Stanley Meyer technology entering the age of the computer

[Tecstatic]

Hi Bussi,

The parts are needed in order to reduce noise in the AD-conversion. So you can use jumpers instead, as we at the moment don't use the AD converter.

One more question: in your part list there is a 74VHC573, V means very fast. And V means difficult to get. Can it be substituted by 74HC573?

@all
Best wishes for the new year

greetings,
bussi04

Sorry no, if you want to have a low number of wait cycles during external memory cycles.
If you want to. you can study the external memory timing requirements in the AT90CAN128 data sheet.

If you use the slow 74HC573, you have to accept 3 wait cycles if I remember right, so it slows the program execution speed.

But no worries, you have 4K internal RAM memory, which is sufficient for some time. Let us use that first.

I guess you will need more stuff later, so maybe Santa will be eager to not being late in the new year

Happy new year to all, let it be the year of breakthrough.

Eric

[bussi04]

[bussi04]

more error description details

4x20 characters are filled after startup with cryptic chars.

but the last 3 chars in line 4 (=char 78 to char 80) are different:
when turning the digital pot char 79 changes between 3 B C E F, when switching left switch char 78 changes between 5 and 7. when switching right switch nothing happens. Cursor remains at char 80.

hope that helps a bit.

greetings,
bussi04

[Tecstatic]

Hi Bussi,

Good progress

What we have here is probably only a display problem, as the LEDs are flashing.

As I clock the shift registers with 1MHz, don't use more than 20cm cable between MB and SUI (at least not until it works).

Examine the SUI PCB, and check you have all vias mounted, that RS, R/!W, and
DB4...DB7 are not shorted to other tracks nearby, the same for the display adapter PCB.

Check for continuity from U2 pins to P6 pins on the display adapter PCB (with cable mounted between P5 (correct, yes) to P4.

When the display is running OK, then you can check the buttons more easily.


If you check the instruction set for the HD44780 LCD controller, you will see that all instructions needed for changing the interface from 8-bit to 4-bit has don't cares on bits 0..3.
You are correct about the connector, thanks.

As you have experienced no smoke or bad smell, use your Dragon to debug the program.
No display mounted.

In file main.cpp, function before line 54 insert:

while ( 1 )
{
   lcd_disp.SPI_MasterTransmit( 0x55 );
   delay_us(50);
   lcd_disp.SPI_MasterTransmit( 0xAA );
   delay_us(50);
}

If you have a scope, this is another way to test for non-shorted signals, if two adjacent pins are shorted you get a voltage (or two because of asymmetric drive capability)somewhere between "0" and "1" definitions with double duration.

If you don't have a scope, consider the digital storage oscilloscope I gave a link to some previously. Although I have a nice 350MHz scope, I bought the 32 euro scope. and put a BNC connector on the scope. The probe will be at least as expensive, but it is hard to miss.

This small inexpensive scope works well up to 300kHz, and would be fine for this error checking, especially as we are in control of the program ( ) so we can slow it down to where the scope works perfectly. If the supplied junk op-amps (for this use) are changed to adequate specked types, then it will be OK to 1MHz as claimed.

The other frequency counter functionality works quite well up to 8MHz (5MHz claimed).
If the crystal is accurate, the frequency counter is accurate. I have checked the counter function with my HP frequency counter.

IMHO this little scope is good value for a low budget.

You can also put two breakpoints on the two delay_us(50) lines to stop the program, so you have all the time you need to measure the pin voltages.

Be alerted if you see a logic voltage between 0.8V and 3V, that is no good.

You can also in the not modified program add a breakpoint (click the Grey area next to the line number) at the beginning and end of:

file lcd_20x4.cpp, write_display4()
on lines 287 and 308.

Add a check mark for Debug->debugging windows->watches

Place the new "Watches" window in the lower right corner by click and drag the top bar of the window.

Click the triangle next to "Function Arguments" so you can watch the data argument when the program stops in the first breakpoint.

Let the program continue to the next breakpoint. Now you should have the data on SUI U2 pins.

Maybe your display has a slower clock frequency than mine, in

file lcd_20x4.cpp, function InitDisplay()

Change all delay_us( 100 ) to delay_us( 200 ).

Good luck

Eric
PS. I assume you have taken ESD (Electrostatic Discharge) precautions so you have not "zapped" anything, at least grounded your solder iron. The bad thing about a discharge is that sometimes "nothing" happens at first, but later the IC fails.

[bussi04]

[Tecstatic]

Hi Bussi,

Often it is the most basic that fails, as it is so "obvious"

Try exercising the buttons while monitoring the logic levels on SUI U1 pins.

Maybe you get it to work this year

I will check with intervals if you post more the next hours.

Happy new year

Now I will have a toast with my wife.

Eric

[bussi04]

success!!!

Hi Bussi,

Often it is the most basic that fails, as it is so "obvious"

Try exercising the buttons while monitoring the logic levels on SUI U1 pins.

Maybe you get it to work this year

I will check with intervals if you post more the next hours.

Happy new year

Now I will have a toast with my wife.

Eric

[Tecstatic]

Hi Bussi,

CONGRATULATIONS

You probably have it right, The numbers I mentioned was just an enumeration of the possible numbers, not the sequence. Maybe you still have a short between some button signals, maybe by a not precisely positioned resistor.

