Ampeg SVT-CL Fault Detection Circuit Theory of Operation

[nevetslab]

I'm in the process of writing a service manual for the Ampeg SVT-CL and SVT-VR family of amps (SVT-CL, SVT-AV, SVT2-Pro, SVT-VR), and one of the tricky circuits they created is the Power Tube Fault Detector Circuit. See Power Amplifier PCB Schematic # 07S419-04 attached. What I've written thus far is:

All six power tubes feed a Diode OR circuit to feed the nominal cathode currents as discrete voltages to the Fault Detector buss, which is pre-biased from the 6.3VAC heater supply voltage, 120Hz half-wave rectified to a comparator circuit at IC2B, which sits at around 2.65V peak. The trip threshold is set by the resistor ratio R48 input (100k), and R49 feedback (220k), which sits between the negative side of the charged cap C13 (positive side at +15VDC) and R50 (10k, the bottom side of which is Ground) and the output voltage of the comparator IC2B, which is nominally +14V. I think that sets a trip threshold of 4.67VDC on the Fault Detector buss, after which it toggles and produces a Fault signal (-14VDC) which both turns off the HT Power Transformer at the AC Mains PCB Assy and turns on the front panel Fault indicator (flashing RED LED).

I haven't lately opened up one of the rental SVT-CL's to get empirical voltage readings to confirm this....right now, just going off of calculations from the schematic. On the schematic, there are six lines from each cathode of the power tubes, coming down from the ribbon cable (from the power tube PCB), where power tube V1 cathode is on line 5, V2 cathode is on line 1, V3 cathode is on line 3, V4 cathode is on line 11, V5 cathode is on line 9, and V6 cathode is on line 7. The nominal DC voltage when biased for 23mA Cathode Current will be 230mVDC on all these lines. This would have the GRN LED's lit on both upper and lower Bias LED meters.
Tubes V1, V2 & V3 are the positive half, V4, V5 & V6 are the negative half.

The assumption for the comparator trip threshold voltage is based on capacitor C13 having charged up, so the negative side of it is now between 10k (R50), 100k (R48 & 220k (R49). As powering up sets the comparator to +14VDC output nominal, you have a voltage divider at the non-inverting input of IC2B of 220k (R49) & 110k (R48 + R50). That's where I calculate 4.67VDC at IC2B Pin 5. The Fault Detector buss, fed from each power tube cathode thru a 2.2k resistor, 10uF cap to ground, and diode, forming a Diode OR Logic Buss (so any tube can trip the Fault Detector circuit). There is an attenuated half-wave rectified 6.3VAC signal by way of D3 & D4, and attenuated by voltage divider R45 (47k) and R22k (22k), which applies a reduction ratio of 0.319 to the 120Hz half-wave rectified AC. That's where I calculated a peak value of around 2.65V.

Does this second paragraph make sense that I composed above?

Power Amplifier PCB Schematics (419xxh2).pdf

[Enzo]

I am a little in the cups at the moment, but I did write an analysis of the comparator circuits on here some time ago for someone.

[Helmholtz]

...to the 120Hz half-wave rectified AC.

I didn't analyze the circuit but that statement sounds contradictory. I think you meant full-wave rectified.
As the 6.3V heater voltage is symmetrical wrt to ground (by way of R41/R42), D3/D4 provide two-phase/full-wave rectification.

[Dave H]

There is an attenuated half-wave rectified 6.3VAC signal by way of D3 & D4, and attenuated by voltage divider R45 (47k) and R22k (22k), which applies a reduction ratio of 0.319 to the 120Hz half-wave rectified AC. That's where I calculated a peak value of around 2.65V.

Where do FIL1 and FIL2 connect? If it's to each side of a centre tapped 6.3V winding then it's full wave rectified (as Helmholtz says) and the peak voltage will be lower.

[Helmholtz]

Where do FIL1 and FIL2 connect? If it's to each side of a centre tapped 6.3V winding then it's full wave rectified (as Helmholtz says) and the peak voltage will be lower.

Heater are balanced by way of R41/R42 providing a "virtual center tap".

[nevetslab]

I didn't analyze the circuit but that statement sounds contradictory. I think you meant full-wave rectified.
As the 6.3V heater voltage is symmetrical wrt to ground (by way of R41/R42), D3/D4 provide two-phase/full-wave rectification.

