Electronic Circuits and Projects Forum

What a coincidence - I just setup my CNC with the same chip about a month ago. I do notice a ~10 - 15KHz or so high pitched whine, but I never bothered looking into it. I used to work with a relatively high end desktop CNC machine a couple years back, and remember a similar sound so I never bothered to look into it.

I don't have the particular timing values, but I'll dig them up later.

James

Is this whining detrimental to the motors, or the chip?

The whine is mostly just an annoyance and shouldn't be detrimental to the motor. I have run into similar issues while using the A3984 and found a few ways to reduce the problem.

The first thing is to decrease the fixed off time to the minimum possible while still meeting your current ripple requirements. This could increase the power dissipated by the chip itself due to an increased switching rate.

The other thing I did was increase the voltage to the motors. Increasing voltage decreases the time it takes for the current to ramp up through the inductance of the motor. This effectively increases your switching frequency and if you raise it enough, it will become inaudible.

Thanks Phalanx.
I tried cranking up the voltage from 24V to 35V, while keeping the logic voltage at 5V.

I was using a variable power supply. As I was cranking the supply up from 24V to 35V, I could hear the frequency changing, I believe it was getting higher. But over all, it did not do much for me.

"Decrease the Off time"
Is that done by the resistor//Cap. I believe Allegro calls it as the RC1/RC2 inputs?
Allegro specs a 30Kohm & a 1nF cap.
I tried lowering this to 20Kohm... Maybe I should go the other way?
Like towards 40K?

I saw a demo schematic by a company called "E-LAB".
They have 36Kohm//1nF for the RC# inputs.

It's driving me nuts here.
All 4 motors are just humming away....LOUD.

Basically, you are hearing the pwm frequency. (ok, pseudo-PWM frequency). Decreasing off time Toff increases the frequency and can push it above the hearing threshold. so what if it annoys the dogs?

Increasing the voltage slightly decreases the time it takes for the current to reach the Itrip level. you still have a fixed off time. thus the slight effect on the frequency of the noise.

The resistor and cap don't form an RC oscillator. Toff = RC, according to Allegro so decrease it. the cap is also used for the comparator blanking (to prevent false shut off due to current bounce) so I wouldn't change that.

I would not run those chips at 35V. While it is at the upper limit of the spec, I've heard of people blowing the chip at that voltage. best to stay around 30V.

So you think I should try lowering those two resistors?
The two found at Pins 10 & 15 of the 44PLCC version?

I should put in some 10Kohm and see what it does.

Thanks for all of your help!!

Originally Posted by philba

I would not run those chips at 35V. While it is at the upper limit of the spec, I've heard of people blowing the chip at that voltage. best to stay around 30V.

I've heard of that happening as well but have never seen it first hand. It seems to involve running motors at 35V with a current at or near the limit of the chip. I think it has something to do with inductive kickback and the associated heat being dissipated by the chip. If you are running at high voltage and current and have a relatively large winding inductance, I would suggest putting schottkys around the motor.


Matilda...

If you are not using external diodes, make sure that synchronous rectification is enabled on your chip. Otherwise make sure you have schottkys around the motor.

I have attached a little photo of my board, showing only one of the 4 channels.
I'm using these isolated 5v:5v regulators, and I made the footprint wrong in orcad. Ignore that mistake! Corrected for REV;B board

Anyway, I added the 8 "Ultra Fast Avalanche Sinterglass Diodes" to the output/motor leads.
I also have the SR line pulled LOW(Active Mode-Sync Rec Occurs).

I also kept the grounds isolated.
As in,
To the A3977 I have 5V & 24Volts going to the chip.
Pins 44,1,2 22,23,24 33,34,35 are connected to the 24V supply ground.
Pins 11,12,13 are connected to the 5V supply.
With the chip off the board, these two ground paths are isolated.
The chip connects them internally.

That was a tip from Allegro. It really helped minimize noise seen on the 5v and 24V supply lines while the motor was running.

the external diodes need to be faster than the internal ones or they are useless. the idea is to have the chip dissipate less energy. I'd hope schottkys would work.

everything I've seen says that allegro was a bit optimistic in spec'ing the upper limits on that chip.

I've not mucked with Toff but since the app note has the formula Toff = RC, I'd say that decreasing R should reduce off time.