Recommendations - Model Train Controller

[Darrel Taylor]

skimask,

Aw rat's, gave you a chance, but you missed it.

You were supposed to say "BAM".

Bit Angle Modulation.

For devices that rely on the "Average Current" resulting from a PWM signal, the PWM signal doesn't need to be exactly symmetrical waveforms based on consecutive ON and OFF cycles.

With BAM, you only need 8 interrupts per cycle. Each time period is "weighted" by it's bit position in the DutyCycle variable.

This way, you could turn every pin on a PIC18F8720 (all 67 of them) into PWM (BAM) outputs (at pretty high frequencies).

And of course, with DT_INTS, it's a piece of cake.
<br>

[Darrel Taylor]

I assume that using a 100 Hz frequency for the PWM would be ideal... what PIC did you use, and is having an internal clock an issue (I'm using a 12F675 for experimential use)

Malcolm,

I was using a 16F877A, I guess it was really comparing Apples to Oranges.

I figured I better check with a 12F675, but didn't have one handy so I used a 12F683. It was close enough to figure out what was wrong.

This get's everyone, so I should have seen it right away, but nope.

CMCON = 7 ; disable the comparators
ANSEL = 1 ; AN0 is analog, rest digital

also, You have a PWM output on GPIO.0 ... The same pin as the POT is connected to on AN0. That'll cause problems with the A/D reading.

Hope you're still interested?

EDIT: 100hz is no problem. But you'll need a faster crystal/resonator
INTRC won't cut it.

HTH,

[malc-c]

Darrel,

Sorry I missed that... I've remmed out the PWM line fro GPIO.0 and have got a reasonable result, especially for slow speed - The crawling speed of an 18 year old tank loco is really amaizing !!

Here's the code

DEFINE OSC 20
CLEAR

DEFINE ADC_BITS 8	' Set number of bits in result
DEFINE ADC_CLOCK 3	' Set clock source (rc = 3)
DEFINE ADC_SAMPLEUS 50	' Set sampling time in microseconds

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

DEFINE SPWM_FREQ  100                ; SPWM Frequency
DEFINE SPWM_RES   256               ; SPWM Resolution

D var byte

DutyVars   VAR BYTE[3]              ; DutyCycle Variables
  ;DutyVar1 VAR DutyVars[0]          ; group them in an array for easy access
  ;DutyVar2 VAR DutyVars[1]          ; with FOR loops etc.
  DutyVar3 VAR DutyVars[2]

ASM
SPWM_LIST  macro                    ; Define Pin's to use for SPWM
    ;SPWM_PIN  GPIO, 0, _DutyVar1  ; and the associated DutyCycle variables
    ; SPWM_PIN  GPIO, 1, _DutyVar2  ; Notice the underscore before variables
     SPWM_PIN  GPIO, 2, _DutyVar3
  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, D 
    DutyVar3 = D   
GOTO Main

Using a close loop fedback output stage and a 12v regulated supply for the track as per the diagram attached, I get a nice smooth slow speed control. This is using the 12F675 with its internal osc.

The loco does growl a bit when at really slow speed... and at maximum speed its fine for a goods train, but not an express. I susspect that using a higher track supply voltage will resolve the speed issue (my DVM reads 6.5v across the track at full speed), but wondered if there was anything I could do to reduce the growling at slow speed.

Please bear in mind that I'm still somewhat new to PICs and programming.. so If I've still overlooked something in the settings then please advise what I did wrong and how to correct it.

[Darrel Taylor]

Hmm, I don't see the CMCON and ANSEL in your program.

There's a R-M-W (Read-Modify-Write) problem when OUTPUTting to a pin set to Analog. It adds a high frequency component to the PWM signal.

I don't know if that's what is causing the growling or not.

6.5V at full speed? That doesn't sound right. The current is getting limited somewhere. Maybe if you increased the value of R4 ????
<br>

[malc-c]

Hmm, I don't see the CMCON and ANSEL in your program.

There's a R-M-W (Read-Modify-Write) problem when OUTPUTting to a pin set to Analog. It adds a high frequency component to the PWM signal.

Would the following be correct

TRISIO.2 = 0   'Set GPIO.2 to output.
ANSEL.2 = 0    'Set GPIO.2 to digital		
GPIO.0 = 1      'gpio.0 set to input
VRCON = 0        ' Voltage reference disabled
OPTION_REG.7 =    0

6.5V at full speed? That doesn't sound right. The current is getting limited somewhere. Maybe if you increased the value of R4 ????

