Spent an hour ransacking my book case, unfortunately I can't find the original source of my assertion, and while I think it true, please don't quote me on it as for now at least it remains unsubstantiable. 
Kevin & Tweeker,
tnx2u4 your efforts!
I am sure to have read something along this somewhere "official" but didn´t remember exactly where either 🙁
(I mean, everybody knows it, nobody doubts it)
Tom
tnx2u4 your efforts!
I am sure to have read something along this somewhere "official" but didn´t remember exactly where either 🙁
(I mean, everybody knows it, nobody doubts it)
Tom
Yup,
Arriving a little late here ....
Firstly, a bi-polar electrolytic is substantially better than an "ordinary" one (explanation will be found in analysis dealing with electrolytics). In an excellent series of articles in Electronics World about 5 years ago, Cyril Bateman published graphs of relative performance.
Secondly, get away from low voltage electrolytics; they can die. In these days of small components there is no reason to go less than 16V.
Thirdly, nobody mentioned tantalum. I have found them very good (as did Bateman), though capacitance tolerance can be a problem.
But I am going to be somewhat of a rebel. This Marantz is one of very few amplifiers using an unloaded output transformer feedback winding. Resultant smaller effect from leakage reactance allows substantially more NFB without the normal h.f. problems. I did not calculate (the winding ratio is not known), but the trimmer adjustment for h.f. gives the impression that stability can be compromised by component tolerance. Another plus is that Marantz recognises that leakage reactance can vary by about 20% in a batch of the same output transformers - important for designers to keep in mind.
Now to my unorthodox recommendation. Yes, that capacitor is in the signal line, but with a time constant of about 8,5Hz, to what degree? It is almost a short at higher frequencies. If one wants to be consistent, then all power supply (electrolytic) bypass capacitors are equally in the signal line. (Respectfully, I appreciate that there are some strong opinions about this, which debate I would not like to open now.) My suggestion would be, increase it to about 150uF (electrolytic), and just check with a scope (can you?) for possible low frequency instability in the region of a few Hz. (Attach your scope probe to the feedback windings, not the voice coil ones, but with and without a l.s. load.) Looking at the rest of the circuit I would say instability is unlikely - there are time constants quite higher up. But just to be careful about the possibility of instability should the NFB be high as said above ....... It can do no harm if you are careful.
Regards.
Arriving a little late here ....
Firstly, a bi-polar electrolytic is substantially better than an "ordinary" one (explanation will be found in analysis dealing with electrolytics). In an excellent series of articles in Electronics World about 5 years ago, Cyril Bateman published graphs of relative performance.
Secondly, get away from low voltage electrolytics; they can die. In these days of small components there is no reason to go less than 16V.
Thirdly, nobody mentioned tantalum. I have found them very good (as did Bateman), though capacitance tolerance can be a problem.
But I am going to be somewhat of a rebel. This Marantz is one of very few amplifiers using an unloaded output transformer feedback winding. Resultant smaller effect from leakage reactance allows substantially more NFB without the normal h.f. problems. I did not calculate (the winding ratio is not known), but the trimmer adjustment for h.f. gives the impression that stability can be compromised by component tolerance. Another plus is that Marantz recognises that leakage reactance can vary by about 20% in a batch of the same output transformers - important for designers to keep in mind.
Now to my unorthodox recommendation. Yes, that capacitor is in the signal line, but with a time constant of about 8,5Hz, to what degree? It is almost a short at higher frequencies. If one wants to be consistent, then all power supply (electrolytic) bypass capacitors are equally in the signal line. (Respectfully, I appreciate that there are some strong opinions about this, which debate I would not like to open now.) My suggestion would be, increase it to about 150uF (electrolytic), and just check with a scope (can you?) for possible low frequency instability in the region of a few Hz. (Attach your scope probe to the feedback windings, not the voice coil ones, but with and without a l.s. load.) Looking at the rest of the circuit I would say instability is unlikely - there are time constants quite higher up. But just to be careful about the possibility of instability should the NFB be high as said above ....... It can do no harm if you are careful.
Regards.
Johan,
Thanks for your description of how this feedback loop works. As far as which cap to use, I already have a 15uf film cap (both checked at 15.1uf on cap. meter) in place and expect my new tubes to arrive this afternoon. I am assuming this should be fine?
I did not understand your abbreviation "l.s. load". Your explanation did make me wonder the proper proceedure for making final adjustments to the amplifier once I have the new tubes in place.
