card on enclosure with these.

[PaulH]

Non-standard? Hard to find? The company has been in business for over 100 years. They have distributors all over the world. If you have ever seen the plugs on a golf cart or an electric forklift, chances are they are from Anderson.

In your country, Arrow Nordic is the distributor.

[Xerxes]

Ok, they're probably are common enough
But still they are almost nonexistent compared to D-sub which is a non-manufacturer specific standard and available everywhere.

[PaulH]

Agreed, but you're asking a small connector to carry an awful lot of current. Even if you wire six pins in parallel, it's going to come back to bite you later. I won't argue any more. Do what you think is best and let us know how it works out.

[Al_The_Man]

I applaud your efforts, but the title of the post was 'Describe your dream drive', to me having to make 18 motor connections is more of a nightmare.
I agree that for encoders, pre-wired sockets are maybe an asset because of the amount of connections etc.
But in the case of high current motor connections where there is only three connections (or four if gnd included), what is wrong with a screw terminal, or any of the other choices that are out there in the way of higher current connectors?
If the object is to make it easier for the amateur CNC builder, I don't see how this avenue will be easier or avoid the chance of mistakes.
It is not like these items are going to be removed and replaced on a regular basis, requiring quick disconnect.
I know what I would prefer to install.
Hopefully accepted as constructive criticism.
Al.

[Xerxes]

Criticism is very wellcome

I know it sounds bit odd to carry large currents in D-sub connector but there should be nothing wrong in it. I think it is very easy to solder thick wire in multiple pins if it's done wisely. It's probably as easy as soldering thin signal wires in single pins.

Or maybe I'm going in wrong tracks here and maybe my brain just takes some time to change it's opinion about it However, fortunately choosing wrong will not lead to catastrophic consequences (both will work). If I choose inpractical connector, I can change it in the next revision. Or maybe we could even publish both at the same time and see which one is more popular.

Attached: half-designed layout of the new VSD.

[Xerxes]

Oh, one more benefit for D-sub: it is very easy to find molded and shielded D25 cables for couple of bucks. Those are meant to be data cables but probably they would be suitable for small motors. Motor cable shielding is important in PWM output drives.

[jahonen]

One thing came on my mind: How is the ripple rating of the bridge filter capacitor, does it handle the full ripple current of the bridge? If you have 10A continuous output on the stepper motor, and the current is assumed to have no ripple on the motor, then the filter capacitor will see worst case 14 A ripple current (45 degrees point on the sine table). This can be reduced somewhat if the PWM timebase would be in different phase in each coil, but I guess that it is difficult due to the sync requirement of the ADC sampling to the PWM timebase. But it would still require 10 A of ripple.

The exact ripple value will depend how much ripple current is handled by the feeding power supply, but I think that it would be EMC-wise to handle the ripple locally, instead of spreading it on the power supply feed.

Regards,
Janne

[Xerxes]

Janne, you're right that capacitor is has the role of absorbing ripple. I have selected high ripple current type of capacitor.

However, the problem is worst only in non-recirculating drives where capacitor sees non-stop current pumping. In recirculating drive the capacitor sees ripple only during short periods of time so heating is much reduced. In VSD-A, capacitor has ripple rating of about 1600 mA but in my test capacitor has got only barely warm.