I think that it could be a good candidate for modding. As Chris said it did not perform as well as the competition at the time, but in my view that just gives a nice starting point from which every mod improvement (or not) can be measured and heard.
Start with modding one channel and use the other as reference as far as possible.
Having said that, one could suspect that there are some e-lytes in the power supply that probably are now ready for retirement, and replacing these will of course help both channels.
But after that there are good chances to see what each mod to the LT input pairs does, etc.
Whether this Pioneer will be a winner is a different question and, as you say, probably best for another thread.
Also, the discussions on NFB probably belongs to a different thread under that specific name, so people interested in it can find it.
Cheers,
Per
Per,
Yeah, the collector current is set high from the factory, about 13mA on the RB-970BX-ii. It didn't seem worth mentioning before, but I also reduced VAS current from 13mA to 6mA by increasing R623 and R625 from 100 ohm to 220 ohm.
Spice shows no reduction in stability for this change. The triple EF doesn't need 13mA to drive it, even 6mA has headroom with the dominant current draws being the compensation caps C611, C613, C615.
My intent was to reduce the power dissipated in the un-heatsinked VAS devices from 600mW to <300mW so they can cook a little slower, and to minimize the voltage drop across the rail filters.
In hindsight, maybe this change is audible. To produce a given change in VAS collector current, the VAS base must swing through a larger voltage now, so the input stage is making a larger current swing. If nothing else, that shows up as modestly worse distortion performance at HF. Whoops. Maybe I'll put the VAS back to 13mA next time. More research is needed
Cheers
John
My intent was to reduce the power dissipated in the un-heatsinked VAS devices from 600mW to <300mW so they can cook a little slower, and to minimize the voltage drop across the rail filters.
In hindsight, maybe this change is audible. To produce a given change in VAS collector current, the VAS base must swing through a larger voltage now, so the input stage is making a larger current swing. If nothing else, that shows up as modestly worse distortion performance at HF. Whoops. Maybe I'll put the VAS back to 13mA next time. More research is needed
Cheers
John
Hi John,
Lowering VAS transistor temperature is definitely a good thing (says someone that has burnt his fingers on several occasions).
And, all things equal, reducing the VAS collector current should reduce distortion due to less Early effect. I would think that if (when) you decide to implement some sort of EF VAS circuit you will find that this improves matters significantly, in particular at higher frequencies and your 6mA will do fine.
Cheers,
Per
Doesn't seem to be a very big effect though. I picked a 2N3391A in LTspice (VAF = a low 35-ish V), and the difference in distortion between 5 and 20 mA at same steady state voltage is pretty marginal. Nonlinear loading by the output stage should generally be more relevant at this point.
Hi nania,
The electrons drift very slowly, but the energy travels at close to the speed of light for our purposes. The medium where the energy flows does have some effect as you've noted, but not to the extent where feedback at audio frequencies would have any hint of a problem. Now, the leakage in all chips running at minimum geometries in silicon or other is an entirely different thing. I don't need or want to get into that discussion, but the energy lost is substantial. These effects have zero to do with the speed of an electric charge (not the electron, the charge) in wire. You are always welcome to your opinion without being attacked for it. All I can do is offer my own opinion in response, perhaps with some reasoning behind it - like you are doing. No problem.
The Pioneer 9500 II. It's a good amplifier, but not a great amplifier *compared to some other products in that power class*. I also understand that if it is the best thing you have seen yet, it should be the best amplifier you can think of. Your ideas on what equipment is better is probably going to change as you experience other amplifiers that are better.
Pioneer has a characteristic sound trademark. There is something funny going on in the mids it seems. Ones I have worked on, I've been able to get rid of "that sound" and customers are always extremely happy with the change. I would have to look at mine again (some day) to determine where the regulator section of the power supply is located, but with most of their receivers, the regulators are mounted upside down under the chassis. So, where does the heat go? Right, it pools on the board and cooks it. The usually need complete regulator PCB reconstruction to make them reliable again. Only for a few years as the board continues to cook. I have had to write off a few receivers because the regulator PCB was cooked so badly that it couldn't be fixed.
So, for restoration you want to quickly survey a couple problem areas and look for places that have seen a lot of heat. Amplifier boards that are upside down will suffer the same fate, and worse and the temperature changes are far greater than for a normally designed amplifier. If you see these two things under the chassis, it's time to put some serious thought to what you are planning to do.
