Recommendations - Model Train Controller

malc-c · - 16th February 2007, 20:42

Ok I think I now understand the settings for ANSEL. Reading the datasheet again

ANSEL — ANALOG SELECT REGISTER (ADRESS: 91h) (PIC16F676 ONLY)

ANS7 ANS6 ANS5 ANS4 ANS3 ANS2 ANS1 ANS0
bit 7 bit 0
bit 7-0: ANS<7:0>: Analog Select between analog or digital function on pins AN<7:0>, respectively.
1 = Analog input. Pin is assigned as analog input.(1)
0 = Digital I/O. Pin is assigned to port or special function.

So if I wanted all analogue I set ANSEL to 11111111 - all digital 00000000

Ergo to set AN0, AN1, AN2 and AN4 Ansel should = 00010111, which is exactly what Darrel suggested. However when compiled with this setting the 4th output still runs at full speed with no control.

Could it be that the issue is with the declaration of the array - as it has [3] and I would assume that there are 4 channels so maybe this should be [4] ??

Originally Posted by malc-c

Could it be that the issue is with the declaration of the array - as it has [3] and I would assume that there are 4 channels so maybe this should be [4] ??

Looked like that was teh cause - I now have 4 working channels... just need to resolve the issue of the PIC locking on at full throttle and I'm done !

**Darrel Taylor** · - 16th February 2007, 23:46

Boy. Try to get some sleep, and people start talking to themselves.

Looks like the guy you were talking to, helped figure a couple things out though.

Good catch on the array size. Didn't see that.

Heard the video. That's just nasty sounding.
Is it any better yet after the changes?

I know you were told to change SPWM_RES to 255, but it really should be 256 for 8-bit resolution. Maybe the cause of the "PIC locking on at full throttle". I'll play with it here too and see.
<br>

**Darrel Taylor** · - 17th February 2007, 01:41

After further playing.

I can't make it "Lock".

Tried all 4 channels, tried 255/256 RES.

Incidently, 255 does work better. It's not the way I thought I wrote it. But...

That way 0 is always off, and 255 is always on.
<br>

malc-c · - 17th February 2007, 09:49

Darrel,

If using an LED its hard to make it lock. I think its either related to the power supply I'm using, or some other factor within the circuit as it does it on several loco's. I have a 0.1uf spread accros the power pins of the PIC, and power rails in a bid to decouple the supply. I've also tried changing the 1K base resistor between the NPn and PIC output as I though that the driver circuit may be drawing more than the 25ma the port can take, but that made no difference

Is the PWM outout a square wave, sawtooth or just spikes that are increased in frequency? I would like to stick with PWM as there is less heat to dissapate in the circuit and it provides far better slow starts than varible voltage type controllers. Maybe I should look at using a PIC with hardware PWM, like the 16F690 ??

**Darrel Taylor** · - 17th February 2007, 12:59

Definately Squarewave. This is about a third of the way on the POT.

Well, before you change pics. Can you give this program a quick try.
It's just a single channel, no interrupts. It runs at about 3.8khz

It should give an indication whether the higher frequencies of hardware PWM will help or not.

Code:

@ __CONFIG _HS_OSC & _WDT_ON & _PWRTE_ON & _MCLRE_OFF 
DEFINE OSC 20
CLEAR
ANSEL=%00110001
CMCON=7
TRISA=%11111111        'set PORTA as all input 
TRISC=%00000001        'set PORTC as all output apart from RC0

PWM1       VAR PORTC.2
DutyCycle1 VAR WORD
GoDone     VAR ADCON0.1
ADCON1 = %00110000
ADCON0 = %00000011
while GoDone : wend

Main:
  if ADRESH > 1 then
    high PWM1
    DutyCycle1 = ADRESH
    GoDone = 1
    pauseUS DutyCycle1
    low  PWM1
    pauseUS 255 - DutyCycle1
    while GoDone : wend
  else
    low  PWM1
    GoDone = 1
    while GoDone : wend
  endif    
goto Main

**mister_e** · - 17th February 2007, 16:58

as long as the POT impedance meet the maximum recommended.. it shouldn't be much of a problem.