As far as adjusting the amp, I have two main adjustments available (after tube bias has been set). Those are A.C. balance for each channel and the two trim caps in the circuit (C12A & C12B) - Is the proper way to adjust the amp to first set the A.C. balance so that I get the 9.8 VAC on each half of the circuit (with 1000hz input signal as described in the notes on the schematic), then use the trim caps to achieve the best square wave performance (what frequency should I use for this?) and then check for low frequency instability on the feedback tap? As you can tell from my questions I do not have much experience with tube amplifiers... I own a scope and DMM so any pointers would be appreciated.
On a side note, the whole point for this maintanance was due to a low level popping sound that stayed with the amp channel (not the tubes). I have since tracked that down to two white wires that went from the EL34 plates to the 1.5pf caps. Turned out that the copper wire inside the white insulation was completely corroded. Luckily no other wires seem to be affected. Must have reacted with something in the white insulation - if you own one of these amps, I recommend you check these wires...
Thanks,
Jeff
Thanks for your description of how this feedback loop works. As far as which cap to use, I already have a 15uf film cap (both checked at 15.1uf on cap. meter) in place and expect my new tubes to arrive this afternoon. I am assuming this should be fine?
I did not understand your abbreviation "l.s. load". Your explanation did make me wonder the proper proceedure for making final adjustments to the amplifier once I have the new tubes in place.
As far as adjusting the amp, I have two main adjustments available (after tube bias has been set). Those are A.C. balance for each channel and the two trim caps in the circuit (C12A & C12B) - Is the proper way to adjust the amp to first set the A.C. balance so that I get the 9.8 VAC on each half of the circuit (with 1000hz input signal as described in the notes on the schematic), then use the trim caps to achieve the best square wave performance (what frequency should I use for this?) and then check for low frequency instability on the feedback tap? As you can tell from my questions I do not have much experience with tube amplifiers... I own a scope and DMM so any pointers would be appreciated.
On a side note, the whole point for this maintanance was due to a low level popping sound that stayed with the amp channel (not the tubes). I have since tracked that down to two white wires that went from the EL34 plates to the 1.5pf caps. Turned out that the copper wire inside the white insulation was completely corroded. Luckily no other wires seem to be affected. Must have reacted with something in the white insulation - if you own one of these amps, I recommend you check these wires...
Thanks,
Jeff
Yah, that was a rather unclear description on my side. SY is right. What I was trying to cover was (a) any low frequency tendency toward instability would probably be very low, in which case measurements directly on the output might reflect less than measured across the feedback winding(s) themselves. On second thought this might not be serious. (b), test with the output loaded and unloaded, preferably with an 8 ohm resistor or whatever impedance you are using. Again this might be a severe test, because many amplifiers do not purport to be stable with open output - a little difficult as I do not know this amplifier. Also, furthermore (though you might guess this) use the scope dc input.
About the sequence of things, I would say first both (if that is not a contradiction!) watch for instability low and high - but you did not alter any low frequency time constants - to make sure things are not wildly out of kilter. If there is a problem set the square wave approximately right. Then you can set the 1 KHz balance and recheck the trimmers for optimal square wave form.
But I took another look when you mentioned the 1,5 pF caps. That somewhat disturbs me: Compensation that depends on such a low value .....? I hope the caps are right on the grids and the wires are on the EL34 anodes' sides, otherwise stray capacitances will act. Was this popping sound of a low frequency - which would indicate some low frequency instability to begin with. You will perhaps know that not all stability is evidenced by something of simply sinusoidal nature. A sharp spike can occur at low repetition rate because of temporary overload of something - could this have been your problem? You would have been able to see this on a scope. Anyway, if these wires were off (open) from corrosion then that compensation would not work, probably causing a repetitive h.f. spike of sorts.
Not to criticise the Marantz - I like much of the design, but when stability calls for such low value components being wired in ..... oh well!
Keep us informed of what the square waves look like.
Regards.
About the sequence of things, I would say first both (if that is not a contradiction!) watch for instability low and high - but you did not alter any low frequency time constants - to make sure things are not wildly out of kilter. If there is a problem set the square wave approximately right. Then you can set the 1 KHz balance and recheck the trimmers for optimal square wave form.