YUP.....you're right. Oversight, as usual. I should have also included the SVT-CL power & heater xfmr drawings with this thread, though I did just post them on my recent SVT-CL/AV/VR...etc service notes post that is in editing at the moment. And, correctly observed, there is NO center tap on the heater winding of this family's xfmrs (nor on the SVT-VR xfmrs). I've added the SVT-CL xfmrs on this post below.

As seen on the ribbon cable diagram on the Power Amp schematic, the heater circuit comes in from the Power Tube PCB via pins 6, 8, 10, 12, 14 & 16 off that ribbon cable.

I had just begun writing up the basic theory of operation on both products, and in re-reading the Protection Circuit text, I knew I needed some smarter minds and eyes on the text to assist in this task.

I'm still puzzling over the an additional 6-channel circuit that also feeds this Fault Detector Buss...that being diodes D29, D30, D31, D39, D40 & D41 (all 1N4007's), C14 (1uF/50V), then the attenuator network R82 (33k) & R46 (22k). From there, it adds to the Fault Detector Buss. Theses six diodes are coming directly from the power tube's cathodes. My guess is, THIS is the path for the real Fault Current from any power tube, as it's being routed thru the 1A Rectifier diodes, while the other path which is tapped off of each tube's cathode resistor (10 ohm), that circuit path then feeds a 2.2k resistor, followed by 10uF35V cap, then feeds the signal diode 1N3070 high conductance fast 500mA diode and feeds the Fault Detector Buss.

Ampeg SVT-CL Power & Heater Xfmrs.pdf

[Dave H]

I'm still puzzling over the an additional 6-channel circuit that also feeds this Fault Detector Buss...that being diodes D29, D30, D31, D39, D40 & D41 (all 1N4007's), C14 (1uF/50V), then the attenuator network R82 (33k) & R46 (22k). From there, it adds to the Fault Detector Buss. Theses six diodes are coming directly from the power tube's cathodes. My guess is, THIS is the path for the real Fault Current from any power tube, as it's being routed thru the 1A Rectifier diodes, while the other path which is tapped off of each tube's cathode resistor (10 ohm), that circuit path then feeds a 2.2k resistor, followed by 10uF35V cap, then feeds the signal diode 1N3070 high conductance fast 500mA diode and feeds the Fault Detector Buss.

I think the main fault current path is the 10 ohm resistors but 1A diodes are used to withstand the peak charging current of C14.
My guess is that the D49,50,51 etc. diode circuit trips if the average current of any power tube exceeds a certain value whereas the D29,30,31 etc. diode circuit trips if the peak current of any power tube exceeds a certain (higher) value, higher because of the divider R82,46

[Dave H]

Heater are balanced by way of R41/R42 providing a "virtual center tap".

Thanks, I looked all over the schematic but missed that.

[nevetslab]

I think the main fault current path is the 10 ohm resistors but 1A diodes are used to withstand the peak charging current of C14.
My guess is that the D49,50,51 etc. diode circuit trips if the average current of any power tube exceeds a certain value whereas the D29,30,31 etc. diode circuit trips if the peak current of any power tube exceeds a certain (higher) value, higher because of the divider R82,46

I loaded the basic Protection Circuit into my NI Multisim 11 schematic program, and set up the circuit simulator, having just one source to simulate the positive-going signal voltage as seen at the cathode of the power tube, added a sinewave source for the 6.3VAC, added some multimeters to measure Comparator Output, Comparator Voltage Threshold at the non-inverting input pin, Fault Detector Buss, and a scope to see the signal source and attenuated heater signal. But, so far, I'm not getting the simulator to work correctly. I was hoping perhaps I could get that working so I could probe around to see what I might see if I open up one of our rental SVT-CL amps. I may just resort to that tomorrow morning, as watching the simulated multimeters start reading at near 0VDC, then progressively sail well beyond the +/- 15V supply voltages applied by the simulator. And, not getting ANY output from the 6.3VAC Source on the simulator. Ain't technology grand?!!

[nevetslab]

I finally got the circuit simulator & fault detection circuit I composed to work....sort of. The first op amp I had used...an NJM2060 seemed to have an issue in the simulator. Ended up using a NJM5532 for my precision rectifier circuit to simulate the positive half-wave signal coming from a tube, and a TL072 for the comparator, as is used in the Ampeg SVT-CL circuit. In the simulator, the comparator isn't yet working as it should, so still got to look at it by forcing the trigger to force the output HIGH, then trigger it to go low. In the simulator, I guess having a 2-ch scope and one or more multimeters active, everything reads slow.

