Proper Adjustment of a CD-4 demodulator

QuadraphonicQuad

Help Support QuadraphonicQuad:

This site may earn a commission from merchant affiliate links, including eBay, Amazon, and others.

The Quadfather

1K Club - QQ Shooting Star
QQ Supporter
Since 2002/2003
Joined
Mar 7, 2002
Messages
1,587
Location
Dixie
I have written on this subject many times before, but it was suggested to me that the thing to do would be to put it in it's own thread to make it easier to find for newcomers. I will try to make this explanation informative and complete, but I will not address extra features on some demodulators such as crosstalk cancellation that are not on most demodulators. So here it is:

To start with, it helps to know what a CD-4 demodulator does, and that helps you understand what you are doing when you adjust one. The CD-4 record is essentially a stereophonic record with added supersonic carriers. Each groove wall is modulated with it's own audio and carrier signal. The cartridge separates the left and right channels the same way it does a stereo record, by each channel's coils being only sensitive to movement in the plane of direction of the signal that it's trying to pick up. In actuality, a CD-4 demodulator is in fact two demodulators in one box. Each demodulator separates it's channel into signals, front and back. Now, the original signal was in fact four channels, and it was mixed into two channels during the recording process, plus two carriers mixed with the two channels. Let's look at one of these channels:

Left front and left back signals enter the CD-4 modulator and are both split into two signals, each identical to it's counterpart. Let's call them "A" asd "B" for the sake of clarity. So then, "A" front is mixed with "A" rear. This is then sent through RIAA equalization and to the output mixer. This signal is low pass filtered to remove anything above 15KHz and recorded on the groove wall as left channel total, and if played with stereo gear, will yield all of the left channel information in the left speaker. So how do you get that into the original left front and left rear channels? The secret is in the carrier. A carrier is a sine wave frequency that "carries" a signal within it. A good example for this purpose is an FM radio signal. In the FM signal, the frequency is shifted in accordance to the modulating audio But the FM radio signal is transmittable, whereas, the audio is not, for all practical purposes. But on the CD-4 record, we're not trying to transmit anything. But what we are trying to do is to include another signal on the record. This signal will be the left front and left back signals that we are discussing now. So now we mix together the front "B" signal with the rear "B" signal. But before we do this, we do something with the rear "B" signal that we didn't do with the "A" signals. We invert the signal so that the positive peaks are now negative and vice versa. then the two "B" signals are mixed together which we will call "B" total. This represents the difference signal with which we will modulate the carrier. But first, it must go through an ANRS noise reduction process. ANRS is JVC's version of Dolby B, it's not the same, but it does something similar and we won't go into that for the purpose of this discussion. It's not germaine to understanding CD-4. So, once the B total signal is ANRS processed, we send it to the modulator. This is a phase locked loop system, and actually produces three types of modulation according to audio frequency band,, but they are similar and for this discussion we will call it FM. FM is actually one of the three types. The carrier is a 30 Kilohertz (KHZ) signal and is above the range of human hearing. The modulating "B" total signal will cause this signal to swing form 15 KHZ to 45KHZ. This modulated 30KHZ carrier is then sent to the output mixer where it is mixed with the "A" total signal and on to the cutting head amplifier. There is an additional process called Neutrex, which is corrective predistortion added to correct stylus tracking distortion. But again, that is not germaine. Now we have the recorded CD-4 signal. The right channel process is identical to the left, except perhaps in the Neutrex, as it is recorded on the other wall of the groove.

