OT, I want to make a simulator.

[ndzied1]

Forgot about a board I got to make a foot pedal interface for my son’s analog synthesizer.

Arduino MKR Zero. Has 7 analog inputs up to 12 bit and one analog out that is 10 bit (actual analog, not just PWM)

Looks like FreeRTOS has a port to run on it.

 

[NetNathan]

When you can buy an excellent temp controller that supports inputs for all TC curves, plus volts and mA, 8 PIDSs assigned by zones, an excellent auto tune algorithm, and comms, plus more for around $500..... why start over?

 

[Peter Nachtwey]

When you can buy an excellent temp controller that supports inputs for all TC curves, plus volts and mA, 8 PIDSs assigned by zones, an excellent auto tune algorithm, and comms, plus more for around $500..... why start over?

The simulator isn't for me and I would like to be able to simulate motion/hydraulic systems. I can already simulate just about anything in software using python.

 

[RVaughan]

When you get the pieces in place, are you planning to make the package available to others?

 

[Peter Nachtwey]

When you get the pieces in place, are you planning to make the package available to others?

That was my goal but I haven't found a cheap solution that would be cheap enough were people wouldn't mind spending the money. I want something that the PLC program can connect his PLC to and simulator simulates his temperature, level, or motion control system.



The Raspberry PI 4B Canakit cost $160 from Amazon. That is cheap enough to buy just to see what it does but there isn't any decent analog I/O. The Beagleboard controllers have the same problem. NXP is very close but it doesn't have the required analog resolution.


Another option is for me to copy some of our analog I/O design and make a board that is compatible with the Raspberry PI header. It would also be compatible with my Nvidea Jetson header. This seems the best option so far. However, the analog board would not be cheap unless I did this for no profit at all or donating my time.


For simulating motion control the Parallax Propeller 2 is the best option so far but again, I would need to design an analog input and output board.


Internally we use Raspberry PI, beagle board and our motion controllers because our motion controller are cheap to us.


Way back around 2005 Ron Beaufort used wood burning irons to simulate a temperature system. They cost only $20. Even if you add the temperature sensor, I can't beat that price.

 

[ojz0r]

From what i've seen there is not many options for chips with 16-bit AD/DA resolution. That is usually handled by some peripheral AD/DA chip on the board, either communicating with I2C, or SPI if you want faster performance.

 

[drbitboy]

Analog signals (temperatures, pressure, level) are typically slow; although that is not always the case, I would think very few would need anything faster than 10Hz or so. So we could forego any physical analog outputs and simply have the simulator write analog data into the PLC's input map. If the PLC program doesn't have an input map, it could be refactored to add one, with a boolean to tell it whether to use its physical I/O or the simulator targets.

Digital signals will need to be much faster, of course, especially for motion and encoders, but I suspect adding physical digital outputs of a RasberryPI or BeagleBoard is much simpler and cheaper than analog outputs.

Just a thought.

 

[Peter Nachtwey]

From what i've seen there is not many options for chips with 16-bit AD/DA resolution.

Yes, but there are no add on boards that meet my specifications either. That means I must design a board

Analog signals (temperatures, pressure, level) are typically slow;

Pressure/force signals on hydraulic systems can be extremely fast. That is why spec sensors with at least a 4KHz band width, 8KHz is obviously better.

Digital signals will need to be much faster, of course, especially for motion and encoders,

That was a little outside my scope. Remember, I want to make a simulator. What you suggest is that I simulate quadrature encoder outputs. Not easy to do right.


BTW, I did find this. It is a design that will do what I want from TI.
It provides what is needed to make the PCB and the bill of materials.

https://www.ti.com/lit/ug/tidu392/t...90658&ref_url=https%3A%2F%2Fduckduckgo.com%2F

 

[drbitboy]

Analog signals (temperatures, pressure, level) are typically slow; although that is not always the case,

Pressure/force signals on hydraulic systems can be extremely fast. That is why spec sensors with at least a 4KHz band width, 8KHz is obviously better.

d'Oh! I knew slow was the typical case, although there are other non-slow cases, but I forgot that the non-slow cases are your core business!

BTW, I did find this. It is a design that will do what I want from TI.
It provides what is needed to make the PCB and the bill of materials.

https://www.ti.com/lit/ug/tidu392/tidu392.pdf?ts=1633880190658&ref_url=https%253A%252F%252Fduckduckgo.com%252F

Hmm, basically a PLC input/output card, so it will need a device to drive it. Why not go with a PLC, or for that matter make the simulation a task of the PLC controlling the simulation and bypass the analog signals altogether?

 

[Peter Nachtwey]

d'Oh! I knew slow was the typical case, although there are other non-slow cases, but I forgot that the non-slow cases are your core business!

Yes.
I can see you have been avoiding your DMD duties on another thread about pumps and pressure.

Hmm, basically a PLC input/output card, so it will need a device to drive it. Why not go with a PLC, or for that matter make the simulation a task of the PLC controlling the simulation and bypass the analog signals altogether?

A PLC isn't fast enough for the motion control and fast pressure stuff.

BTW, I did write a simulator in ladder for MicroLogix 1500 and SLC5/05. it is in the download/misc/hotrod.zip
Basically it simulates Ron Beaufort's hot rod ( wood burning ) temperate training units. This was around 2005. I simulated a SOPDT temperature system.

I have another file called SMC.zip for sliding mode control
No one seemed to use these.


BTW, I wrote simulator code in C for one of our programmers. He used the NXP board which is fast enough but it didn't have very good analog I/O.

 

[AustralIan]

Peter, a Beckhoff CX7000 + EL3161 + EL4102 + EL9012 may suit your needs.

 

[Peter Nachtwey]

Peter, a Beckhoff CX7000 + EL3161 + EL4102 + EL9012 may suit your needs.

That would work but it wouldn't be cheap so that people would just buy it to simulate something.


I was hoping it would be cheap enough for people besides me to buy and use. If I want to simulate something now I would use one of our motion controllers. However, like the Beckoff solution, our motion controllers are too expensive.


So far the best option is to us the TI design I posted a link to but modify the header so it fits on a Raspberry PI board. It would still need a 24 volt power supply but the PLCs can supply that. Soldering all that surface mount stuff will be a pain.

 

[Gene Bond]

That TI board does indeed look like what you are searching for. With the help of a buck/boost power supply, a 5v power bus may be all you need.

Obviously, designing the board and producing it will cost more than the processor board, due mostly to volume.

 

[Peter Nachtwey]

That TI board does indeed look like what you are searching for. With the help of a buck/boost power supply, a 5v power bus may be all you need.

I would stick to the 24 volt supply since PLCs and motion controllers already use 24 volts. Did you notice the TI board has optical isolation.

Obviously, designing the board and producing it will cost more than the processor board, due mostly to volume.

Most of the design work is already done. There are files for the Altium designer which we already have
[/quote]
There are two main problems with the TI solution.
1. buying the parts.
2. soldering those little parts on the board.

One more thing. The Raspberry PI header is 40 pin and the TI board is 50 pin.
I haven't figured out which 10 will not get connected. It would be nice if the TI board just plugged into a Raspberry PI board.