LM2917 Speed Switch Help!!! PLEASE!!!

[hpfiend]

Hi all,

I am a complete newbie... After lots of research I have purchased lots of parts and tried several thing and still can't get it to work!

Using the schematic on page 9 of the pdf here
http://www.ti.com/lit/ds/symlink/lm2917-n.pdf,
figure 7 "Speed Switch" I have assembled a solderless breadboard.

I used an LM2917N-8 and connected it as follows:

OK first of all- I am assuming positioning the little half circle notch to the left with the text reading right side up from left to right is how you orient it? The bottom left most pin I call 1. and the bottom right most pin is 4

pin 1 to frequency pulse positive
pin 2 a 10uF capacitor and then to ground
pin 3 a jumper to one side of 5.1K resistor (reads 4.97 on MM) then other leg jumper to ground
pin 3 a 1uF capacitor in row of jumper and then other leg to ground
pin 4 jumper to ground
pin 5 ground side (outside) of 12 volt light bulb positive side of light bulb (center) is to the positive row.
pin 6 one leg of 5.1K resistor and other leg to pin 7
pin 6 jumper in same row and then to positive source (Vcc)
pin 7 one leg of first 5.1K resistor (from pin 6) and then first leg of second 5.1 K resistor, the other leg goes to pin 8
pin 8 final leg of pin 7 resistor and then a jumper in same row to ground
I then have a 9 volt battery hooked up to the positive and ground rows
Finally I have a jumper from the top row ground to the bottom row ground and then a ground jumper to ground of the frequency source

Using 1/ (2RC) the threshold frequency to turn the bulb on should be 10 Hz.

I tried a 10k linear potentiometer at R instead of the 5.1K resistor and set it to 5 with my multimeter and it did not work either

I generated a known frequency using my arduino uno I just bought with a simple on pulse of 100 ms and then off for 100 ms in endless loop which if I did that math right should be 10 Hz. I checked it was pulsing with an LED. My arduino board is powered currently by USB so it should be 5 volts but I don't think it is bleeding through somehow... I wrote a program to make it pulse twice as fast at 50 ms on 50 ms off and it still didn't work. Oh yeah I am putting the positive frequency pulse jumper in the pulse pin and the ground jumper to ground on the arduino

9 volt battery reads at 7.88 volts right now. This multimeter doesn't read capacitance, but my dad has one that does... light bulb works when connected directly.

Should I be using the 2907 instead I bought one of those too as I couldn't figure out what the zener diode had to do with it and where it was...

Please help!

 

[davenn]

hi ther,
welcome to the forums

a pic is worth 1000 words 2 pics even better. please do a couple of sharp and well lit pics of your breadboard layout. Us guys may see a wiring error that you missed

you battery sounds like its going flat, or its being asked to supply more current than its able to and causing a voltage drop

your last comment about a zener diode ?? what zener diode? there isnt one shown in fig 7

Dave

 

[hpfiend]

hi ther,
welcome to the forums

a pic is worth 1000 words 2 pics even better. please do a couple of sharp and well lit pics of your breadboard layout. Us guys may see a wiring error that you missed

you battery sounds like its going flat, or its being asked to supply more current than its able to and causing a voltage drop

your last comment about a zener diode ?? what zener diode? there isnt one shown in fig 7

Dave

Thanks for your reply!

That's a good point about the voltage drop- apparently if it drops below 6 it wont work- I will change it out to be sure that isn't it... Nope I switched to a new 9.2 volt battery...

The 2917 apparently has an integrated zener diode for some reason while the 2907 does not... I don't know where it is or why they have incorporated it as it does not show in that datasheet clearly enough for me to understand.

I am also going to check these jumpers for continuity... nope not that either...

 

[geekygenius]

I think I found it. Check the connection between pin 7, pin 4, the 10k resistor and ground. Pins 7 and 4 should both be going through the same 10k resistor into ground. In the last picture, it looks like pin 4 is connected straight into ground, and pins 7 and 8 have two resistors in parallel between them, which could be problematic.

 

[KrisBlueNZ]

How much current does that light bulb draw? I think the IC is only specified to sink up to 50 mA into pin 5.