What happen to your supply when you plug your loco directly?

Now, If you add 1 or more?

Now do the same but add your transistor to drive them.. what happen?

i got my few motor here, i'll do some test later.

**mister_e** · - 17th February 2007, 19:10

Well well well... running 3 different Motor type from 500mA to 10A @12Vdc.. and a pretty messed up breadboard + Jumper

it works

2 channel PWM... all square...

3 pwm ... still the same, life is good

BUT... when you add ADCIN 4, DutyVar4... somethings goes bad....

Do one test with anything matched with ADCIN 4... something go really bad... where it comes from... i don't know for now.

Same thing happen if you do something with DutyVar4...

scratch scratch scratch

**mister_e** · - 17th February 2007, 19:23

Everything just fine using...

Code:

@ __CONFIG  _HS_OSC & _WDT_ON & _PWRTE_ON & _MCLRE_OFF 


ANSEL = %00010111    ; - dt suggested value



DEFINE OSC 20
CMCON=7
TRISA=%11111111        'set PORTA as all input 
TRISC=%00000001        'set PORTC as all output apart from RC0

DEFINE ADC_BITS  8     ' ADCIN resolution  (Bits)
DEFINE ADC_CLOCK 2     ' ADC clock source  (Fosc/32)
DEFINE ADC_SAMPLEUS 11 ' ADC sampling time (uSec)

INCLUDE "DT_INTS-14.bas"            ; Base Interrupt System
INCLUDE "SPWM_INT.bas"              ; Software PWM module

DEFINE SPWM_FREQ  150                ; SPWM Frequency
DEFINE SPWM_RES   255               ; SPWM Resolution

DutyVar VAR byte[4]

ASM
SPWM_LIST  macro                    ; Define Pin's to use for SPWM
    SPWM_PIN  PORTC, 2, _DutyVar   ; and the associated DutyCycle variables
    SPWM_PIN  PORTC, 5, _DutyVar+1   ; Notice the underscore before variables
    SPWM_PIN  PORTC, 4, _DutyVar+2
    SPWM_PIN  PORTC, 3, _DutyVar+3
  endm
  SPWM_INIT  SPWM_LIST              ; Initialize the Pins
ENDASM

ASM
INT_LIST  macro    ; IntSource,        Label,  Type, ResetFlag?
        INT_Handler   TMR1_INT,  SPWMhandler,  ASM,  yes
    endm
    INT_CREATE                      ; Creates the interrupt processor
ENDASM

@ INT_ENABLE  TMR1_INT              ; enable Timer 1 interrupts

;_____________________________________________________________________________

Main:                               
 ADCIN 0, DUTYVAR[0]
 ADCIN 1, DUTYVAR[1]
 ADCIN 2, DUTYVAR[2]
 ADCIN 4, DUTYVAR[3] 
 GOTO Main

GO GO GO FIGURE IT

malc-c · - 17th February 2007, 19:44

Originally Posted by Darrel Taylor

Well, before you change pics. Can you give this program a quick try.
It's just a single channel, no interrupts. It runs at about 3.8khz

It should give an indication whether the higher frequencies of hardware PWM will help or not.

Code:

@ __CONFIG _HS_OSC & _WDT_ON & _PWRTE_ON & _MCLRE_OFF 
DEFINE OSC 20
CLEAR
ANSEL=%00110001
CMCON=7
TRISA=%11111111        'set PORTA as all input 
TRISC=%00000001        'set PORTC as all output apart from RC0

PWM1       VAR PORTC.2
DutyCycle1 VAR WORD
GoDone     VAR ADCON0.1
ADCON1 = %00110000
ADCON0 = %00000011
while GoDone : wend

Main:
  if ADRESH > 1 then
    high PWM1
    DutyCycle1 = ADRESH
    GoDone = 1
    pauseUS DutyCycle1
    low  PWM1
    pauseUS 255 - DutyCycle1
    while GoDone : wend
  else
    low  PWM1
    GoDone = 1
    while GoDone : wend
  endif    
goto Main

Interesting..... works fine, no drop out or locking up after 20 or so tests. However the noise from the armature when starting is very loud and high pitched, with harmonics as you increase the throttle.

