hi, i am using CY8C29466-24SXI kit, and i am trying to convert the RPM speed of motor to a frequency from the monitor "n", i have problem with the code, the output is constant and should change when i change the speed of motor

the equation should be : Freq(monitor) = [n ( speed[rpm]) * Zpol ( which is 8) ] / 20

thanks 

When you do an ISR you have to modify boot.tpl in root project directory to add

the vector jump.

You have currently -

org 34h ;PSoC Block DBB11 Interrupt Vector

This should be -

org 34h ;PSoC Block DBB11 Interrupt Vector

Also you did not use the #pragma qualifier when declaring the C intterupt, look at Imagcraft manual.

Regards, Dana.

Note the interrupt vector comes from the last block in a timer or counter chain, in your case DBB11.

Regards, Dana.

Sorry, forgot to add this -

#pragma interrupt_handler foo

void foo(void) {

..

..

}

Note "foo" is not used with an underscore in the declaration, just in boot.tpl.

Regards, Dana.

Hi Dana,

i have the freqcode.asm, and tickhander.asm and called them in the h file. i am not sure if i have to use interupt for the timer because it should capture from the comparator. can you please be more clear what exactly i am doing wrong?

Thanks

Do you have a schematic you can post ?

Regards, Dana.

Hi Dana, here is schematic for my project. basically, i want to take the frequency and put it in the RPM equation to measure RPM of the motor with the voltmeter ( manually ). the thing is the output from monitor is a digital signal (high/low) and is equivlant to a third of the commutation frequency.

the equation is: RPM = 2.5 * Freq (monitor)

Thanks

Basically to measure PW you use a timer, start it on positive edge of

input pulse, and either do a capture on 1 edge or an ISR on negative

edge. You could also use a counter to do this.

One simple method is use positive edge of pulse to start timer, and negative edge

to indicate end of pulse period,. That can be input pin set to ISR on negative edge.

Here is an ap note with several methods. Including an anlog method.

That works if freq is a constant.

http://www.psocdeveloper.com/docs/appnotes/an-mode/detail/an-pointer/an2103/an-file/119.html

Regards, Dana.

Hi Dana,

i tried to fix a project to have a function generator as input signal. and have compare out and terminal count out as outputs,, and connected LED on port 2-0 to check if it is working. for some reason when i connect the generator it doesnt want to compile, but once i take the generator out it compiles but cant test it beacuse of no signal in.

FYI, the link you sent earlier, i cant open any of these projects to look at it.

Thanks a lot for your time

it works, but the terminal count out i have nothing comes out, and compare out is same as signal i am putting in.

I am at a remote site, do not have my cube with me for debugging, but here

is what I see -

1) boot.tpl. root directory you have to modify vector table to do the ISR vector to TachTimer16_MSB_ISR()

Edit boot.tpl vector table to reflect this change

org 24h ;PSoC Block DBB01 Interrupt Vector

2) In TachTimer16INT.asm I think the call name TachTimer16_MSB_ISR

must be changed to _TachTimer16_MSB_ISR, not 100% sure about this,

but I think that is the case.

3) The programming pins on the 29466 are Vdd, Vss, XRES, SCLK, and SDAT, so if you

are using any of these in design they have to be disconnected to program.

4) Function generator putting out 0 - 5 V, not some negative voltage to 5V ? You

have offset its output waveform, right ?

Regards, Dana.

Some other observations -

1) Any reason for running the part not full speed at 24 Mhz ? Why waste MIPs (unless trying to save

power) available.

2) If you have a port with mixed I/O, you need to use shadow registers -

http://www.cypress.com/?id=4&rID=28251

http://www.planetpsoc.com/psoc1-articles-digital/13-basics-of-psoc-gpio.html?start=5
 

3) ISR notes -

http://www.cypress.com/?id=4&rID=36720

]http://www.planetpsoc.com/component/content/article/43-writing-a-c-isr.html

http://www.planetpsoc.com/psoc1-articles-digital/13-basics-of-psoc-gpio.html?start=7

4) I was wrong on call inside TachTimer16INT.asm,  you do not need the underscore as

you are calling an asm(), not a C f().

Regards, Dana.

One last, I am so used to writing C interrupts that I forget when not to use

the _ preceding a call.

When I discussed boot.tpl, I indicated to put the underscore preceding that call,

but you are calling an asm function, so leave it off, should be -

ljmp TachTimer16_MSB_ISR

Bonehead I am.

Regards, Dana.

Hi Dana,

i think i am too close now, i have it working but my output ( Compare out ) on PORT0-0 is always zero, it doesnt update with the frequency.

i wont be able to debuge it by LCD, however can i tested by Multimeter?

thanks

here is the file

Maybe start at the basics for a moment.

1) To measure PW we use counter or timer.

2) Since all PSOC 1 counter/timers (C/T) are down counters, we load
them with a full scale value, in your case you have chosen 65000.
Normally for a 16 bit C/T you would use 65535.

