TMC428 with Allegro driver chip

[krunal_299]

Hi!
I am going to experimenting with one of the allegro DMOS chip as stepper dirver with TMC 428 as controller.

But allegro chip requires 19 bits datagram via SPI interface.
So with TMC428 , is it possible to send 19 bits datatgram(per dirver) if 3 drivers are in the chain?

[Andrew]

In general, the format of data must be the same. Not just the communication interface.
As far I remember, a TMC428 chip cannot change its format. So, it can not change the quantity of data either.

You would need to drive your driver from you microprocessor directly. So you would need to devise a whole custom motion control FirmWare.

Or devise an intermediate processor, which takes data from TMC428, and converts into something else.
For example, this is done internally in a TMCM302 module. Its Step/Direction interface converts data from SPI output of TMC428.

[krunal_299]

I think it is possible to send 19 bit long datagram(per driver) to the driver chain of 3,
because 428 has cover datagram register(JDX=%0011) of 24 bit length.

Any suggestion?


Wait.....
I think I should read 428 application note (A3972)before having further query.

[krunal_299]

After going through the application note of A3972 ,
I have clear idea on how to manipulate A3972, but still confused with

cover datagram register

issue for transmission of WORD1 in A3972.

If more than 1 driver are in the chain than how can the Word1 transmission (via Cover datagram register) of successive driver be performed?


More precisely i would like to say: For WORD0 transmission there is NXm bit to salect the next driver in the chain but how can i select the next motor for WORD1 transmission by using Cover Datagram Register?

[krunal_299]

Hi to all!
I hsvr read the AN again but yet not cleared this stuff
So please let it get cleared to me so i can advance my project .
Hope to have favorable response at the earliest.

[stewi]

Hi folks,
I'm interested in the same IC combination.
According to this article http://www.trinamic.com/tmc/media/Downl ... be20267ce6
recommended driver chips were TMC288 or TMC289, which I couldn't find in the product list. Also, the target market was meant for 42V application.
TMC drivers go up to 34V, but the Allegro can do 50V motor voltage

[Andrew]

I guess nobody answers this post, because all is either already answered, or is already located in documentation.

I gave a general answer. If Krunal is not satisfied with it, then perhaps one has to dig documentation by himself.

I don't know the exact answer. I am sure I could give him an answer. But in order to do that I would have to dig documentation - just what he himself can do as well.
I prefer him do that.
He needs to dig not only documentation of TMC428, but also that of A3972, which he mentions.

My opinion remains the same. TMC428 over SPI makes sense as long as data formats coincide between TMC428 and A3972. The number of bits is not the only consideration. The formats must be the same. Is that clear?
A cover datagram has nothing to do with stepper motion. It is just a configuration for the driver. If you want to use it for stepper motion, then how do you suppress TMC428? You can suppress TMC428 by never instructing any motion. But then why use TMC428 at all??? When you instruct a motion, it sends its own standard datagram, which you cannot change.
That's the situation as far as I remember.
And why am I writing this? It took me a few moments to look at documentation, and remember some things. You could do the same. Krunal has the chip TMC428. He simply could do some tests quickly. But no - he goes on a prolonged question spree.
If you want to use TMC428 as a mere SPI pass-through, then I see no point in using TMC428 at all. Connect the microcontroller to your driver directly over SPI. If you use your own microcontroller directly with A3972, then you have no products from Trinamic. Then there's nothing to ask us about. Unless you want to order some consultancy service from us. But, as I mention below, you are not prepared to pay. You only use us as a free university.

Perhaps someone from Trinamic also saw his post (I am not at Trinamic). If they felt a need to answer, they would have done so. If they did not, perhaps their opinion concurs with mine.

Regarding the 50V, which you mention, I suggest you use TMCM035. It has the required rating. And you spare yourself lots of time on design.

Currently Krunal has wasted lots of our time, and his design does not look to be ready any time soon.

If he wants us to design his system, then the system is already there. It is a TMCM110, or TMCM303, or TMCM301 with TMCM035. Simply take it, and use it right away.

If he wants to design a system himself, then he is the designer. Not us.

I do not deny though that Krunal has spotted a few bugs in the products or in documentation, if I remember well.
I think Trinamic is grateful for that.

But the habit of asking without learning by himself takes the better of him. He has done much more of such rampant asking.

I have one particular customer, who invested 1 year into design. And he asked me not a single question. There were questions only before he bought the chips. I gave him a good starting guidance. Afterwards - no questions.
I myself also have not asked a single question regarding my design.
On the contrary. My customer gave me one, or two suggestions, or bug reports. I myself have produced many bug reports, and suggestions.
Does this tell you anything?

I have other customers. Some of them do ask too many questions. But also they buy products from me.
They ask far less that Krunal does. The rest they learn by themselves. Those, who can not be designers, simply ask me to do the design for them, instead of asking questions how to do it.
Even when they ask me something, they pay me.
You usually pay to a catalogue distributor, who takes most of the money, and does not answer any questions, and even does not care.
Then you ask for our support, but never pay me. You do not order from me anything. You just ask how to do the job. Then you take the profit (whatever that is). You rather run to that catalogue, and buy from them more.
Perhaps Trinamic gets some of your money. But not much. You paid them for a bit of hardware. But not for the sytem design, or system integration, which you are asking for.
My business is system integration. So, I know it warrants an extra pay. Not a bundle. By the way, that is why most of other Trinamic's partners are absent from this forum. - Because it is free.
This my rant also explains why Trinamic themselves do not rush to answer some questions. Some questions is not their business.
We do it gratis.

