Sound Quality Vs. Measurements

Status
Not open for further replies.
My speakers have to fill with sound a room of some 110 sq.ft., i.e. small. They are nominally 8 Ohms, minimum 6.5 Ohms, worst case phase shift is -25 degrees, effciciency is 92 dB/2.83V/1m.

My speaker cables are van den Hul Hybrid 352. That's a mix of silver plated OFC cable, mixed with carbon fibre (vdH has been a professor of metallurgy at a Dutch university for 40 years now). Cable has 2 x 256 strands and is capable of passing through well over 200 Amps.

My current range of readily available amps consists of cheap'n'cheerful low level (a bit above absolute entry level) 2x50W to very much High End 2x180W amps, from Toshiba, Marantz, Sansui, Harman/Kardon and Karan Acoustics ( Karan Acoustics | Manufacturer of High-End Audio ).

The point is not to go watt mad, I already have more watts available than I know what to do with. The point is twofold:

1. It has to be mine, I have to design and make it, for the hell of it, and
2. As John put it so aptly, it has to be done right.

At this time, I have a simulator working model which will nominally deliver 28.3 Vrms into 8, 4 and 2 Ohms, with minimal voltage loss, on a continuous basis. It will not clip up to around 175 W/8 Ohms, due to relatively high voltage lines, +/-63 regulated for everything including predrivers and drivers, and +/-56V for the output trannies alone. There should be two versions.

One should use 3 pairs of Motorola/ON Semi MJ 21195/21196 (metal can TO-3 trannies, rated at 250W each, 15 Arms, 30 Apeak, Ft > 4 MHz).

The other should use 4 pairs of Motorola/ON Semi MJL 3281/1302 (platic pack TO3, rated at 200W each, 15 Arms, 30Apeak, Ft > 30 MHz).

The two versions have a double purpose. One is to check out for myself a theory I have, which is that in many cases, older power devices, developed over decades, actually sound just as good and often better than the new(er) ones, even if their spec sheet says different. I have a nagging feeling that older power devices appear to be more stable than the new ones, and I want to check it out.

The other is to produce two solid power amps, which I can then use to biamp my speakers. If it turns out as well as I hope it will, perhaps a third unit later for the wife, who currently has a Harman/Kardon 680 integrated powering her JBL Ti600 floorstanders.

That's it in a nutshell. No binding peconceptions, my starting point is a fully complementary topology, but I do not insist on it.

The only thing I do insist on is fully electronic protection against overvoltage, overcurrent, overheating and excess DC. I do not trust fuses alone, I've seen far too many burnt out amps with "protection" fuses intact. I believe the fastest fuse we have is one made up of transistors.

That's it.

One of my amplifier is a PS audio 200CX it has a 3 pr output mounted on copper
Very good sonics , can even drive the amp killa ...:)

So i have one amp with 24 outputs / ch , one with 20, another with 12 and this PS audio with 6. Not sure if 6 is some kind of magic number , but it works , gold plated fuses and all ....

I still suggest you stay away from sonic killing protection , current limiting et al , KILL SONICS , if you are serious about it , don't do it ....

Regards ....
 
Last edited:
One of my amplifier is a PS audio 200CX it has a 3 pr output mounted on copper
Very good sonics , can even drive the amp killa ...:)

So i have one amp with 24 outputs / ch , one with 20, another with 12 and this PS audio with 6. Not sure if 6 is some kind of magic number , but it works , gold plated fuses and all ....

I still suggest you stay away from sonic killing protection , current limiting et al , KILL SONICS , if you are serious about it , don't do it ....

Regards ....

"3 pairs" is, at best, half information. I have seen just about anything in terms of 90, 100, 110, 130, 150, 180, 200 and 250 Watts rated devices in pairs. I think we'll agree pretty easily that it's not quite the same thing whether they are of the 90 or 250 Watt persuasion.

