# Picking Resistors/Capacitors for Tachometer using LM2907 and LM3914

Discussion in 'The Projects Forum' started by Tickers, Jun 29, 2010.

1. ### Tickers Thread Starter New Member

Dec 24, 2009
17
0
I'm trying to build a circuit to display engine speed as a bar of 6 LEDs using an LM2907 and LM3914. I'd like first LED to illuminate at 8000rpm and for all LEDs to be lit at 13000rpm.

I've found a few example circuits (including those in the datasheets) which I can tell are what I need, but I'm having trouble picking the correct resistor and capacitor values for the chips to give me the range I want. So far I've got values of 33.3kR, 29nF and 73nF for R1, C1 and C2 on the LM2907 using the application notes pdf. The capacitor values aren't standard, so I'd like to know how I can change them to get a range from the LM2907 suitable for the 3914.

The supply voltage can be either 5V or 12V.

2. ### SgtWookie Expert

Jul 17, 2007
22,183
1,728
Do you have Linear Technology's LTSpice installed?

I posted some LTSpice macromodels for the LM29x7 and LM29x7-8 a month or so ago.

You can simulate it using LTSpice to get an idea of how it'll work.

3. ### Tickers Thread Starter New Member

Dec 24, 2009
17
0
I have those models and I started building a circuit around them but I couldn't work out how to input a frequency. I tried playing with using the voltage supply with a .wav file but it wasn't really playing.

4. ### SgtWookie Expert

Jul 17, 2007
22,183
1,728
OK, are you using the 14 pin version of the LM2907?

How many cylinders does the engine have?

Last edited: Jun 29, 2010
5. ### Tickers Thread Starter New Member

Dec 24, 2009
17
0
It's just started to work now in LTspice.

Given a Vcc of 12V and input freq of 433.33Hz (13000rpm for a 4cyl car).

Using C1 = 0.01uF
C2 = 1uF
R1 = 95.3kR
R2 = 10kR

I get a Vo between 4.88-4.945V. Is a ripple of 0.07V acceptable in this situation? I'm guessing so, considering I'm only using 6 LEDs to cover such a wide range of frequencies.

How do I get the LM3914 to illuminate the LEDs at equal points between 8000-13000rpm (266.66-433.33Hz)? I'd like them to be very bright.

6. ### SgtWookie Expert

Jul 17, 2007
22,183
1,728
OK, if your highest voltage out is about 4.8v, then on the LM3914, use a 1k resistor from pin 7 to pin 8, and about 2.67k from pin 8 to ground.
Also, connect pin 7 to pin 6.

You will then need a resistor from Pin 4 to ground.

Ahhh, wait a minute - you said 6 LEDs, not 10. :/

7. ### Tickers Thread Starter New Member

Dec 24, 2009
17
0
Does using a different number of LEDs severely complicate things?

Could I get away with fitting the 6 LEDs to the pins for LED 1, 3, 5, 7, 9, 10? The last two would be unequally spaced but the driver shouldn't be spending much time around 13,000rpm anyway.

Also, does that resistor combination only display between 8-13k rpm (266-433Hz)?

You don't happen to have a spice model for the 3914 do you?

8. ### SgtWookie Expert

Jul 17, 2007
22,183
1,728
Only because I'm really tired today.

Nope; use only the last 6 outputs for the LEDs (pins 15, 14, 13, 12, 11, 10).

Not without another resistor between pin 4 and GND.

There is an internal 10k Ohm divider, in 1k increments per output. This value is approximate, and I don't know what the tolerance is. Your mileage may vary.

If you used a 10k pot between pin 4 and GND, you should have plenty of adjustment. A 10- or 21-turn trimpot would be a good thing to use.

National Semiconductor never released a model for it.

9. ### Tickers Thread Starter New Member

Dec 24, 2009
17
0
There's been a change of resistor/capacitor values now due to a change in how our ECU will be sending the frequency.

Where it was 2 pulses per revolution it's now 1 pulse per revolution, meaning 13000rpm is now equivalent to 216.67Hz.

R1 has become 191k, and C1 has become 0.5uF to make up for the length of time it was taking to reach steady state. There's a voltage swing of 0.1V but that should still be manageable.

At 13000rpm, the output is now 4.98V. Does that affect the resistors you recommend for use with the LM3914? It seems so close that I doubt it has an affect but I thought I'd check

10. ### SgtWookie Expert

Jul 17, 2007
22,183
1,728
Well, 191k seems high. Better decrease the resistance and increase the capacitance to compensate, otherwise you may experience interference from all of the electrical "noise" that's omnipresent around internal combustion engines that use high-voltage sparks for ignition.

Be aware that 0.5uF is not a standard value for capacitance. 0.47uF (470nF) is a standard value.

If you wish to decrease ripple on the output, then try adding another filter order. You'll get less of a phase shift (delay) and lower ripple if you use more elements with lower values of R and C than a single order with large R and C values. (I'm talking about the low-pass filter circuit between OUT and OUT2.)

Your results may vary. The model is simply a macromodel that is a rough approximation of the LM29x7 series. Once you get a SPICE simulation working to something approaching what is desired, you will then need to build the circuit and test it.

Please note that I do not have any of these ICs to experiment with.

Important: note that Mr. Kincaid indicated that the real ICs will saturate at Vcc/2, where the macromodel does not.

Note that the LM2914 and LM2914EP have a Zener diode incorporated; the value of the Zener is approximately 7.5v, +/-150mV.

Since the saturation voltage is Vcc/2, your output limit is approximately 7.5v/2 or 3.75v. I suggest that you should be more conservative and de-rate the expected output by 10%, which gives you 6.75v/2 = 3.375v maximum expected output.

If you want your output to actually reach 5v, then you need to use the LM2907-8 or LM2907-14 along with a regulator that is capable of producing a stable output of at least 10v up to perhaps 12v with a 10v to 60v DC input.

11. ### Tickers Thread Starter New Member

Dec 24, 2009
17
0
I've built the circuit now but I'm having a lot of trouble getting it to work.

There doesn't seem to be any change in the output of the LM2907 as frequency varies. It always hovers at less than a volt. I've changed the value of the resistor used and there is no affect.

The output at each of the LED pins on the LM3914 is also the same, at a few volts.

I've been testing it using a signal generator to make a square wave at 216Hz. What sort of voltage should I have on the input?

12. ### SgtWookie Expert

Jul 17, 2007
22,183
1,728
Without seeing your schematic, all I could do is make wild and inaccurate guesses.

Post a .png format image of your schematic, along with your LTSpice .asc file.

13. ### Tickers Thread Starter New Member

Dec 24, 2009
17
0
I think this shows what I have, it took a while to make a schematic based off the file I was using.

My LTSpice file only includes the LM2907, not the LED driver

• ###### schematic.png
File size:
26.5 KB
Views:
195
Last edited: Jul 12, 2010
14. ### Tickers Thread Starter New Member

Dec 24, 2009
17
0
I found that I had missed off the ground on one of the capacitors. I've sorted that now.

However, with a Vcc of 12V the output never really changes from 9-10V, no matter what the frequency input/amplitude is.