But I took another look when you mentioned the 1,5 pF caps. That somewhat disturbs me: Compensation that depends on such a low value .....? I hope the caps are right on the grids and the wires are on the EL34 anodes' sides, otherwise stray capacitances will act. Was this popping sound of a low frequency - which would indicate some low frequency instability to begin with. You will perhaps know that not all stability is evidenced by something of simply sinusoidal nature. A sharp spike can occur at low repetition rate because of temporary overload of something - could this have been your problem? You would have been able to see this on a scope. Anyway, if these wires were off (open) from corrosion then that compensation would not work, probably causing a repetitive h.f. spike of sorts.
Not to criticise the Marantz - I like much of the design, but when stability calls for such low value components being wired in ..... oh well!
Keep us informed of what the square waves look like.
Regards.
I don't think those caps have a significant effect on stability. IIRC they are there to allow tweaking of the sq wave response at 10kHz or so.
The settings are specific to the individual output transformer and compensate for variations in leakage inductance or stray capacitance in the windings. (or both, can't remember exactly what Marantz said about this) Factory service notes do not indicate a need to adjust these caps with routine tube replacement. I would just leave them alone. In the dozen or so of these amps I have repaired over the years I have never seen one that need them adjusted.
The settings are specific to the individual output transformer and compensate for variations in leakage inductance or stray capacitance in the windings. (or both, can't remember exactly what Marantz said about this) Factory service notes do not indicate a need to adjust these caps with routine tube replacement. I would just leave them alone. In the dozen or so of these amps I have repaired over the years I have never seen one that need them adjusted.
Kevin, no fine.
If you had that experience then so be it. (I take it you are talking about the 1,5pFs?). As said I have no practical experience of these amps.
My comments were in general. If caps, especially low values, are needed to "correct" a square wave, (i.e. influence h.f), I presume they were necessary, i.e. made a noticable difference. In such a case overcompensation could result in instability. But Yup would then probably find with his scope that the Marantz square wave diaplay would look right before (or without) any adjustment.
Thansk for your experience.
If you had that experience then so be it. (I take it you are talking about the 1,5pFs?). As said I have no practical experience of these amps.
My comments were in general. If caps, especially low values, are needed to "correct" a square wave, (i.e. influence h.f), I presume they were necessary, i.e. made a noticable difference. In such a case overcompensation could result in instability. But Yup would then probably find with his scope that the Marantz square wave diaplay would look right before (or without) any adjustment.
Thansk for your experience.
Hi Johan,
Actually I was talking about the variable caps on the output feedback winding. 😀
The 1.5pF caps are extremely small and seem to effect the phase margin well above 20kHz. I am not sure whether or not the amplifier would always oscillate without them, but I seem to recall some did..
Note that there are a number of caps and resistors in the feedback loop that were factory selected to optimize sq wave response and presumably stability margin individually for each channel.. Look at the * parts.. The specmanship game had already started by the time this amplifier was designed. The 8B had very good specs by the standards of the day and was unaffordable to most enthusiasts.
The Marantz 9 is my favorite commercial monoblock of the time.. These days they are ridiculously expensive. I have a friend who picked up a pair for $100 in the late 1970's. He's been offered as much as $10K for them.
Actually I was talking about the variable caps on the output feedback winding. 😀
The 1.5pF caps are extremely small and seem to effect the phase margin well above 20kHz. I am not sure whether or not the amplifier would always oscillate without them, but I seem to recall some did..
Note that there are a number of caps and resistors in the feedback loop that were factory selected to optimize sq wave response and presumably stability margin individually for each channel.. Look at the * parts.. The specmanship game had already started by the time this amplifier was designed. The 8B had very good specs by the standards of the day and was unaffordable to most enthusiasts.
The Marantz 9 is my favorite commercial monoblock of the time.. These days they are ridiculously expensive. I have a friend who picked up a pair for $100 in the late 1970's. He's been offered as much as $10K for them.
Kevinkr & Johan,
Thank to both of you for your comments. I have put in the new tubes and measured the operating voltages... I guess I will start a new thread since I am changing subjects... please give me your comments there.
Thanks,
Jeff
Thank to both of you for your comments. I have put in the new tubes and measured the operating voltages... I guess I will start a new thread since I am changing subjects... please give me your comments there.
Thanks,
Jeff
Kevin,
Your post #29 noted. I did rather like the design, but $10K!! If not too off-thread and you are able, can you list the specs, please?
Yup,
I will look out for the thread.
Regards.
Your post #29 noted. I did rather like the design, but $10K!! If not too off-thread and you are able, can you list the specs, please?
Yup,
I will look out for the thread.
Regards.
- Status
- Not open for further replies.