How the reset occurs you will have to measure.

I was asking for the layout to know if there is a gnd wire directly from the ATm board to the driver board or if the ground is routed only through the black wires seen in the foto.

I would connect a power supply to the IR2110. The from there I would run a power supply to the Atmega. The ground wire of theis power supply will also serve as the return path for the PWM signal from Atmega to driver.

Is it connected like I described?

 

Almoast, i've got 2m power supply cable to the table, form there it splits up and connects to headers on each board. Cables are not more than 20cm long from where they split. But every board is also grounded trough the bolts to the aluminium frame that is grounded to the household ground. The power supply and the battery pack negative pole are grounded to households power supply also.
I don't have a scope so i'll have to use ambiguous methods to test this.
i'll connect some voltmeters on the circuit. It is strange that reset happened while the motor ran.

So you don't suggest any other means of protecting the chip from the IR2110's.

One more thing, i noticed that IR2110 are couple of amps rated, and i'm using 7812 which is 1A for 12V supply. Could this be an issue? i can pull 12V from the battery directly if necessary.

Regards

 

It is strange that reset happened while the motor ran.
Regards

No it's not strange. Actually it's exactly when the motor runs you will encounter EMI related problems.
I say it's a layout issue.

One more thing, i noticed that IR2110 are couple of amps rated, and i'm using 7812 which is 1A for 12V supply. Could this be an issue? i can pull 12V from the battery directly if necessary.

Regards

Rated a couple of Amps? Where do you see that? You mean maximum output current? That would be peak current to charge/discharge the gates.

7812 is more than enough. The peak current comes from the bypass and bootstrap caps.

 

Havent added these schottkys yet. So before i make another board what would you change regarding the layout?

 

I'd connect the driver and Atmega board the way I described above.

To analyze the reset we need to have a closer look at the atmega PCB.

 

Hi, sorry for bothering you, it was false alarm.

It is working fine. Thanks again for helping me make this work.

Any experience with debouncing rotary encoders ? i'm using these mechanical encoders for the testing http://hacker.instanet.net/Kits/Sensor/RotaryEncoder.jpg as this is closed loop control.
100nF across the pins?

 

This is a rotary switch, right? A rotary encoder is a device to measure angular position, either absolute or incremental.

The best way to debounce them depends on where and how they are mounted in the circuit but yes, typically a capacitor or a capacitor resistor combination will do the job.

 

Any experience with debouncing rotary encoders ? i'm using these mechanical encoders for the testing http://hacker.instanet.net/Kits/Sensor/RotaryEncoder.jpg as this is closed loop control.
100nF across the pins?

For low res encoder, why not use a slotted opto sensor, in conjunction with a slotted disc, you can also get dual types to set up quadrature style if needed.
Keep it all electronic!
Max.

 

It's incremental rotary encoder, graycode output, and push switch. picture is taken from the side where switch pins are visible only, there are 3 pins on the other side for encoder output.

I will switch to double slotted optosensor(i'm only testing with these), something like this you meant? http://datasheet.octopart.com/EE-SX1031-Omron-datasheet-71659.pdf
Hey these are transistor receiver, so i will not need additional electronics to connect them to controller?

I'm still thinking how to "produce" the slotted consistent slotted disc. This is DIY project, i don't have machining tools available and the shaft end is M8 so ones from the old ball mouse are out of the option.
And do the sensors need to be shielded from ambient light?

These mechanical are not even bouncing at these rpm,(530 rpm for maximum no load feed rate of 270 in/min) but i think they will wear off quickly in everyday use.

 

That is what i had in mind.
They are TX & RX in one and they are open collector so they are easily interfaced.
They are IR so not affected by ambient light.
Max.

 

Thanks,

But, there is an issue of making a slotted disc, and how to design it. The size of the slots and the size of the non slotted area. To create QE output in respect to dual photo-diode sensor.

I'm guessing there are some rules of thumb online i will Google that, i don't want to bother you with such details. I've registered here to solve electrical problems beyond my capabilities.

