Re: Motorcycle flasher problem.

On Fri, 28 Dec 2007 08:30:46 -0800 (PST), steve
<stevesemple@xxxxxxxxx> wrote:

Well I'm back with some results Bob. There seems to have been a bit of
discussion in my absence. Unfortunetely I'm not Electronically
intellegent enought to add anything into any of the debate. However I
can tell you my results on the tests you suggested on your circuit.

I have put them in context of your text below. I have done 7 tests.
Please see them numbered.


Go right to the bottom of the my post - that should get you up and
running right away

You show around 2 volts on the collector - that is not enough voltage
to turn on the mosfet - you want 4 or more volts there.

With it turned off (lamp in the off state), the voltage at the emitter
should rise to greater than 4 volts, the voltage at the base should be
clamped to around 3-4 volts - if you can remove the transistor - where
the emitter connects the voltage should be ~12+ volts,

** 1. Connector 14.70volts (basically the AC adapter output) So this
seems to be OK.

the base should be 3-4 volts,

** 2. Base 1.6 -- This is too low based on what you are saying.
However I tried two other transistors one the exact same and another
that has a little lower Hfe (because I had it on hand) and it really
didn't change.
The base should only be 3-4 volts with the transistor out of the
circuit - otherwise it will vary.

That reading (if indeed with the transistor out) is wrong. So that's
where I'd look first (I'm assuming the cap charged to ~ 14 volts in
the previous test). That part of the circuit (sans transistor) is
nothing more than a resistor feeding a zener and establishing a
voltage reference. Only two parts - not much to go wrong. A resistor
that is too high or zener that is leaky or wrong value will establish
a wrong voltage there.

There are two other ways to establish a voltage there other than a
zener if you don't have the right one or spare

1. Use a simple voltage divider instead (the 220K will be changed for
a 4.7 K and zener will be replaced with a 1.5 K) circuit operation is
virtually the same with some degradation in flashing rate tracking the
power supply voltage (rev the motor flashes faster) With a 14V
supply that should be a little over 3 volts with no transistor in the
circuit.

2. A string of 6 series connected forward biased silicon diodes to
replace the single zener diode (arrows or body bands pointing to
ground - backwards from the single zener) - each diode gives a drop of
..6 volts and six should be 3.6 volts (again, sans transistor).

These are ways to substitute the zener and establish the 3-4 volts the
base connection requires - and the circuit should work that way when
the transistor is replaced (but check to see that there is 3-4 volts
there first to save time and effort)


and the collector (with no transistor in the circuit), should be 0 V

** 3. Collector is 0.0

I'd suspect a bad PNP transistor (or wrong transistor or connected
wrong).  Or just a transistor with poor gain - the gain should be on
the order of 100-200 in that circuit.

** 4. I tested the transistor using the little thing on my Volt meter
I get HFE 189

That's good

Second suspect is the timing cap being leaky - in order to get 4 volts
out you have to have >4 at the emitter - and you don't.  With the
transistor removed if you got 12+ Volts there after about a second for
the cap to charge - the cap is good.

** 5. I tried this, removed transistor and I get voltage of >14 volts.

OK that clears up my other concern - you DO have ~14 volts at the +
terminal of the cap.
I try to discharge the cap and then try it again and it seems to very
quickly go from low 0 to >14volts.
So that seems OK. I changed the cap just in case. Same result.

Yes that's what it should be doing
The other thing that is likely cause, assuming the other things
checked out, is if the current to the gate of the mosfet is way too
high.  That would pull down the other voltages.  


A good mosfet has almost no gate current - the gate is an insulator -
that 100 K gate to ground is for the purpose of discharging the gate
capacitance.  That has to be 100 K or so, you don't want a lower value
there, so check that.  (band next to the gold or silver is yellow)

** 6. Tested all resistor they are all right.

To check the mosfet - disconnect the gate (and you may want to
temporarily  ground it so you don't destroy it with static) - with the
gate disconnected, the voltage at the collector of the PNP should rise
to >4 volts - if it does and does not when connected, the mosfet has
already encountered static.

** 7. I get about 1.88volts. Interestingly the Light dimmly lights up
untill I put my probe on the gate. I also tried to test the mosfet
using the following ...

Well that 1.88 should be greater than 4 volts up to a maximum of
around 7 volts.

With the gate of the mosfet floating in air it is a very high
impedance (very sensitive to voltage) input and will pick up stray
radio waves and power line frequency floating in the air (especially
with a test lead on it) and that will cause the lamp to come on and
off too rapidly to see and you see it glow dim.

http://www.uoguelph.ca/~antoon/gadgets/mostest.htm
It seems to test ok.

Im sorry to be such a dolt on this. Its a simple circuit and yet there
is something Ive done that just isnt right. By the way on my original
schematic I put in the orientation of the transistor and Triac are
they right?
I assume they are right that's how mine are based (some manufacturers
can vary the basing and still give them the same part number but that
is very rare) - the other convention for that style transistor is
E-C-B (but if it already checked in your meter with a gain of 189 -
you had it plugged into the meter with the correct orientation (!)
The meter test also checks the gender of the transistor in case a NPN
snuck into your parts bin. (both of these things you should be in the
habit of checking - gender and base connections when using a
transistor tester, just in case)

eg Transistor 2N3906 Holding legs from bottom legs pointing down,
looking at flat part of TR. with round part at the back EBC
Mosfet IRFZ34 Holding legs from bottom Looking at "front" of mosfet
Heatsink to top
GDS.

