PlutoP compatible Quad servo control with integrated H-Bridge (L298 or equivalent)

[nullbert]

Hello,

I'm currently an EMC2 user on 3 separate machines using simple step/direction drives.

After doing some research I (like others) became excited about the possibilities of an inexpensive closed-loop servo drive system for my mill using the Pluto-P board and H-Bridges.

Having read the other threads on building a high-power servo drive system using discrete mosfets (Large servo drives on the cheap...) I've decided I'd like to tackle a more modest goal of designing a board that integrates cleanly with the following features:

* Clean integration with the Pluto-P board
* 4 H Bridge drives using one L298 per channel (or other integrated H Bridge)
* Encoder inputs
* Limit switch inputs
* On-board power supply (bridge rectifier, capacitor) - optional

I was initially considering a single L298 per channel, which would offer 4A @ 46V- potentially enough to run my mini-mill.

However, I started browsing mouser for some similar parts - looking for integrated H-Bridges like the L298, possibly newer-tech. I came across a candidate that seems very exciting:

VNH2SP30-E

http://www.st.com/stonline/books/pdf/docs/10832.pdf

It's a fully integrated H-Bridge driver designed for automotive applications.
It's rated at 41V / 30A, uses mosfets internally, with a scant 19mOhm Rds On. at around $9 a piece it seems like a bargain.

Take a look at figure 32 in the datasheet. With the high level of integration it seems like a more sophisticated, next gen version of the L298, with built-in current limiting, thermal-limiting- even includes flyback diodes- CMOS/TTL inputs, everything!

At first glance the chip should be compatible with the pluto-P outputs- I believe all of the equivalent enable/pwm channels map to this device.

My primary concern is the voltage ratings for this device- it shows the absolute maximum Vcc at +41V, but the max operating supply voltage at +16v. Seems like a large discrepancy- possibly there to conform to automotive environments. I do want to know if this device would work reliably near its absolute maximum voltage.

Would like to hear the experts weigh in on this- especially on the viability of the VNH2SP30-E.

In any case, parts are on order for an L298 and VNH2SP30-E configuration. I'll follow-up with more information as I proceed.

Thanks,
Brent

[skullworks]

You should take another look at the data sheet for the L298N. Specifically look under applications (bottom of page 7, top of pg 8.)

For continuous DC operation you need to de-rate the chip to 3-3.5A and parallel both H-bridges. Note that while the chip is a dual H-bridge, the motor supply (Vs) is all fed from just Pin 4. PWM is not straight DC but during rapid moves it will pull full current at near 99% voltage.

I have seen a few home brew PCB's that used 3x L298 in parallel and they worked quite well. There was still the issue of current limiting... The work around in that config was to have power supplies which maxed out just under the peak current.

[nullbert]

Thanks for the feedback. Yes, the idea is to dedicate an L298 per servomotor channel by paralleling the driver channels of each L298's.

A friend of mine designed a stepper motor driver with 3 L298's in parallel. In that project, a major concern was properly balancing load across the L298's- even to the point where he was trying to create matched sets of L298's.

If I go with L298's on this project, I would probably limit it to a pair of L298's per channel.


Another device similar to the L298, but just one H Bridge- offers slightly higher current capacity (5A Peak)
http://www.st.com/stonline/products/...1373/l6203.pdf

[skullworks]

Nice thing about the L298 is it does not have the transient diodes internal - so you can gang them up and still use a single set of 4 diodes.

The L6203 has internal diodes and better switching so would be a good choice for a sub 5A application.

You might be able to pair them up if you added external diodes?

I have been exploring using the L6506 with the L298's ganged for current limiting, but have not gotten around to ordering test samples yet. Other projects have been more pressing.

[epineh]

Nice find on the VNH2SP30-E ... there is certainly some confusion about the supply voltage, if it was a max of 41V as it first states then this would be very cool, I am running 24V servo's at 30V at the moment and they have a lot of power for a hobby size machine, this chip would run my servo's all day long (so long as the VCC is 41volts not 16volts )

The other chip I was looking at was the LMD18200, but it is still just a little low on the current side...

Russell.

[epineh]

I have sent a request to ST to clarify the datasheet, we will see what they come back with, it does look like it is a 16V part though, Digikey lists it as such.

Cheers.

Russell.

[nullbert]

Just as I was committing my mouser shopping cart on the VNH2SP30-E I saw in the datasheet it has a over-voltage shutdown at 16V. Nuts!! It would've been a nice one-chip solution.

For the power levels I want I'm starting to lean towards a discrete H-Bridge using mosfets. I'm looking at this Intersil driver chip to integrate some of the H-Bridge driving circuitry to preserve most of the convenience of the L298.

http://www.intersil.com/data/fn/fn3658.pdf

I'm sampling two sets of mosfets:
http://www.fairchildsemi.com/ds/FD%2FFDB045AN08A0.pdf
and
http://www.necel.com/nesdis/image/D16852EJ1V0DS00.pdf

The NEC part has a tantalizingly low RdS on, but a lower voltage than I would like. Plus the gate capacitance is rather high. Since they both share the same pinout, the parts could be swapped for each other.

The mosfets's aren't the cheapest, but with their low on resistance one could save the expense of heat sinks and/or forced air cooling.

Realizing that I'm somewhat replicating existing work in this forum, I want this effort to be more tailored as a low-cost, single board Pluto-P compatible servo control for 4 axis small mill setup, with all of the standard input/outputs handled (at least the ones I care about. ) Ideally, I'd like the Pluto to plug into this board without jumper cables to make it extra clean.

I'll post some schematics for review soon.

[epineh]

Well I got a reply from ST...quote :

"There are a number of parts in this range, differnect technologies. The VNH2SP30 is in MO4. Vcc Abs Max rating is 41 Volts. them means the the device can be subject to 41 V and it will not be damaged.

However the device is programmed only to opperate up to the Opperation voltage and then will shut down. In this case Min is 16V.

If you need higher voltage considera different device. eg VNH3SP30 if possible.

Best regards,

Trevor Pike"

I looked up the other part he listed and it is 30 amp continuous and 36 volt supply, but only 10KHz PWM frequency close but no cigar, but it may be worth sifting through to see if they make one with the same volt/amp rating with 20 KHz PWM freq.

As for FET's that is a whole world of choices, I use IRFZ44's for my low voltage bridges (50volt 50 amp) they are about 60c each, IRF540's are an old favourite for about the same price, and I have ordered some IRF3710's that are a whopping 90c each, 100volt, 57 amp with 23mOhm Rds on that I want to try out on a BLDC drive if I ever get around to it, would also be good for the Pluto/EMC/H Bridge config, I will be using IR2110's for convienience, I can get them here easily and cheaply, but the Intersil chips are nice as well, I was looking at the 4081, but it is a little harder for me to get.

Cheers.

Russell.

[nullbert]

Thanks for checking with them. Too bad that part only PWMs to 10KHz, what were they thinking? Not sure how loud that would be with a typical servo motor, but don't want to find out and require earplugs while running the machine.

I'm enamored by the ultra-low RdS on mosfets out there. I do want something that would at least not require heat-sinking- at least for the currents I'll be running, and I don't mind paying a bit more for them.

Most, if not all of the D2Pak devices seem to have the same pinouts, and likewise with the T0xx devices- so the board design could have spots for both and accomodate a variety of devices.