I'll try that and see what happens, thanks for the suggestion

[Acetronics2]

Hi, Malc

I do not know where you found R4 and D1 ... but your problems come from there.

No real use for them ... so, back to their drawer will be the best.

... or use another Pic pin if you want to do something with the freewheel emf !!!

a simpler output stage would be welcome ( your NPN Emitter to ground and a series 1 k in the base connection i.e. )

Alain

[malc-c]

Alain,

Thanks for the suggestion... the output circuit was taken from a book and is actually a cloose-loop feedback controller... I just fed the output from the PIC directly to the base of the NPN tranny... but I'll try your suggestions and see if it improves the max speed settings. However in its current state it makes for an excellent branch line controller where slower trains would run...

To give you some idea of the excellent slow speed the PWM gives have a look at this video I've uploaded http://www.micro-heli.co.uk/controller.avi (9mb DIVx format)

[malc-c]

Omitting the feedback section (R4 and 1N4148), and adding a 1K base resistior I now get a good range of speed and excellent slow speed crawling. Just need to resolve the growling in the armature

[Acetronics2]

Hi, Malc,

The only way is to increase the PWM frequency ( 500 or 1000 Hz might be enough ).
also Think to the minimum PWM ratio zone to skip ( ~ 10 % )

you'll find a problem here : some motors do not like high frequencies ( > 3000 Hz ) due to their high inductance ... and might overheat.

Do you think I could make a test with my Big Boy, die cast "crocodile" or JEP 231 ... or some oldies ???

Hé,hé ... my father gave its life and brain to our national railway company ...and still builds custom models ( 84 years ...the papy !!! )

Alain

[malc-c]

Well I've been playing with the values for SPWM frequency and resolution and found the highest setting for the 12F675 using internal osc was around

DEFINE SPWM_FREQ  200                ; SPWM Frequency
DEFINE SPWM_RES   256               ; SPWM Resolution

Its not possible to up the resolution without using an external OSC. However I intend to use a different PIC and may opt to use it at is max xtal rating, especially if I use the one PIC to provide all 4 channels. Maybe the higher resolution will reduce the grumbling in the armature... mind you if this was on old 0-6-0 shunter it would look and sound just right

One thing I did notice is that when using the PWM version the power tranny runs cold, where with the close loop version it warmed up (slightly)

Anyway, I'm having fun learning

[malc-c]

There's a R-M-W (Read-Modify-Write) problem when OUTPUTting to a pin set to Analog. It adds a high frequency component to the PWM signal.

That's interesting... whilst I was reading up on the use of your code I could hear this high pitched (estimated 4 - 8 Khz) faint sound like you often hear eminating from a TV... placed my ear near the loco (which is standing still on the track) and its singing away !!

I presume that this is due to the fact that GPIO.2 is still set to analogue and not digital ??

What config settings should I use for CMCON and ANSEL to resolve this ?

[mister_e]

in theory the right setting have to be

ANSEL=%00110001
CMCON=7
GPIO=0
TRISIO=1

and later i just change the main loop to...

Main:                               
        ADCIN 0, DUTYVAR3
        GOTO Main

AND NOW it's working pretty good and smooth... well here...

For once i found a use of the Left-Justified results... Never thought ADCIN handled it

Make sure your Pot impedance meet the maximum recommended on.. 2.5 k if my memory serves me well. 10K here (EasyPic 4) still work fine..

Also, as the ADC results is a 8 bit result, you should change the SPWM_RES to 255.

[malc-c]

Steve, Thanks for your input. I made the changes you suggested and the performance is still great and the growling sound produced by the PWM in the armature appears less noticable. The high pitched sound is still noticable, but nothing appears hot

[mister_e]

Unfortunately i don't have anything like that here to test, but Maybe the PWM have to be smooth first...

what happen if you place a big capacitor (470 & up) in parallel with the motors?

What happen if you place a capacitor (few uF) on the mosfet gate pin? (transistor base)

What happen if you use a variable voltage regulator (or PSU)? instead of PWM?

[malc-c]

Unfortunately i don't have anything like that here to test, but Maybe the PWM have to be smooth first...

what happen if you place a big capacitor (470 & up) in parallel with the motors?

Tried both high (1000uf) and low (100uf) - the loco won't move

What happen if you place a capacitor (few uF) on the mosfet gate pin? (transistor base)

Tried, 1uf to 10uf. with a 1uf the growling was less noticable, but the slow speed control became less. Increasing this to 10uf and the loco had almost no ability to crawl.

What happen if you use a variable voltage regulator (or PSU)? instead of PWM?