So, can they be done? Sure they can. Is it a good plan to do one? Yes, as long as the unit is in good shape electrically, it doesn't show signs of unusual overheating, it doesn't use special parts (unobtainium) and / or you have a bond with that particular product. It goes without saying that the unit should be in good cosmetic shape.
As a technician, your job would be to protect your customer from themselves in case they are doing something that isn't wise. Certainly from spending money on a device that won't last. After telling your customer why it should or shouldn't be restored, it's up to them. Pioneer represents a lot of work - which is fine. Sometimes it represents a project that will never reach your goal. So inspect it for those big problems, and if they don't exist, do a good job on it.
Best, Chris
The electrons drift very slowly, but the energy travels at close to the speed of light for our purposes. The medium where the energy flows does have some effect as you've noted, but not to the extent where feedback at audio frequencies would have any hint of a problem. Now, the leakage in all chips running at minimum geometries in silicon or other is an entirely different thing. I don't need or want to get into that discussion, but the energy lost is substantial. These effects have zero to do with the speed of an electric charge (not the electron, the charge) in wire. You are always welcome to your opinion without being attacked for it. All I can do is offer my own opinion in response, perhaps with some reasoning behind it - like you are doing. No problem.
The Pioneer 9500 II. It's a good amplifier, but not a great amplifier *compared to some other products in that power class*. I also understand that if it is the best thing you have seen yet, it should be the best amplifier you can think of. Your ideas on what equipment is better is probably going to change as you experience other amplifiers that are better.
Pioneer has a characteristic sound trademark. There is something funny going on in the mids it seems. Ones I have worked on, I've been able to get rid of "that sound" and customers are always extremely happy with the change. I would have to look at mine again (some day) to determine where the regulator section of the power supply is located, but with most of their receivers, the regulators are mounted upside down under the chassis. So, where does the heat go? Right, it pools on the board and cooks it. The usually need complete regulator PCB reconstruction to make them reliable again. Only for a few years as the board continues to cook. I have had to write off a few receivers because the regulator PCB was cooked so badly that it couldn't be fixed.
So, for restoration you want to quickly survey a couple problem areas and look for places that have seen a lot of heat. Amplifier boards that are upside down will suffer the same fate, and worse and the temperature changes are far greater than for a normally designed amplifier. If you see these two things under the chassis, it's time to put some serious thought to what you are planning to do.
So, can they be done? Sure they can. Is it a good plan to do one? Yes, as long as the unit is in good shape electrically, it doesn't show signs of unusual overheating, it doesn't use special parts (unobtainium) and / or you have a bond with that particular product. It goes without saying that the unit should be in good cosmetic shape.
As a technician, your job would be to protect your customer from themselves in case they are doing something that isn't wise. Certainly from spending money on a device that won't last. After telling your customer why it should or shouldn't be restored, it's up to them. Pioneer represents a lot of work - which is fine. Sometimes it represents a project that will never reach your goal. So inspect it for those big problems, and if they don't exist, do a good job on it.
Best, Chris
Hi Svitjod,
I rebuilt a hybrid amplifier with many problems. For that unit, 9 dB of feedback was about right. Higher levels of feedback (>20 dB) were not possible due to the phase margins you are talking about. Not the fault of feedback, but rather a fault in the design.
Design decisions are up to your personal comforts and experience. So you use it as you see fit. It comes down to your experience and comfort level.
Best, Chris
Mostly because you can run the stage at higher bias current that also reduces cross-over distortion. Why use feedback in that case? Why not, if the rest of the circuit is up to it, using classic feedback will reduce output impedance, widen bandwidth and reduce non-linearities. These are all very good things. So you're trading gain for bandwidth and distortion. Since most circuits have very high gain, you have to lose it somewhere. Making that trade to feedback becomes a very intelligent decision.
Those problems existed before you closed the feedback loop. However, as you have noted, moderate amounts of feedback can actually make things worse, but those things were already there, they weren't created out of thin air.
I rebuilt a hybrid amplifier with many problems. For that unit, 9 dB of feedback was about right. Higher levels of feedback (>20 dB) were not possible due to the phase margins you are talking about. Not the fault of feedback, but rather a fault in the design.
