Troubleshooting the ST-6 demodulator

Complete troubleshooting instructions are given along with voltage measurements and theory of operation

By

Irvin M. Hoff, W6FFC

The article on the ST-6 in the January issue was somewhat general in scope. There are several topics remaining which are of interest to the serious enthusiast, including voltage charts, troubleshooting comments and more detailed analysis of circuit operation.

voltage measurements. These were all made with a Heathkit vtvm with the standard 10 megohm dc probe. The power supply had +/-12.0 volts output. On each op amp, the voltages at pin, 4 were about -11.9 volts, and at pin 7, about +11.9 volts. If you are reasonably close to these measurements and the ones that follow, you should expect normal operation in that part of the circuit. Other voltage measurements are shown in table 1.

Some latitude on these measurements is acceptable. The voltage with mark signal at test point 2 might vary from unit to unit, but should be 6 to 9 volts. All measurements on U1 should be pretty close to those given. On U2 most of the measurements will be pretty close, if the voltage at pin 6 is not between 7.5 to 9 volts, adjust resistor R’D’ as explained later. On U3 the measurements will be very close again, with voltages at pins 2.3 and 6 slightly less than at pin 6 of U2. U4 should be quite close to these. On U5 the voltages at pins 2 and 3 may vary a little, but the rest should be very close. On U6 and U7 the voltages should be pretty close to those shown in table 1.

On Q1, the voltages at the base should match very closely, although those on the collector can vary some. That is, with a mark signal, you might get from 0.5 to 2.0 volts; the voltage on space will depend on your loop supply and transformer, but probably will be between 140 and 200 volts. Q2 and Q2 voltages should match closely. The voltage on the base and emitter of Q3 may vary somewhat. The voltage on the collector of Q4 may be anything from almost zero to 0.2 volts. The voltages on Q5 and Q6 should be about the same as those in table 1. On Q7 the collector voltage with mark signal will be from almost zero to 0.2 volts, and the voltage on space might be anything from 8 to 10 volts.

troubleshooting

The first thing to check is power supply output; make sure you are getting close to +/-12 volts. Then start with the U1 stage. Disconnect the input to the ST-6 and balance the pot for 0 Vdc output; then connect the input and insert a mark signal. Check the voltage at test point 2; it should be around 7 to 9 volts. Check the voltage at test point 3, it should also be around 7 to 9 volts. If more than 9 or less than 7, you might want to change resistor RD on U2 until you get close to but not more than 9 volts. Then pick an appropriate capacitor for CC from table 2.

(Note: RD is R212 and CC is C206 on Circuit Board 2, 850 Discriminator-L.P. Filter)

By keeping the voltage close to 9 Vdc maximum dynamic range will be obtained for limiterless linear operation. If the voltage at test point 3 is allowed to exceed 9 volts, the input to U4 could (with RTTY being received) go higher than 4.5 volts and possibly damage U4. as the maximum differential input voltage for a 709C op amp is 5.0 volts.

If the output at test point 3 is around 8 or 9 volts, then the output at pin 6 of U4 should be approximately +10.8 volts. This should cause the printer to stay in mark. With a signal, the motor should have come on within 4 or 5 seconds or less; if not, the autostart system is not working properly or has not been adjusted correctly.

theory of operation

automatic printer control. Let's approach this from two different aspects. Let's say we have been printing an incoming signal; just this instant it quit and we are getting random noise from the receiver, the motor is still running, and the receive lamp Is still on.

With a random noise input, the voltage at test point 2 will vary from one moment to the next, but in any event it should be less than the 7.5 volts or so you had when the signal was present. Thus, the fixed bias on the non-inverting input of U5 will be greater than that coming to pin 2 of U5. (The two 68k resistors, R66 and R67 reduce the voltage at test point 2 to one-half so that the 5.0 volt input limit of U5 will not be exceeded.) Thus U5 is now controlled by the voltage at pin 3, and the output is approximately +11 volts. This is passed by CR27 and charges the 350 uF capacitor C21 through R61.

Since C21 is in parallel with a resistor network, the capacitor takes 0.8 to 1.0 second to charge. As it charges, the voltage at terminal 3 of U6 rises to about 4.7 volts. As it goes above the 2.2 volt fixed bias level on pin 2, it takes charge of U6, which then flips to positive output of about +11 volts. This is blocked by CR25 but passed by CR24, and applied to the standby line, which causes Q1 to conduct and puts the printer into markhold.