Don't know as I don't have a varible regulator to hand.. what would be the advantage of using one over using a pot accross the supply with the wiper connected to the base of the NPN tranny ?

[mister_e]

what's your actual schematic?

how much current your motor need?

[malc-c]

Typically these loco's draw around 300mA - 600mA ( a quick test showed mine drawing 380mA with no coaches at full throttle )

The schematic as it stands is attchaed

[Darrel Taylor]

Saw the video.

I've seen snails move faster than that, when it's raining.

Cool.

Hope someone can figure out the Growl.

[malc-c]

Saw the video.

I've seen snails move faster than that, when it's raining.

Cool.

Hope someone can figure out the Growl.

LOL - thanks !

I'm not too bothered about the noise the PWM makes, but as this will be used on a club lauout, other members might complain...

[Acetronics2]

Typically these loco's draw around 300mA - 600mA ( a quick test showed mine drawing 380mA with no coaches at full throttle )

The schematic as it stands is attchaed

Hi, Malc,

R3 is no use at all here ... short it !!!

... but a 1N 540x cathode at the + 12v, anode on the V+ output could be rather useful ...

Alain

PS ... for Growling, may be you could try to " unsharpen" the PWM to somewhat with increased rise and fall time ... by inserting a filtering coil in series in the output line ( a LITTLE smoothing capacitor too ??? )

[malc-c]

Hi,

Don't have that Diode suggested, have plenty of 1N4148's or 1N4001 to hand, would they do ?

[Acetronics2]

The 4001 will be a bit too weak ... for serious use.

try to find a 3A rated model ... or use a PNP Darlington with diode Built in !!! like BD678/680

Alain

[Acetronics2]

Hi, Malc

Some more food for your brain :

http://www.manhattancontrols.com/for...wtopic.php?t=4

Alain

[malc-c]

OK I thought, lets move this on and use a PIC capable of providing all 4 seperate PWM outputs. So after consultation via a few PMs with Steve and Darrel I've opted for a 16F676.

I've used a 20Mhz xtal, with 22pf caps between pins 2 and 3, wiper from 10K pot goes to pin 13 and PWM output on pin 8.

I've modified the code used for the 12F675, and compiled - no errors or warning, and loaded it into the PIC.

@ __CONFIG _HS_OSC & _WDT_ON & _PWRTE_ON & _MCLRE_OFF 

ANSEL=%00110001
CMCON=7
TRISA=%11111111        'set PORTA as all input 
TRISC=%00000001        'set PORTC as all output apart from RC0


DEFINE OSC 20
CLEAR

DEFINE ADC_BITS 8	' Set number of bits in result
DEFINE ADC_CLOCK 3	' Set clock source (rc = 3)
DEFINE ADC_SAMPLEUS 50	' Set sampling time in microseconds

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

DEFINE SPWM_FREQ  100                ; SPWM Frequency
DEFINE SPWM_RES   255               ; SPWM Resolution

DutyVars   VAR BYTE[3]              ; DutyCycle Variables
  DutyVar1 VAR DutyVars[0]          ; group them in an array for easy access
  DutyVar2 VAR DutyVars[1]          ; with FOR loops etc.
  DutyVar3 VAR DutyVars[2]
  DutyVar4 VAR DutyVars[3]

ASM
SPWM_LIST  macro                    ; Define Pin's to use for SPWM
    SPWM_PIN  PORTC, 2, _DutyVar1   ; and the associated DutyCycle variables
    SPWM_PIN  PORTC, 3, _DutyVar2   ; Notice the underscore before variables
    SPWM_PIN  PORTC, 4, _DutyVar3
    SPWM_PIN  PORTC, 5, _DutyVar4
  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, DUTYVAR1
 ADCIN 1, DUTYVAR2
 ADCIN 2, DUTYVAR3
 ADCIN 3, DUTYVAR4
 GOTO Main

I then tested the code / hardware and I don't get the same performance as I did with the 12F. There is no responce from the loco until the pot is at mid point, at this stage the armature sings with a high pitch noise. Advancing the pot increases the speed, but I don't get anywhere like the slow crawl speeds I got with the 12F. I can also assume that the armature is warming as it smelt a little (carbon smell) after a short run.

I think its probably down to my poor understanding of Ansel etc.. so an constructive comments would be welcome.

Oh and I've not changed the driver stage yet as Alain suggested.

EDIT:: EDIT::
Changed the Xtal for a 4Mhz and the frequency of the noise has dropped.. but its still there and not the low growl that was produced by the 12F.. so it it a timing issue ??