I tried reming out the last two channels as Steve suggested with the original code and I still experienced problems... but this new code works.

**mister_e** · - 17th February 2007, 20:22

Malc,
it's little bit more than rem ADCIN line. i'll place all the changes i made in RED

Code:

@	__CONFIG  _HS_OSC & _WDT_ON & _PWRTE_ON & _MCLRE_OFF 


	ANSEL = %00010111    			; - dt suggested value



	DEFINE OSC 20
	CMCON=7
	TRISA=%11111111        			'set PORTA as all input 
	TRISC=%00000001        			'set PORTC as all output apart from RC0

	DEFINE ADC_BITS  8     			' ADCIN resolution  (Bits)
	DEFINE ADC_CLOCK 2     			' ADC clock source  (Fosc/32)
	DEFINE ADC_SAMPLEUS 11 			' ADC sampling time (uSec)

	INCLUDE "DT_INTS-14.bas"            	; Base Interrupt System
	INCLUDE "SPWM_INT.bas"              	; Software PWM module

	DEFINE SPWM_FREQ  150                	; SPWM Frequency
	DEFINE SPWM_RES   255               	; SPWM Resolution

	DutyVar VAR byte[4]

ASM
SPWM_LIST  macro                    	        ; Define Pin's to use for SPWM
    SPWM_PIN  PORTC, 2, _DutyVar                ; and the associated DutyCycle variables
    SPWM_PIN  PORTC, 5, _DutyVar+1              ; Notice the underscore before variables
    SPWM_PIN  PORTC, 4, _DutyVar+2
    SPWM_PIN  PORTC, 3, _DutyVar+3
  endm
  SPWM_INIT  SPWM_LIST                  	; Initialize the Pins
ENDASM

ASM
INT_LIST  macro    ; IntSource,        Label,  Type, ResetFlag?
        INT_Handler   TMR1_INT,  SPWMhandler,  ASM,  yes
    endm
    INT_CREATE                      	        ; Creates the interrupt processor
ENDASM

@ 	INT_ENABLE  TMR1_INT              	; enable Timer 1 interrupts

	;_____________________________________________________________________________

Main:                               
 	ADCIN 0, DUTYVAR[0]
 	ADCIN 1, DUTYVAR[1]
	ADCIN 2, DUTYVAR[2]
 	ADCIN 4, DUTYVAR[3] 
 
	GOTO Main

Now for the noise, to me, as i'm not a motor pro, there's something normal in that. Your motor have a comfort zone in which he can operate. So if you're under that (let's call it threshold, lower possible voltage or else), for sure he will yell. It has to fight the inertia.. (OK OK probably not the right term... but it's how i understand

)

Now, is this louder using PWM or not? this is why i asked for a variable supply before.

If it's better, we could still redesign the whole thing... same thing if you find a PWM frequency which reduce the noise.

Time for few tests on your side... even if you use few battery in serie

malc-c · - 17th February 2007, 21:41

Guys,

Firstly may I say that I admire both DT and Steve's skills in programming, and want to thank you both for staying with this. - I'm feeling way out of my depth here, and all I can do is test the code you offer, and report back.