table 1. DC voltage measurements In the ST-6 RTTY demodulator.
Op Amp U1
Op Amp U2
Op Amp U3
Pin#
Mark Sig.
Pin#
Mark Sig.
Pin#
Mark Sig
Space Sig.
1
7.8
1
8.2 .
1
8.2
na
2
0
2
0
2
8.5
-8.5
3
9
3
0
3
8.5
-8.5
4
-11.8
4
-11.9
4
-11.9
na
5
-11.4
5
-11.4
5
-11.5
na
6
7.7Vac
6
8.4
6
-11.5
-8.5
7
11.8
7
11.9
7
11.9
na
8
9.4
8
8.7
8
8.5
na
Op Amp U4
Op Amp U5
Op Amp U6
Pin#
Mark Sig.
Pin#
Sig.
No Sig.
Pin#
Sig.
No Sig.
1
na
1
7.4
8.2
1
7.5
8.1
2
0
2
3.9
0
2
2.2
2.2
3
2.0
3
3.4
3.4
3
0
4.7
4
na
4
-11.8
-11.8
3
-11.9
-11.9
5
na
5
-11.0
-11.9
5
-10.8
-11.9
6
11.0
6
-10.8
11.4
6
-10.8
10.8
7
na
7
11.8
11.8
7
11.9
11.9
8
na
8
11.5
7.8
8
11.4
7.7
 
U7
Q1
Q2
Pin
Mark
Space
Pin
Mark
Space
Pin
S3 On
S3 Off
1
7.5
8.1
B
0.6
-0.7
B
11.2
12.0
2
2.3
2.3
E
0
0
E
12.0
12.0
3
0
4.2
C
1.0
170
C
11.9
0
4
-11.9
-11.9
Q3
Q4
5
-11.9
-11.9
Pin
Sig
No Sig
Pin
Sig
No Sig
6
-10.8
10.8
B
-10.3
0
B
-0.7
0
7
11.9
11.9
E
-9.5
0
E
0
0
8
11.5
11.5
C
-12.0
-12.0
C
-0.03
-12.0
 
Q5
Q6
Q7
B
-11.3
-12
B
-11.9
-11.3
B
0.7
-0.7
E
-12.0
-12.0
E
-12.0
-12.0
E
0
0
C
-11.8
+10.2
C
+12.0
11.8
C
0.05
9.3
 
Test point 2: Mark signal, 170 shift discriminator 7.8Vdc; Input to St-6 disconnected, 0 volts.

 

At the same time, the 20-uF capacitor no longer has any voltage to keep it charged, so its 9-to 10-volt charge starts to bleed off slowly. When it has dropped to less than -0.6 or -0.7 volts, Q3. Q4 and Q5 stop conducting, causing the relay to open and the motor to turn off.

At this time Q6 conducts pullingabout the same current through the 500-ohm resistor R47 that the relay had been pulling through Q5. Thus the drain on the power supply remains rather constant, and better regulation is possible. In fact, at the time this particular part of the circuit was developed I had not Intended to use any type of regulation in the power supply at all.

table 2. Correct values of CC for dIfferent values of RD (see fIg. 5, reference 1).
Resistor RD Is shown for 9 volts at test point 3 wIth a mark input (see text for discussion).

RD
CC
300k
.018uF
270k
.02uF
240k
.022uF
220k
.025uF
200k
.027uF
180k
.03uF
160k
.033uF

For the next configuration we have the motor off, and there is no signal. Now we suddenly get a signal into the ST-6. The voltage at test point 2 goes to approximately 7.5 volts. This becomes around 3.8 volts at pin 2 of U5 and as this is somewhat more than the fixed bias (3.2 volts) at pin 3, the output switches from +11 volts to -11 volts. (Pin 2 is called the "Inverting input," and a positive signal becomes negative at the output.) This voltage is blocked by CR27, so the 350-uF capacitor C21 starts to bleed off through R59 and R60.

As the voltage at pin 3 fails lower than about 2.2 volts, the fixed bias on pin 2 takes over, and the output of U6 flips from positive to negative output. This is blocked by CR24, so now the standby system is removed and the printer Is free to follow the incoming signal. At the same time the negative output of U6 is passed by CR25, so the 20-uF capacitor C20 is quickly charged through the current-limiting resistor R55.

As this happens, Q3, Q4 and Q5 all conduct, closing the relay and tuming the motor on. This causes Q6 to stop conducting, so again the current in this part of the circuit is similar to what it had been when the relay was not being used.

