Way OT!!!! How to kill some mV leakage? Trust me it's OT!

[BillyJ]

I have a circuit that leaks some millivolts when it shouldn't.
I need it to be 0volts most of the time except when there is a triggered 1/2 volt running through it. When this happens that voltage needs to get through.
Is there an easy way to bleed off this small voltage and let the larger one through when needed?

I hate to ask here cause this is for a downhill skateboard timer but I do not know a better place.
Maybe someone can recommend a forum where this will be welcomed?
Maybe some cool skaters want to help me make this thing work?
I am getting the shaft because I made some boards for a guy exactly as he wanted them. I have had to correct a number of his mistakes already and I am getting really bummed.
He expects me to fix his design. I am getting no pay for this bit of engineering but he is holding out on the 50 bucks he hwas to pay me for the boards.
I am about one step away from smashing these and telling the guy to bugger off for good.
If I was a violent guy....but I am not...
I would rather just make these dumb ass things work than explain to the guy why he sucks.
He has one that works but has been not letting me borrow it so I can see what it is doing. I think he @#$%ed up in the schematic and is trying to bullshit me now.
He has only given me a mostly hand drawn schematic.
I have checked the layout time and again and I cannot for the life of me see anything wrong. I need someone to verify.
I happen to think the guy is jerking me around but all I need to do is make these things stop leaking and I am done and can get back to my hobby and forget this very bad headache!!!
Please help me. This thing is killing all my joy!!

Please forgive me for the way offtopic post. I don't know anywhere else to go with this.....

[Mike Burgundy]

let's start with seeing the circuit, that'll give a lot of people a more clear idea of where to go.
What are you switching with? TLL and such logic leaks, so you might want to consider sticking in a MOSFET as a switch between 0 and some kind of +.

[Rob Strand]

Presumably you have done a layout and made some boards for a circuit designed by someone else?

If you can *prove* to yourself there are no layout bugs then argue that with the guy, end of story.  You are absolutely crazy solving his problem to cover your butt - it's not your problem.  *You* are already looking for enginnering solutions to *his* problem, otherwise you would be only checking your layout to prove your part - you job doesn't entail engineering.  Ideally both parties should work together to solve the problem and converge to a solution and pay accordingly.  However, if the guy won't make attempts to solve the problem at his end and wants to pass the buck then forget it, don't deal with him - you can't work with people like that, it never works.  If you want to get involved in this then I would tell him up front that if you have to do engineering to fix *his* problems then you will charge him for engineering time and you want 50% up front payment for that time.  I'm not being hard nosed about this  it's actually the proper way to divide the work and the problem.  I do this stuff for a living and believe me you don't want to go where you are going, the way to do it is to spell out the situation to the customer and if an agreement  cannot be reached end it and keep the stuff you did - it's yours until it's paid for in full.

[Paul Perry (Frostwave)]

I can't say anything without seeing the circuit (or the part at fault) but, if the only problem is the output needs to come down a few mv, perhaps you could jack up the whole circuit (with a bypassed cap in the ground cnnection) then tap off along a resistor from output to "true ground", like  getting bias in a valve amp.
What problem is the 'few v' causing?

[gez]

I have to agree with Rob, but if you do feel compelled to fix this thing, try and incorportate some built in obsolescence (just enough to get your money first!)  

[Peter Snowberg]

I agree with Rob 100% too. I've been in your shoes and I know it's no fun. I won't develop anything or do any engineering work anymore without a contract up front.

If he won't provide you with a proper schematic he's blowing it. If your layout matches the hand drawn schematic he gave you then the question is how badly do you need that $50.

Since you said you've already corrected some mistakes you've ventured into a hybrid land of your work with his.    Now it's hard to be sure of anything and you can't point to something that is 100% his design.

People always yell at me for this, but I ususally try to muddle through and make sure at the end I'm not going to catch flack for something that doesn't work; even if it's not my fault. I usually learn along the way too.

Post the schematic and I'm sure a conclusion can be found. If you don't have a web site, you can get a free one with almost zero advertising at http://www.freewebs.com/

-Peter

[Paul Perry (Frostwave)]

If you have a reed switch, ther eisn't going to be any leakage.. sometimes a quasi-mechanical solution is best!
But if you are trying to do some kind of timing by integrating a voltage, then you have plenty of other problems, believe me!!

[BillyJ]

This afternoon I will draw up the schematic and get it posted for you all to look at.

Please note too!!!!
I am having trouble with my meter too which is why I think this happened and I didn't catch it.
My analog meter doesn't read any mV leakage at all!!!!
I have a digital one but it reads mV in the air?
It also seems to be completely inaccurate. I can get a reading, remove the probe and the reading stays. Use the probe again and the reading will jump up again. I can pump this reading like that all day.
That cannot be corrcect can it?
I have no idea how I will deal with this with a meter I cannot even trust.....rats!!!....

