Allegro A3979 board

[veryevil]

Hey, I've been looking to build a 3-Axis CNC router for ages now and ive got myself some samples of the Allegro A3979 (see http://www.allegromicro.com/en/Produ..._Numbers/3979/ ) and was wondering if anyone had a circuit schematic and/or a pcb layout designed for this chip.

I have seen a couple of threads that mention his new chip and people saying they are working on designing things but there doesn't seem to be any working examples on the internet.

If you've got anything you would like to share that would be great.

Thanks Steven Richardson

P.S. Great forum look forward to sharing my CNC build!

[epineh]

Interesting chip, I have played around with the A3986 and working with those TSSOP chips isn't much fun at all, unless you are set up for it. You would have to be careful with the layout as the chip drives the stepper directly, they are tiny legs on the chip.

Have you looked at the A3977? Much larger package size and similiar specs, Phil has PCB's already done and boards available :

http://pminmo.com/ss3977/ss3977.htm

Good luck !

Russell.

[veryevil]

hey, ive got the equipment to deal with ttsop as im at uni and they have it all. Also i already have the A3979s

Thanks

[BCwanderer]

Try Looking on Ebay
There are 3979 chip boards already in production
They are not expensive and it would be worth buying a premade board to see how a proven design works before you invest a lot of time and setup by rolling your own.

Here is a board with a A3977 chip Tssop which is very similar
http://www.stepmasternc.com/stepmasternc_006.htm

Another alternative is to download the circuit of a A3977 at Tero's site
Then delete the A3977 and place the A3979 tssop chip in it's place
The chips are quite similar. You will then have to rebuild the connections for the tssop chip.
It is not identical but it is very similar.
http://www.students.tut.fi/~kontkant/a3977.php.html

Another interesting chip is the A3982
It has a 24-pin SOIC package package which is easier to solder

Their newest chip
A4983
DMOS Microstepping Driver with Translator
has very few extra connections.
There is a working unit for sale on ebay, very small,

Originally Posted by veryevil Hey, I've been looking to build a 3-Axis CNC router for ages now and ive got myself some samples of the Allegro A3979 (see http://www.allegromicro.com/en/Produ..._Numbers/3979/ ) and was wondering if anyone had a circuit schematic and/or a pcb layout designed for this chip. I have seen a couple of threads that mention his new chip and people saying they are working on designing things but there doesn't seem to be any working examples on the internet. If you've got anything you would like to share that would be great. Thanks Steven Richardson P.S. Great forum look forward to sharing my CNC build!

[pminmo]

My experience with the 3979 is it's tough to get it well enough heat sinked to get near the maximum current. I'm sure it's possible...........

[mechatronicsman]

Originally Posted by veryevil I have seen a couple of threads that mention his new chip and people saying they are working on designing things but there doesn't seem to be any working examples on the internet. If you've got anything you would like to share that would be great.

Hi veryevil,
late answer to you post, but in case you are still looking; I have just finished my CNC conversion, and for that project I built a stepper driver board using the A3979s.

On my first version of the board the focus was getting it to work, not layout issues wrt heat transfer. The board worked well (unfortunately I have lost the layout files so they are not available anywhere), and my machine is now running, see http://hobbymechatronics.com/mechani...cnc-conversion. But the A3979s get very hot, due to the layout not being optimal for low thermal resistance.

As I lost my previous layout files, I have made a new layout from scratch, and this time focusing on lowest possible thermal resistance from the junction of the A3979 to the surroundings to avoid using heat sinks. See http://hobbymechatronics.com/electro...g-motor-driver. This layout should cope with 1,5-2Amps without heat sink. The boards are not back from production yet, but I will post back an update on thermal performance when the boards have been successfully tested.

[ssureshtronics]

Dear members, Pnimino, and all experts.
I tried with A3979 to drive my 2A NEMA 34 stepper motor with 24V. All the waveforms at the OUT1A,OUT1B,OUT2A,OUT2B are perfect as per the datasheet. The Rs i used was 0.15E 2Watts and i fixed Vref = 3V as a Reference voltage. The RC1 and RC2 values are Rt = 12K, Ct = 470pf. The PFD pin voltage was 2V. All the HOME output and STEP i/p waveforms are fine when motor was not connected. The MS1 and MS2 was in HALF step mode (I tried with all mode also). The motor makes hiss noise when it was connected but not running. At very low frequency of STEP i/p (500 Hz), the motor was moving slowly but the IC getting abnormal heat. when i tried with increasing frequency the motor stalls and making noise. Here with i need some clarifications from u experts.
1. Is it possible to run the 2A/coil motor at high speed.
2. what will be the maximum speed achieved using A3979.
3. Is it normal that the IC getting over heat?
4. Any body worked with this IC and achieved the fine (speed) running, if yes what is the motor current and at what voltage.