At this point it is worthwhile to mention how the circuit could be changed to a "fail-safe" type. The relay could be placed in the collector of Q6, and resistor R47 would then be in the collector circuit of Q5. The relay would be activated any time the motor was supposed to be off (the back contacts on the relay would be used for the motor), and if the relay or any part of the ST-6 failed, the motor would come on automatically. However, with solid-state circuits there is not too much reason to worry about a fail-safe system

fast-slow switch. This consists of a 150-uF capacitor C22 in series with the 350uF C21. The total capacitance in "fast" then becomes about 85 uF, and the circuit responds in about one-fourth the normal time, making attended fastbreak operation possible, yet retaining the features of automatic printer control. Switch section S4B operates at the same time to keep the motor running. This type of circuit is not suitable for unattended operation due to the short time constants.

anti-space. This circuit is quite interesting, and some knowledge of how it works is beneficial. Basically it puts the printer back to markhold if the input signal goes to space for longer than a normal RTTY character. This also prevents the autostart from activating if a signal appears in the space channel. Therefore, if a station is playing around with his shift and going between mark and space, your machine will not run open.

With a mark signal a positive voltage will appear at the output of U4, causing Q7 to conduct, Its collector then goes to less than 0.2 volts, effectively short-circuiting the 1O-uF capacitor C19. The 330-ohm resistor R40 is there to limit current through the transistor as it suddenly shorts this capacitor. The voltage at pin 3 of U7 now becomes approximately 0.1 volt or less, and the 2.5 volts fixed bias on pin 2 takes over, putting negative voltage out of U7. This is blocked by CR21 and CR22, so nothing at all happens.

With a space signal the output of U4 negative, so Q7 stops conducting and is biased off the voltage held to -0.7 by the protective diode CR20 in its base. The 10-uF capacitor is now charged to about 9 volts. This becomes about 4.0 at pin 3 of U1 as the network is provided to keep this less than the 5.0 volts maximum allowable input. It will take about a quarter-second for the voltage at pin 3 to build up more than the 2.5 volts fixed bias on pin 3. When it does, U7 flips to positive output, and the e voltage is passed by both CR21 and' CR22. CR21 goes to the standby line, and puts the printer into markhold. At the same time the voltage through CR22 goes to the automatic printer control line and starts to charge up the 350-uF capacitor C21. This will take 0.8 to 1 second, putting the system into "autostart off."

As this happens much faster than the incoming signal could discharge that capacitor, the motor would never turn on should an incoming space signal be received, and if a signal suddenly goes to space after printing authentic RTTY, the autostart would soon be disabled, and 20 to 30 seconds later the motor would turn off. The anti-space works equally well whether the limiter or autostart are on or off, or whether the incoming signal is being "straddle-tuned" due to incorrect shift.

No provisions were provided to disable the anti-space. However, if you need this feature for any reason, just short the collector of Q7 to ground. A switch could be permanently added for this purpose, if needed.

fsk output. With nothing connected to the J3 keyer jack, the voltage there should swing from approximately -35 volts on mark to +35 volts on space. Without the 12k resistor R31, this voltage would be around 70 volts or so, which is enough to destroy most germanium diodes used in typical fsk circuits. When the fsk system is plugged into J3, the current on conduction is held to 4 or 5 mA by the 8.2k resistor R30, offering suitable saturation for good keying of the transmitter.

the threshold corrector. I have not said much about this circuit, and indeed it is complex enough to merit a complete article of its own. Perhaps a few words will help you understand its operation. If a steady mark signal is present, the voltage at test point 3 will be about 8.5 volts. Let’s call this 8.0 volts for the time being. This voltage would charge C6 and also be passed by CR 10, eventually going through R18, R19 and CR13 to ground. Thus, the voltage at the input of the normal-reverse switch would be roughly 4.0 volts. (There is another network in the normal-reverse switch which further drops this 4.0 volts to 2.0 volts at pin 3 of U4: this will be explained in a moment.)

With steady mark of 8.0 volts, the output of the atc section is 4.0 volts and the input to U4 is 2.0 volts. If you now flipped suddenly to space you should get the same 8.0 volts at test point 3; however it would no be -8.0 volts. This would be passed by CR11, and on to ground through R19, R18 and CR12, again putting -4.0 volts s at the output of the atc system. This negative -8.0 volts would also charge C5.

The interesting thing that happens on a quick reversal such as this, however, is that the previously charged capacitor C6 now discharges, adding another -8.0 volts to the system, which is reduced to 4.0 volts by the action of R18 and R19; thus you get the -4.0 volts from test point 3, plus this capacitor discharge voltage of an additional -4.0, making a total voltage at the output of the atcc system of -8.0 volts.

With steady input of mark or space, you would have half that voltage appearing at the output of the atc circuit, but with steady reversals (RTTY characters), the voltage on the output would be approximately equal to the original input. Therefore, the new voltage at the input of U4 would be about 4.0 volts. The reason for the 220k resistor network now becomes more apparent, if you remember that the input of U4 cannot safely exceed 5.0 volts.

If copying mark-only signals but with RTTY reversals, it can be shown that the voltage at test point 3 would be roughly 10.0 volts for mark and close to zero volts for space if you assume for a moment that the system doesn’t particularly respond to noise alone. In this case you would have an on-off voltage instead of the plus-minus swing previously mentioned with normal two-channel RTTY.