The circuit is to be triggered by a tapeswitch which is run over by the wheels.  But the leakage is at the other end of the circuit where it is to be coupled with the stopwatch on/off switch.
The leakage is causing the stopwatch to not be able to switch as it is seeing some voltage all the time. I guess the stopwatch needs to see 0V between switchings.
The reason this circuit needed to be built was to turn on or off the circuit with one set of wheels. The problem is one set turns the timer off but the second set turns it back on.
The circuit uses a 555 timer set to put out voltage for somewhere in the 2 second range. This is long enough for the second set of wheels to pass over the tape switch and not 'switch' it again.
That part seem to work ok.
Next comes a 3041 optoisolator which is supposed (as far as I have been made to believe) to have 0 volts until triggered and then it is supposed to let out around .5V.
It lets out the .5 V but leaks mV.
The pins of this opto are suppose to be hooked up to the watch switch.
One of the mistake I have corrected already were the pinout on the opto.
He drew some pinout that was whack. I made one like that and of course it did not work. I looked at the working model and sure enough he drew a bogus pinout.  There were other corrections, small but annoying.
I am a novice. I can follow a schematic but engineering.....
I am so bad at dealing with people in this manner but I have no choice but to get walked on.
I will get the schematic online by tonight and maybe with a little help from you folks I will be able to show this guy that I only did what he drew.
Or crawl under a rock.......I don't mind being wrong....It would be better than to feel what I feel right now.

I will get this schematic up and then at least I will be in a better place to deal with whatever I must deal with.

I assumed I could put a diode on wherever the mV leakage was and thought that the voltage drop would soak it up.
I can't prove myself wrong or right (though I must be wrong) because of these damn meters...
I am so sorry for dragging this here.
I promise I will never mess with anything not 100% audio fun!!!!
Want to time your skating? Have a car follow you ;O)
Thanks everyone you are truly the coolest around!!!
Check back tonight fopr the schematic!!
I thank you!!!!

[Paul Perry (Frostwave)]

A problem with digital meters is that, because the imput impedance is incredibly high, the probe can get 'charged up' and read when it isn't connected.
And on the AC range, it is very confusing because there is enough stray capacitance between transformer windings, that it can give a reading on a broken or isolated lead!
Easily fixed by putting say a 1 meg resistor across the probes.

[BillyJ]

Ok here it is.



If that didn't work here is a url
http://www.freewebs.com/billyj/timerschematic.htm


Thank you all so very much!!!!
If any of it is unclear please ask or point it out to me.

[Paul Perry (Frostwave)]

For a start, you want to use the emitter and collector as output to control the stopwatch (assuming the stiopwatch can be controlled by shorting to ground).

[Peter Snowberg]

Okay.... now for some analysis. It sure helps to see things.

The first question I have is about the opto.

You show connections between pins 4 and 6. You should actually be connected between 4 and 5 (collector and emitter).  Leave pin 6 floating.

Hooking up to other circuits like a stopwatch can be tricky when you're using bipolar transistors. You have to know something about how the stopwatch works. Polarity is important, but most importantly, the battery powered stopwatch may not like the additional 0.5 volts of forward voltage drop in the opto transistor.

When that opto is off, you should see no continuity between pins 4 and 5. When the LED is on, pins 4 and 5 act like they have a diode across them. There are also optos that use FETs for output instead of bipolars. These act like a small resistor when on instead of a diode. This makes interfacing to some battery circuits easier. Anyway....

You should be able to easily test the opto by removing the resistor to pin 3 and connecting a 1K resistor between +9V and the anode of the opto. I would use the 1K instead of the 470 ohm off pin 3 too. If the stopwatch doesn't trigger, try reversing pins 4 and 5.

Assuming that now works, add another 1K resistor and red LED in parallel with the opto so you can monitor the output pulse and reconnect pin 3.

I see a small problem with the component values. The 100uF capacitor is a bit big for the 555 to discharge with the trigger pulse. Try using 10uF and increasing the charging resistor (to maybe 470K?) to keep the timing constant the same. You want to use the smallest capacitor you can. Maybe 4.7uF and a 1M? Just a guess. Try a 1M pot with a 10K safety resistor in series and substitute cap values to test.

The other thing you could do would be to increase the trigger pullup from 10K to maybe 47K, and then maybe even add a switching transistor like a 2N2222 or a 2N3904 to amplify the trigger pulse. I don't know how your switch works so I'm assuming here. First you need to know if the problem is with your signal conditioning, or the output switch.