I'll test the circuit's on the loaded shaft, and then make boards for all three axis and then start coding closed loop control.

Since i will be working on this for few more weeks, don't be surprised if i ask more questions and for more help.

Many thanks to everyone for all the help so far.

Regards

 

The spacing between the two optical axis is 7mm so it would appear that the slot size would be a minimum of the distance required to block or clear both axis simultaneously.
This could easily be tested by a simple material such as card etc with a small number of slots.
A permanent solution could be a clear plastic disc, it could be sprayed black with the slots masked out during paint or spray procedure.
Max.

 

The encoder disc patterns are available on Google. Such as - http://groups.csail.mit.edu/mac/users/pmitros/encoder/ Or - https://www.google.com/search?q=opt...la:en-US:official&client=firefox-a&channel=sb

 

By my reckoning, you can use what the link terms 'Regular' for non quadrature types to register in a quadrature fashion with the dual IR sensor if the spacing is slightly larger than the sense distance between the two IR paths.
Max.

 

Thanks guys,

It never crossed my mind to paint the thin acrylic with specific pattern, i always taught about drilling(or milling) the non transparent plastic.

You helped me a lot.

I'm testing developing the closed loop control with these mechanical ones a the moment. I will switch to making the optical ones when i make some progress with this.

Regards

 

Hi,

First, you have to have a decoupling 10uF ceramic cap on Vcc of U1 and U2.

Second, it would be good to use a 2.2R resistor in series with D2, D3, D5 and D6 to be able to control the fall time of the MOSFETs. It also limits the current through the turn-off diode within the diode's peak current rating.

 

Hi,

First, you have to have a decoupling 10uF ceramic cap on Vcc of U1 and U2.

Second, it would be good to use a 2.2R resistor in series with D2, D3, D5 and D6 to be able to control the fall time of the MOSFETs. It also limits the current through the turn-off diode within the diode's peak current rating.

Thanks Mik, you joined a bit late, but better ever than never.

I did solve an issue of burning diodes on H-bridge, we shifted discussion to the self made rotary encoder discs. However what do you mean by 10uF cap on Vcc of U1 and U2? what are U1 and U2? i didn't use them on my or any other schematic i posted here.

Also guys if you don't mind i would like to post one generic question here.

So this project i'm working on is a hobby closed loop cnc controller with all accompanying electronics. Reason i'm working on closed loop is that i can use very cheap an powerful DC motors. Also the entire electronics part of the project will cost me less than 100$.
On the way i'm developing microelectronic(atmega) software.
It seems like a shame to develop and perfect all this and not share it with other hobbyists. Do you think this project has the perspective, or is it worth sharing and making an open source? And are there many similar projects posted online already?

 

Do you know about CNCZone.com - http://www.cnczone.com/ ?

No disrespect meant by the following statement. Personally I don't understand why you and others that post on this forum, are trying at this late date in home cnc, to reinvent the wheel. Using a micro controller as an interface to motion control of the machine seems like just one more level of complication that isn't needed. You still need a PC and CAM software to do the control. There are so many fairly inexpensive motor drivers available that are 'plug and play', and proven to work. Starting a new type this late in home brew cnc just seems like an up hill battle. Again no disrespect meant to you or anyone else.

 

I would imagine that in Croatia it is possible there may not be a great source of DC motor drives available even on the used market?
If using the free downloadable front end such as Mach3 or Linux, an intelligent drive is still required to close the PID loop.
Max.

 

No offences here, i wanted honest opinion.

However it seems i haven't been detailed enough in explaining:
The controller receives, interprets and executes g-code send from PC(including control), and reports back telemetry just FYI. So it's not interface, it's complete cnc controller.

I went for brushed DC motors because they are about 10 times cheaper per wat of power than stepper motors. Eg. (250W cordless drill DC motor costs 10$). And i need high feed rates, meaning power in the whole range. Closed loop with some encoder is the only way of control for such motor. So while i'm working on it, i'm wondering if it's worth sharing. That's all.

I appreciate the cnc zone link.