Ok power supply:
240v primary - 12-0-12 secondry 500mA transformer. Rectified using 4 x 1N4004 diodes, regulated via 7812, 7912 1A regulators, with 1000uf capacitors to give smoothed +12 and -12v supply. Then a 7805 is used off the +12v to give +5 line. 0v taken from the centre tap. Actual voltages measured are +4.94v, +11.84 and -11.86.
+5v used for the PIC, +12v used for the motor supply - 0v commoned to both

Code.
I compiled Darrels code, and no matter what I tried I couldn't replicate the latching or drop out (ie the loco would just stop and not respond) when advancing and retarding the 10K lin pot in quick succession. However even with a 4 Mhz xtal the whine from the motor was loud

Compiling Steves code I got the low growl rather than a whine, but it was a lot less noticable than DT's original code. But advancing and retarding the pot as before still caused the lock up / drop outs.

Voltage measurements:
When advancing and retarding the pot quickly between min and max the voltage drop on the +5v line to the pic was less than 0.03v (4.94 - 4.93 was read), so its not enough to cause the PIC to stop running. The drop on the +12v line was a little more, but not drastic, with the lowest reading being 10.56, so the peak drop of around 1.4v.

To varify that it is the PIC that is locking up and not the output stage, I placed an LED on the output of the PIC, and this would stay illuminated when a lock up occured, even when the loco was removed. I also removed the connection between the PIC and the base of the NPN tranny to see if the LED would go out and it had no affect, so it would appear that the driver side isn't causing the issue.

Hope this feedback helps.

malc-c · - 17th February 2007, 22:37

I feel so embarrased... but I think the problem of the drop outs etc is down to the powersupply.

To rule the PSU out I've just used the previously described PSU to power the 12v driver circuit, but used a 5v regulated switch mode wall PSU for the PIC. This was running at 5.12v. I had to common up the ground lines to get the loco to move, which also caused a bit of pulsing in the armature even with the pot at zero, but I couldn't get the lock up or drop out no matter what I tried.

So a re-think on the design of the PSU is in order.

**Darrel Taylor** · - 17th February 2007, 22:55

Grreat!

One by one, knocking out the problems.

As for sticking with it, I'll stick with anyone that put's in the effort. Like you.
Thanks for making it fun.

Steve,
Do you hear the whining etc in your motors?
Just wondering if it's the type of motor that makes a difference.
<br>

**mister_e** · - 17th February 2007, 23:35

Yeah that's what i thought too, hence why i buy few different model.

Well, yes i hear some little noise which is more or less louder depending of the model... but there's always some noise when you play around the minimum low-voltage threshold.

I tried higher PWM frequencies and it may cure some part of the problem on some voltage range and on some motor model. but it's certainly not perfect on ALL motor i have here.

To me, in this specific case, a better choice could be to use some kind of linear regulation. Something around the following.

it have some good point and some bad ones. But do you really a 0-100% regulation? mmm

Malc, don't worry... it's an interesting and funny thread.

An idea... we could skip the PWM and use some I2C DAC...

malc-c · - 18th February 2007, 00:38

Originally Posted by Darrel Taylor

Grreat!

One by one, knocking out the problems.

As for sticking with it, I'll stick with anyone that put's in the effort. Like you.
Thanks for making it fun.

Thanks Darrel,

I'm always willing to put effort in to these projects, and I've learnt a lot over the course of this thread. I just feel a tad embarrased that I didn't try two seperate power supplies earlier and kept pointing the finger at the code...

Mind you, hopefully the use of your PWM routines in this application may of hi-lighted some areas such as the armature noise, that you may want to look at for future verions.

As for the motors, I have four 2mm scale locos, two are circa 18 years old with an open frame 5 pole motors, the others are less than a year old and both have small can motors... guess which ones produce the better running... yup the older ones !!

As a side note, when I tried your last bit of code I was using a 4mhz Xtal, I then tried it with a 20mhz and it was quite painful, for my ears and the motor !!

Anyway, thanks once again to you and Steve for all your help. To everyone else who's been reading this, sorry if it bored you at times

**Acetronics2** · - 18th February 2007, 10:20

Originally Posted by mister_e

Yeah that's what i thought too, hence why i buy few different model.