[malc-c]

I've also tried the other channels, and get the same performance and noise from each one. (I've also reduced the frequency down to 100 to make sure I get 4 channels) - and I'm using the int osc to compare it against the 12F

I've noticed that the growling from the armature has a second resonating sound that I can only describe as the sound you get with a multi-engine prop plane when the engine is slightly out of sync. I can only assume this is due to the PIC having to work hard outputing to 4 channels rather than the 1 on the 12F.

If all else fails, I'll use a 12F675's to drive each of the 4 controllers... overkill I know, but the performance I was getting with the 12F was just what I was after.

I'm hoping the issue with the 16F676 is due to a conflict of configuring the ports to the correct analogue / a2d / digital combination, and with your help, will work as good as the 12F !

[mister_e]

There's something bad in the schematic indeed. R3 have it's use... but not at the good place. Should be in serie with the transistor base. If you short it, you'll burn the transistor...

I'm curious.. i'll try to find some 12V motor (1A or more) 'round here today, and test the whole thing. Sure it's possible.

I'll dig in the transistor datasheet... sure there's something bad in their HFe mathe here....

Anyways, i'll be back!

[Acetronics2]

There's something bad in the schematic indeed. R3 have it's use... but not at the good place. Should be in serie with the transistor base. If you short it, you'll burn the transistor...

Anyways, i'll be back!

Hi, Steve

Did you notice a 1K resistor had magically appeared in series with the trans. base ??? ...

Strong coffee needed ... Malc is a one not so easy to follow !!!

The output schematic can be enough for this time ... let's see one problem at a time !!!

Inserting some PAUSE ( 100-200 ms ) between inputs pollings might do something interesting here ... let's try that first

Alain

[malc-c]

Hi, Steve

Did you notice a 1K resistor had magically appeared in series with the trans. base ??? ...

Alain, the 1K base resistor was added way back after you suggested in post #46 that I removed the R4/D1 combo and and a series 1 k in the base connection

The nearest shematic I have is as detailed in post #57 - apart from the fact that I'm using a 16F676 rather than the 12F675 - the output stage is the same, I've not changed anything since, and wouldn't edit posts to suit...

[Darrel Taylor]

Malcolm,

I've got it running here on a 16F676. But...

You have a POT input on AN3, which is also the OSC2 for the crystal.

If I comment out the ADCIN 3, DUTYVAR4 then it seems to run fine.
But with that statement in there it goes wacko.
<br>
Added: ANSEL needs to be changed too, but I'll wait to see where you move the AN3 to.

[malc-c]

Whoops !

Thanks for spotting my deliberate mistake !!!!

Thanks for the PM, I've created another copy of the code and downloaded the DT_int 18 file, the following compiles OK but I have to use a different programmer for pics with 18 pins or more and I've never had much luck with 18F2550's and this programmer as Steve will vouch !!

I'll try some changes for the 16F676 and see how I get on.. Thanks to everyone again for your assistance...

[malc-c]

Malcolm,

ANSEL needs to be changed too, but I'll wait to see where you move the AN3 to.

I actually marked it on the datasheet pinout to use AN4 (pin 10) on the 16F676 - I should pay more attention to what I plan !!

[Acetronics2]

Alain, the 1K base resistor was added way back after you suggested in post #46 that I removed the R4/D1 combo and and a series 1 k in the base connection

The nearest shematic I have is as detailed in post #57 - apart from the fact that I'm using a 16F676 rather than the 12F675 - the output stage is the same, I've not changed anything since, and wouldn't edit posts to suit...

We're in sync Malc ... just R3 NOT to be shorted !!! ... but placed in series with the NPN COLLECTOR ...

MEA Culpa ...

Alain

[Darrel Taylor]

AN4, ok.

Make ANSEL = %00010111
<br>

[malc-c]

AN4, ok.

Make ANSEL = %00010111

Darrel, can you confirm that I have understood the variables for SPWM_PIN right, and that I'm using the correct channels for a/d, as I only had 2 out of 4 outputs working right, and now made changes I get that high frequency noise from the armature (where as before it was a low growl, even with a 20 Mhz xtal)

I want to use the following pins on the 16F676 for inputs from 10K pots, and their corresponding PWM outputs

