[Mariss Freimanis]

The G380 board gerbers went out for fast-turn prototype boards. I generated a rendering of what the finished drive will look like using those boards. It looked cool so I thought I'd share the picture.

Mariss

[guru_florida]

I'm curious. What EDA software do you use? I'm going to take a guess at Altium DXP?

[the4thseal]

so when do you think we can get this?

[Mariss Freimanis]

Considering the CPLD is a Xylinx CoolRunnerII I'm using Xylinx's ISE development software. For everything else I use ACAD2000 (schematics, board layout, 3D modeling, etc.).

Mariss

I'm curious. What EDA software do you use? I'm going to take a guess at Altium DXP?

[TOTALLYRC]

It looks like another winner from the stable of Gecko drive.

At one point, IIRC, you were considering offering an analog version of this drive.
If so, would it be sooner rather than later and can the drive be set to no pid since the pid is done in my dspmc/ip controller. I don't know if having pid inside pid would pose any problems.

Mike

[Mariss Freimanis]

Here is a scope picture of how the G380 reacts to an encoder failure or disconnect. The motor is turning at 3,000 RPM when the encoder CH_A (yellow trace) is disconnected.

The motor is dynamically braked at the CURRENT LIMIT set value, 12A in this case (green trace). Note the 'flat' current at 12A. The motor brakes to a stop 0.08 seconds later. The G380 FAULT LED lights immediately on encoder disconnect and stays lit until the encoder connection is restored.

Mariss

[TOTALLYRC]

What happens at very slow speed or at rest and the encoder is either disconnected or dies?

I know that perfect encoder failure detection not that easy, but it is nice too see that you are working on it. It also doesn't hurt to dream.

Mike

[fatal-exception]

Just a suggestion, and I'm probably not the only one who thinks this, but why do you use such small terminal blocks? They are a PITA to work with 'in the real world'.

I doubt size of the unit is a big consideration for most people who buy your products.

Just my opinion.

Paul

[Mariss Freimanis]

The encoder fail detect senses illegal voltages on the encoder inputs (1V > illegal > 3V) using window comparators on both encoder channels. This is both a static as well as dynamic measurement so it applies from zero to any RPM. The reaction time is nearly instantaneous (< 2uS) so there is no delay from fault to shutdown.

The comparator levels are set to the minimum acceptable encoder input voltages. The secondary advantage of the circuit is it will reject sub-standard encoders from the get-go instead of allowing them to operate and produce faulty results.

Mariss

[MrWild]

The terminals are fine with the right screwdriver. I'm curious how the drive knows how much reverse current to give in order to stop the motor. Does it sense back emf? How can it do that while the drive is pumping juice to the motor?

[Mariss Freimanis]

Paul,

Size has an effect on price. Printed circuit boards price-out by the square inch, metal by the pound, inventory stock by the cubic foot and shipping by weight. The connector is the single most expensive item in the drive; larger means even more expensive.

Mariss

[Mariss Freimanis]

MrWild,

Any error (following error, encoder fail, short-circuit, over-temperature, under-voltage, disable input, etc.) results in the motor leads being shorted together by the bottom pair of power MOSFETs. This 'shorting' (called dynamic braking) causes the motor to act as a generator and quickly dissipate kinetic energy. Motor current still passes through the current sense resistors and the current limit circuit (set by the LIMIT trimpot) regulates the current to not exceed that value. This refinement has the added benefit of controlling motor current during a hard-stop (running into a brick wall) and preventing over-current failure of the drive.

Mariss

[TOTALLYRC]

The encoder fail detect senses illegal voltages on the encoder inputs (1V > illegal > 3V) using window comparators on both encoder channels. This is both a static as well as dynamic measurement so it applies from zero to any RPM. The reaction time is nearly instantaneous (< 2uS) so there is no delay from fault to shutdown.

The comparator levels are set to the minimum acceptable encoder input voltages. The secondary advantage of the circuit is it will reject sub-standard encoders from the get-go instead of allowing them to operate and produce faulty results.

Mariss

I will definitely be watch for these to come out as I have had problems with some of those infamous heds modules. They have now all died or been removed and replaced.
When they went bad, they would get stuck in the middle at about 2-3 volts IIRC and would not turn on and off like they were supposed to.

This circuit should save you a bunch of tech support calls and prevent a lot of broken machines.

Time to plan a conversion to put them on.

Mike

[David Campbell]

I am curious to how many servo motors will actually require the derivative action.

My day job has nothing to do with motion control (I build training simulators for chemical plant operators) and rarely do I see the derivative enabled on PID controllers (eg: temperature control loops). The only case where derivative is typically used is where there is some dead time present.

I am struggling to see where dead time could be present in a typical CNC system unless there is some backlash (eg: a gear head arrangement)

I think I might of just answered my own question...

David Campbell