Timer solution -

3) We start the Timer on the positive edge of the input signal whose
pulse width we want to measure. So we have the starting value of the
period register, which is 65000 in your case. We start it with an ISR
from the positive edge of the input signal (via pin configuration), and
 read the timers count register in this ISR opr set a flag to do so.
 Remember Timer runs continuously unless you stop it.

4) Timer continues to run, counts dropping in its counter register, then
input signal drops, triggering a capture event (another ISR). Now we read
timer count register giving us counts at end of its pulse. So we take start
counts and subtract capture counts. Note, there is a problem, if timer over-
flows, ie counts to 0 then next count starts back at full count (65000 in your
case) we have to take care of this subtraction problem.

Example, counter starts at 65000, counts down to 30000, PW = 35000. If
Clock to timer is 1 Mhz, 1 uS, then PW = 1 uS X 35000 = 35 mS.

The Compare register has no functionality in this solution.

Counter solution -

5) We load max period into period register, 65000 in your case.
6) Input signal goes to enable pin on counter.
7) Input goes high, counter starts counting down.
8) We have the input pin configured for an interrupt on falling edge.
9) Input falls, ISR triggered, we read the counter register, not the
period register, and get the counts at the end of the pulse.
10) We subtract the ISR counts from the full scale counts we loaded (65000
in your case)  and we have the PW counts. We get the PW time using same
calculation we did for timer case based on clock.
11) We stop and reload counter with 65000 in this ISR
12) We have to take care of the same “rollover” problem as the timer.

The Compare register has no functionality in this solution.
When to use the compare register. Say you want to know when the pulse has exceeded
3000 counts, in the case of a 1 Mhz clock, 3 mS width. Then you would set the compare
Register to 65000 – 3000 = 62000. When the C/T reaches 62000 the compare output
goes high. That can trigger an ISR to handle your pulse has exceeded 3 mS.

Hope this helps.

Regards, Dana.

The reason your compare output never went high is the counts never dropped below

200, the value you wrote to the compare register.

As an aside, the core of all C/T blocks is a counter, counting clock inputs pulses.

The COMPARE and PERIOD register are separate registers from the core counter in the C/T.

 The compare has logic, that looks at whats in the core counter, and compares its value against

whats in the compare register, generating a compare output and an ISR if you enable it.

The compare logic you select, < or <=.......

Regards, Dana.

//----------------------------------------------------------------------------

#include "PSoCAPI.h" // PSoC API definitions for all User Modules

#pragma

Hi dana,

thanks for your respond. i used a tach timer, i am testing it with function generator as input signal, and measuring the output by Multimeter as output.

Correction to my prior post. I forgot, frankly, the Timers do not have a separate capture register,

so was trying to advise you the compare register was pointless in the design to calc PW.

That being said I fooled around trying different approaches, looking at jitter, and seem to

think a slightly different approach produces a better result.

So I set up a timer, to capture on falling edge, and the input pin to the timer to ISR on rising edge.

In ISR for pin I simply read timer register. In capture ISR I do not read the compare register,

I execute again a simple read on timer register. So I have two values, and keep the timer running

continuously. Reason I did not read the capture register for second value is I felt that the machine

cycles, latency, is same for both ISR calls. So my observed jitter was better with this approach.

I worked with pulses up to 10 Khz with a 1 uS timer clock. And PW down to  30 uS in width,

where error seemed too high. I have not tried faster Timer clock, say 12 Mhz, to see what that will do.

I then tried using a counter solution, just one ISR on GPIO negative edge that feeds counter enable.

Pulse coming in goes high, starts counter counting. ISR stops counter, reads it, reloads it, restarts it.

Net result seemed I could handle much narrower pulses, 10 uS reading accurately.

Food for thought, Dana.

I omitted a couple of points.

1) Counter requires only 1 ISR.

2) Counter enable serves two purposes, going high we start

counter with a known value, going low ISR to indicate done counting.

3) For real accurate count one would look at read counter register

ASM listing and count instruction cycles to get latency.

Regards, Dana.

Hi dana, can you take a look at that project, i used PWM to get RPM from tachtimer.

i am using function genrator connecting to pin 0-5 and have the out put that connect to Multimeter to measure RPM on pin 0-2

Thanks

Soime questions -

1) You are changing thePWM16 pulse width, not its frequency, with -

PWM16_WritePulseWidth(curr_rpm), is that what you intended ?

Most multimeters only measure frequnecy.

Hi Dana,

the curr_rpm is from the equation of the freq, so when i change the freq, it should change the rpm. and take that rpm value and measure thro the multimeter. as freq.

Mina

Dana,

thats what i am trying to do, take the freq from output and put it in the RPM equation and measure the freq that equivlant to RPM.

equation is RPM( freq) = 2.5 ( monitor freq)

so if i input freq 200HZ, i want to read on freqmeter 200*2.5= 500HZ

Can you advice me if my way is right or wrong ?

Thanks

1) You are changing thePWM16 pulse width, not its frequency, with -

PWM16_WritePulseWidth(curr_rpm), is that what you intended ?