Digital to analogue convertor

Thread Starter

ThatOtherGuy

Joined May 30, 2016
18
Hi everybody I am a Diesel mechanic and while I have basic auto electrical and electronics understanding I don’t have the ability to design my own component in this instance and wondered if someone may be able to help or point me in the right direction of how to build or weahter the is an off the shelf solution.
I need a digital to analogue converter for a project I am currently working on , I understand you may need more information but I’ll do my best to describe what I am after. I am needing a DAC to convert a signal from 2 separate speed sensors form a 0.65v- 1.325v square wave digital signal to an analogue A/C sine wave signal that I have measured to be 5v peak to peak with the shaft speed at 190rpm and it needs to read between 0 and 1000 rpm. The pulse wheel on the shafts has 48 pulses per revolution. Obviously, the amplitude varies relative to speed of the shafts. The system voltage is 12 volts. I’d greatly appreciate any information you can provide.
 

Papabravo

Joined Feb 24, 2006
18,964
The only thing that comes to mind is a DSP (Digital Signal Processor) that can read and store the information from the incoming pulse trains, and then after some delay synthesize the AC output waveform and run it through a low pass filter. The programming would not be all that challenging for somebody who works with DSP type stuff day in and day out. This might be a real challenge for you without some assistance. I would not think that there would be an off the shelf solution to this problem, but I'll confess that my knowledge of the auto and truck industries is limited to the automation of their factories.

Can you provide a link to the datasheets for the sensors and for the device that is receiving the AC sine wave?
 

MrChips

Joined Oct 2, 2009
26,782
If I understand your requirements correctly, you wish to convert the frequency of a square wave to an analog voltage.
There are off-the-shelf solutions that can do this, for example, LM331 frequency-to-voltage converter.
 

Thread Starter

ThatOtherGuy

Joined May 30, 2016
18
The only thing that comes to mind is a DSP (Digital Signal Processor) that can read and store the information from the incoming pulse trains, and then after some delay synthesize the AC output waveform and run it through a low pass filter. The programming would not be all that challenging for somebody who works with DSP type stuff day in and day out. This might be a real challenge for you without some assistance. I would not think that there would be an off the shelf solution to this problem, but I'll confess that my knowledge of the auto and truck industries is limited to the automation of their factories.

Can you provide a link to the datasheets for the sensors and for the device that is receiving the AC sine wave?
Unfortunately I can't get any manufacturer data regarding the sensor or the module receiving the signal as such, only what I can measure. What this is is a Toyota rear axle that I am swapping out of a later model Toyota that has a much stronger final drive. The new axle has the digital 2 wire wheel speed sensors and the vehicle I am fitting it to requires the A/C signal, the signal is sent to the ABS module in the vehicle. Toyota don't supply any information on their equipment other than what is required to diagnose faults. I could as a last resort modify the axle to take the analogue wheel speed sensors and the pulse wheel to suit, however I'd like to avoid this if possible so as readily available part will still fit if required.
 

Thread Starter

ThatOtherGuy

Joined May 30, 2016
18
This is the only information I have regarding what the signal should look like, this comes from the diagnostic manual of the vehicle that I am upgrading the axle in,
 

Attachments

Irving

Joined Jan 30, 2016
2,979
Just to be clear... you want to convert an unknown digital pulse to an analogue sine wave of a similar frequency (or some known linear relationship)?

How did you get the voltage of the digital signal - is this from manufacturer data or measured with a voltmeter? Normally digital signals have fixed voltage and either vary the frequency or the ratio of ON to OFF (duty cycle). Normally measuring a digital ON/OFF signal with variable frequency on a DC meter gives ~1/2 the ON voltage whatever the frequency. The fact you get variable voltage suggests a variable duty cycle at a fixed frequency OR, more likely, a fixed ON time and a variable OFF time.

If you rotate the wheel slowly can you find a point where you get a definitive transition between an ON and OFF voltage?

You're obviously a competent mechanic to contemplate swapping axles, however I'm duty bound to say that modifying a safety-critical element, eg the ABS will invalidate any road-use insurances you may rely on - unless for off-road use on private land.
 

Thread Starter

ThatOtherGuy

Joined May 30, 2016
18
Just to be clear... you want to convert an unknown digital pulse to an analogue sine wave of a similar frequency (or some known linear relationship)?

How did you get the voltage of the digital signal - is this from manufacturer data or measured with a voltmeter? Normally digital signals have fixed voltage and either vary the frequency or the ratio of ON to OFF (duty cycle). Normally measuring a digital ON/OFF signal with variable frequency on a DC meter gives ~1/2 the ON voltage whatever the frequency. The fact you get variable voltage suggests a variable duty cycle at a fixed frequency OR, more likely, a fixed ON time and a variable OFF time.

If you rotate the wheel slowly can you find a point where you get a definitive transition between an ON and OFF voltage?