Now having said all that, now we have to separate out the four channels. Like the CD-4 modulator, The demodulator is actually two demodulators one for left and one for right. For the purpose of this discussion again, I will only discuss the left channel. The left channel signal is picked off the record by the cartridge and sent to the demod's left input. It is preamplified and brought to "line" level. It then goes through a band splitter, which splits the signal into two bands, one is everything below 15KHZ and the other is everything above 15KHZ. The below 15KHZ band is the left channel total signal referred above as the "A" total. It is sent through the RIAA equalization and then to the output mixers. The above 15KHZ signal is the carrier, and is sent to the phase locked loop discriminator to be demodulated to audio. This is the "B" total signal referred to above. It is sent to the ANRS processor to undo the ANRS encoding, and then to the output mixers. In the output mixer, the "A" total is split in two identical parts which we will call "A1" and "A2". The same is done with the "B" total signal. Now the A1 signal is mixed with the B1 signal. Here's where the majic takes place. Since the A1 and the B1 signal both contains the Left Front signal, the left front signal is doubled. But the A1 signal also has the rear left signal. The B1 signal also contains this signal, but as you recall, during the encoding process, in the B1 signal, this signal was inverted. So what happens is that when A1 and B1 are mixed together, the rear left signals cancell each other out, being exact opposites. So now you have only the front left signal which is sent to the output. Now to get the rear left signal, we invert the B2 signal and mix it with the A2 signal which is left uninverted. Remember that the A2 signal is the total left channel audio. Also, the B2 signal contains the left front uninverted, and the left rear inverted. But once this signal is inverted, we now have the left front signal inverted and the left back signal is uninverted all mixed together. We then mix that with the "A2" signal, and the A2 front channel signal mixes with the B2 front channel signal and since the B2's front channel signal is inverted, it is opposite and cancels. The A2 rear channel signal mixes with the B2 rear channel signal, and because the rear signals are both uninverted, they add together, and become the left rear signal. Once again, the same process occurs in the other demodulator for the right channels.

Are y'all still with me? Now for the actual setup.

OK, so how do you adjust the demodulator? On most demodulators, you will find three adjustments. They are Left channel separation, Right channel separation, and carrier level. After you get your turntable cartridge aligned, (a whole 'nother subject) you start with putting on a CD-4 record. I assume the demod is connected properly and switched to CD-4 mode. As before we will discuss only the left channel, which mirrors the right. The carrier level control is actually a ganged pot, and controls the left and right together. So once this setting is done, it is good for both channels. We'll start with it. Start by turning down the separation pots all the way counter clockwise. Turn up the carrier level all the way. Put on a record, I would suggest Doobie Bros, because it has good separation and will make adjustment easier. What you will hear, is the difference signals for both left and right signals. It will sound hollow, this is proper and normal, but it may also sound harsh and scratchy, which is not good, and indicates that the phase locked loop circuit is having trouble tracking the signal. Back down the pot slowly until you get a hollow sounding signal that does not sound harsh or scratchy. The more natural (except for the hollowness) the sound, the better. You may have to go through the whole range of the pot to find the best spot, and I recommend doing this. When you get it close, move the pot slightly, and then listen after you move it. Once you've obtained the least possible distortion, you have it set. Note that if you have a new stylus, it has to be broken in, and you might not get all the distortion out until this occurs. The best way to break in a stylus is to play about five stereo or SQ records with it. That way, you won't have to put up with scratchy sound while the stylus is breaking in. If your demodulator has two carrier level pots, one for left, one for right, adjust them with the amplifier's balance controls set for left front or right front channel only depending on which channel you are adjusting. If your demod has crosstalk cancellation pots, turn these down all the way counter clockwise to turn it off, unless you have the adjustment record that came with the demod. (The Technics SH400 is a good example of this type demod) You cannot properly adjust these controls without the record, and if you aligned your turntable correctly, you won't need it.