Design decisions are up to your personal comforts and experience. So you use it as you see fit. It comes down to your experience and comfort level.
Best, Chris
Hi sgrossklass,
You are right, but at least there is no need for jpc to be concerned about this.
On this topic, I redesigned the VAS proto PCB circuit for my mods as I wanted a direct plug-in replacement for the VAS transistor without the need for connecting extra flying ground airwires. Please cf. the enclosed.
The pnp emitter follower isolates well from output stage loading, but I found the best results when (most of) the Miller compensation included the follower. Don't mind the transistor choices, I just lazily picked those from the library to get the right package.
Any comments - or indeed any suggestions for best SOT23 transistor selections?
Per
Can anyone tell me what nania means by circuit "momentum" and why he differentiates voltage and current feedback? I've never heard this before but it sounds interesting. Perhaps he can offer an explanation why current feedback amps tend to perform closer to their sims (sometimes better) than voltage feedback designs.
anatech
As I understand it, the electrons themselves are not the signal wave propagated by them when they are excited but by the same token, the whole wave is not moving at a constant, uninterrupted speed. Even things as minute as molecular imperfections in the conductive material are known to deflect or fracture these energy waves to an alternative path so it isn't unreasonable to deduce some time anomalies being caused by the semi-conductive parts it goes through. It makes sense to me and I think a blanket statement like "the energy travels at close to the speed of light for our purposes" is a bit simplistic. Is it close enough for horseshoes? Maybe but definitely not something to accept without consideration of nania's counter argument. By the way, I'm a big fan of your posts. There isn't anyone on this forum who can shoot the elephants down faster or better than you and AndrewT do.
AngelP
Duuude! Great job on the mods man. You must have been "bored as excrement" (sorry for the awkward auto censure evasion) to start fiddling with that amp. It ain't even pretty. LOL
anatech
As I understand it, the electrons themselves are not the signal wave propagated by them when they are excited but by the same token, the whole wave is not moving at a constant, uninterrupted speed. Even things as minute as molecular imperfections in the conductive material are known to deflect or fracture these energy waves to an alternative path so it isn't unreasonable to deduce some time anomalies being caused by the semi-conductive parts it goes through. It makes sense to me and I think a blanket statement like "the energy travels at close to the speed of light for our purposes" is a bit simplistic. Is it close enough for horseshoes? Maybe but definitely not something to accept without consideration of nania's counter argument. By the way, I'm a big fan of your posts. There isn't anyone on this forum who can shoot the elephants down faster or better than you and AndrewT do.
AngelP
Duuude! Great job on the mods man. You must have been "bored as excrement" (sorry for the awkward auto censure evasion) to start fiddling with that amp. It ain't even pretty. LOL
Yes I have done some
1. Kenwood KR-9400. You know it's a fun project when the factory service bulletin is written in "Deliverance" typeface
. The amp section is now a Honey Badger, more or less, on the original PCB. Still has its original drivers and outputs as they were never bad. The preamp has major simplifications: I bypassed the tone board and upgraded the one opamp to a modern one, and deleted some terrible tantalum coupling caps.2. Sanyo JCX-2600K. Amp section is a Wolverine, more or less, on the original PCB. All new transistors. Preamp is recapped but otherwise stock. Power supply recapped.
3. Sherwood HP-1000. Full custom home-cooked amp PCBs with Wolverine clones. Minor simplifications to the preamp: the high and low filters were irredeemable, everything else was great with a couple modern opamps.
4. Sherwood S-7300. Full custom home-cooked amp PCBs with wolverine clones. All high-current halfwave-rectified currents kept off the amp PCBs, they're on the other side of the heatsinks. Full custom new preamp PCB with a modern dual opamp. Redid gain distribution so the preamp is +9db, followed by the volume pot to drop it down, then the amp is +26db. That gets the noise floor low enough to use it with DIY fake klipschorns and not hear noise.
5. Sansui 5000A originally with the famous flammable WMD driver boards
. Replaced the amp PCBs with home-cooked custom Wolverine clones. Added a protection board that wasn't there in the original design. (I do NOT recommend adding a large board in a location where none was present from the factory. Took too long to figure out how to mount it. I am not a MechE or industrial designer. 