Pot 1 - Pin 13 - AN0 - controlling output on pin 8 - RC2
pot 2 - Pin 12 - AN1 - controlling output on pin 5 - RC5
pot 3 - Pin 11 - AN2 - controlling output on pin 6 - RC4
pot 4 - Pin 10 - AN4 - controlling output on pin 7 - RC3

so I've set up SPWM_PIN as

SPWM_LIST  macro                    ; Define Pin's to use for SPWM
    SPWM_PIN  PORTC, 2, _DutyVar1   ; and the associated DutyCycle variables
    SPWM_PIN  PORTC, 5, _DutyVar2   ; Notice the underscore before variables
    SPWM_PIN  PORTC, 4, _DutyVar3
    SPWM_PIN  PORTC, 3, _DutyVar4

And have the main program as

Main:                               
 ADCIN 0, DUTYVAR1
 ADCIN 1, DUTYVAR2
 ADCIN 2, DUTYVAR3
 ADCIN 4, DUTYVAR4
 GOTO Main

I'm assuming the value of ANSEL is incorrect as its ommitting that high frequency noise. Given the above info, which I hope is clear, what would be the correct value for ANSEL, or have I made another boo boo in assignig the pins ?

By the way, prior to the change I had it running really well on CH1 - you thought that on a 12F @ 4mhz the loco was slow.... wait to you see the video for it running on a 16F @ 20 mhz !! - its just amazing !!

[malc-c]

Darrel,

Here is that video - http://www.micro-heli.co.uk/controller2.avi

I've left the audio element in so you guys can hear the growling the loco makes, the downside is that this has made the file 24MB

[Acetronics2]

No secret, Malc ...

PWM freq too low.
( may be also a poor quality motor too ... but you can't choose which will be used.)

Alain

[Archangel]

Tried both high (1000uf) and low (100uf) - the loco won't move
Tried, 1uf to 10uf. with a 1uf the growling was less noticable, but the slow speed control became less. Increasing this to 10uf and the loco had almost no ability to crawl.

4 - 8 kc whine? have you tried going even smaller? say .1 or .01 uf ? There is an outside chance you are not hearing the direct frequency in those motor coils but rather a beat frequency from the difference between 2 frequencies in which case you may be looking at bypassing RF.<br> I would bypass the motor from each power lead to the case too.

[malc-c]

Thanks for the suggestion, a tad over my head, but I think I follow your drift.

Darrel / Steve,
I've noticed that if I advance the pot fully (thus taking the A/D input pin to supply) and then turn it back down the loco stays at full speed, and the only way to correct this is to disconecct and reconnect the supply. - any suggestions (could this be down to the wrong setting for ANSEL)

[malc-c]

No secret, Malc ...

PWM freq too low.
Alain

The max setting I can use seems to be 190 for freqency, 255 for resolution, with an OSC value of 20

[malc-c]

I've been trying to phathom out what I need to set ANSEL to or whats wrong with the code to get this to work 100%.

Referring to my post #74, I have three PWM outputs, but when I connect pot 4 to pin 10 and the corresponding output (pin 7 RC3) the loco just shoots off at full speed with no control - Here's the code

@ __CONFIG  _HS_OSC & _WDT_ON & _PWRTE_ON & _MCLRE_OFF 


     ;ANSEL = %00010111    - dt suggested value
ANSEL=%00110001     ; original 
CMCON=7
TRISA=%11111111        'set PORTA as all input 
TRISC=%00000001        'set PORTC as all output apart from RC0


DEFINE OSC 20
CLEAR

DEFINE ADC_BITS 8	' Set number of bits in result
DEFINE ADC_CLOCK 3	' Set clock source (rc = 3)
DEFINE ADC_SAMPLEUS 50	' Set sampling time in microseconds

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

DEFINE SPWM_FREQ  190                ; SPWM Frequency
DEFINE SPWM_RES   255               ; SPWM Resolution

DutyVars   VAR BYTE[3]              ; DutyCycle Variables
  DutyVar1 VAR DutyVars[0]          ; group them in an array for easy access
  DutyVar2 VAR DutyVars[1]          ; with FOR loops etc.
  DutyVar3 VAR DutyVars[2]
  DutyVar4 VAR DutyVars[3]

ASM
SPWM_LIST  macro                    ; Define Pin's to use for SPWM
    SPWM_PIN  PORTC, 2, _DutyVar1   ; and the associated DutyCycle variables
    SPWM_PIN  PORTC, 5, _DutyVar2   ; Notice the underscore before variables
    SPWM_PIN  PORTC, 4, _DutyVar3
    SPWM_PIN  PORTC, 3, _DutyVar4
  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, DUTYVAR1
 ADCIN 1, DUTYVAR2
 ADCIN 2, DUTYVAR3
 ADCIN 4, DUTYVAR4
 GOTO Main

Here's hoping someone can point e in the right direction