Now, you have a nice hollow sound oozing from your speakers, but it's not quite right sounduing, and it's definitely not quad. So now pick the left channel. Set the amplifier controls so that only the front left speaker is heard. Now start slowly raising the left separation control. You will hear the sound get louder as you bring in the left total signal. But loudness is not what you are going for. as the level of the left total signal catches up with the level of the left difference signal, you will notice that some of the instruments fade away while some get louder. This is what you are listening for. set the pot for maximum cancellation. Now, here's where you have to make a choice. Do you want maximum separation, or would you be willing to sacrifice a little separation for better fidelity? If the answer is better separation, then you are there. If you want better fidelity, then increase the pot just a little more past the null point. The more you turn it, the louder it gets and the more separation you lose, so don't go too far. You can get all the loudness you want with the amplifier's volume control. Then set the amplifier's controls so that you hear only the right front signal and repeat the same procedure for the right separation pot. Once done, if the demod is working right and you did it right, the left front and the right front channels should be about the same levels, and ditto for the back channels. Set the amplifier balance controls to center,, pop a brewsky, and sit back and enjoy some great quad.

The Quadfather

P.S. I invite discussion and questions on this subject.
 
Thanks for this great help!!!!
Are y'all still with me?
LOL on myself. Still trying to catch up. I guess it will take a lot of rereading your message untill it sinks in. It seems so complicated for a nontech guy like me.
Please forgive me for being so redandant, If using the demodulator as a phono preamp for a stereo playback, with the CD demodulator turned on, I am like using a stereo phono preamp and the sound does not lack in anything during the process of demodulator except the 15KHz up? When we say 15Khz is this the same as the frequency of the mucsic as in 30Hz is bass and 15Khz treble?
 
The human hearing range is said to be from 20HZ to 20 KHZ. Yes, it's the same value. Hertz is a measure of frequency and is otherwise known as cycles per second. The AC line frequency is 60 hertz. A Kilohertz is a thousand hertz. A cycle is one transition of the voltage from zero to it's positive peak then to it's negative peak and back to zero at which time another cycle begins. Voltage is a measure of electrical pressure.

In reality, most people cannot hear above 15KHZ, Young people can hear at 15 KHZ, but anyone that has ever heard a malfunctioning horizontal oscillator in a TV set will say it's not a pleasant sound. It will literally give you a headache. But that's not music. I knew my hearing had diminished over the years, but recently I was playing with an audio oscillator at the studio, and was shocked to discover that I could not hear anything above 11KHZ. Yet my enjoyment of music is not diminished. I could once hear the horizontal oscillators in broken TVs. So the 15KHZ cutoff in CD-4 does not concern me.

A CD-4 demodulator is very much like the stereo demodulator in an FM radio. It's the same principle. The main difference is with FM radio, there is only one demodulator, and there is no ANRS. BTSC TV stereo is the same way, only it uses DBX instead of ANRS. One thing I didn't mention in my explanation is the fact that the cutting heads of the day could not record up to anywhere near 45KHZ, the top limit for CD-4. To solve this problem, once a four channel master tape had been made, the tape was played back into the CD-4 modulator at half speed and the turntable on the cutting lathe was run at half speed. This cut all the frquencies in half, allowing the cutting head to handle the bandwidth. Of course the carrier frequencies and the RIAA encoding and the ANRS encoding had to be adjusted also, as these are all frequency dependent. This actually improved the audio portion of the records such that there was a time when stereo records were cut this way and sold for a premium price as audiophile records.

The demod, if that is indeed what you have, can be used as a preamp, it will amplify stereo to line level just fine. Occasionally you might run across a record with high frequency content that will cause the phase lock loop circuits to try and lock to it, which will cause a major splatter sound when this happens. Elton John's "Goodbye Yellow Brick Road" will do this. (To avoid this, remove the input capacitors to the phase locked loop chips. You can solder wires into the holes and mount a double pole double throw switch on the rear panel. Mount the capacitors to the switch by the leads. I will write more of this if you are interested) But first, we have to determine if what you have is really a demodulator. If you can get a picture of the rear panel, the front, and a shot of the circuit board with the top removed. Then maybe someone on this board might be able to tell you what you have.

The Quadfather

P.S. If you want a CD-4 demodulator, Lou Dorren is building a new design. You can sign up for one if you want one, but you have to get your name in the pot before he goes to production. It will sell for a ballpark figure around $500.00 . Lou's demodulator operates on a different principle from other demods, and therefore will not have a carrier level control.
 