As for using copper for heat sinking, the best scheme I have ever seen so far is using a 3 mm thick copper plate in full back contact with a hefty aluminium fin heat sink. Copper has the advantage of faster heat transfer, but making entire heat sinks of copper must be ridiculously expensive, and I don't know anyone selling these.

However, substituting wiring with thick copper bars to link the PCB output points with the back side speaker binding posts is quite possible and is a good idea.

Regarding the number of output devices, this is both a matter of power requirements and personal choice. Some people I know swear that the total number must be an odd one (3,5, 7, etc pairs), others swear it must be an even one, and so forth.

Personally, I feel one needs to be practical about it. I do need several pairs because I am expecting significant power capability into 2 Ohms, and that power has to come from somewhere. On the other hand, increasing the number of output pairs complicates my life a lot due to the requierement of as close matching as possible. And I am not exactly Sony or some such to directly approach a manufacturer and order 100,000 matched pairs.

4 pairs of 200W devices is in potential power terms about 2.5 or 3 TIMES what the industry in general uses, but is still managable in terms of matching. Their work just fine, so mine should as well, only (hopefully) better.

As for protection circuits, I guarantee they will stay dormant until the load impedance starts to dip below 2 Ohms. In measurement terms, with or without them, I get exactly the same results even at 2 Ohms. So they stay.
 
As for using copper for heat sinking, the best scheme I have ever seen so far is using a 3 mm thick copper plate in full back contact with a hefty aluminium fin heat sink. Copper has the advantage of faster heat transfer, but making entire heat sinks of copper must be ridiculously expensive, and I don't know anyone selling these.

That is how it's done ... I will add this amp was built in 1990 ...
 
That is how it's done ... I will add this amp was built in 1990 ...

In 1990, just as in 10,000 BC, copper was still faster at transfering heat than aluminium was. :D :D :D

Wayne, there are so many things that can be done, it boggles the mind. The reason why many are seldom done is the usual suspect - price. And the added hassle.

I was thinking about having a copper plate machine bent (so there are no creases, deformities, etc) into an L shape, then bolting it to a hefty heat sink from the back side and then adding another half heigth sink on the front side. 20 years ago I wouldn't have wanted to do that, but today, with the proliferation of affordable CNC machines, it shouldn't be too much of a problem. Precision drilling is no longer a big, bad issue.

That should give me very good heat sink capability, admittedly WAY over what I need in everyday use, but - sort of - just in case I want to run it at high power levels for a longer time.

Given the areas involved, I'm going to need like a pound of thermal compound. :p

At this time, my current upper range of filters uses pure copper for its internal second bottom plate. I don't know if it helps, but I know it sure can't hurt. The point is, I already have some experience with copper; most materials require some hands-on experience to be comfortable with them.
 
diyAudio Member RIP
Joined 2005
One of the difficult aspects of copper, particularly the OFHC variety, is in the machining. One machinist I worked with at UCLA described it as "gummy". Of course it has also gotten very expensive.

Once machined, it's not very rugged, and then you also need to plate it to prevent corrosion. A lot of work. But, thick enough it is a pretty good magnetic shield as well as a thermal conductor.
 
Member
Joined 2002
Paid Member
One of the difficult aspects of copper, particularly the OFHC variety, is in the machining. One machinist I worked with at UCLA described it as "gummy”

One of the difficult aspects of copper for home diyers is power drilling.
Trimmed (shallow) cutting angle on the drill bit, rpm adjustment and less axial thrust is needed. And clamp component securely.
Otherwise, stacking (breaking) of the drill and damage of the copper component will occur.
Some hand fingure may be badly injured.

George
 
In 1990, just as in 10,000 BC, copper was still faster at transfering heat than aluminium was. :D :D :D

Wayne, there are so many things that can be done, it boggles the mind. The reason why many are seldom done is the usual suspect - price. And the added hassle.

