Hilbert transformer

[epilot]

Hello,

This circuit is a Hilbert transformer from ARRL Handbook, that I want to use it with MC1496 balanced modulator to generate SSB from DSB via phasing method.
The Hilbert filter generates a 0 , 90 , 180 and 270 degree phase shifting for the op amps and finally we will have a 0 and 90 degree phase shift at the output of op amps(which I want to use a TL072 op amp at 12V single supply).

Now I wanted to know if the wiring is correct for the input of op amps or not?
I think there is a problem because it seems that the + and – of an op amp is connected too 0 and 180 degree and the input of the other op amp is connected to 90 and 270 degree pin? So how we can have a 90 degree phase shift at V3 and V4(outputs of the op amps)?

Another question is that I have heard that a Hilbert transformer is working fine at the range of 300 to 3000 or 3500Hz, if so, can an expert let me know how it works on voice Communications!?

Thanks for any help.


C1: 0.044 uF, C2:0.033 uF, C3: 0.02 uF, C4: 0.01 uF, C5: 5600 pF, and C6: 4700 pF AND all resistors are 12K ohm 10%

 

[JimB]

The top opamp in the circuit is taking the 0deg and 180deg signals and amplifying the difference of those two signals.
Similarly with the lower opamp, it is amplifying the difference between the 90 deg and 270 deg signals.
The result is two signals one at 0deg and the other at 90deg.

The frequencies used for voice communication are 300 to 3000 hz, this is fairly standard for any radio or telephone voice comms.

JimB

 

[audioguru]

The 75k resistors to ground make the non-inverting input voltages very close to ground. Since you want a single supply of +12V then the input voltages are much too close to the negative supply voltage ( ground in your circuit) to work. The fix is to have a positive and a negative supply voltages, or use a 3rd opamp to make a V/2 reference voltage for the 75k resistors, and capacitor-couple the inputs.

 

[epilot]

The top opamp in the circuit is taking the 0deg and 180deg signals and amplifying the difference of those two signals.
Similarly with the lower opamp, it is amplifying the difference between the 90 deg and 270 deg signals.
The result is two signals one at 0deg and the other at 90deg.

The frequencies used for voice communication are 300 to 3000 hz, this is fairly standard for any radio or telephone voice comms.

JimB

Thanks Jim,

Difference between 0 and 180 degree is = 180 degree,
Difference between 270 and 90 degree is = 180 degree.
so How we can have a 0 and 90 degree?!

But 3000Hz is too away from the upper limitation of human hearing(20kHz).
sorry I can not understand you?!

 

[epilot]

The 75k resistors to ground make the non-inverting input voltages very close to ground. Since you want a single supply of +12V then the input voltages are much too close to the negative supply voltage ( ground in your circuit) to work. The fix is to have a positive and a negative supply voltages, or use a 3rd opamp to make a V/2 reference voltage for the 75k resistors, and capacitor-couple the inputs.

thanks audioguru,

No I can not to use a negative supply because it makes me to change whole my circuit.

Ohh, now that I think I see that the reason that a lot of my circuits based upon op amps did not work is that I did not care about V/2 and negative supply(I used them with single supply).
Please can you let me know more about that?
can you let me know the best choice for make V/2?Is that way perfect and works just fine like a negative supply at all Frequencies?

 

[Hero999]

But 3000Hz is too away from the upper limitation of human hearing(20kHz).

He's talking about voice frequencies. The range of the human voice is only 300Hz to 3kHz so there's no point in wasting 20Hz to 20kHz (20 times the bandwidth) on transmitting the entire audio range when only the human voice needs to be transmitted.

 

[audioguru]

I drew opamp circuits that work nearly the same with a dual supply as with a single supply. But not at very low frequencies nor at DC.

 

[audioguru]

He's talking about voice frequencies. The range of the human voice is only 300Hz to 3kHz so there's no point in wasting 20Hz to 20kHz (20 times the bandwidth) on transmitting the entire audio range when only the human voice needs to be transmitted.

All the vowels are near 300Hz. Other important voice sounds are the high frequency "ss", "th" and "ff" sounds that extend up to 15kHz and higher.

The 3kHz high frequency cutoff was decided when old speakers produced nothing but distortion and noise at higher frequencies.

1) How many people say, "What?" on a narrow-band (300-3kHz) telephone call?
2) The TV announcer reads the news with his (her) microphone hidden down on their chest where there isn't any high frequency voice sounds. It sounds awful, doesn't it? Hard to understand what they are saying?
3) Have you ever heard an underwater throat microphone? Vowels only. Very hard to understand what is being said.

