Unwanted oscillation in a blocking oscillator (tachometer)

Thread Starter

graybeard

Joined Apr 10, 2012
99
I am trying to repair a second generation Smiths tachometer from a 71 Volvo P1800. It consists of a blocking oscillator in a one-shot configuration driving an ammeter. The owner complains that it goes wonky when it is hot.

Here's the circuit:

1690145716737.png

At room temperature, the circuit puts out a nice, clean square wave as it should, but when I heat the circuit up with a hair dryer, it starts to oscillate after the end of the one-shot pulse. When I hit the components with freeze spray, none of the components appear to be temperature sensitive other than the transistor and cap. They are too close together to cool independently.

I have replaced both the transistor and cap with new ones, but the problem persists. I have reflowed all of the solder joints, cleaned the pot, cleaned the connection between the pot and the PCB and cleaned the connection between the PCB and the meter frame to ensure a good ground.

The photos show the calibration signal (top) and the collector signal (bottom). The first shows the tachometer working correctly at room temperature:

1690146444017.png

The second shows the tachometer reading very high after heating it up with the hair dryer:

1690146589217.png

I am very bad with oscillators, so I have no clue as to what is causing these oscillations and how to fix this poor guy's tachometer. Any guidance would be greatly appreciated.

Mark
 

LowQCab

Joined Nov 6, 2012
4,186
That's because it's made by Smiths.
You would be ahead of the game if You would gut-it-out and
install a modern Tach behind the original Dial-Face.

I see that You have found a "modern" replacement for the Transistor.
Replace it.
Also replace the Capacitor. p/n https://www.digikey.com/en/products/detail/panasonic-electronic-components/ECW-F4245JL/656399

Are You absolutely positive that the Engine-Ignition-Circuitry is in perfect working-order ?

None of the other components should cause any problems.
.
.
.
 

Thread Starter

graybeard

Joined Apr 10, 2012
99
Thanks for the reply.

I did replace the transistor and the cap. My lab setup generates a 1A square wave current pulse through the primary winding with the frequency controlled by a Digilent AD2. My bench setup is able to duplicate the problem the owner is seeing in his car.

If it were my tach, I probably replace the guts, but you would be surprised at how many people want to keep their vintage cars as close to original as possible, including tachometer innards.

From the scope traces, one thing I can see is that the negative overshoot at the end of the one-shot pulse is larger when the problem occurs. Could that spike be triggering the oscillation? Would clamping the collector to ground with a diode be a reasonable fix if I can't figure out the root cause?
 

LowQCab

Joined Nov 6, 2012
4,186
These ancient crude Tachs are not likely to work properly without being
connected to the actual, original, Ignition-System in the Car.
.
.
.
 

Thread Starter

graybeard

Joined Apr 10, 2012
99
I have fixed a ton of these tachometers with my bench setup, just none have had this strange oscillation failure. My setup is reliably triggering the one-shot. The failure symptoms are consistent between the car and the bench.

I wish I could remember my schooling well enough to understand why the pole is getting into the right half-plane when warm...

Reminds me of the old joke about why Polish airliners don't have seats on the right side of the plane. :)
 

Poyntat

Joined May 24, 2022
60
Is the ammeter part of your clients setup or part of your test rig?
If it is the clients, is it possible that the ammeter is somehow faulty and is producing a back emf when the drive at the transistor collector transitions to 0v?
 

Poyntat

Joined May 24, 2022
60
Just had a look at ’Understanding and adjusting the Smiths tacho Part 1‘ on YouTube in which there is a schematic of the two transistor variant of this tachometer which as you would expect has more components but works in a similar way to yours. Is the schematic at the top of your thread complete or could it be a simplified version of the actual circuit?
 

Thread Starter

graybeard

Joined Apr 10, 2012
99
The ammeter is the tachometer's meter with an RPM scale.

The schematic in the you tube video is Smiths' Gen 1 design which is not a blocking oscillator design used in the 1960s. The schematics above show the Gen 2 design which is a cost reduction used for a couple of years starting in about 1970. After that, Smiths made Gen 3 and Gen 4 units that used a TI IC.

The Gen 1 circuit utilizes an RC network to time the one-shot pulse. This blocking oscillator seems to use LR for the timing.

I tried clamping the collector to ground with a diode with no luck. I am studying every web page that explains how blocking oscillators work.

From reading that, this is how I think the circuit works. As current starts to flow in the transformer primary, current will also start to flow in the red/green secondary (transistor base) which starts to turn the transistor on. As the transistor puts collector current through the white/blue winding, that current induces more current through the red/green windings, turning the transistor on even harder. That process continues until the transistor saturates. At that point, there are no more changes in the current in the primary winding or the white/blue winding.

What I don't understand is the mechanism that causes the transistor to turn off. What happens with the transformer when the current is no longer changing? The one-shot timing is not affected by changing the cap value, so I know that the cap has nothing to do with the timing. It will charge to the germanium junction voltage while the transistor is on, that I know. But what turns the transistor off?

I do appreciate all thoughts on this. I am super curious and would love to be able to fix this tachometer.
 
Last edited:

Dodgydave

Joined Jun 22, 2012
11,318
I am trying to repair a second generation Smiths tachometer from a 71 Volvo P1800. It consists of a blocking oscillator in a one-shot configuration driving an ammeter. The owner complains that it goes wonky when it is hot.

Here's the circuit:

View attachment 298917

At room temperature, the circuit puts out a nice, clean square wave as it should, but when I heat the circuit up with a hair dryer, it starts to oscillate after the end of the one-shot pulse. When I hit the components with freeze spray, none of the components appear to be temperature sensitive other than the transistor and cap. They are too close together to cool independently.

I have replaced both the transistor and cap with new ones, but the problem persists. I have reflowed all of the solder joints, cleaned the pot, cleaned the connection between the pot and the PCB and cleaned the connection between the PCB and the meter frame to ensure a good ground.

The photos show the calibration signal (top) and the collector signal (bottom). The first shows the tachometer working correctly at room temperature:

View attachment 298918

The second shows the tachometer reading very high after heating it up with the hair dryer:

View attachment 298919

I am very bad with oscillators, so I have no clue as to what is causing these oscillations and how to fix this poor guy's tachometer. Any guidance would be greatly appreciated.

Mark
What happens if you remove the thermistor?
 

Poyntat

Joined May 24, 2022
60
I’m struggling with the TS’s explanation of the circuit operation.
I had thought that the circuit operated on an RC time constant but this doesn’t explain the operation of the second secondary winding, WHT/BLU.
Can someone explain the operation of this circuit in detail?
 

Dodgydave

Joined Jun 22, 2012
11,318
Only thing that i can suggest is to change the transistor and capacitor, you could try to apply heat with the soldering iron to the transistor and capacitor individually to see what happens?
 

Thread Starter

graybeard

Joined Apr 10, 2012
99
I replaced the transistor and the cap. I used a 2.2uF non-polarized cap. I also added different caps from 0.1 to 2.2uF in parallel. The only effect was that the frequency of the oscillations dropped a little with the extra 2.2uF. The one-shot pulse width did not change, so it must be an RL time constant or something completely different.

What a mystery...
 
Top