Abgriff für Spannung des Gaspedalpotis... | Posts 16+

hello Jürgen,
Do you use AVRs?
greetings

Yes
There's nothing better for those small tasks...
There's no simpler or faster way to develop.
An Atmega128 could easily handle the complete control of the TDI.
128KB of on-chip flash memory and a clock speed of 16 MHz are more than sufficient for this application.
For my project, I am using the AT90S8535.
While it has been replaced by the ATMega8535, it is still used in larger...
I have the required quantities in stock. Additionally, the newer version can be easily implemented without any modifications.
in the old circuit.

'Are you sure about using AVRs, specifically the 8535? I also do a lot of work with AVRs. I don't know exactly what you're trying to achieve, but 8k of flash memory can be quickly exhausted, and the speed becomes a limiting factor for relatively complex mathematical operations. I tried to simulate a lambda sensor (LMM) using an 8535. I used the motor speed and boost pressure as input, and calculated the air mass, converting it to a voltage according to a characteristic curve. The problem is the conversion of the air mass into a voltage, which resulted in a function of the 10th order, and that puts the AVR under a lot of strain! I think you might encounter similar problems! In my opinion, an MC16 or something with more processing power would be a better choice!'

Are you sure about the AVRs and specifically the 8535? I also do a lot of tinkering with AVRs. I don't know what you're trying to achieve or how, but 8k of flash memory can be quickly exhausted, and the speed becomes a problem with relatively complex mathematical operations. I tried to simulate an LMM (likely mass airflow meter) using an 8535. I used the motor speed and boost pressure as input, and I calculate the air mass and convert it to a voltage according to a characteristic curve. The problem is the conversion of the air mass into a voltage. This resulted in a function of the 10th order, which overwhelms the AVR! I think you might encounter similar problems! In my opinion, an MC16 or something with more processing power would be a better choice!

I completely disagree.
I have already created a large number of projects on the AVR platform, and in one project with approximately 1000 lines of code, I almost completely used the available 8KB of memory.
On the other hand, the AVR is a RISC processor that executes 95% of all instructions in a single clock cycle.
It is approximately 20-30 times faster than an 8051 at the same clock frequency.

I don't know what you programmed on the AVR, but everything I've tried with it...
'Having defeated 8535 and other types, I was nowhere near my performance limit.'
this uCs.

The new ones even run at twice the clock speed (16 MHz) of the older chips.

I even managed to do it (just for fun, because sometimes it's nice to play around...).
To read an 8-bit video signal from an ADC and then output it to a DAC.
The sample rate was 4 MHz, and the image was almost indistinguishable from the original. Of course, no data processing was performed.

'Various hobbyists even use the AVR microcontroller to recreate a CRT and display a video image with it.'
To create a 12x8 character display.

To implement the basic functions for my project on the AVR microcontroller, I need...
Please provide the text you would like me to translate from German to English. I will keep the translation under 1KB.

I'm absolutely certain that this won't even come close to utilizing the 8535 (8MHz) to its full potential!

For comparison:

AVR with 8 MHz quartz crystal:
intern werden die 8 Mhz direkt in der CPU verwendet. That means,
When using single-cycle instructions, the processing speed is exactly 8 million instructions per second.

8051 with an 8 MHz quartz crystal:
To generate different phase alignments for the CPU, the quartz frequency is used.
first, divide by 12.
After that, the 8051 requires approximately 3-9 clock cycles per instruction.
It then operates at 1/12 of the clock speed and further at approximately 1/3 of the instruction rate...
It achieves approximately 0.2 mega-instructions per second.
(old scrap )

The AVR also features 32 'general-purpose registers'.
beneficial, whereas the 8051 only has 2 or 3 (I'm not entirely sure).

Furthermore, the AVR can form 3 16-bit registers using 6 registers and process them directly.

No problem for the AVR!!!