 

[Dr.EM]

I never had a problem understanding people using those tie clip mics? Mabye they are reinforced with shotgun mics or something?

 

[epilot]

He's talking about voice frequencies. The range of the human voice is only 300Hz to 3kHz so there's no point in wasting 20Hz to 20kHz (20 times the bandwidth) on transmitting the entire audio range when only the human voice needs to be transmitted.

OoPs!?
I am being confused again!?

Can you explain more about human voice range?
I want to play music via Hilbert transformer too, not only human voice.
why are you thinking that I want to transmit only the human voice?

 

[Dr.EM]

Music won't sound very good at 300hz-3khz bandwidth at all. The design is only really suitable for voice transmission it would seem

 

[epilot]

Music won't sound very good at 300hz-3khz bandwidth at all. The design is only really suitable for voice transmission it would seem

If so, then i have to go for an alternative

I have found some of 40kHz crystals(the carrier freq is 40kHz too) so what about filter method for generating SSB(there is no problem if I have VSB at the output too).
any recommendation?

 

[Hero999]

Other important voice sounds are the high frequency "ss", "th" and "ff" sounds that extend up to 15kHz and higher.

Which is the upper cut-off unless you're under 25, your high frequency hearing is one of the first thing to go, when I was 16 I could hear up to 21kHz, now it stops at about 17kHz and I'm only 24, I suppose it's all that going clubbing listening to loud music.

Anyway, if the higher frequencies are so improtant than why do most telephones only have a 3kHx bandwidth? I remember doing commumications awhile ago at college an was told that they shift the voice spectrum up the scale allocating 3kHz for each channel, so they can fit 100s or more channels down one piece of cable, of course it's all changed now with digital fibre optic an lines.

epilotI have found some of 40kHz crystals(the carrier freq is 40kHz too) so what about filter method for generating SSB(there is no problem if I have VSB at the output too).
any recommendation?

The only applications for this is transmission using ultrasound or VLF, both are really only used by submarines.

 

[JimB]

Thanks Jim,

Difference between 0 and 180 degree is = 180 degree,
Difference between 270 and 90 degree is = 180 degree.
so How we can have a 0 and 90 degree?!

But 3000Hz is too away from the upper limitation of human hearing(20kHz).
sorry I can not understand you?!

The 0 and 180 are added together.
The 90 and 270 are added together.
You then have two signals at 0 and 90, (or is it 0 and 270, or 180 and 270, all the same thing!).

3000Hz = 3kHz
3kHz < 20kHz

JimB

 

[JimB]

I want to play music via Hilbert transformer too, not only human voice.
why are you thinking that I want to transmit only the human voice?

Probably because no one in their right mind would use SSB to transmit music! Why do you want to do this??
If you want to transmit music use FM.

JimB

 

[audioguru]

I think the Hilbert transformer has a top end of only 3kHz to 3.5kHz because it needs many more RC stages having capacitors with reduced values for it to phase-shift higher frequencies. Then the values of the 75k resistors will need to be increased for the gain to be the same as now.

 

[epilot]

I think the Hilbert transformer has a top end of only 3kHz to 3.5kHz because it needs many more RC stages having capacitors with reduced values for it to phase-shift higher frequencies. Then the values of the 75k resistors will need to be increased for the gain to be the same as now.

Yes audioguru, I will need A LOt of RC component to do the job for 20kHz too


I am thinking if I can get a SSB or at least a VSB via a filtering method(to get very more bandwidth) with my 40kHz crystals?

P.s audioguru what software you used to drew the op amp circuit?

 

[audioguru]

Yes audioguru, I will need A LOT of RC component to do the job for 20kHz too

It would need too many parts.

I am thinking if I can get a SSB or at least a VSB via a filtering method(to get very more bandwidth) with my 40kHz crystals?

I made an SSB voice scrambling system with an MC1496 and a switched-capacitor lowpass filter IC that worked very well.

What are you going to use for a carrier that is at the extremely low frequency of only 40kHz? It is far too low a frequency for radio and also too low for infrared if you want response to 20kHz.
Wireless infrared headphones and speakers frequency-modulate carriers of about 100kHz for one channel and 200kHz for the other channel.

what software you used to drew the op amp circuit?

I do all my schematics with Microsoft Paint. I copy and paste symbols from other schematics or from datasheets.

 

[Hero999]

Wireless headphones also use PFM, not SSB.

 

[audioguru]

He wants to use SSB to save bandwidth.
Maybe he wants Dolby 5.1 channels.