Pre emphasis

Hi Radio Joe.

There will be more educated comments from the trained radio engineers in the audience, but as a self-taught semi-knowledgable guy with opinions, here is what I think...

I believe the specific values of resistor and capacitor in a pre-emphasis circuit requires knowing the impedance of the audio line being pre-emphasized.... input impedence and output impedence.

From theory, the roll or slope of a single RC network is 6 dB per octave. As alluded by Carl the R includes all resistances transformed to a single equivalent resistance for accurate calculation, thus the source and load R for a given network needs to be accounted. This equivalent R will affect the corner (cutoff) frequency but will not affect the slope or roll of the network. The time constant determines the corner frequency so if this is held constant with changes in R and C there will be no difference. Changing the slope will require additional RC stages, each producing an additional 6 dB per octave.

In other words, it is the number of poles in a network which determines the slope and not the values of the R and C used and the time constant sets the corner frequency. A pole is essentially the number of RC pairs in the circuit.

A web search on "Bode plot" should yield detailed information.

Neil

Most of the chineeze transmtters have a 50 micro second pre emphasis, I built a nice box with a variable LC 3.3 nano farad caps and 10 k varaible resistors these are set so I get 25 micro seconds, or resistors set at 7.5 k and caps are at 3.3 nano farads that gives 25 micro seconds.
I use this at the inputs of the transmitter, it sounds perfect now, but now im concerned that I am running 12 db slope instead of 6... I have changed the pre emphasis circuit in the past on the boards, by installing a 3.3 nano farad caps and removing the stock 2.2 caps, but its a hassle since these are SMD caps.
I suppose i did wrong, I dont know it works -
Dont tell anyone if I did wrong

A quick review is that the time constant sets the corner frequency with a shorter time constant representing a higher frequency. If another RC high pass network is added before the input of the transmitter then the net result will be a combination of the added network and that which is inside the transmitter. If the time constants are different then there will be two corner frequencies. Starting low in frequency and going higher the response will be flat up to the first corner and will increase at 6 dB/oct. When the frequency goes above the second corner the slope will be 12 dB/oct.

You mentioned that the transmitter signal sounds good so maybe it is not worth worrying about but for technical purity you could perhaps short the on board caps to defeat the preemphasis and use your external box to set this.

Neil

Which Premphsis would have the first corner, the 25 micro second or the built in 50 micro second? any idea what audio frequencys those corners would occure at?

Ive built the networks for two different transmtters and they both offer a very smooth response to the ear, much better than stock. I wonder if I would take and audio sweep into the transmitter and watch the output on a scope from a good receiver if I could see if there are any bad peaks or what the response is in reguards to the modifactions... I have an audio generator that has both pink and white noise besides a 20-20K sweep- what do you think would that show me any thing? if the receiver had a liner response etc.. at any rate Ive did what i set out to do and I dont have to run an eq into the transmitter. Most of the transmtters ive woked on as mentioned Iv changed the 2.2 with a 3.3 nano caps and that works out fine but wanted to have an out side box to share with what ever I was using or testing.

I think a sine-wave is the correct type signal to apply, and it should be observed on an FM receiver with the de-emphasis disabled.

Or, this might be logical, if the de-emphasis in the radio is in circuit, I think you would see a FLAT signal in the pre-emphasis region IF the pre-emphasis circuit is doing its job.

I have a pioneer receiver it has a 50 or 75 us pre emp switch on the back, I can see what the orginal network does by setting the switch to 50 micor seconds it should be flat when I go sine wave 20-20 k if the oscillator is flat, I have a dual trace scope so can montor input and out put to compare.

I mean like 50 to 15khz -

The corner frequency can be calculated as Fc = 1/(2*pi*R*C) but since the time constant TC = RC then Fc = 1/(2*pi*TC) so for TC = 50us Fc = 3183 Hz and for TC = 75us Fc = 2122 Hz.

From this if the frequency goes from low to high the 75us TC will kick in first.

Neil

(Unsolicited advice.)

Care must be used when transmitting pre-emphasized audio waveforms.

For example, with 75 µs pre-emphasis the amplitude change of a "flat" audio waveform is about +17 dB at 15 kHz, compared to 50 Hz.  With 50 µs it is +13.7 dB, and with 25 µs it is 8 dB higher at 15 kHz than 50 Hz.  A voltage increase of 17 dB is a difference of 7.08 X in amplitude.

This means that without due care in the system design/setup, higher frequency audio easily can overmodulate AM and FM transmitters, causing distortion at the output of a receiver and (possibly) interference to radio stations on adjacent carrier frequencies.

Ideally, audio processing (expansion/compression) for AM/FM broadcast systems should be done by equipment designed and properly adjusted for the pre-emphasis curve in use.

Yes, I have spent a lot of time observing the very modulation issue Rich is talking about.

In the July 1964 issue of Electronics Illustrated an FM Micro Transmitter project led to a long period of experimentation in which I was able to really tinker with the basics of FM transmission.

The completed transmitter had such a weak signal I spent many hours hoping to find an error I might have made during construction, but the output power is a side issue.

Using three tubes, the transmitter put a very beautiful signal on the FM radio with a pre-emphasis circuit as being talked about above.

The transmitter featured an audio VU meter which was well designed to be accurate for controlling modulation level.

But I went a step further, and put a VU meter on the output of an FM receiver, which I calibrated to reflect the modulation of a high-power professional station, and then tuned to my frequency and saw that my transmitter level was matching that of the big station.

Little certified transmitters like the C.Crane have a calibrated red LED that flashes on over-modulation, and indeed it's the high frequency portion of an audio signal that sets it off.

The processing equipment that makes the apparant modulation seem much higher without over-modulating the highs is a whole other ball-park, beyond just talking about pre-emphasis.

Thank you everyone for all of your information you have given me, I have found my answers in your posts and also found this topic can get complex! In addition I have done more research and found this on the web for anyone else who wants to look into this further http://www.radiomuseum.org/forum/fm_pre_emphasis_and_de_emphasis.html
Its another look at preemphasis and compars 100 75 and 50 microsecond boost.
What I found out and what I was also lookng for, was at what frequency the different preemphasis starts,(by the graphs from the web link) the greater the time constant, the lower the frequency things start happening, so with 75 us you also get more mid range boost than with 50 us, and I also found out that adding another 25 us circuit will kick up the top end and not the mid range to some degree. I also realize that you can go too high on the pre emphasis and cause transmitter modulation overload with HF audio. I havent experienced this in the things I have been working on but I realize the possibility. Working preemphsis on AM transmitters has to be more critical in this respect.

I might add I suppose the best thing for people to do if they want to boost treble is to change the orginal pre emphsis circuit in the transmitter, but from my experience burning up circuit boards, in the hobby anyway, it seems what ever works will work too as long as there are no negative effects. I have used an equilizer to boost treble a bit on some projects, but it never seemed to have the correct balanced sound on the receiver as when changing the pre emp circuit correctly in the transmitter. Maybe im suffering from pre emp syndrone, and could be I should go out and get back to work painting the neighbors fence and get away from the transmitter room for a while !
Thank you again for your help on the subject!
Radio Joe the fence painter