I could show that the voltage at the output of the atc would go from +4.0 to about -4.0 volts and the voltage at the input of U4, instead of being +/-4.0 volts with normal RTTY, would be +/-2.0 volts with mark-only RTTY. Since the slicer is operated in an open-loop configuration, anything more than a few microvolts plus-and-minus will cause it to operate suitably. The important thing is to keep the information properly centered plus-and-minus, not how much it swings one way or the other.

The atc is a most fascinating circuit, and goes a great way in counteracting for signals that, drift, for shifts that are incorrect, and even to some extent, for signals that have distortion on them when received.

An important thing most RTTY enthusiasts overlook is the diversity effect offered by such a system, as it actually samples both the mark and the space signal and uses either or both to provide proper information to the slicer. A system such as that used in the ST-6 actually offers diversity reception with only one antenna, one receiver and one converter. At one time commercial stations had ‘to go to dual-diversity reception to allow f or selective fading. This required two antennas, two receivers and two converters that fed into a diversity combiner. All this duplication was beyond the ability of most amateurs (and many commercial and military installations as well) so other means of improving the signal had to be found. The atc represents one of the best of such solutions.

Another system which is much more sophisticated but built around the same concept is called dtc (decision threshold computer) and was patented by Page Engineering for use on the Dew-line defense system. As long as the limiter is used, for all practical purposes atc works as well as dtc. With the limiter off, the dtc has some advantages; particularly in slow hand-speed transmission, but it does not work well at all if mark-only handsent transmissions are being used. (Neither system works well on slow-speed space only.)

Thus, from a practical standpoint atc was selected for the ST-6 rather than dtc. With the simple linear discriminators used in the ST-6 (and most TT/L-2 units), limiterless copy has little to offer. You also lose stability of tuning indicator information, whether using the meter or an external scope, and you lose the ability to include automatic printer control. Finally, with low voltage solid-state circuits, the dtc is not at all practical in the form most amateurs have seen it.

I did develop a dtc system for 12-volt systems, but it took a total of 8 op amps, and to be compatible, you would have to add active detector circuits that require two more op amps. All this seemed too stiff a penalty to pay for the occasional limiterless operation an individual might use. If going to that extra work, then it seemed silly to retain simple linear discriminators. As a consequence, the ST-6 is presented in a form that represents Excel lent performance, with practical considerations rather than going for the very complex unit that could have been designed. As an example, if you were to build a really top-of-the-line unit with sharp mark and space filters (three toroids each) then you would find it difficult to use the unit at all if somebody’s shift were off a little bit. As you are no doubt aware, few amateurs actually are within 50 Hz of 850 to start with, so practical considerations almost immediately rule out using first-rate filters anyway. Other examples could be used to show why, with present circumstances, you might not want such a unit even if it was offered.

no automatic printer control. While I think you ought to build the ST-6 pretty much as presented, some people still insist they do not need the autostart features. They would probably not want the anti-space either in this case, in order to save additional construction costs. The ST-6 was laid out so that all you need for similar performance is a loop supply, a +/-12 volt supply, the Ul, U2, U3, U4 and

Q1 circuitry, and a switch from the collector of Q1 to ground for standby. That would do it. If that is still more than you need for R V work, forget the ST-6 and build the ST-5 presented in the September, 1970 issue of ham radio.2

summary

For those of you interested in learning more about solid-state circuits, this brief review should help In any event you will better understand how the ST-6 works and what the various systems are supposed to do. It h taken me about six years of development to reach this phase, and there were ma y interesting thoughts picked up along the way. I feel the ST-6 will not be obsolete or antiquated in any way for some years to come, although some designers m y wish to use more exotic filters, or newer op amps, or add digital logic for autoprint.

Incidentally, Sel-cal devices (3) attach to the ST-6 with only a 10k resistor and a silicon diode. Since only a few dozen of the Selcal units have been built, write to me if you need that information. Large, easy-to-read schematics are still available from me for $1, postpaid anywhere in the USA or Canada; add $1 for air mail to other areas. Over 400 amateurs have requested schematics as of this writing, so it seems evident that the ST-6 promises to be one of the most popular RTTY demodulators of all time.

references

1. Irvin M. Hoff, W6F FC, "The Mainline ST-6 RTTY Demodulator," ham radio, January, 1971, page 6.

2. Irvin M. Hoff, W6FFC, "The Mainline ST-5 RTTY Demodulator," ham radio, September, 1970, page 14.

3. W. Mallock WA8PCK, and T. Lamb, K8ERV, "The Selcal -- An RTTY Character Recognizer," 73, May, 1968, p. 58.