Edit: OMG, WHERE did you get those resistors? I haven't seen resistors like those in 30 years!

 

[davenn]

I think I found it. Check the connection between pin 7, pin 4, the 10k resistor and ground. Pins 7 and 4 should both be going through the same 10k resistor into ground. In the last picture, it looks like pin 4 is connected straight into ground, and pins 7 and 8 have two resistors in parallel between them, which could be problematic.

pin 4 is supposed to be going straight to ground
look at the schematic in the datasheet

what 10 k resistor ? they are 5.1 k resistors

dont worry about the components shown inside the chip, for practical purposes they are not important, they are shown just to give you the basics of what functions are inside the chip ( thats any chip info)
As long as you correctly wire the components outside the chip the circuit will work within its specifications.

Now Kris brought up a point that I forgot to comment on in my earlier post.

there's a good chance that the chip's driver transistor cannot handle the load that that lamp is placing across the output. it is now possible that you have killed that transistor in the chip. I would be trying a LED there instead with a ~ 1k series resistor for current limiting through the LED

Dave

 

[davenn]

ohhh ... also meant to comment

you have used a few more jumpers than needed too....
you have one that comes off pin 3 to the resistor then another from the resistor to the -V rail

tidy the circuit up by getting rid of those 2 jumpers and take the resistor directly from pin 3 row to the -V rail just like you have with the capacitor from pin 3

makes it much easier to follow

D

PS I had to giggle as kris's OMG comment .... yeah they are really old carbon resistors

 

[hpfiend]

How much current does that light bulb draw? I think the IC is only specified to sink up to 50 mA into pin 5.

Edit: OMG, WHERE did you get those resistors? I haven't seen resistors like those in 30 years!

Thanks for your reply!

The 1893 bulb supposedly uses around 4 watts- that is 330 milliamps at 12 volts oops! My newbie self figured the 9 volts would cook the LED... I actually tried that first and it wouldn't light and I heard a pop.. should have used the resistor...

I will switch it to an LED with resistor now. Hopefully the 2907 will work in place of the 2917... as genius me only bought one of each...

They must be old stock as I got them from Mouser the same time I bought the rest of the parts...

 

[hpfiend]

I think I found it. Check the connection between pin 7, pin 4, the 10k resistor and ground. Pins 7 and 4 should both be going through the same 10k resistor into ground. In the last picture, it looks like pin 4 is connected straight into ground, and pins 7 and 8 have two resistors in parallel between them, which could be problematic.

Thanks for your reply! I think 4 is correct from my understanding which isn't much. The resistors do look to be in parallel from the way I took the pictures but they are in series as far as I can tell.

 

[hpfiend]

ohhh ... also meant to comment

you have used a few more jumpers than needed too....
you have one that comes off pin 3 to the resistor then another from the resistor to the -V rail

tidy the circuit up by getting rid of those 2 jumpers and take the resistor directly from pin 3 row to the -V rail just like you have with the capacitor from pin 3

makes it much easier to follow

D

PS I had to giggle as kris's OMG comment .... yeah they are really old carbon resistors

I had it that way as I originally wired it with a 10K potentiometer (linear) and couldn't straddle to ground but you are right the less connections the better- I have changed it...

Thank you!

 

[hpfiend]

dont worry about the components shown inside the chip, for practical purposes they are not important, they are shown just to give you the basics of what functions are inside the chip ( thats any chip info)
As long as you correctly wire the components outside the chip the circuit will work within its specifications.

Dave

Dave,

The only reason I was concerned as the paragraph above those schematics on page 9 talks about the 2917 which has the zener and not the 2907 which does not but I am not sure if that makes a difference in this circuit. I am going to swap them out right now and I will let you know if anything else cooks..

Thanks!

 

[hpfiend]

Well-

The fresh 2907 didn't work either Maybe I should order a new 2917?

Do you think I am creating the frequency incorrectly? Is 100ms on 100 ms off 10 hz with a digital square wave output? I doubled it to 50 ms on and off with the new chip also and it still wouldn't light.

Perhaps I am way off on my 1/2RC calculation? I put the 10K pot back in and turned it slowly lock to lock and the LED never lit... I dimmed the lights as well to verify and I hooked the arduino output to the resistor and LED separately and it flashed properly....

