Solved: Please Help (Fried Pins) I think i am slowly killing my STE

[dml]

Yes although the resistor values are wrong, especially the one on the left, for 22v. It will toast the IC.

The layout is ok though - for the simplest version of it.

[AtariOwl]

THe 22v shouldnt touch it.
Its not going to be connected across the rails. I still intend to use the GND rail and the separate rail that runs necxt to the GND which i would feed with 5v



You can just about see the separate rail.

I assume we probably need to have some capacitors etc to protect the 4N25 ?

[dml]

Yep ok - but I'd still calculate the resistor from the optoisolator datasheet vs 5v. Don't just copy something from the internet without checking stuff.

e.g. for 10mA to the optoisolator @ 5V:

5V / 0.010A = 500ohm.

But use the mA rating from the actual datasheet - somewhere between switching threshold and maximum rating.

The optoisolator should cope with typical noise from the motors but adding a small cap across it should help with noisy switching. Too large a value will suppress or prevent detection though.

If switching is still flaky (getting multiple on/off signals for a single event) you can debounce it by adding more bits to filter the switching. But see if it works at all first.

[AtariOwl]

Hi Doug

You are very right.

I now have a cheap oscilloscope on its way. I think this will help me check things a lot better than my simple DVM.

I can then approach this in a much less panicky way.
I have some caps of varying sizes now (up to 10uF) i can experiment with those too.
Also I now have some PC817s as well as 4n25s

[AtariOwl]

Hi Everyone

Well after a series of setbacks, fried HDs, dead laptops etc. None connected with this issue, i returned to the problem at hand.

Regarding the timing i have a system working now using Opto Isolators

Specifically...

I am using


Where
A is 470 Ohm
B is 1K8 Ohm
C is 1K8 Ohm

This seems to be running really well, very solid, but my numbers are somewhat arbitrary, how do i figure out what the best resistors will be? I've been trying to keep the numbers up to reduce current. But i just dont know.


OK now onto the Transformers


I'm not at all sure how to attach them across the rails without short circuiting things?
Plus its DC, so i'm not sure what to do in that respect, i think in fact instead of being proper DC its a half wave thing which might mean we have effectively AC to transform, but i'm not sure.
I've torn apart an old dead VCR for coils so hopefully i have something to work with.

All advice gratefully received.

[AtariOwl]

Double post

[AtariOwl]

I switched to linear opto isolators and for a while that seemed to work, but now it is not.

In essence, what is happening is that the IL300 seems to be deadening the input of the LM741

The input goes into the LM741 just fine and comes out with the waveform i expected. A DC offset with half sine on top.

This is all fine until i connect it to the IL300, then instead of transferring the signal across as it did yesterday it deadens it to a slightly lower DC.

The same think happens with all my IL300s and did not happen yesterday.
I have checked my circuit several times but the problem remains.



Any ideas?

[dml]

Hi,

The optoisolator input is basically a LED, which is basically a diode.

A diode has a fixed forward voltage. i.e. it acts like a clamp.

So if the LED has a forward voltage of e.g. 1.5V, then applying 1.2V will result in 1.2V across the diode and ZERO current flowing. It's not enough to meet the diode's forward voltage.

Applying exactly 1.5V will produce 1.5V across the diode, with still no current flowing.

Applying 3V will still produce 1.5V across the diode, plus a *lot* of current flowing.

So applying signals to a LED will turn the (excess) voltage waveform into a current waveform, and clip the voltage waveform flat. Which is what you are seeing.

The fact your signal is completely flat suggests the DC offset is above the LED forward voltage, and the LED is therefore always on. You need a lower offset voltage to get the LED to turn off in the 'troughs'.

e.g you have:


/\/\/\/\/\/ 2v-5v (led always on, voltage always clipped to 1.5v by diode)

------------- 0v


...when you want more like...


/\/\/\/\/\/ 0v-3v (led only on when voltage > 1.5v, and is clipped at 1.5v)
------------- 0v


These are example numbers only - I don't know the forward voltage of the IL300 LEDS, or the output volage range of your circuit. But this is how the two things would behave. Diodes are clamps, which turn the excess (clamped) part directly into current (and/or smoke = DEDs) which is why a small resistor is always needed to limit the current.

Also be aware opamps have limited current output -LM741 considers >25mA as a short circuit. This is probably ok as it is, but driving anything more current-hungry than a LED+100R could fault the LM741.

D

[AtariOwl]

Hi Doug

As always your help is invaluable and i am making much progress.

I am now generating a very nice response from the system however once i connect the input to the EJP it goes wrong.

Is this because the current is too small from the LM741?

When i do not connect to the ADC i get a nice smooth line (when it is off i am running from 5v down)

When the car runs i get a nice smooth have sine DOWN from this line which is NOT clipped. I am having difficulty getting a screen shot at the same time as this. I am happy with it though

When i do Connect to the ADC i see this,


And if I have a cap across this to smooth it

[AtariOwl]

OK and running with the throttle on and a car on track



Connecting to the EJP does something odd to teh signal as you can see

[dml]

I am now generating a very nice response from the system however once i connect the input to the EJP it goes wrong.

Is this because the current is too small from the LM741?

It's an optoisolator so the input side shouldn't be influenced by the output side (or vice versa). There's an 'air' gap between the LED and the photodiode.

However if you have connected both GNDs together then you might get some issues depending on the rest of the circuit. Keep the ground and power lines isolated at both ends. Hopefully you're not *taking* power from the computer somehow. Also you should probably have a resistor in series with the computer/EJP side of the circuit.

BTW I assume EJP = joystick port?

