- TVS diode across the output is on my list of ordered parts
- I am not sure what you mean by output network? The output caps are 4.7uF and the input impedance of the power amp of the Marantz PM8005 is 15k
15k ohm
I used an online calculator to check the 4.7 caps and the low frequency roll off. I picked 4.7 because it appeared like it would give me low frequency response down to ~20hz
Last edited:
Great advice about muting circuit, RF traps, etc. These features are mandatory for a properly functioning preamp circuit.
I assembled a preamp board that had an intermittent defect. It would play fine and all of a sudden the power amp safety circuit would trip. When the power amp came back online there was only hiss. If I powered down the preamp and powered back up it worked fine again, until next time.
One op amp pin was not soldered to the board. This caused the output bias voltage to swing to the rail. It had all the safety features of course, but the spike caused by the failure mode was nasty enough to trigger the amplifier safety circuit.
The moral of this story is to be really careful what source you connect to your power amp. Even a dodgy RCA cable can mean a blown speaker.
I assembled a preamp board that had an intermittent defect. It would play fine and all of a sudden the power amp safety circuit would trip. When the power amp came back online there was only hiss. If I powered down the preamp and powered back up it worked fine again, until next time.
One op amp pin was not soldered to the board. This caused the output bias voltage to swing to the rail. It had all the safety features of course, but the spike caused by the failure mode was nasty enough to trigger the amplifier safety circuit.
The moral of this story is to be really careful what source you connect to your power amp. Even a dodgy RCA cable can mean a blown speaker.
4.7uF into 15k gives 2.3Hz. Meanwhile 1uF into 300k in the PSU means that only frequencies above 0.5Hz will be attenuated. Both will be first-order rolloffs i.e. 6dB/octave. Suppose the mains has a voltage bounce with significant energy in the 1Hz region. This will be attenuated by around -6dB in the PSU, and -7dB in the preamp, giving -13dB in total. A 5V jump in mains could send 1.1V into the power amp. At the very least this could make your woofers flap a little. These are rough calculations, but they suggest there is a weakness in this design and your component change has made it worse.
So changing the 1uF cap in the PSU to 3.2uF will attenuate frequencies above ~.17Hz?
I am new to this, how did you calculate/estimate the attenuated dB drop?
Are you referring to the change of the output caps from 2.2uF to 4.7uF?
A first order low pass filter gives a response which goes as 1/sqrt[1+(f/f0)^2], where f is the signal frequency and f0 is the corner frequency. A first order high pass filter is 1/sqrt[1+(f0/f)^2]. At the corner frequency the response will be -3dB. One octave away (into the stopband) it will be -7dB, but -6dB is a close enough approximation.
You need to widen the gap between the two filters. Your change from 2.2uF to 4.7uF narrowed the gap.
You need to widen the gap between the two filters. Your change from 2.2uF to 4.7uF narrowed the gap.
With shorted inputs, and the COM probe of my voltmeter attached the ground through a 390pf capacitor, and the red probe attached to the output RCA jack, I get an AC voltage reading of about .16v.
This is using an autoranging meter. I tried my Craftsman analog meter, on 10Vac range, and the needle doesn't move.
This is using an autoranging meter. I tried my Craftsman analog meter, on 10Vac range, and the needle doesn't move.
I installed a bunch of upgrades today.
Including the following:
1. RF caps across the inputs
2. TVS Diodes across the outputs
3. 100ohm screen stoppers
4. Upgraded grid stoppers on 6V6 to carbon comp
5. Increased 1uF cap in power supply to 2.2uF
Currently, I am playing louder than usual on an older Pioneer amp. Seems to be performing well. Sounds good. If I hook the autoranging DMM to the output, the vDC still jumps around like in the linked video in post 1, but the range of jumping is diminished. Although, I am not really sure what is measured in that test.
I tried looking for oscillation with my analog meter per the post above, and I couldn't get any reading other than zero. This may be due to poor quality meter or operator error. Unsure. :/
Including the following:
1. RF caps across the inputs
2. TVS Diodes across the outputs
3. 100ohm screen stoppers
4. Upgraded grid stoppers on 6V6 to carbon comp
5. Increased 1uF cap in power supply to 2.2uF
Currently, I am playing louder than usual on an older Pioneer amp. Seems to be performing well. Sounds good. If I hook the autoranging DMM to the output, the vDC still jumps around like in the linked video in post 1, but the range of jumping is diminished. Although, I am not really sure what is measured in that test.