Start by turning down the separation pots all the way counter clockwise. Turn up the carrier level all the way. Put on a record, I would suggest Doobie Bros, because it has good separation and will make adjustment easier. What you will hear, is the difference signals for both left and right signals. It will sound hollow, this is proper and normal, but it may also sound harsh and scratchy, which is not good, and indicates that the phase locked loop circuit is having trouble tracking the signal. Back down the pot slowly until you get a hollow sounding signal that does not sound harsh or scratchy. The more natural (except for the hollowness) the sound, the better.

I looked at few CD-4 schematics. The Carrier level pot does the only one thing - turns on or off the demodulated (after the FM demodulator!) signal. Meaning making the output 2ch or 4ch. There is almost no quality change with pot rotating as long as the "Radar" light stays on. The only noticable change is near the "radar" turning OFF - the audible signal is being attenuated without any change in quality. Since the attenuation (on/off) is done AFTER the FM demodulator the position of the pot has no effect on the quality of the demodulation.

Since CD-4 demodulators are quite old it is highly recommended to replace electrolythic caps at least in critical areas. There is a cap in series with the FET(s). It must be replaced with new one. The decoupling caps in the decoding matrix MUST be replaced. This results in DRAMATIC improvement.

My recommendation - crank carrier level all the way up and leave it like that. The only potential issue with it is that if the demodulator output is quite high it might cause the signal clipping. Which I do not believe is the case - the signal is quite weak. And if the clipping does occur (capture the signal with PC to check for it) it must be solved by other means - by adding extra pot to attenuate the signal (in parallel with the FET). Using rectified carrier level to control the FET is plain silly.
 
I set the carrier signal w/ a JVC CD-4 set-up record. I set up L-R separation by ear over the course of 2 or 3 records to get a best case scenario.
 
There's more to it than what appears to be. The carrier level is not directly in the signal path, but it is in a path that affects the loading on the output of the chip. I will have to look into this further. But I know it has more of an effect than to just turn off the light. I never thought much about it, but yes, this is odd, now that I've taken a good look at the schematic. I would be very interested to learn more about this.

The Quadfather
 
I think I was not quite right. The demodulator is based on CD894 (NE565) chip. The output (pin 7) has a series resistor of 3.6K. With an external resistor (FET) it will make up a voltage divider. The voltage at pin 7 will control the VCO inside the chip. Hence it will control the gain of the feed-back.

Assuming the PLL is locked, i.e. VCO follows the input frequency the output (pin 7) should be the same for the same frequency sweep. Regardless of the resistance of the FET. The lower resistance of the FET the higher the voltage at the internal amplifier should be (before 3.6K resistor) in order to maintain the same voltage at the pin 7. It still appears to me that open FET will result in least required voltage at the internal amplifier and as result in cleaner signal.

What is missing in CD894 is the limiting pre-amp. It seams that feed-back gain control is trying to achive similar functionality by increasing loop gain when the carrier is getting stronger. Which is a lousy way of doing it IMO. I wonder if the mod with limiting preamp before the chip will be beneficial here. It is easy to do with op-amp and two diodes in the feedback.
 
The more I think about carrier level the more it seems to me that FET was placed there to break the feed-back when carrier is not present. When carrier is not there (LP is not played) FET is closed (shorted) and the loopback is broken. Phase detector is not trying to track the noise and the VCO runs exactly at center frequency. When carrier becomes present FET becomes open and feedback is in place for the VCO to track the input signal. Thus FET control is there to make a lock on the carrier easier.

I just wish I understood FET operation better. It seems to be N-channel FET. I believe when Gate and Source are shorted (or Gate is more positive than Source) the FET is closed (lowest resistance). To make FET open the Gate voltage has to be made negative with respect to the Source. If I am right it means No Carrier - FET is closed and Carrier present - FET is open. Therefore Carrier Level pot will control how much FET is open.