Well, yes i hear some little noise which is more or less louder depending of the model... but there's always some noise when you play around the minimum low-voltage threshold.

I tried higher PWM frequencies and it may cure some part of the problem on some voltage range and on some motor model. but it's certainly not perfect on ALL motor i have here.

To me, in this specific case, a better choice could be to use some kind of linear regulation. Something around the following.

it have some good point and some bad ones. But do you really a 0-100% regulation? mmm

Malc, don't worry... it's an interesting and funny thread.

An idea... we could skip the PWM and use some I2C DAC...

Hi, Steve

Your design is to try with a low input RC constant in order to produce some kind of trapezoidal wave ... instead of the beautiful square wave.

the problem lays in the slew-rate ... remember lentz's law ... the motor windings oppose an "electrical inertia" proportionnal to the slew-rate ...

that will create some dissipation in the ballast ... but make less noise and heat.

may be some "Mid freq" rectified sinus, used in burst mode could also do the job with minimal issues ( seen on the motor side ... LOL !!! )

Alain

**mister_e** · - 18th February 2007, 10:46

My motor theory is far far away... and never felt the need to understand it in the past

interesting... thanks.

Few interesting points in the followings ...
http://homepages.which.net/~paul.hil...llersBody.html ok ok it's talking about a starter but anyways..

http://ww1.microchip.com/downloads/e...tes/00770b.pdf

malc-c · - 21st February 2007, 19:03

Guys,

Just thought you might want to see the prototype controler (only two channels)

The 10K pot and reversing switch is connected via the 6-pin header infront of the power transistors. I've also left room for a small heatsink, but in tests so far these don't get hot enough to warrant one !

The only problem is I'm still waiting on some 16F676's as I fried the only one i had in the breadboard... (got distracted by a phone call and connected the +12v to the PIC rather than the +5v !!! - it was glowing and had to cut short the call to pull the plug !!) - Tested the output stage and it works fine so can't see any reason for it not to work when the PIC's arrive.

**mister_e** · - 21st February 2007, 19:08

The only problem is I'm still waiting on some 16F676's as I fried the only one i had in the breadboard... (got distracted by a phone call and connected the +12v to the PIC rather than the +5v !!

Looks good! Keep us posted!

**Sean_Goddard** · - 25th February 2007, 02:26

Why not talk to a PIC in the module using I2C?? After all, the track is two wires.

You could modulate this on top of a standing DC voltage then control a motor using hard coded addresses on the loco module PCB and an H bridge driven by the PWM from the PIC.

Just a thought. I tried it and it worked fine. Be careful of the need to filter the I2C signals HEAVILY but not so much that you lose them.

malc-c · - 25th February 2007, 10:46

Interesting concept, sort of a cross between modern DCC and traditional control. However, its not practical in this instance as the contoller is for a club layout and would mean that every member would have to have the same system installed on their own layaouts at home. Also with the scale, fiting modules to small loco's is a pain in the ***

**Acetronics2** · - 8th March 2007, 08:17

Hi Malc,

I Had my father home this week-end ...

We talked about supplies and it seems the today's "must" is to use switching supply during speed transitions ( or accelerations ) and DC during constant speed running ( of course for noize !!! ) ...

That meets one of Sougata's threads about signal generation with a Pic ... and opens interesting reflexion and work about what output stage to use !!!

new challenge for Darrel !!! ( He Looooooves !!!! )

Alain

Thread: Recommendations - Model Train Controller

Thread Tools

Search Thread

Display

Seems I'm the only one posting :)

I think I solved it

a small last trail ???

Update

Food for Brain ...

Similar Threads

Microstepper controller

problem with the GSM controller

6-Digit, 7-segment, serial LED display controller IC

Model Train controller - revisited

Dedicated LCD Controller question

Members who have read this thread : 1

Bookmarks

Bookmarks

Posting Permissions