Until last year I had some 99 dB per watt baffles about the same size as Quad ESL 63s. With massive EQ 30 Hz flat was possible. A Quad 303 was excellent with them and not at all slow. In some ways better than the 63s and not too coloured. On moving house they had to go 3 watts typical measured about 115 dB two metres. 40 watts would be required for 30 Hz.
1.7 watts on my I don't want a sound bar is unbelievably loud. It's not loud enough perhaps for opera. 17 watts would get there. After that a house upgrade would be best
1.7 watts on my I don't want a sound bar is unbelievably loud. It's not loud enough perhaps for opera. 17 watts would get there. After that a house upgrade would be best
Naim nap 250 sounds better than the small Naim amplifiers. Mostly PSU. Is better might be better English and is the key difference in Naim of old. The PSU are often manifold in the same amplifier. My dyslexia and rushing might sometimes mix my words. I don't get paid for this so I seldom worry.
I have spent!y my life reassuring other dyslexic engineers. It seems many engineers are wired to use spelling ability for rotating objects in their minds. As my son says spelling came from spells.If you think about it who but the most perverse would treasure the ability to remember an illogical set of rules. Look at Pearson as a surname. What idiot came up with that. Nigel is also a French misunderstanding of a Norse name that the monks thought meant black as in Nigeria etc. We then have rules like A Historic or An Historic. The latter refers to it being of French origin or poetic. An hotel being not uncommon for that reason.
George Stephenson seems to have been dyslexic and especially Alan Blumlein.
George Stephenson seems to have been dyslexic and especially Alan Blumlein.
Member
Joined 2009
Paid Member
That's like all things. As a dyslexic boss I probably would have a slight preference for a dyslexic worker. This happened by accident as I learned many of the people I managed were dyslexic. They always were willing to please. It was rare that their disability was not counted against them. Like all things if you are not that person you can not imagine how it is. I was listening to a deaf blind lady who has just graduated from Harvard. She defines herself by landmark steps she has made. Although that's obvious I had never thought it. I have no idea how she can be so upbeat. I wouldn't want my disability to be more obvious than it is. I give to a charity called Smile It repairs children's hairlips. For almost no money you can change a life. Where they live the impediment is horribly real.
Can someone comment on the thermal stability of a quasi complementary output stage relative to the JLH output which can obviously be biased quite high? (With four NPN output devices on the heat sink along with the Vbe/bias device while the four drivers are mounted on the PCB without any heat sink)?
Would it be improved with the four drivers mounted to a little common heat sink?
Would it be improved if the drivers were mounted on the main heatsink?
I ask since I am dealing with a quasi complementary output stage that seems to have pretty poor bias stability at bias currents far below class A. One indicator is that the inner pair of the four NPN outputs biases much higher than the outer pair. The outer pair biases at about +/- 30 mA and the inner pair at +/- 50 mA. Emitter resistors are 0.22 Ohm.
Would it make any sense for me to disassemble this amp and measure the hFE and Vbe of the outputs and drivers? (It was purchases assembled on eBay so I did not have the chance to measure the hFE and Vbe like I did with my MA9S2.)
Perhaps I should replace the drivers and outputs with hFE and Vbe matched parts from an authorized distributor?
Would it be improved with the four drivers mounted to a little common heat sink?
Would it be improved if the drivers were mounted on the main heatsink?
I ask since I am dealing with a quasi complementary output stage that seems to have pretty poor bias stability at bias currents far below class A. One indicator is that the inner pair of the four NPN outputs biases much higher than the outer pair. The outer pair biases at about +/- 30 mA and the inner pair at +/- 50 mA. Emitter resistors are 0.22 Ohm.
Would it make any sense for me to disassemble this amp and measure the hFE and Vbe of the outputs and drivers? (It was purchases assembled on eBay so I did not have the chance to measure the hFE and Vbe like I did with my MA9S2.)
Perhaps I should replace the drivers and outputs with hFE and Vbe matched parts from an authorized distributor?
Attachments
John Ellis has covered this. Hopefully he will join in. All I know is that the heatsink must be large enough. The driver might benefit by being close to the outputs. Large seems to be twice as large as maths might suggest. Even for the more modest version 0.25 degrees per watt stereo.I didn't have to calculate mine as I had some heat sinks like that.
0.22 at the emitter is 7-10mV across the resistor. Ube spread can be larger. We need a selection of drivers and output transistors in each leg.
In class A, the current is 1-2 A. Base current limited. With the heating of the transistors, thermostating occurs (stabilization of the collector current).
Try to set the current to at least 50mA. The radiator should be of a large area with sparse vertical ribs and a thick base.
Use thermal grease and pads with good thermal conductivity.
There may be stability issues (generation).
In class A, the current is 1-2 A. Base current limited. With the heating of the transistors, thermostating occurs (stabilization of the collector current).
Try to set the current to at least 50mA. The radiator should be of a large area with sparse vertical ribs and a thick base.
Use thermal grease and pads with good thermal conductivity.
There may be stability issues (generation).
Last edited:
I suspect that the devices were not selected and that I have mismatched Vbe and/or hFE. Perhaps I should just disassemble it and pull (and measure) the drivers and outputs. I guess I should stick with kits instead of assembled boards in the future.
But I do not know if there is a design related issue, such as perhaps the lack of any thermal connection between drivers, no driver heat sink and/or no thermal coupling between drivers and the output heat sink.
What is the mechanism that results in thermostating instead of runaway?
Last edited:
The layout is for one common heat sink - you can look up the temperature rise per watt specifications in relation to the needs - power rating of the amplifier module.
There has been some recent discussion about staggered cross-over points having advantages where one set of outputs is still in moderate conduction where the other is in transition into cut off.
This discussion might have been in the thread on Bob Cordell's book.
Also being dyslexic and somewhere on the spectrum near autistic, I have learnt that my way of looking at the world is an advantage for me and to be treasured.
Nigel, have you ever taken the Myers-Briggs personality test as, in my experience, people with dyslexia seem to gravitate towards what are called 'Intuitive' thinkers who can see both detail and the bigger picture - not a common trait.
Kind regards
Mike