) I did something crazy and built a direct coupled amp around a transformer with no center tap, which is of course impossible. Yet it works great-- http://www.diyaudio.com/forums/solid-state/284073-modern-take-sansui-f-1040-amp.htmlThe preamp on the Sansui is stock, except that I had to revise the grounding architecture to quiet some hum.
6. Yamaha CA-810. With surprisingly few mods on the stock PCB, the amps on this one can be turned into Wolverine clones. It helps that the factory already gives you current mirrors. Preamp is stock except the bass pot had major physical damage so that's removed and defeated by soldering iron; the treble control still works. If I ever find a matching pot the bass knob goes back in.
The CA-810 has a nice feature from the factory: the gain structure is volume pot -> preamp -> muting switch -> power amp. This one is also quiet enough to use with Khorns with the muting switch set to -20db. Whereas at 0db and no signal playing, I get a bit of AM radio through it...
Predictably, I love them all. They sound great. I'm not set up for distortion measurements so never took any. Maybe building Cordell's distortion magnifier is the next project.
If anyone's interested in modding one of these types let me know. I've got LTspice models, build notes, and debug notes from these builds. Some were a lot more work than others-- the Yamaha was the easiest, it might be the easiest way to build a Wolverine.
That makes sense. I know spice doesn't show Early effect well.
It looks like reducing VAS current on the RB-970BX improves stability margin. Once I remembered to remove the RF filter cap from the amp input section, aha!, a frequency response peak around 1MHz showed up. At lower VAS current the peak is lower and less sharp.
So reducing VAS current appears to trade away some linearity in exchange for better stability.
Hi dellama,
BTW, I didn't dismiss nania's comments out of hand. I tried to figure out what was being said and answered that question best I can without researching the topic.
Thank you for the compliment. Now I have to try and live up to it!
Best, Chris
It comes down to questions of scale. Electric current has some irregularity, like granules flowing in a way. Get down close enough and you can detect the irregularities in flow. However, if you are filling a truck, these small irregularities cannot be detected. When you get current flow down in the atto-amperes, you can detect each charge, but up in the pico-ampere ranges, these effects are averaged out so you have a good approximation of a constant homogeneous flow. The currents we deal with in audio signals are probably going to be as low as pico-amperes in places, fractions of milli-amperes in some circuits and right up to multiple amperes in power amplifiers. We simply don't have the resolution to measure atto-amperes within currents of even milli-amperes. Close enough for horseshoes and hand grenades in my book.
Well, the speed question is also a comparison of frequencies. The inverse of jitter length or offset translates to frequency easily enough. Basically, if you do convert the maximum imagined jitter into frequency, it ends up so high in frequency that it is well beyond our ability to sense it. I used the word sense to be overly optimistic with the range of hearing we have in a perfect scenario. Some folks are trying to prove we can sense high frequency sound higher than the 20 KHz limit that s generally accepted these days. Maybe some confusion arises because we talk about jitter in a CD player. But there we are talking about 44.1 KHz, or 96 or 196 or ... KHz. Off the CD itself, the clock is running at 4.3218 MHz. We are not talking about huge phase shifts here, just minor ones that can confuse an imperfect decoder whether to sample now, or somewhere in the waveform close. We are talking about a tiny amount of time shift at radio frequencies and trying to bring that difference into the range where we can detect it. Consider the system I just described. It's a mechanical system spinning a plastic disc. When you talk about jitter on a circuit board, the frequencies are much higher, in the GHz range where jitter becomes a monster problem. How far away from 4.3 MHz is that? Looking at the differences between frequencies, or wavelengths if you would rather, you simply cannot detect the differences in scale. Look from the PCB trace jitter and our hearing could probably be proved a fixed point in time. Looking at it from the scale of our hearing, it could be proved that the wavelengths are too small to be detected. Not with our instruments today. To look at wavelengths that high up, we examine interference patterns - or the effects of these frequencies. I think Keysight makes an oscilloscope that can show those high frequencies, but the cost is a fair bit north of $250,000.00 from the last email I got from them. Horseshoes and handgrenades again. It's all just a question of scale.
BTW, I didn't dismiss nania's comments out of hand. I tried to figure out what was being said and answered that question best I can without researching the topic.
Thank you for the compliment. Now I have to try and live up to it!