I expect your valuable replies, i seen many post regarding A3979 but no one specified the motor current and RPM.

Thanks

[pminmo]

"The motor makes hiss noise when it was connected but not running. At very low frequency of STEP i/p (500 Hz)"

Normal for choppers, stopped is typically the worst, but they will also at real slow rpm's.

"when i tried with increasing frequency the motor stalls and making noise."

What your describing is mechanical resonance. It's worst at full step, and improves as you decrease step size. All stepper motors need an acceleration and deceleration step stream. Adjust the acceleration and deceleration rates by starting slow and increasing to tune the motor/electronics to the mechanical mechanism.

1. Is it possible to run the 2A/coil motor at high speed.

It's spec'd to handle 2A, but getting the heat from the part via the thermal pad IMHO is troublesome. To run 2A you need a lot of copper and design methodology to handle the thermal issue. I don't believe it's practicle with 1 or 2 oz copper. However if you cut out the pcb material under the pad so you could effectively solder a heat sink assembly created to mate the pad then you would have a better chance.

2. what will be the maximum speed achieved using A3979.

No way to say because speed is dependent on many factors.

3. Is it normal that the IC getting over heat?

YES, see above.

4. Any body worked with this IC and achieved the fine (speed) running, if yes what is the motor current and at what voltage.

I worked with it and ditched it because of the thermal issue.

[MechanoMan]

If you'll use the forum's Search function, you'll find an EXTENSIVE thread on the A3986:
http://www.cnczone.com/forums/showthread.php?t=24699

It is NOT good for CNC work. Its high ratio of Blanking to Toff creates a minimum duty cycle which makes it IMPOSSIBLE to resolve microsteps with a high source voltage. In fact it can barely resolve the fullsteps at zero speed! That high source voltage is essential to running a stepper quickly.

Without proper resolve of microsteps, the drive current and stepping motion is all distorted and prone to resonance and stalling.

There. I just saved you 51 pages of reading.

The A3979 is essentially the same problem.

BTW, it is only of marginal use to heatsink the back of a plastic-case chip, because the plastic insulates the die. These chips have a thermal pad underneath that is directly connected to to the die. Copper's thermal conductivity is crazy high, and when connected DIRECTLY to the die, even though a PCB copper fill is not finned like a heatsink and gets hotter than a heatsink, it'll effectively keep the die temp down.

You need a thick trace running underneath the chip and a wide area on the PCB on both sides, fanning out to something like 1/2 sq in of copper. Which generally means electrical connections cannot be routed under the chip on the topside. The thicker copper-foil PCB weight is better. If you can get better than standard 1oz copper, that'd be great.

[pminmo]

Originally Posted by MechanoMan If you'll use the forum's Search function, you'll find an EXTENSIVE thread on the A3986: http://www.cnczone.com/forums/showthread.php?t=24699 It is NOT good for CNC work. Its high ratio of Blanking to Toff creates a minimum duty cycle which makes it IMPOSSIBLE to resolve microsteps with a high source voltage. In fact it can barely resolve the fullsteps at zero speed! That high source voltage is essential to running a stepper quickly. The A3979 is essentially the same problem.

The 3979 doesn't suffer from the same current control issue as the 3986.

[MechanoMan]

Well, let's look at the spec sheet:

Taking typical values, A3986's min duty is 0.068. A3979 is 0.024. OK, that IS better.

If you were to run a Keling 282 on it with a 1.2ohm coil at 2A with a 35v supply on 1/4 steps, that'd require a minimum of 0.026 duty. OK, on paper, it does seem to work. Assuming you don't use 1/16 microstepping, which is probably not productive anyways.

Staying within a 2A and 35v limit may not yield a speedy driver. But numbers do suggest it'll work.

[pminmo]

Duty cycle of blamking to on time isn't the issue. Motor coils aren't a resistor. There are plenty of 20khz fixed frequency choppers that run approximately the same ratio that are outstanding performers. Read up on decay modes, then think about how that relates to rising and falling currents, then you will understand half of the issue with the 3986.