That's all well and good what you're writing! But I can only share my experiences. Even though it's a risk and a large part of the commands are processed with a single clock cycle, in many cases, there's an 'accumulation' of multiple basic commands to achieve the result. The conversion of the air mass to voltage in my example took a total of about 400,000 clock cycles. At 4 MHz, that's about 1/10th of a second. The other 'fluff' that was related to it brought the total time to 0.15-0.2 seconds, and that could be a much too long time for the acceleration process. Of course, you'll probably bring up frequency and 'household' and 'flint' again. But just imagine: what you're planning is safety-relevant! And it shouldn't be taken lightly!

It doesn't make sense to discuss this here.
You have your opinion, and I respect it.
I, however, do not agree with your opinion .
I could provide counterexamples and ask you a question like, 'Would you drive from Rome to Cologne if you read this?'... it doesn't make sense.

For the task I have in mind, the AVR is more than sufficient, and the job is certainly not safety-critical; at least, it's not more safety-critical than a simple potentiometer in a foot pedal .

Best regards,
Jürgen

Why are you being so aggressive and arrogant?

'I may not have the same understanding of MCs as you present, but I am a software developer and have some knowledge in certain areas. A function of the 10th order (I'm not sure if you even know what that is) probably can't be solved with just 5 processor cycles, as you suggest. That's why I'm going directly to Cologne, not via Rome!'

Okay, I'm logging off now! As I can see, you've already caused problems in this thread with your way of discussing things!

why so aggressive and arrogant?

'I may not have the same understanding of MCs as you present, but I am a software developer and have some knowledge in certain areas. A function of the 10th order (I'm not sure if you even know what that is) probably can't be solved with just 5 processor cycles, as you suggest. That's why I'm going directly to Cologne, not via Rome!'

and now I'm logging out! As I see, you've already caused problems in this thread multiple times with your way of discussing!

I'm sorry if it sounded arrogant or aggressive; that was definitely not my intention!
I don't deny that your method of mimicking the LMM is so computationally intensive that it would be too much for an 8-bit system (but did you drive from Rome to Cologne? ), to calculate it in the required time. However, there are always other ways to achieve the goal that are not necessarily worse, but simpler.
Often, we get so caught up in the details that we fail to see the forest for the trees and overlook a simple, effective way to achieve our goal. This allows us to extract a lot more performance from a system than simply following the straightforward, conventional path.

But exactly what we are discussing now is what I wanted to avoid. There is still a lot to be said, and each person is trying to convince the others of their own opinion.
That won't do anything!
Sure, here's the translation:

'Again:'
I respect your opinion, and you have your reasons for thinking that way.
Do the same for me. Our two opinions can exist at the same time .
Throughout my career, I've always faced challenges and learned to look both ways. You're certainly a better mathematician than I am, but I could recite the AVR and its capabilities in my sleep. Capturing a speed reading with an AVR requires approximately 10 lines of code and a Schmitt trigger, thanks to its absolutely brilliantly designed hardware.

So, please don't be offended if I don't turn this into a uC forum .

Best regards,
Jürgen

and now I'm logging out of here! As I see, you've already caused problems in this thread multiple times with your way of discussing things!

Yes, it's obvious that not everyone can tolerate my direct and straightforward manner.

No offense...

Calm down!

Sure, but using a microcontroller right away seems like overkill to me - and what about...?
Other people collect stamps!

If it's fun, that's perfectly fine!
And frankly, it doesn't really matter which microcontroller it is; I'd probably go with an 8051 derivative because I'm familiar with it and I have one lying around.

Furthermore, there are so many ways to interfere with the motor management system from the outside that it can keep both the microcontroller and the programmer busy for quite some time.

My gas injection system is currently only working on a test basis and manually. If anyone wants to automate it, I can explain how it can be done mechanically.

Hi there,

'No one seems to be getting worked up here. So far, the discussion has been quite objective. Everyone has their opinion, and I respect anyone who says they're working with AVRs. It's probably the simplest and cheapest option. There are certainly thousands of examples where the AVR is far from being fully utilized. However, due to its 8-bit architecture, it also has its limitations. As I mentioned, a 16-bit microcontroller is better suited for complex calculations.'

@jürgen

What exactly do you want, and how do you plan to implement it? Please send me the details via private message or email so we don't 'mess up' this thread any further. I'm very interested in hearing about your ideas and solutions!