-Peter

[BillyJ]

Cool thanks everyone. I am going to try these things today as soon as I can.
Here are some other things that may help things.

The other party gave me a layout from somewhere online. It is blurry so he had to draw in thre values of the R's. The cap values can be read.
There was no opto originaly. This is his addition.
The thing that is bugging me is he has a working circuit.
Of course I have yet to have this in my hands to wrok from so:
I am not sure he has a 3041 in his. He has said he bought some optos froma place that had onluy a few left. He gave me a sk2042 to mess with.
He has said that the one he gave me and the one in the working model are different but that a 3041 was going to be the correct one for this circuit.
The sk2042 does not seem to work any better in this circuit.
I have tried pins 4 and 5 but the same problem exists.
Apperently the stopwatch can take the 1/2 volt but not whatever the output was before the opto.
I wonder Peter if the one he has is uses FETs for output like you mention?
But still I am told the half volt is fine but the leakage when off is the issue.
If his circuit really works and my layout is correct than what is missing is some value or a different opto (I suspect this).

I will start in these things ASAP and let you all know what happens.
If anyone thinks the opto is suspect please share your thoughts.
I suspect that he has some generic opto that may or may not be a replacement for the 3041. I think this for whatever reason is working and that he was simply told the wrong thing for a replacement part or something along these lines.
He has bneen a total weenie about letting me have this working model because it is in use all the time......
I want to be armed well before I go to him and open his one and start pointing to things that are different than what he told me.

Anyway I need to get some real work done and then I will get started on this and get back with results.


Thanks again everyone, you are the best ever!!!

[BillyJ]

First of all let me say that when this is sorted I will be dropping some Paypal funds Arons way..... Thank you Aron, not just for letting get help with this super offtopic thing but just for having this place. The coolest!!


Paul..With a 1m resistor across the leads I no longer read any mA's?
Does this mean they aren't there?
If so, does this mean the circuit functions properly?
If not, could this be used to fudge it to work?

The other party said that his device shows no mA leakage but mine does using his meter........?
If it is false wouldn't his detect some leakage on both circuits?

I was thinking again maybe his opto is using a FET output like Paul said and therfore his doesn't show up because it is acting like there is a resistor across the pins....but Paul said small resistor and the 1M is large....

I will try the other suggestions next but I am still confused if the 1m across the leads of my meter are helping me to see there is no leakage or is it masking it?


Thanks!!

[Paul Perry (Frostwave)]

with any troubleshooting, it has to be broken down to different parts of the circuit, to be sure what is wrong.
If the circuit is working fine up a far as the opto input, then the problem must be the relationship between the opto and the watch.
I just have no idea whatever as to what the watch is expecting (are there any specs or a description of this anywhere? is it a kit?)
Putting a 1 meg resistor across the leads of a meter won't interfere with current readings, but leakage isn't  a problem ever with current readings.
It's only voltage readings that can be a problem.

[Rob Strand]

I still can't get the schematic.  Seems like that free site is one which comes and goes - like many of them!

[BillyJ]

Rats! I am sorry I meant mV NOT mA.
Could you reanswer with that in mind Paul.
Gosh what a nightmare I fallen into.

The watch wants to see the half volt when the switch is activated. After that the watch wants to have 0 voltage through it.

That is what I am being told mind you. I have not been been supplied with a watch to perform any testing on.
I am only going with what he told me and that is..
His circuit does not have any volts when just sitting there.
Mine do, and therefore will not work.....
He has not even stated to me in so many terms that it has been tested with the watch. I asked to try but he hasn't said he did. Only that these will not work with volts leaking.

[Peter Snowberg]

I can't pull the site up right now either. Oh well... you get what you pay for.

That stopwatch is almost surely CMOS while that opto is bipolar. The difference is that CMOS circuits are triggered by voltage and bipolar circuits are triggered by current. That may be a real mess to get working because of the leakage current through the opto. I'm guessing he does have a different optocoupler. Some are known as solid state relays and one of those will almost unquestionably work. I looked but could not find any data on a SK2042.

Why not get the number from the working box? without the opto everything else is moot.

A simple solution is to replace the opto with a 2N3904 or any small NPN (2N2222 etc.), a diode, and a tiny reed relay. Change the pin 3 resistor to 10K and connect that to the base, ground the emitter, and connect the collector to the relay coil and the un-banded side of the diode. Connect the other end of the relay coil and banded end of the diode together, and connect them to +9V. Presto; one electromechanical isolator.

-Peter

[BillyJ]

Sorry folks I blew it.

This circuit isn't leaking current.
That was a mistake on my part.
It is only leaking Voltage....

[Peter Snowberg]

Does that mean it works now?