Best, Chris
Hi nania,
Just to put things into perspective from my viewpoint. Can they be improved - yes. Nothing wrong with doing that. I recently did this with a Pioneer SM-83 tube amplifier. Lot's of hurdles, but it finally got there. Other brands would have been easier and achieved higher performance. This job I even recommended against improving it. It finally got there and at great expense after the owner pushed me a bit. I would still recommend that it not be done simply because other units get to these levels of performance a little easier. I'm not saying that they are not worth repairing at all, and they can be made to sound a lot better than they did out of the factory.
I do this kind of work on a continuing basis, but I'll always tell people the truth. Also, most amplifiers with a J-Fet front end have a leg up on high performance. The Kenwood amplifiers should respond well, depending on exact model.
-Chris
Just to put things into perspective from my viewpoint. Can they be improved - yes. Nothing wrong with doing that. I recently did this with a Pioneer SM-83 tube amplifier. Lot's of hurdles, but it finally got there. Other brands would have been easier and achieved higher performance. This job I even recommended against improving it. It finally got there and at great expense after the owner pushed me a bit. I would still recommend that it not be done simply because other units get to these levels of performance a little easier. I'm not saying that they are not worth repairing at all, and they can be made to sound a lot better than they did out of the factory.
I do this kind of work on a continuing basis, but I'll always tell people the truth. Also, most amplifiers with a J-Fet front end have a leg up on high performance. The Kenwood amplifiers should respond well, depending on exact model.
-Chris
Dellama, I perhaps can shed some light over this "speed of current" issue.
The electrons actually travels far below the speed of light, I don' have any figure but it's only a couple of meters per seconds or so.
But when you say that current travels at the speed of light, it's actually the energy. I think that's what anatech meant. Regard the electrons as small balls stacked up in a long row. If you push at the first ball, the last one will immediately be pushed away and will transfer energy instantly.
But i dialy practice, the current ( or voltage) actually sometimes behaves as if it was very slow. Look at a simple LP filter made up of one resistor and one capacitor. If you apply a voltage spike, it will be delayed disregarding the speed of the electrons. Simply because the capacitor stores both electrons and the resulting energy.
AngelP: Jag ska försöka hålla mig till ämnet
The electrons actually travels far below the speed of light, I don' have any figure but it's only a couple of meters per seconds or so.
But when you say that current travels at the speed of light, it's actually the energy. I think that's what anatech meant. Regard the electrons as small balls stacked up in a long row. If you push at the first ball, the last one will immediately be pushed away and will transfer energy instantly.
But i dialy practice, the current ( or voltage) actually sometimes behaves as if it was very slow. Look at a simple LP filter made up of one resistor and one capacitor. If you apply a voltage spike, it will be delayed disregarding the speed of the electrons. Simply because the capacitor stores both electrons and the resulting energy.
AngelP: Jag ska försöka hålla mig till ämnet
You probably didn't mean to sound as condescending as you did but to clear the record, I have auditioned quite a number of quality amps from very esteemed designers, Charles Hansen, John Curl (hard to avoid his), Steve Dunlap, etc. I presented the Pioneer and Kenwood line because of their easy availability and presentable enclosures. These would be plausible candidates for a frugal path to a true high fidelity system if their innards could be improved like AngelP did with his Rotel.anatech said:
Yes, your arguments are logical when you attribute an orderliness to a current (signal) which is not there. Ditto for the simultaneous, multilevel, multi-frequency waves of music we are accustomed to approximating with orderly/antispetic signal tones. I won't go into the whole psycho acoustic discussion here. We are already way off topic and I apologize to AngelP for this temporary diversion but the human has an amazing audio filter that distills sound to something recognizable. This "feature" impairs our ability to differentiate things that sound similarly. If we look at our evolution, this is a logical genetic development. We would need to recognize our loved ones voices coming from a field or a cave. This may be the reason we do so miserably at double blind listening tests. That said, our mind process is very adept at spatial recognition. Again, there is an evolutionary benefit to this feature that should be self explanatory. This is where my group delay observations above seem to provide evidence. When an audio signal is time slurred via the mechanisms described above, we lose our ability to pinpoint the source of the wave and this distortion does not show up with our examination of generated tones but in our failure to perceive a soundstage. It also shows up in the loss of "realism" of the sound. I recall reading the posts of another member here who tried to produce a framework to "prove" this phenomena and I think he was very close to the target. Unfortunately, he was quite unpopular and was eventually banned from this forum.