Interestingly when I unplugged the resistor to put the potentiometer in the light lit but I guess with no resistance it should... did I burn the chip up doing that? I should have unplugged it first... again complete beginner here- it didn't occur to me until afterwards.

I am going to check the actual capacitance tonight once I get my other multimeter (harbor freight I got for free) as opposed to this nice craftsman that wont measure it.

 

[KrisBlueNZ]

They must be old stock as I got them from Mouser the same time I bought the rest of the parts...

Perhaps Mouser are having a 50 year retro clearout! That style of resistor was used in 1960s-era tube radios.

A 10 Hz signal is 10 cycles per second. A cycle consists of an ON period plus an OFF period. So you need 50 ms ON, 50 ms OFF to get a cycle time of 100 ms which is 10 Hz.

 

[geekygenius]

pin 4 is supposed to be going straight to ground

oh wow, I was looking at the next schematic

 

[(*steve*)]

Do you remember the older style resistors that had dots on them rather than stripes?

If you get any of those from Mouser, let us know.

I do know that some people seem to value those carbon composition resistors for some reason (apparently they "sound better"). Check eBay some day!

 

[hpfiend]

Perhaps Mouser are having a 50 year retro clearout! That style of resistor was used in 1960s-era tube radios.

A 10 Hz signal is 10 cycles per second. A cycle consists of an ON period plus an OFF period. So you need 50 ms ON, 50 ms OFF to get a cycle time of 100 ms which is 10 Hz.

Thanks! I was very excited when I read this as that would definitely not allow it to work- unfortunately I set the arduino to 50 on 50 off and then 25 on 25 off and then 10 on 10 off for 10 hz 20 hz and 100 hz if my calculations are correct and no LED action...

I am half tempted just to have the arduino do it for me but I really wanted to figure this out the hard way first...

oh wait I didnt try the original 2917 again.. hang on- nope

tried different output pins on the arduino in case I burned one of those up- nope

flipped the capacitors- nope. Apparently this HF multimeter doesn't read capacitance- my dad has one that I think does- another craftsman I will check that later..

 

[hpfiend]

I measured capacitors to be 10.26 uF and .985 uF...

Anything above 9.805 Hz should light the LED

Verified arduino output to be 10 Hz and 20 Hz.

The only thing left is the breadboard is faulty, schematic in datasheet incorrect, or I got two dud LM29X7 chips from Mouser?

Thanks for all of the help thus far!

 

[KrisBlueNZ]

Can you post a complete schematic of what you're testing at the moment?

Are you still using the LM2917 that may have been damaged?

 

[hpfiend]

Does this make any sense to anyone??? I found it in more searching...

>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>
The sine wave makes no difference, since the input is just going to a
comparator anyway. However, the sine input must be centered on 0. If you
are giving it 0 to 5V, it won't work, since the comparator in the LM2907
compares against 0. So, put a cap in line with the sensor output, and a
big resistor (like 1MEG) from the chip to center the input on ground
>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>

My circuit is as this one:
http://img313.imageshack.us/img313/7242/speedswitch8xc.jpg
C is 10.26 uF, R is 4.97Kohms and the capacitor at pin 3 next to R is .985 uF. At pin one I am feeding the positive frequency output from the arduino and I am connecting the ground pin on the arduino to the ground bar on the breadboard.
Vcc is a fresh 9 volt battery which I have connected to ground and the positive bars on the breadboard

I have a positive source feed at pin 6 and pin 5 is connected to the ground side of my 1Kohm resistor and LED and the positive leg I have plugged into the positive bar... here are some more high res breadboard pictures:

I moved the chip and all connections to a different location on the bread board and it still would not work.

I touched the safety 1K LED resistor to a leg of the resistor pin 7 and 8 at one point and it did light for a second...

I am using a fresh LM2907 that has never been near that big 330ma light bulb... I don't have a spare LM2917

I am wondering if the frequency has to be derived from a magnetic reluctor wheel like on a vehicle speed sensor instead of my arduino digital perfect all positive square wave output. I think the speed sensor is an induced AC voltage...

 

[hpfiend]

should I be posting this in circuit help or project help instead? I don't mean to muck up the boards....