New build noob questions

hi jr three BTN8982TA is a very interesting driver. I read the specs and must admit it meets the requirements we need. Hope you will have success in testing them and please report asap here . Thank you.

Hi all,

A brief update on my testing. I received two BTN8982TA-based evaluation boards 2 nights ago and quickly proceeded to burn them both I suspect my bench power supply that was inadvertently set to too low current and quickly folded back to under voltage condition. As you probably know, removing power input from MOSFET switches while they're active is a no-no.

I will have more amperage next time. For now, I ordered some spare BTN8982TA chips which seem to be on short supply at the moment.

BTW, I'm using SMC3 with an Arduino board for my testing. SMC3 is an awesome tool! The Arduino code for it is also very good, though it does some funky things to use just one PWM line in an H bridge. I haven't dug too deeply into that but I think that means only one side of the H bridge switches and at some point the high switch of that side does the switching. N-switches typically have less switching losses so I've always preferred to switch the low sides. The Atmel chip does have plenty of PWM channels so I'm not sure what was the issue with using 2 pairs of those.

Regards,
JR

Sorry to hear about that. I hope you succeed with the BTN chips the next time

Hello all,

This is turning into a doom-gloom series of small failures but I figured it would be interesting to someone else working through the same issues. Hopefully it doesn't discourage others from trying.

Here we have the latest in a string of 3 BTN8982TA-based eval boards that I have managed to kill. This latest one was a bit more dramatic. I setup my eval board with two 12V power supplies in series (more on this below) for a +24V rail. These are serious power supplies capable of dishing out 75A each for 24 * 75 = 1,800 Watts out. The same setup but at 12V seemed to work ok but rather slow with my 24V motors. The DC motor controllers are fully spec'd up to 18V with support up to 40V input. The output of the controller was clamped to 30V via fast diodes and TVS.

I ran SMC3 with the Arduino code and, as the French excite: voila, la flambe!



This was no small smoke or current trip. There was a loud pop followed by a 2 inch (50mm) plasma shoot and a 4 second flame that extinguished itself out.

First, this part is obviously not as well protected as Infineon claims it to be. Second, and most interesting, in all 3 cases, it has been half-bridge #1 that has gone out. I can't claim the code in the Arduino is causing these flameouts with some sort of cross-switching, just that it's interesting that it always is switch 1 that goes. Coincidentally, that's the one the code drives on PWM. I do have spare switches now, I'll see if the traces on that board can be soldered back in.

Throughout all of the fireworks, the power supplies remained steady. I'm using a pair of HP DPS-1200FB server units. You can find these monsters on eBay for about $40 a pair. I floated the secondary side (and only the secondary side) of one and connected them serially. This worked to get me 24V at a ~600mV ripple. Since I wanted better than this and also prevent the regen from coming back into the power supply, I added a diode bridge and a large capacitor. This 3-phase bridge I got from Amazon has only ~400mV drop across it and a built-in heatsink.

As for the mechanical aspect, I'm working on both rotating and linear positioners. The rotating one come with a substantial motor - maybe too substantial for my controller - and a 170 RPM gearbox output. The linear one is a Motion Corp industrial actuator that I'm modifying to make quicker. I have a couple of DC servo motors coming in for those, we'll see how they do.

Regards,
JR

The BTN Horror Picture Show...WTH has gone wrong? Which SMC3 code did you use? Perhaps you should have ask @RufusDufus before you t/o this gadget...I did not study the white paper, anyway any differences to the older BTN chips cannot be ruled out.

I am VERY interested in what becomes of this, as I was about to order a pair of those boards. I hope they aren't too good to be true, they seem like a very positive alternative to the Sabertooths. (IIRC these boards have current sensing, which the Sabertooths do not expose. I could be wrong, but I seem to remember the Infineon chips had that).

I'm wondering if the code were modified to offer dead time between reversals if it would still fry the chip? (ie. perhaps a back EMF issue instead?)

Hi all and sorry for the delay in responding. The SMC3 code is v0.7. There might not be anything wrong with the code. After all, the Infineon chips have built-in drivers that make it impossible to shoot-through regardless of logic input commands. It might simply be that my damn wheelchair motors are way too big for the H-bridge, with too much inductance and BEMF.

I fixed one of the (least) burnt boards and will be trying again tomorrow. I've also been working on the actuators and have this so far:

I ordered a worm and gear last night to mount the pot on the back shaft of the motor. Not ideal but the 20:1 ratio of the worm will give me about 5" (127mm) of travel. If it need more, a higher ratio worm can do that. No load speeds at the moment, running on 24V and 2000 RPM is ~10" (254mm) per second. Those are not the original motors that came with the actuator. They are 5-1/2" (140mm) long, brushed servos from eBay, approx $40 US each. The question remains if they'll have the torque to be useful.

As I work through both rotating and linear actuators, I'm finding the latter seem less costly, simpler and better for this application. Because there are rotating forces involved in linear actuators - unless they are designed otherwise - there will be a small loss of quick/short movements while the spherical eye mounts spools up the slack. This can translate is loss of minute vibrations, etc which is something we want to feel. No such losses with rotating actuators. Because they typically travel 180 degrees total, they might not require gearing on feedback pots either.

Regards,
JR

noob questions,
what does a linear actuator look like internally?
What does a rotating actuator look like internally?
I know I have ball screw type on mine, so would that be linear or rotating?
Yes I actually dont know, hence I am asking.

Hey BSFT,

These things vary quite a bit. Linear actuators have screws and nuts with a tube attached to the nut that you can see outside the structure. The best ones are so-called "ball screws" where they have ball bearings in the nuts and can be very efficient (>90%) at transferring rotating forces from a motor to linear motion.

Here's one:

They again vary quite a bit. Some have planetary gears (a "sun" gear in the middle with "planet" ones around it), straight gears or worm gears. You can tell if you have a worm gear because you can't (easily) turn the output shaft and have the motor turn. The gearboxes I use come from wheelchairs. Many of these are massive and the gearboxes are very robust.

Here's a couple of examples:

Typically, only linear actuators have ballscrew drives. If you have wiper type motors and gears, that's a rotating drive. I suppose you can call it linear too because in the end, it's linear motion that moves your sim seat.

It might be hard to make out the details but here's a sample of a wiper motor that uses worm gear:




Regards,
JR

ok, thanks for the clarification, I have linear actuators then.
Although rotating design look to me like a fatter version of a worm gear motor gearbox