If I am correct it still seems that keeping the pot cranked all the way up is the right thing.

Quadfather, I would appretiate your thoughts on this. Thanks.
 
Hey Jason:
I do not know why an op amp and two diodes for limiting the signal before feeding to the phase locked loop wouldn't work. However, I get quite satisfactory performance with it just the way it is. It's all in the quality of the signal being fed in. Good carrier, good lock. But we have to remember, that during recording, the carrier level was modified in accordance with the level of audio. When the audio was soft, the carrier was brought up to decrease the noise level. When the audio was loud, the carrier level was reduced, because any noise produced would be masked by the audio, and there was less groove depth available not taken up by the audio.

Now, we see that this circuit has the carrier level pot feeding a single stage amplifier. This feeds a diode array that I assume changes the carrier AC to DC but I don't see much capacitance for ripple filtering. Of course, at 30KHZ you wouldn't need as much as for 60HZ. This DC voltage is fed through another adjustment pot to the FET. Now the FET seems to act as a signal sink for the output of the chip. There's a capacitor between them and that will block any DC bias. There is also, at the bottom of the pot, a connection to the other channel same point. This is to allow each channel to influence the other somewhat. There should be no RF or audio here, so it is just control voltage. It seems the more signal applied (more carrier) the more the FET is turned on, shunting signal on the pin to ground. With less carrier, the FET would be biased closer to off and there would be less shunting action and the signal on the pin of the chip would be less attenuated. I haven't messed with field effect transistors in a long time, but their action is different than a bipolar transistor. It's more like the way a vacuum tube works, except with advantages a tube doesn't have, like being able to use an AC only signal as it's load, which is the case here. We are using the DC voltage to control flow of an AC current, in this case, to ground.

In an FET, the semiconducting material has no junction between source and drain. But there is a junction between the gate and the others. The gate is usually reverse biased, so there is no current flow. But a voltage will set up a molecular field around the junction, and as the bias voltage increases, this field will increase in size until the whole semiconductor is taken over. The field blocks or restricts the flow of electrons from the source to drain, and vice versa. Therefore a DC on the gate will control the flow of an AC current between the source and drain. Whether that FET is N channel or P channel determines whether that voltage on the gate has to be negative or positive relative to I believe, the drain. But then, like I said, I'm a little rusty on this .

The Quadfather
 
Nice to see my quad stuff being used (y):D

go to www.grizwald.plus.com/quad

there are QS and SQ test tones along with other goodies including an article detailing mods to a JVC 4DD-5 or Marantz CD400 (same demod, different case) that may be of interest.

Malcolm
 
My QC-04 has a CD894 PLL based FM demodulator in it. It also appears to use ANRS with adjustment pots and a level control after the LPF after the FM demodulator. I'm trying to think of ways to test it. The QC-04 I got came from Etco was incomplete with no cover and might not have been adjusted before I got it. What if I made a 30 kHz VCO type generator that can be FMed? I could test the deviation using a deviation meter or maybe using the modified Lissajous figure method with an oscilloscope. Assuming there is limiting of the carrier in the QC-04, the level of the demodulated tones should correspond to the deviation of the FMed carrier used to test it.
 
One idea I had for a test generator was to have two oscillators such as two VCOs, that generate frequencies just above and below 30 kHz and a switch controlled by another oscillator that produces an audio frequency square wave, to switch between each ultrasonic oscillator to produce basically a frequency shift keying signal (carrier FMed with square wave) centered around 30 kHz. I have a frequency counter so I can measure the frequency of each oscillator accurately. The oscillators can then be adjusted to control FM deviation and audio frequency of the modulating signal. The disadvantage is that the signal is modulated with a square wave and not a sine wave. The amplitude shouldn't be too critical since the PLL FM demodulators in the QC-04 should only demodulate the FM and not AM.
 