Regarding the TMC288. This chip does not exist. You must note that the document you are referring to is made in year 2001. Maybe this chip existed then. But I think this is just a draft document of a future product, which it was back then.
Please download the documentation for TMC428. There you find the correct diagram. The main drivers are TMC239/TMC249/TMC236/TMC246.
You can complement the TMC249/TMC239 dirvers with external 4A/30V transistor bridges TMC34.
Again, if you want 50V, take TMCM035

To be moved to the thread for rants.

[stewi]

Hi Andrew,
I get your point, but you don't have to yell at us. Wir sind ja nicht schwerhörig.
Indeed for myself, I was too lazy to read through the specs. None of the Allegro A3972, A3955 or A3992 neither ST L6228 will talk the same language.
The reason why I was even considering to go stray with other vendors was a cost ratio. If I require just a 0.5 A more or increase the supply voltage for more juice out of the motors, I'll have to add multiple MOSFET instead of using more powerful drivers.
The beauty of the TMC428 is in its capability to drive more than one axis. Hands down, I can't find any other IC, which can do that. Let me look at it in more detail to see, which solution could be more cost effective. It may come back to me like a boomerang, if my colleagues refuse to write the firmware controlling multiple axis from one microcontroller.

[Andrew]

Who is yelling?..
Look at Krunal... Look at me...
I have written 279 messages.
He has written 69 messages.
He wrote but 4 times less tham me.
That's not little. But most of his posts is meaningless. Only a few are useful.

(Of course judging by the amount of text, it might be 100 times less. But that's another story.)

Regarding higher current or voltage. I agree: non standard rating may make the cost higher. But not necessarily so.
More of my answer is here:
http://www.trinamic.com/ttdg/viewtopic.php?t=547

[TRINAMIC_BD]

To give some comment:
Yes, you can control the A3972 or other SPI drivers with the TMC428. One special thing about the TMC428 is, that it can freely configure the data it sends to the driver ICs! So, it fits nearly all ICs available on the market, it can even control step-/direction drivers.

Even "old" parallel control Ericsson (now NJRC) like NJM3771 drivers can be controlled by the TMC428, just add an inexpensive external shift register like 74HC595. You find information on how-to-do on the TMC428 product page. This also was realized on an early TMC428 evaluation board.
If you need a higher voltage solution, TRINAMIC can offer the TMC239 or TMC249 with external MOSFETs, like SI4559, which gives you for example 48V / 2A driver capability, full diagnostics and protection.
Please refer to the driver datasheet and / or FAQ for complete schematics.

Concerning the sending of a cover datagram: Some drivers require this for configuration purpose. The TMC428 thus can "cover" part of the datagram with specific data to do this job. Please see TMC428 documentation for more info.

[stewi]

Thanks for clarification. It's all there, I only have to read it:
http://www.trinamic.com/tmc/media/Downl ... 2_v100.pdf
I must admit with !%#^* $$$ against Euro this is a valid option. Not so sure about the required Schottky diodes either. At least from the A3972 datasheet they wouldn't be required.
Nevertheless, this is your website and I'll politely quit promoting other components.

[Andrew]

As I noted here:
http://www.trinamic.com/ttdg/viewtopic.php?p=2747#2747
Perhaps it is not a good idea to stick Trinamic into every pothole.

Trinamic products are good at some things. They are poor at other.
I myself have discovered many bugs, and shortcomings... Some I reported on this forum.
Trinamic products are far from ideal.
That's life.

Not only bugs... But there are also some design flaws.
Consider for example the ludicrous limitation on the range of positions. Why is it 3 byte, rather than 4 or 8?

Another limitation is the microstep wave table. For some mysterious reason it is only a quarter of the wave, rather than the whole period. From this may stem a certain irregularity of microsteps. Also this places many restrictions on the user, who wants more control.

Again, that's life. You can circumvent it, at least.

But you should not proceed with circumvention regardless of costs to your sanity.

It is quite all-right to criticize products of Trinamic here, or adver-tize competitors here. It would at least apply pressure on Trinamic. Without that, they would get fat and lazy.
If they spot malicious adver-tisement, they would remove it anyway.

On the other hand, suggestions are sometimes falling on deaf ears.
Long ago I suggested they make a 6 phase controller... Similar to TMC453. They did not listen, and made a new chip TMC454 with the same 5 phases as before.

Was it so difficult to add 1 phase?.. to make 6...
No.
And they could serve one more market.

Now, if I ever need 6 phases (such motors do exist), I would have to do something similar to what you hoped to do here - adapt somehow the existing Trinamic.

In my case I would hope to adapt a 3 axis controller to 6 phases. Namely, a TMCM303 already has 6 outputs (2 coils per driver TMC246 of an axis). But this is quite cumbersome, especially with some of the mentioned, and not mentioned limitations.

Perhaps it is easier simply to go for another make.