Arduino for variable LPI

[mischmerz]

Howdy :

I am planning to use an Arduino driven PWM circuit to power the motor driving the screw of the over head. I'd like to implement some real time pitch control to increases or decreases pitch width in regard to the low frequencies. This will of course not be perfect as this approach will be easily overwhelmed by sudden spikes - but I don't have a preview yet and I don't want to use a digital delay.

Anyway - my question would be: Is it enough to measure the voltage of the signal and assume "more volt = more pitch width" or should I add a low pass filter before measuring? Any other ideas?

Michaela

[markrob]

Hi,

Since the head is equalized to be flat as a velocity device, you would want to integrate the input audio (low pass is probably best) to have a better representation of excursion. However, the RIAA EQ approximates a differentiator and you would want your sensing to be post this EQ to take its action into account. So, you could well get by with just using the raw audio (pre-RIAA) as a crude gauge of excursion. Or, use a integrator (always use a "leaky" integrator) after the RIAA if you want better accuracy.

Mark

[mischmerz]

Thanks Mark. I was planning on going directly from the cutting signal (after RIAA). I thought about coupling both channels with a high resistance to mono and re-amplify to measure voltage. So you suggest a leaky integrator in that mono-fied signal path that would increasingly drop the higher frequencies. Makes sense. Thanks.

Michaela

Hi,

Since the head is equalized to be flat as a velocity device, you would want to integrate the input audio (low pass is probably best) to have a better representation of excursion. However, the RIAA EQ approximates a differentiator and you would want your sensing to be post this EQ to take its action into account. So, you could well get by with just using the raw audio (pre-RIAA) as a crude gauge of excursion. Or, use a integrator (always use a "leaky" integrator) after the RIAA if you want better accuracy.

Mark

[mischmerz]

I was thinking along the line of this, where the low pass is calculated with a cutoff frequency of 10hz which would put me within the stop band and the low-pass would more or less act like a leaky integrator. I could measure the + / - values at L-C or and R-C . Now I am neither an ee nor have I ever played one - and I could of course be completely wrong about it.

cutterconnect.png

R1 and R2 around 250K so there would probably be an op-amp necessary.

Michaela

[dubcutter89]

All points by markrob are absolutely right!

But I'm not sure what you mean by "cutter" in your scheme. Do you want to connect it direct to the drive leads of the head?
Keep in mind that most/all cutters do not have a flat velocity vs voltage response and therefore are compensated with another eq so the drive voltage is not a true response of riaa emphasized music...

you could also use the input signal from your computer/cd source and use a simple low pass at around 500Hz (where the 2nd RIAA turnover is, there will be an error below 50hz but maybe not so problematic?) and sum it with 2 resistors. add coupling capacitors and a bias network to get the signal to vcc/2 for the adc of arduino and then let the uC do the rest...
or for all analog heads: you could also insert a diode between R and C (red, where the summing is) to make a simple peak detector and use this signal to control the pitch motor (feed to an amp...)

Lukas

[mischmerz]

Thanks Lukas. Actually - I was thinking of connecting the circuit directly to the cutter head feed as their is currently no additional eq'ing to compensate for the head. My VU meters are connected to this feed and - well - I thought it would make sense to connect the pitch-regulator there as well. But you are right - the pitch regulation doesn't require this - so I can hook it up somewhere else in the chain. So I am going to sum up the channels, low pass and feed it into the uC.

Michaela

you could also use the input signal from your computer/cd source and use a simple low pass at around 500Hz (where the 2nd RIAA turnover is, there will be an error below 50hz but maybe not so problematic?) and sum it with 2 resistors. add coupling capacitors and a bias network to get the signal to vcc/2 for the adc of arduino and then let the uC do the rest...
or for all analog heads: you could also insert a diode between R and C (red, where the summing is) to make a simple peak detector and use this signal to control the pitch motor (feed to an amp...)

Lukas

[dubcutter89]

I also have a Grampian which even sounds ok without riaa (and no damping fluid). the low frequency turnover is implemented in the head inductance and high frequency boost by resonance... well not best sounding, but ok for playing around and scratching discs...

audio VU connected to head is not so good because (if you use inverse RIAA) of the extrem frequency distortion.
I have a normal Volts meter connected to the head - this serves as indicator for power going to the head. an amp meter would be better!

btw. for accurate pitch the head response of course has to be accurate as well. also phase relation etc...
another thing would be that if you insert a delay between the pitch computer and cutteramp you could look ahead of cutting and calculate the needed space in advance - this is how pro setups work.

Lukas

[mischmerz]

another thing would be that if you insert a delay between the pitch computer and cutteramp you could look ahead of cutting and calculate the needed space in advance - this is how pro setups work.
Lukas

I know - but I don't want to use digital delays. Working on a preview head for my tape machine.

Michaela