Yes that's right. Without the four volts necessary to start the
mosfet conducting the circuit cannot work - so that's what you have to
get working - the PNP transistor should be supplying that voltage.

Once that emitter (that has no problem charging to 14 volts with the
transistor out) exceeds the base voltage by point six volts - that
should put about 4-7 volts on the collector (the two 100 K resistors
are a voltage divider that won't allow the voltage to get higher than
7 volts - but at 4 volts it should start flashing already.)

I have replaced the Mosfet and as I mentioned above exchanged the
transistor, tested all resistors to make sure they are the right R
value.
I even am using a different DC adapter. instead of 12v 300 mA Im now
using a 12v 800mA thinking maybe there just was not enough current. No
real change. I have hooked it up in the following way, Using two
diffierent lights. One is a test light that works up to 50v and the
other a 12volt car test light. Both work out of the circuit.

Negative ----- Circuit --- Lamp ---- Positive

One caveat with DC adapters - they don't always contain filter caps so
the DC voltage could have a lot of ripple on it. I don't see how the
circuit should be sensitive to ripple, other than throwing the timing
off maybe, but I use a variable 1.2-18 V regulated supply on my test
bench and battery on the bike. A large cap between the plus and minus
output of the adapter should suffice if you have ripple (2,000 uf
16-25 volts or so)


If the voltage is <4  V at the collector of the PNP with the gate
disconnected (and the 100 K is indeed 100 K) the problem is before the
mosfet.
Im not sure what is considered before the mosfet.

The circuit as it is before the mosfet (to the left of the mosfet)
that should be providing 4 volts of drive and is not.

Mosfet specs contain something called the "threshold" voltage - that
is the voltage (source to gate) that the mosfet begins conduction at.
Most standard high current mosfets work at about 4 volts - "logic
level" mosfets kick in at ~2 volts.

Or do you think the mosfet specs I used are Ok. Im not sure where to
look next.
Again it is probably user error but I seem blind to what it is. Could
some of the resistors be lowered to increase power ??

Regards

Concentrate on the base voltage of the PNP transistor - that is
obviously wrong even with the transistor out of the circuit. It could
be wrong because the zener is not a 3.4 volt one, or is connected
backwards (which should give you point six volts not 1.88 like you
write), or the zener is leaky (in which case using a lower value
resistor will hit it with more current - try something much lower like
1-10 K - the thing should work). (if you go much lower the leds could
ghost "on" during the off cycle - only downside, short of burning up
the zener with too much power dissipation - most zeners are 1/2-1 watt
in small diode packages - so you'd be a long way from burning one out)

!%$#&&^#@ Damn! 1N4744?????!!!!! That's a 15 Volt zener diode it
should be a 3.4 Volt zener diode. for this to work it has to be a
much lower voltage like 3.3 or 3.6 Vz diode like a 1N4728 or so.

I can state with certainty that it won't work with a 1N4744 - so your
options are use a 1N4728 or go up to the post where I suggest using a
resistance divider to get ~3.3 volts or string of 6 silicon diodes for
3.6 volts.

In spite of that - I think your base voltage should have been much
higher with a 15 volt diode in there and the transistor out of the
circuit (like whatever your adapter puts out with no zener to limit
it) so something may still be wrong - but start with the right diode
and check it again. Anode to ground side, cathode to 220K (with the
transistor out of the circuit those are the only connections -220K to
+ (lamp to be flashed) , other side of 220K to zener, zener to ground
- NO OTHER CONNECTIONS until the base of the transistor is connected
and it is in a working configuration.

Take care
bob
--
.

Relevant Pages

  • Re: Motorcycle flasher problem.
    ... The one thing I had to "guess" on was the Mosfet. ... the turn signal is in an off state but provides voltage to ... on the gate of the mosfet falls below its conduction threshold, ... The base should be clamped to around 3 volts (while ...
    (sci.electronics.basics)
  • Re: Motorcycle flasher problem.
    ... to turn on the mosfet - you want 4 or more volts there. ... I'd suspect a bad PNP transistor (or wrong transistor or connected ... A good mosfet has almost no gate current - the gate is an insulator - ...
    (sci.electronics.basics)
  • Re: Question about transistor
    ... Each transistor is provided a potentrial base drive through a series combination of the 1k collector resistors and the 4.7k base resistors. ... So the remaining 2.4 volts is divided between the 1k and 4.7 k resistors, with the 1k dropping about 2.4*1k/=0.42 volts and the 4.7k dropping about 2.4*4.7k/=1.97 volts. ... But as each transistor begins to conduct collector current, that additional current through the 1k collector resistor increases the voltage drop across them, lowering the voltage feeding base current to the opposite transistor. ...
    (sci.electronics.basics)
  • Re: Circuit & Component Check
    ... >> Driving any switching transistor with a large voltage and large series ... > outputs about 5 volts, and the speaker, drawing from the top of the ... > divider through collector gets consirably more. ...
    (sci.electronics.basics)
  • Re: Question about transistor
    ... Each transistor is provided a potentrial base drive through ... remaining 2.4 volts is divided between the 1k and 4.7 k ... If neither transistor collector was collected to ... increases the voltage drop across them, ...
    (sci.electronics.basics)
Loading