I tried looking for oscillation with my analog meter per the post above, and I couldn't get any reading other than zero. This may be due to poor quality meter or operator error. Unsure. :/
Don't know the ohms per volt of the craftsman analog meter. This should show in very fine print on the meter scale. I'm getting good 1 mhz oscillation check on my 5000 ohm/voltAC simpson 266XLPM. (Like I found some oscillation) Which has 2 vac & 20 vac scales.
Cheap DVM produce random numbers on music, my test on the back of a hammond organ demonstrated. The $200 Fluke RMS voltmeter will produce accurate numbers on music up to 7000 hz, but is totally dead on RF waveforms. Waste of money, IMHO when the 266 is about $280.
The $100 scopes tend to be limited to 80 v input, which is incompatible with tube circuits. I've a couple of dead $500-$1000 scopes with hundreds of expired ecaps in them; I view used scopes as a huge waste of time. "PC scope" using the sound card subjects the ground of your PC to whatever faults are in the device under test. A good way to blow up your PC, in my opinion. I've had solid state amps produce 160 vdc with capability of a >1000 A arc out of the 10000 uf capacitors. The simpson 266 has withstood all this crazyness, is 33 years old, has no ecaps in it to expire. If I step on a probe and break it it is not $60, it is $12.
Buy what you want.
No suggestions on your circuit faults, I'm using tried & true dynaco ST70 tube amp and I still have blown 2 tweeters with it in 49 years. (bad luck? oscillation?) There were half a million of these built with pictures of where every wire should go, doubt if there were design problems. I have huge problems with solder joints I have made coming loose in 1 year to 5 years. BANG! Bought a hotter iron this year.
Shocked that bidirectional TVS diode on output is a valid technique! I just thought it up based on the product selection chart in the newark website. Bought some myself for the ST70. The TVS diode has a parallel capacitance of 2 nf, my parts express capacitance meter says.
Peavey protects the inputs of their PA amps from tube guitar amps with a 1000 ohm 1/10 watt resistor from input jack to op amp input, with diodes from the op amp input to both power supply rails. So if the input goes outside the op amp supply rail, the input shorts to it until the resistor blows up. One of their used amps I've bought have these resistors toasted & gone. Great protection technique for the solid state amp input. Doubt if your "hifi" marantz amp has this.
Fixing blown up proff. PA gear has been a great cheap education in design prevention of common problems.
Cheap DVM produce random numbers on music, my test on the back of a hammond organ demonstrated. The $200 Fluke RMS voltmeter will produce accurate numbers on music up to 7000 hz, but is totally dead on RF waveforms. Waste of money, IMHO when the 266 is about $280.
The $100 scopes tend to be limited to 80 v input, which is incompatible with tube circuits. I've a couple of dead $500-$1000 scopes with hundreds of expired ecaps in them; I view used scopes as a huge waste of time. "PC scope" using the sound card subjects the ground of your PC to whatever faults are in the device under test. A good way to blow up your PC, in my opinion. I've had solid state amps produce 160 vdc with capability of a >1000 A arc out of the 10000 uf capacitors. The simpson 266 has withstood all this crazyness, is 33 years old, has no ecaps in it to expire. If I step on a probe and break it it is not $60, it is $12.
Buy what you want.
No suggestions on your circuit faults, I'm using tried & true dynaco ST70 tube amp and I still have blown 2 tweeters with it in 49 years. (bad luck? oscillation?) There were half a million of these built with pictures of where every wire should go, doubt if there were design problems. I have huge problems with solder joints I have made coming loose in 1 year to 5 years. BANG! Bought a hotter iron this year.
Shocked that bidirectional TVS diode on output is a valid technique! I just thought it up based on the product selection chart in the newark website. Bought some myself for the ST70. The TVS diode has a parallel capacitance of 2 nf, my parts express capacitance meter says.
Peavey protects the inputs of their PA amps from tube guitar amps with a 1000 ohm 1/10 watt resistor from input jack to op amp input, with diodes from the op amp input to both power supply rails. So if the input goes outside the op amp supply rail, the input shorts to it until the resistor blows up. One of their used amps I've bought have these resistors toasted & gone. Great protection technique for the solid state amp input. Doubt if your "hifi" marantz amp has this.
Fixing blown up proff. PA gear has been a great cheap education in design prevention of common problems.
Last edited:
The plate resistance of the output tube is 5 kohms. If the Craftsman meter was 1000ohms/volt on the 10 v scale, that is a load of 10 kohm. So if the oscillation into 100k amp input was 1 vac, and you measured it with this, you could see 0.67 volts average noise.
However oscillation of 20 mv may drive your amp crazy. An input stage could increase your impedance at the cost of a 12 v wall transformer ($1 at charity resale shop) a $.33 transistor a $.25 e-capacitor on supply, three $.06 resistors and a coupla 2.2 u coupling caps. Plus a $1.50 project board.