Svitjod
The consensus on the speed of electrons through a normal conduit seems to be about 35mph last I checked. As dellama correctly points out, the signal wave that rides upon them is magnitudes faster and much more variable in speed. That is exactly why this signal is much more susceptible to time distortions than many are willing to believe.
dellama
I've already abused this thread too much with my off topic content but briefly, feedback creates a kind of pseudo mass to the signal wave by repeating its path. This pseudo mass is what generates the circuit momentum I describe above. The reason anatech is under the impression that a signal travels at the speed of light is because the charge traveling on the electrons is almost massless. Being nearly massless, it is easily deflected, fractured, delayed and distorted. That is the primary reason why feedback has such a stabilizing effect on circuits. Without feedback you are witnessing the natural paths of a signal which are more similar to a lightning bolt than an accurate and orderly signal tone. Current feedback is a purer iteration of the signal than voltage feedback. Briefly, voltage predicts the signal, current is the signal.
I've already abused this thread too much with my off topic content but briefly, feedback creates a kind of pseudo mass to the signal wave by repeating its path. This pseudo mass is what generates the circuit momentum I describe above. The reason anatech is under the impression that a signal travels at the speed of light is because the charge traveling on the electrons is almost massless. Being nearly massless, it is easily deflected, fractured, delayed and distorted. That is the primary reason why feedback has such a stabilizing effect on circuits. Without feedback you are witnessing the natural paths of a signal which are more similar to a lightning bolt than an accurate and orderly signal tone. Current feedback is a purer iteration of the signal than voltage feedback. Briefly, voltage predicts the signal, current is the signal.
Guys please, could someone PLEASE start new threads on NFB and on the theoretical physics of electric current conduction. Interested parties would have no way of finding these fascinating subjects on this humble Rotel improvement thread. And as for me, my Uni courses on these subjects are simply too long ago and half forgotten for me to start commenting.
dellama,
Beauty is in the mind of the beholder. I was at times frustrated but never bored during these mods. No dog is ever too old to learn new tricks.
Now, each time I turn on the sweet sounding 820AX I get the satisfaction to think that "hey, I did this!" That's beautiful.
Per
Hi nania,
You're right about timing distortion of a sound source. I wouldn't classify that as jitter though. I do agree wholeheartedly with what you've just said. For a piece of equipment to throw things that far out of whack, there is something seriously wrong with the equipment.
I have had a certain percentage of people walk through my door wanting this or that modification for improvements. Very often, I find their equipment is defective in some way. So I call them and make a deal. I fix the issues and they reconsider the modification after listening to their repaired equipment for a week or so. They mostly opt to leave things where they are and forget the mods. This applies to popular changes that actually don't do anything positive, or are extremely negative towards the proper operation of that equipment. I will, and do modifications that do not impact reliability and do make a positive change to how the equipment is operating.
The Pioneer. Yes, it can be improved to decent sound quality, but it might cost you more to go that route. That's all.
Your take on signals, voltage and current, is novel. Completely alien to me in fact. I don't quite know what to make of it. Interesting, but it predicts a chaotic reality that would suit the designers you've listed. It does fly in the face of what I know from the test equipment and calibration world that dovetails nicely with accepted science.
Best, Chris
No, I really didn't intend to. Sorry.
You're right about timing distortion of a sound source. I wouldn't classify that as jitter though. I do agree wholeheartedly with what you've just said. For a piece of equipment to throw things that far out of whack, there is something seriously wrong with the equipment.
I have had a certain percentage of people walk through my door wanting this or that modification for improvements. Very often, I find their equipment is defective in some way. So I call them and make a deal. I fix the issues and they reconsider the modification after listening to their repaired equipment for a week or so. They mostly opt to leave things where they are and forget the mods. This applies to popular changes that actually don't do anything positive, or are extremely negative towards the proper operation of that equipment. I will, and do modifications that do not impact reliability and do make a positive change to how the equipment is operating.
The Pioneer. Yes, it can be improved to decent sound quality, but it might cost you more to go that route. That's all.
Your take on signals, voltage and current, is novel. Completely alien to me in fact. I don't quite know what to make of it. Interesting, but it predicts a chaotic reality that would suit the designers you've listed. It does fly in the face of what I know from the test equipment and calibration world that dovetails nicely with accepted science.
Best, Chris