To me it looks like you switched left and right digits in some way, try to switch the digits, e.g. the number "FC" to "CF" and what do you get then ?

If you take the sequence numbers and draw the two bits like pictured on a scope for the rotary encoder, then you should see two square wave signal with one offset 90 degrees to the other.

This is the way you detect incremental positioning. It is the same if you put another type of rotary encoder on a pulse motor, then you know with great precision where in the rotation the rotor is, and from the program you can then fire the coils with signals of your choice, not dependent on slow opto forks and cut holes for timing the light, where you need a new timing plate each time you want to try another timing.

When we get so far we can do that by modifying a number on the display with one of the rotary buttons, and observe the change live. We can even make a setup with an miniature of my 1400Nm electric dyno brake and have a program optimize the motor timing for some desired property, e.g max. torque or max "efficiency".

But for now I have to solve the last problems with downgrading my SW from very flexible graphics to alpha numeric. I'm not used to downgrading SW, normally functionality is ever growing.

As I'm bound to some other activities until next week, you will get time for your freedom board. (have to see the positive side ).

You have now passed a lot of the learning curve to use a uC for development.

Let 2010 be a good year with successful circuits

Best wishes from my wife also.

Eric

[Tecstatic]

Now I got more time, although not as much as I would like for doing work here.

I wonder if it would be interesting, if I make a small PCB circuit with the following:

1. TO220 or TO247 case N-channel MOSFET, e.g. IPW60R045
2. a MOSFET driver IC, the TC4420
3. a current limiting circuit, which protects the MOSFET, and has an "overcurrent" logic output.
4. A fixed (by resistors) or programmable (by PWM) current limit.

This small PCB could ease the application of MOSFETS.

If wanted by some I will do that first.

Although a MOSFET gate is insulated at DC, then it is another case with AC, which can require a large but short duration gate current. The TC4420 can deliver 6A, which is what come in handy to drive the gate capacitance for fast switching.

Eric

[bussi04]

Hi Tecstatic,
your inbox for PMs seems to be full. Please clean up so that I can address you.
greetings,
bussi04

[Tecstatic]

@bussi

Sorry for that, not full now.

Eric

[Tecstatic]

Find 5 applications (at least)

Hi everyone,

Today I designed a power stage circuit with a MOSFET.

Oops, accidentally I "dropped" in a uC also, which can be connected to a SUI.

This circuit has a bit in common with the "find 5 errors" pictures.

Here some components can be omitted (even the uC) or shorted. By doing this the circuit can be used for several applications.

The PCB has not been made yet, so your ideas could still have a chance of being included.

This circuit is meant to be simple, inexpensive, flexible and easy to assemble, having a PCB.

How many applications can you find for this circuit ?

Please step forward and tell what you think it can be used for.

You do not need to find more than one new application of the circuit to post.

Eric

[bussi04]

...
The PCB has not been made yet, so your ideas could still have a chance of being included.
...

Eric

[bussi04]

[Tecstatic]

As all PCBs are different, then I have had no succes reusing the UI layout even though it was the same diagram. No need to do the same over and over again, so I made a separate SUI PCB.

Let me start saying, that I published the diagram as soon as the last line was drawn, no "afterwork" to remove errors in the detail, just the big picture.

You found the GND label was missing

So you got two applications for the SPS which includes the uC, and they are not the simplest ones.

Good thinking

Before we are home safe for this intended application, there are a few details, which need attention. So do not consider the diagram as the final solution.

I have not included more MOSFETS as I want the PCB to be small, as the cost will be low, it could be better to consider this a LEGO block, and build something using more than one PCB with or without uC on them, just mounting the components needed. More components can aways be added later on if needed.

I have updated the diagram in the post above with the details corrected so far.

Having a PNP or NPN transistor with the base as the pin in the middle, do you think you could mount the transistor + wires into connector P4 or P5 ?

Could that be used as the EEC transistor ?

Also could a thyristor be connected to one of the opto couplers, and what other applications could that be used for ?

Take a look at uC pin12 and 11. Thats the inputs for a comparator. one of the comparator inputs can be an internal voltage reference, so only one pin is needed, and pins are a limited resource on this second smallest pin number package in the AVR family.

Note the option of using a fixed current limit (not programmable) to free the I_LIM pin for other uses.

Remember we have a phase also in the PWM output signal, and the tiny84 ADC (analog to digital converter has a multiplexer which can route one of the pins in PORTA to the analog comparator pin AIN1 if we don't use the ADC.

However as other applications could use both AIN0 and AIN1 I have extended the circuit.

bussi, could you make a diagram showing the SPS as an IC with pins having the naming convention <connector> + "_" + <pin number>, e.g. P4_1.

Then put the outside connections to VIC power supply etc. Then it should become more clear how the VIC control is implemented.

Same for the resonant heater.

If we have two identical VICs and we have current limit control on one, then we have almost protected the other in normal use.

Unfortunately I'm out of time, but I will follow up later.

I have also considered a few posts to demystify some of the inner workings of a uC, And for this I think the tiny84 is a good choice, more can be added later for the larger uCs.
Crawl before waking.

The datasheet for the tiny84:

http://www.atmel.com/dyn/resources/prod_documents/doc8006.pdf

To sum up the applications so far:

1. VIC + EEC controller
2. VIC resonant heater

What about other apps, e.g. Bedini and Hector stuff.

Eric