Build the first stage of a TGM8, or the first stage of an AX6 or AX8. TGM8 thread is on the first page of solid state. If you only have npn small transistors, build the jerluwoo simpleAB amp or something like that. Then measure with some gain on the front end of your $20 Sears analog meter.
If you have 4558 or better op amps, you could make a gain 10 or 20 circuit with the same # of parts and a dip project board. Fake the analog ground with a coupla similar 20-47 k resistors 12v to zero, couple in & out with 2.2 uf disk caps. Gain is Rf/Ri. I like the $2 DIP project boards from newark, & the TE connectivity phosphor bronze dip sockets. Board spreads the soldering out to where you can see it for individual 28 or 30 ga wires.
If you don't want to wait on a project board, have a drill motor & a vise, 1/16" or #44 or .5 mm drill bit, build on masonite, plywood or plastic. I don't like lexan, the solder tends to stick to it and bridge the DIP pins. PVC would probably do the same. I like #28 or 30 solid core wire, which is a $10 ebay item or comes from one of the surplus houses like electronicssurplus.com in NY or apexelectronics.com in LA. I got a 100' roll of the stuff from Olsen radio when there were part stores in strip malls. A miller stripper, reading glasses, there you are. instant circuits, no nasty chemical waste to dump illegally.
However oscillation of 20 mv may drive your amp crazy. An input stage could increase your impedance at the cost of a 12 v wall transformer ($1 at charity resale shop) a $.33 transistor a $.25 e-capacitor on supply, three $.06 resistors and a coupla 2.2 u coupling caps. Plus a $1.50 project board.
Build the first stage of a TGM8, or the first stage of an AX6 or AX8. TGM8 thread is on the first page of solid state. If you only have npn small transistors, build the jerluwoo simpleAB amp or something like that. Then measure with some gain on the front end of your $20 Sears analog meter.
If you have 4558 or better op amps, you could make a gain 10 or 20 circuit with the same # of parts and a dip project board. Fake the analog ground with a coupla similar 20-47 k resistors 12v to zero, couple in & out with 2.2 uf disk caps. Gain is Rf/Ri. I like the $2 DIP project boards from newark, & the TE connectivity phosphor bronze dip sockets. Board spreads the soldering out to where you can see it for individual 28 or 30 ga wires.
If you don't want to wait on a project board, have a drill motor & a vise, 1/16" or #44 or .5 mm drill bit, build on masonite, plywood or plastic. I don't like lexan, the solder tends to stick to it and bridge the DIP pins. PVC would probably do the same. I like #28 or 30 solid core wire, which is a $10 ebay item or comes from one of the surplus houses like electronicssurplus.com in NY or apexelectronics.com in LA. I got a 100' roll of the stuff from Olsen radio when there were part stores in strip malls. A miller stripper, reading glasses, there you are. instant circuits, no nasty chemical waste to dump illegally.
Last edited:
Just wanted to add a conclusion to this thread. After the upgrades to the preamp listed in post #51, I haven't had any issues with the preamp like described in original post.
I am still considering adding a muting circuit to the preamp, but just want to listen to it for a bit and enjoy the music. I have some different output caps coming, Clarity Caps, to replace the cheaper Panasonic caps I am using. Not sure if I will notice a difference. I usually give the preamp minutes to warm before turning on the pre, as it takes me a few to find a record and get it spinning.
Thanks again to everyone who lent their expertise and advise.
I am still considering adding a muting circuit to the preamp, but just want to listen to it for a bit and enjoy the music. I have some different output caps coming, Clarity Caps, to replace the cheaper Panasonic caps I am using. Not sure if I will notice a difference. I usually give the preamp minutes to warm before turning on the pre, as it takes me a few to find a record and get it spinning.
Thanks again to everyone who lent their expertise and advise.
I still haven't gotten around to adding a muting circuit. It is still on my list of projects, though.
I still use this preamp all the time. Currently, it feeds a solid state power amp, or a Bob Carver 275 tube power amp.
If I built this again, I might choose a different tube sockets and change up the enclosure a bit. But the circuit is simple and sounds great. As I have it arranged now, it would be hard to neatly add the muting circuit. (Presently, I just make sure all power amps are off before shutting down the pre.)
Hope that helps.
I still use this preamp all the time. Currently, it feeds a solid state power amp, or a Bob Carver 275 tube power amp.
If I built this again, I might choose a different tube sockets and change up the enclosure a bit. But the circuit is simple and sounds great. As I have it arranged now, it would be hard to neatly add the muting circuit. (Presently, I just make sure all power amps are off before shutting down the pre.)
Hope that helps.