D

[dml]

BTW a smoothing capacitor isn't always a great way to debounce switching noise because it reduces amplitude (filters). A better way is to use something like a Schmitt trigger IC for digital, or an opamp arranged as a comparator (or a dedicated comparator IC like LM397) for analog circuits.

These accept fuzzy signals in, and put something like digital out. They has some hysteresis so there's a region in the middle which doesn't trip easily - it requires a slightly higher voltage to turn 'on' and a lower voltage to turn back 'off' with a cushion between them. This turns messy switching signals into nice square output.

I wouldn't complicate matters with this just now but I'd look into it after you sort out the other issues.

[dml]

I've been looking at your 741 circuit and realise I don't know what its meant to do. I thought it might be configured as a comparator (wobbly in, digital out) with a resistor-configurable threshold but it's not that.

It looks more like its using feedback from one photodiode to regulate activity to the LED. Like a current regulator or something. Is it there because the supply voltage for the LED is really unstable?

Where did you find the diagram for this and how is it explained?

[AtariOwl]

umm

Those are good questions

It came from a place called mikrocontroller.net as a Unipolar Photovoltaic Amplifier. It contained an Op Amp on the output but i was already getting levels i liked from the opto isolator and i havent included it yet.

This is what i am NOW running...


I fixed the diagram

The output of the IL300 is connected to the 5V and GND of the Joystick port. I Assumed maybe wrongly that it would work like a potentiometer for an analogue joystick.

The other GND and 5V are entirely independent of the STE

The Signal i am watching is between pin 9 and 5
EJP is Enhanced Joystick Port

My scope is between pin 9 and the point between R4 and Pin 5 so that i can still monitor if pin5 is connected or not
If i do not connect Pin 5 i get the to two traces
If i DO connect pin 5 i get the bottom two


The 741 seems to to have something to do with the IL300, apparently the matched photo diodes at 3/4 and 5/6 keep it stable, feeding back from the 3/4 photodiode ... i would imagine its a little sensitive because its a linear coupler instead of a simple on/off thing.



I should put the other opamp in shouldnt i? Instead of the cap?

[AtariOwl]

OK New circuit with buffer OpAmp instead of cap


Different results

[AtariOwl]

Hmmmmmmm
Well i changed my circuit slightly and the response is much cleaner but the ADC is not giving teh correct values.
If i connect the +5v to pin5 (via a resistor of course (I'm not crazy) i get about 5 from the ADC, if i dont connect anything i get about 140, if i connect GND i'm still getting 140 hmmmm

[AtariOwl]

It is solved

Thank you so much for your help.

Essentially the ADC in the EJP cannot be used as a voltage follower, in essence what it does is measure resistance between pin5 or 15 and pin 7 (i've been writing 6 i think but it is 7)

Therefore i do not need a buffer op amp. By connecting the STE to pins 5 and 6 of the IL300 i get a signal which covers 70% of the range (values 40-220)

This is perfect for my purposes and now keeps my STE fully isolated from the track.. and safe.

Cheers.

[piku]

I know, rising the thread from the dead... You have nerves of steel! You are experimenting on a $300+ mostly irreplaceable machine when you could instead get a $5 microcontroller, build up the circuit perfectly and then connect it with serial as someone else suggested. Sorry to not offer practical help but..

[AtariOwl]

That's OK. You're very entitled to say it and really i am NOT someone with steely nerves... i didn't really understand the dangers before i started and if i had i might not have taken the risks i ultimately did. I also didn't realise how valuable STEs had become. I picked up mine for £15 from a car boot sale years ago.

Thankfully, it's all ok now anyway.

I have the timing part working and the throttle following part working and a part to control an external counter which switches diodes and resistors in and out of the power circuit... all isolated from the STE with opto isolators. There is no part of the power circuit or the timing circuit which shares any direct connection with the STE. All connections are opto-isolated.

I also have LEDs running from the STE for the pits and Start Gantry.

All from the original STE.. admittedly i did kill 3 pins on port A but i have killed no more pins since i started the thread and now everything is working fine from Port B ... doing everything i set out to do (well the sound is set up but its not really doing anything though it will)... and more.

Plus of course i can run the timing, lap counting, position monitoring, lap by lap charts, power limitation to simulate fuel load and tyre wear, etc in real time on the STE. I hope ultimately to add sampled sound also.


At least i wasn't using my Falcon

[Foxie]

I also didn't realise how valuable STEs had become. I picked up mine for £15 from a car boot sale years ago.

This surprises me too. I opened a thread about it here: viewtopic.php?f=30&t=32962

All from the original STE.. admittedly i did kill 3 pins on port A but i have killed no more pins since i started the thread and now everything is working fine from Port B

The good news is that it should be easy to fix. The joystick circuit consists of three 74LS244s (for the digital inputs), one 74LS373 (for the row select outputs) and two LM556 (for the analogue inputs). Those parts are very cheap and easy to find, and they're through-hole so easy to replace. I suspect you've blown one of the 244s. That really takes some doing, because TTL chips are hard as nails!

[AtariOwl]

yes .. i have blown the Voltage on pins 11,12 and 13 of port A (they have all dropped from 4.95V to about 1.2V) i suspect one 244 covers all 4 on the bottom row.

So currently i'm
1 - controlling the LEDs from teh parallel port
2 - controlling the output power with pins 1-4 on EJP port B
3 - reading the throttle position with pins 7, 5 and 15 EJP port B (connecting the input pin and 7 across an IL300 works very well)
4 - reading the timing onpins 11-14 on EJP port B

At some point i'll fix pins 11-13 on port A