I have the Quadradisc sample record with test tones - the first two tones are for left and right pot adjust. It says for minimum output on the rear left speaker adjust till you get minimum output - then white noise. When I do this i get minimal output but then after doing the right as well it sounds awful.
So what IS the procedure for using these tones, or are they a waste of time? Many thanks. J
 
I have the Quadradisc sample record with test tones - the first two tones are for left and right pot adjust. It says for minimum output on the rear left speaker adjust till you get minimum output - then white noise. When I do this i get minimal output but then after doing the right as well it sounds awful.
So what IS the procedure for using these tones, or are they a waste of time? Many thanks. J

Cool! I adjust my pot (s) too! :yikes
 
I have the Quadradisc sample record with test tones - the first two tones are for left and right pot adjust. It says for minimum output on the rear left speaker adjust till you get minimum output - then white noise. When I do this i get minimal output but then after doing the right as well it sounds awful.
So what IS the procedure for using these tones, or are they a waste of time? Many thanks. J

I am not a big fan of CD-4 test records, I have never had much luck getting the optimum performance from them. I would suggest you do this:

Turn down the separation pots all the way. Set the carrier level pot in the middle. I like to use Doobie Brothers "Toulouse Street" to do this, because it's recorded well, and has good separation. Put on the record.
Listen: The audio is the separation signal from the subcarriers. It will sound weird and hollow, but not distorted. If there is distortion, adjust the carrier level for the best sound. If the distortion cannot be eliminated, then the stylus is not tracking the record properly, and improvements will have to be made to the turntable pickup system, assuming the demod is not malfunctioning. If the signal is not distorted, then you are ready to set the separation pots. set your amplifier(s) so that only the left front channel is heard. Now, bring up the left channel separation pot and listen for some of the instruments to disappear. There will be a place in the travel of the pot where maximum cancellation is heard. That is the null point. go higher and the separation decreases. I like to set mine just above the null point to maximize fidelity over separation, but that's personal prefernce. set your amplifier for the right front channel and do the same with the right separation pot. Now turn on all the channels and you should hear great quad.

If you find yourself needing to improve the pickup system, an inexpensive cart is the Audio Technica AT440MLa. It will do a very good job with CD-4. But don't read the specs, they will only cause unwarranted doubt.

P.S. I wrote these words from an Email posting, not realizing that I had already written the CD-4 setup procedure on this thread. But I decided to let this second telling stand, because it is worded differently, and might contribute to understanding. There is a discrepancy in where to set the carrier level pot between the two tellings, both will work, since the object is to then readjust the pot for the best signal.
The Quadfather
 
Thanks Quadfather - most helpful. I have spent the weekend rewiring the decoder so that I can plug my turntable output easily from my Whest phono amp to my Grundig decoder easily. It took hours and hours to get it right - I settled on XLR plugs, with allowances for the fact (not many know this) that the output of a cartridge is a form of balanced audio, with ground. This works well using XLR plugs and its easy to utilise ground wires too. This worked fine for the phono amp, but getting audio into the Grundig CD-4 decoder proved to be much harder. Capacitance issues are REALLY CRITICAL on cabling, I have learned to my cost. I spent the night thinking about how to resolve this issue and what I could use that didn't cost a fortune to use as low capacitance interconnect. I hit on some spare coax used for aerial feed - which I use for FM. It worked! After all FM coax is used for extremely low level high frequency signals. The Grundig has 4 sets of din plugs which proved to be unreliable, and the signal has to go through a 40 year old switch, plus old cable etc.. so I elected to strip all that out and directly inject the phono audio into the PCB, and also run some new output cables, again direct from the PCB and ditched the output DIN plugs. The result was spectacular - and I have been listening to amazing quad all afternoon. (I set up as you suggested). My only thing I need to do is to renew / improve the power supply which occasionally can be a little buzzy - but that comes and goes. Its quite likely the smoothing capacitor is tired. Anyway a couple of discs I bought are brilliant - I got the Doobie Brothers, A wonderful Barry Manilow, and Tomita's Bermuda triangle is a MUST.
 
Back
Top