I was thinking about having a copper plate machine bent (so there are no creases, deformities, etc) into an L shape, then bolting it to a hefty heat sink from the back side and then adding another half heigth sink on the front side. 20 years ago I wouldn't have wanted to do that, but today, with the proliferation of affordable CNC machines, it shouldn't be too much of a problem. Precision drilling is no longer a big, bad issue.

That should give me very good heat sink capability, admittedly WAY over what I need in everyday use, but - sort of - just in case I want to run it at high power levels for a longer time.

Given the areas involved, I'm going to need like a pound of thermal compound. :p

At this time, my current upper range of filters uses pure copper for its internal second bottom plate. I don't know if it helps, but I know it sure can't hurt. The point is, I already have some experience with copper; most materials require some hands-on experience to be comfortable with them.

Just make sure it can drive this ......:)
 

Attachments

  • image.jpg
    image.jpg
    745.1 KB · Views: 124
One of the difficult aspects of copper, particularly the OFHC variety, is in the machining. One machinist I worked with at UCLA described it as "gummy". Of course it has also gotten very expensive.

Once machined, it's not very rugged, and then you also need to plate it to prevent corrosion. A lot of work. But, thick enough it is a pretty good magnetic shield as well as a thermal conductor.

Agreed, but Brad, I don't need any fancy work, just simple holes and that one bend at the bottom for the version using TO-3 metal can trannies. That's it.

@magnoman

True, however, as pointed out above, it's just plain boards with me. And, for once, where I live works in my favor - copper is mined and refined locally, the town of Bor in eastern Serbia is Europe's biggest copper mine, so prices are low enough not to matter (for plates and cables).

With aluminium, it's the other way around, it has to be imported, no local resources.

@gpapag

George, all I need is simple drilling, polishing and plating. Trust me, I've done this before, it's no problem at all, just a bit of hassle, because I have to scoot all around town - buy in one place, drill at another, plate at a third place. A tour of around 30 miles/50 km all told.

Obviously, I anticipate three amps, so I'll do the plates for all three in one go.
 
Just make sure it can drive this ......:)

Wayne, good buddy, that's my new reference standard for evil loads. :D :D :D

That's no loudspeaker, it might just as well be a welding machine. In my case, your minimum of say 1.3 Ohms is 6.5 Ohms, I quite simply do not need what you obviously need.

It will not be able to drive that loudspeaker on a continuous steady state basis; it could be done, of course, add two more pairs and watch it go, but I'd also have to beef up just abut everything, and especially heat sinking.

However, if we come back to Earth, and recognize that we simply cannot run a power amp at full rated steady state power all of the time, as we must provide at least some headroom for transients, assuming we leave just 6 dB of headroom, then it should be able to drive them, bless your cotton socks.

As for those transients, remember that the protection will stay dormant for about 40 mS, so even larger peaks can be accommodated. Meaning you wouldn't have to settle for whisper level sound only.

Where DID you find those speakers, anyway? :cool:
 
All this talk reminds me of 1975 or 1976. A good friend had a pair of loudspeakers which were also bad news for many amps, especially in those days.

Then he bought a Studer/reVox A740 power amp. It was rated at 100/175W into 8/4 Ohms, and used three pairs of custom Motorola 250W devices per side, mounted on a VERY hefty heat sink. 2 x 33,000 uF caps per side. Big transformer, discrete high power diodes as rectifiers, ona separate heat sink running all the way from left to right. Said to be a low TIM level design. See it here: Prospekt Revox A 740

And suddenly, all his problems with both micro and macro dynamics were gone. Nada, just music. Thankfully, his speakers (I forgot the manufacturer's name, it wasn't a big name, and that was like 37 years ago) were fairly effcient, if memory serves something like 93 or 94 dB/2.83V/1m.

This was the case that gave me on object lesson of what drive capability is all about.