I did at least see the 'tube signal' coming in by the two paths...one across the 10 ohm cathode resistor, followed by the RC network before passing thru the fast signal diode, then the slower higher path thru the 1A rectifier diode followed by the 1uF cap. Just seeing varying charged DC levels of the positive half-wave sine. What does remain on the Fault Detector buss is the attenuated full wave rectified sine from the heater winding. This weekend, I'll open up an SVT-CL to look at the circuits for real, as this simulator kind of seemed like an exercise in futility.....for as much as I use it.

The Precision Rectifier circuit obviously is not part of the Ampeg circuit. The rest was whittled down to just have one channel of the six that are used in the Ampeg circuit.

SVT-CL FAULT DETECTION CIRCUIT -1.pdf

[nickb]

The think the purpose of the rectified AC input is over voltage sensing. If the threshold is 4.67V then the RMS voltage on the filaments would be approx (4.67+ 0.65) * 0.707 *2 = 7.5V RMS which seems a sensible number. I don't think it's used as a bias since the normal voltage from it is less than the threshold and so has little effect on the other functions.

[REV-AMP]

I'm in the process of writing a service manual for the Ampeg SVT-CL and SVT-VR family of amps (SVT-CL, SVT-AV, SVT2-Pro, SVT-VR), and one of the tricky circuits they created is the Power Tube Fault Detector Circuit. See Power Amplifier PCB Schematic # 07S419-04 attached. What I've written thus far is:

All six power tubes feed a Diode OR circuit to feed the nominal cathode currents as discrete voltages to the Fault Detector buss, which is pre-biased from the 6.3VAC heater supply voltage, 120Hz half-wave rectified to a comparator circuit at IC2B, which sits at around 2.65V peak. The trip threshold is set by the resistor ratio R48 input (100k), and R49 feedback (220k), which sits between the negative side of the charged cap C13 (positive side at +15VDC) and R50 (10k, the bottom side of which is Ground) and the output voltage of the comparator IC2B, which is nominally +14V. I think that sets a trip threshold of 4.67VDC on the Fault Detector buss, after which it toggles and produces a Fault signal (-14VDC) which both turns off the HT Power Transformer at the AC Mains PCB Assy and turns on the front panel Fault indicator (flashing RED LED).

I haven't lately opened up one of the rental SVT-CL's to get empirical voltage readings to confirm this....right now, just going off of calculations from the schematic. On the schematic, there are six lines from each cathode of the power tubes, coming down from the ribbon cable (from the power tube PCB), where power tube V1 cathode is on line 5, V2 cathode is on line 1, V3 cathode is on line 3, V4 cathode is on line 11, V5 cathode is on line 9, and V6 cathode is on line 7. The nominal DC voltage when biased for 23mA Cathode Current will be 230mVDC on all these lines. This would have the GRN LED's lit on both upper and lower Bias LED meters.
Tubes V1, V2 & V3 are the positive half, V4, V5 & V6 are the negative half.

The assumption for the comparator trip threshold voltage is based on capacitor C13 having charged up, so the negative side of it is now between 10k (R50), 100k (R48 & 220k (R49). As powering up sets the comparator to +14VDC output nominal, you have a voltage divider at the non-inverting input of IC2B of 220k (R49) & 110k (R48 + R50). That's where I calculate 4.67VDC at IC2B Pin 5. The Fault Detector buss, fed from each power tube cathode thru a 2.2k resistor, 10uF cap to ground, and diode, forming a Diode OR Logic Buss (so any tube can trip the Fault Detector circuit). There is an attenuated half-wave rectified 6.3VAC signal by way of D3 & D4, and attenuated by voltage divider R45 (47k) and R22k (22k), which applies a reduction ratio of 0.319 to the 120Hz half-wave rectified AC. That's where I calculated a peak value of around 2.65V.

Does this second paragraph make sense that I composed above?

Power Amplifier PCB Schematics (419xxh2).pdf

Please be aware :
If you trace the PCB layout of the main board (the real world) of the SVT-CL and compare it with the Ampeg cct diagram (the intention of the designers), that R38m and R40 are transposed - they marked up the circuit diagram WRONG. Other components in the area are correct WRT the layout on the board. At first I thought the linkage from the header was drawn wrong but it is just the numbering of these two resistors that a re wrong.