AMC1280USB Motion Simulator Controller

Here is another connection diagram showing the important connections for use with sabertooth 2x60 drives:

Each sabertooth should have its own ground wire to the amc1280usb, to avoid creating ground loops.

Another video that may help test the 12bit analog outputs on the AMC1280USB barebone and help set it up correctly for use with Sabertooth drives:

Hi @Thanos , I cannot seem to get the PID to work like it should . I can wind the Proportional all the way up to 95 and while tracking improves 'slightly' it never starts to overshot like it should. If I add any Integral at all, even just 1, it starts to oscillate . Any ideas?

As I was trying to find out if it was a problem with the Sabertooths, decided to spend the morning rewiring in my old 1280, which I know work well with my old MM drivers.

So it works perfectly , with PID values and response that makes sense and now all my detail is back with rumble on the ripple strips which I just could not get on the new 1280 without turning everything up were it was just picking up allot of unwanted noise as well. I'm happy that I now have a working sim, but I'm quite certain that you have a problem with the PID setup in this new 1280.

I'll check it tomorrow in a small fast linear actuator I have. I know I didn't quantize the PID calculations for 12 bit resolution, so I assume it's 4 times less than it should. I'll test it and make the PID calculations dynamic for each resolution.

Maybe it's also time to upgrade the resolution of PID to more than 255 values...

Thanks
Thanos

I have a very different PID that the other AMC ( green )
for this one I set 40 .0.0
if I change I and D , it is not works correcly.

Because the 12bit resolution still has 8bit pid quantization... if you increase the Integral term it will kick the motor disproportionately as of the wrong quantity of set of values for normalization. As i said ill be adding a dynamic set patch to it.

Really? You send out your product without a manual and needing firmware updates to work properly too? Have you considered how much of peoples time you have squandered because of this? If you were a business you probably wouldn't be in business long because your customers would get ticked off and look for an alternative or just go back to what they were using before. Oh yeah, they already have.

ps I just couldn't restrain myself.

I was in (successful) sales management for over 35 years. Quality / Service-Support / Cost................................ANY one of those items that are not 100%, will cost you:

1st - Loss of business 2nd - LOSS of your Business.................

You could not have said it better BlazinH - From ALL of US here at the forum, thanks for watching out for us - AND letting Thanos that we are not happy customers!

Tom

Yes, luckly I'm retired . Spent allot of time trying to work out what was wrong and finally worked out it was the PID calculation . People nowadays are so easily offended and don't take criticism well, so I've tried to be a bit more mellow and only respond low key and were I have to .

Hi Tom,
I think you were waiting for this...



I hope it's easy to follow, I had space only for two VFD inverters on the drawing, but I think it's easy to figure out the connections for the third VFD inverter. If you need connection diagram for the position sensors too, remind me what sensors you are using...

Thanks
Thanos

I'm sorry you feel this way.

Apparently you have no idea of the complexity of the internal functions of the Amc1280usb firmware. Making the pid code work with multiple size inputs and outputs while keeping the same dynamic response on the motor is not easy. And I don't have many motor setups to test these every time I touch the code of the firmware . All development relies mostly to feedback from the users, which results to things being fixed.

Let me guess, simtools 2.0 just appeared from thin air, all perfect with everything ...I don't think so. It gets feedback and reports on bugs and it gets improved a little bit at a time.

Everything works like that now, have you noticed how many weekly updates for apps and system you get on your smartphones and computers? Not to mention Simulation Games, always need patches eve now and then... they are not perfect, the whole world is their beta tester all the time... it's in the fine print...

I found the issue, as I thought it's the different size output which is much bigger for the 12bit. I increased the PID levels 4 times and seems to respond better now. Tomorrow morning I'll apply the fix in all six PID subroutines and post the new revision for you to try.

Thanks
Thanos

Unfortunately I'm now on the old 1280, which is working well and therefore I have no reason to change it out. I'm sure there are other users out there who can test it.

@Thanos

Wrong.

Disagree

Agree but only after alpha testing. This is still in alpha as you acknowledge issues that you knew of but didn't even attempt to add the code needed to fix it. A 255 pid count? Come on.

Of course not. Its author included all the features and code that were necessary to perform its intended functions then thoroughly tested it himself first. Then he asked for volunteers to beta test it and gave them a profession copy for free for their efforts. Most issues were due to the fact that it had to run on multiple platforms where yours does not.

Again, you're not even to the point of beta yet as far as I'm concerned.

B.S. I know you're trying to be cute but show me where you informed anyone that this was largely untested and would need to be troubleshooted by those that bought it first before it could be relied on and safe to use.

One reason to change it out may be the 255 resolution pid. SMC3 is 1024 and fully functional and thoroughly tested already.

Maybe, but I have already spent too much time on this, its now working as I expect, so I'm happy and will be moving on to other things/projects . My goto test is Assetto Corsa GT86 at Magione. If it can give me a expected smooth ride with a good realistic rumble only when I go over the ripple strips then I am happy (FI this is something that I could not do on my old sim and wormdrive motors). My Wife has a GT86, so I am quite familiar with its ride and how it feels on a track. Sorry a bit off topic, but it makes me think back to when I first built a sim. You love that it just shakes you around all the time like a amusement ride, but after a while this novelty wears off and you strive for realism.

I don't need to repeat, but like most of the values you set on the LCD menu, PID settings are percentages as well not absolute 255 values. They are adjusted internally for the the range and the resolution of each input/output setting. When I say to increase the value is to increase the entry points for the LCD menu for more fine approximation values. But then you will have to turn the encoder knob lots of times to set the the value you want (12 pulse / rev encoder).

Here is an OVER-SIMPLIFIED diagram that shows some of the conversions and crossover calculations happening inside the PID code alone.

SMC3 is great for the few types of motor/ sensors it supports. Not much work is needed to test it and verify it. The AMC1820USB needs more verification every time a code change occurs in the PID code. I'm sorry, that is true.

Your methods appear excessively overcomplicated. SMC3 has 10 bit inputs from Simtools and from its adc with 10 bit pwm output so its pid resolution is 10 bit also. Sabertooth packet serial input is 8 bit though so the output is simply divided by 4 when using it.

Since your boards maximum input is 12 bit from its adc why not keep it simple and do the same as above but using 12 bit? Using 16 bit input from Simtools is just a gimmick since it doesn't make the overall resolution any better. Then when 10 bit adc input is used instead all you need to do is multiply it by 4 to bring it up to 12 bit equivalence. And all I see a need for is 12 bit outputs since pwm doesn't care if its 8 or 12 bit nor does anything that accepts an analog input.

I'm a firm believer in K.I.S.S.

Its called multiple option support, not only for sensors and feedback types but also for motor control types.

The resolution is 16 bit because i plan to add 16bit SPI and i2c ADC expansion hardware to that. Already have appropriate connections designated for it on the AMC1280USB PCB if you noticed. Already tested those to be working before kickstarter boards production.

And the Encoder2Position adapter will have optional 12bit or 16bit SSI digital outputs.



So you saying that the need for 16bit is gimmick, is baseless.

The AMC1280USB is a huge modular project, and not limited in use for small dc motors only. Its being mostly used for industrial AC motor, or Linear Servomotor platforms the last years.