The case was sealed by a Harman/Kardon receiver, I think it was a model 930 (a alter evolution model of this one: http://wegavision.pytalhost.com/harman76/hk1.jpg ), nominally just 45 W/8 Ohms per side, but in a true dual mono topology, two big (for the power) transformers, two rectifiers, 2 x 2 capacitors, etc. 45 WPC is hardly ane exciting power level, nor was it exciting even then, but that receiver somehow managed to sound way bigger than it really was. And it refused to be caught out by any transient thrown at it. A fine machine I wouldn't mind owning even today.
 
Last edited:
I once thought of placing the output transistors in water . It might work if the contamination was kept low ( water filter element from a filter jug ? ) . The specific heat capacity of water from memory is 4.2 KJ / KgK . Deal with the heat as we do in motorcars . Oil might be more practical although not as good a transporter of heat . Water is a poor conductor of heat . Not a problem if it is kept on the move . I suspect even convection circulation might work . Anyone ever built a heat pipe ? We accept water in motorcars , why not in amplifiers ? It only has to be as well safeguarded as a central heating system . DF 96 says he likes facts . Here are some to chew over .
 
I once thought of placing the output transistors in water . It might work if the contamination was kept low ( water filter element from a filter jug ? ) . The specific heat capacity of water from memory is 4.2 KJ / KgK . Deal with the heat as we do in motorcars . Oil might be more practical although not as good a transporter of heat . Water is a poor conductor of heat . Not a problem if it is kept on the move . I suspect even convection circulation might work . Anyone ever built a heat pipe ? We accept water in motorcars , why not in amplifiers ? It only has to be as well safeguarded as a central heating system . DF 96 says he likes facts . Here are some to chew over .

The way you put it, Nige, it might appear you need some waterproof transistors. :D

Not "in water", but "use water cooling". We sort of had something like that with Sony's now dead in the water heat pipe, which used gasses much more efficient than water. It was all the rage in the early eighties, but now, nobody is using it as far as I know.

It was lighter than classic aluminium, true, but was in fact not really smaller, rather more compact in length rather than depth.

Also, it was hardly practical. If there was a leak, you were done for, you had next to nothing cooling and of course, had to dish out the dosh for a new one, only to discover that pricewise, it may as well have been made of a platinum and titanium alloy, bound together by uranium.

Not that it had to be so, of course, just the good, old Ford logic, accepted without question in Japan and revered more than the Emperor - sell the basic product as cheaply as you can, and then skin them alive in service and price parts.

I'll stick to my caveman aluminium fin heat sinks, thank you velly much.
 
Last edited:
Water cooled heat sinks have been used in industry to cool both tubes and transistors for multiple decades. It is a hassle, BECAUSE you have to used d-ionized water that you buy at significant cost and you still need a radiator the size of an auto, to dissipate the waste heat. We used to use one at Humphrey Instruments to cool a huge control amp for a strobe that we designed.
 
Love that revox , power meters and all, i still dont get it, why do designers keep leaving them out .....:(

POWER METERS = REAL AMPLIFIER !!!!!!


DVV , yeah...madness load coupled with 80db/M/2.83v, it would be easy if the sound was evil , but far from it , yeah they work me , most amplfiers loathe them ....:)


John !!! Power meters would look fantastic on the JC-1 ..:)
 
Administrator
Joined 2004
Paid Member
Water cooled heat sinks have been used in industry to cool both tubes and transistors for multiple decades.
Water cooled film gates were common in very high power movie projectors, such as those used at drive-ins. It was the only way to keep the gate and the film cool. Normal water was used, IIRC.

I haven't seen a water cooled digital project yet, but with the huge increases in brightness, they may not be far away.
 
In did think of TO3 with heads into the water . The PNP and NPN kept apart by a plastic bit , the pipe could be all plastic although daft not to use copper . Getting the water in close seems ideal . The problem with non reactive fluid is heat capacity , water is a very remarkable liquid . I suspect TO3 in water would outperform totally enclosed in oil . As luck would have it where water fails transistors also . I could well beleive transistors could be run hotter if direct liquid cooled , albeit slightly .
 
Status
Not open for further replies.