You're obviously a competent mechanic to contemplate swapping axles, however I'm duty bound to say that modifying a safety-critical element, eg the ABS will invalidate any road-use insurances you may rely on - unless for off-road use on private land.
Hello Irving, thanks for your reply and I appreciate your input.
The digital signal I have verified with and oscilloscope to be 50% on regardless of speed. The sensors recieve a 10v reference signal in and 0.65v when low and 1.325v when high, this is a square wave signal that just alternates voltage high and voltage low and this occurs 48 times every rotation which is the same number of analog cycles per revolution.
 

Irving

Joined Jan 30, 2016
2,979
That's great info. So what you need is a Phase Locked Loop where the output oscillator is locked to the digital input frequency. It's not hard to achieve, but neither is it trivial. I'm sure my colleagues here will have many options on how to approach it. I'll post my solution tomorrow.
 

Thread Starter

ThatOtherGuy

Joined May 30, 2016
18
That's great info. So what you need is a Phase Locked Loop where the output oscillator is locked to the digital input frequency. It's not hard to achieve, but neither is it trivial. I'm sure my colleagues here will have many options on how to approach it. I'll post my solution tomorrow.
Thanks, much appreciated
 

LowQCab

Joined Nov 6, 2012
2,309
The new Sensor may work "as-is" without any modifications.
I would try it with a simple Capacitor to convert the Signal from
a Square-Wave with a DC Bias, to being a Square-Wave without the DC-Bias.

Trying to round-off the sharp corners opens a whole new bucket of worms,
and may not be necessary.
The older Input-Amplifier is designed for a "VR" (Variable-Reluctance) signal,
and so, it has the capability to decipher a Signal surrounded by
all sorts of odd conditions and Voltages.

A ~1uf, 16-Volt, Capacitor in between the Sensor and the Electronics may do the trick.
.
.
.
 
Last edited:

Thread Starter

ThatOtherGuy

Joined May 30, 2016
18
The new Sensor may work "as-is" without any modifications.
I would try it with a simple Capacitor to convert the Signal from
a Square-Wave with a DC Bias, to being a Square-Wave without the DC-Bias.

Trying to round-off the sharp corners opens a whole new bucket of worms,
and may no be necessary.
The older Input-Amplifier is designed for a "VR" (Variable-Reluctance) signal,
and so, it has the capability to decipher a Signal surrounded by
all sorts of odd conditions and Voltages.

A ~1uf, 16-Volt, Capacitor in between the Sensor and the Electronics may do the trick.
.
.
.
Will this provide an alternating positive/negative signal?
 

Irving

Joined Jan 30, 2016
2,979
Yes, @LowQCab Has a good point, though I suspect the level might still be a little low. If so that's easily dealt with by a simple amplifier.

What did you reckon were the highest and lowest frequencies you needed to deal with and how do they relate to road speed?
 

LowQCab

Joined Nov 6, 2012
2,309
Unless it's a Race-Car, the Rear-Wheels will never get over 200rpm,
times 48 pulses, if the sensor is on the axle,
and 48 X the Rear-Axle-Gear-Ratio if the Sensor is on the Driveshaft or,
in the tail-housing of the Transmission.

200 X 48 = 9600 pulses per minute,
9600 / 60 = 160hz,
or, 160hz X Rear-Axle-Ratio = ???, maybe 3.7:1 ???
3.7 X 160 = 592hz.
.
.
.
 

LowQCab

Joined Nov 6, 2012
2,309
DC can not pass through a Capacitor.
The Capacitor will average the pulses around zero-Volts, ( or Ground ).
So what You get is an AC waveform.
.
.
.
 

LesJones

Joined Jan 8, 2017
3,773
If these are ABS sensors do they produce the same number of pulses per revolution as those on the original axle ?
Also will the same diameter wheels be used as on the original axle ?

Les.
 

Thread Starter

ThatOtherGuy

Joined May 30, 2016
18
Unless it's a Race-Car, the Rear-Wheels will never get over 200rpm,
times 48 pulses, if the sensor is on the axle,
and 48 X the Rear-Axle-Gear-Ratio if the Sensor is on the Driveshaft or,
in the tail-housing of the Transmission.

200 X 48 = 9600 pulses per minute,
9600 / 60 = 160hz,
or, 160hz X Rear-Axle-Ratio = ???, maybe 3.7:1 ???
3.7 X 160 = 592hz.
.
.
.
The sensors are on the axle itself and with a tyre diameter of 33.03 inches and I've calculated to be a maximum of 180km/h the be around 1100rpm. So 1100 x 48 ÷ 60 = 880Hz.
 

Thread Starter

ThatOtherGuy

Joined May 30, 2016
18
If these are ABS sensors do they produce the same number of pulses per revolution as those on the original axle ?
Also will the same diameter wheels be used as on the original axle ?

Les.
Yes they both produce the same number of pulses per revolution. The pulse wheels have 48 teeth or segments.
 
Top