If using really large MJL4281A/MJL4302A but only a single pair, what kind of power can you expect. Not worried about years of reliability, just what max you can get away with on a single pair on an average range of cheap heatsinking.
How much power @ 4 ohms and what rails
How much power @ 2 ohms and what rails
and @ 1 ohm?
*I'm planning out a small 2 channel amp for some people with 2 channels, 1 pair per channel on a 12.6/0/12.6V transformer to give me approx 18V-0-18V rails. Low rails, but high current.
I want the ohms impedance to be VERY flexible because many average people like to parallel a bunch of speakers on their amps, not knowing any better. many times getting less than 4 or 2 ohms per channel. I'm planning 1 ohm stable, so he can wire whatever combo of speakers he likes. Would one pair of MJL4281/4302 handle these rails at 1 ohm?
How much power @ 4 ohms and what rails
How much power @ 2 ohms and what rails
and @ 1 ohm?
*I'm planning out a small 2 channel amp for some people with 2 channels, 1 pair per channel on a 12.6/0/12.6V transformer to give me approx 18V-0-18V rails. Low rails, but high current.
I want the ohms impedance to be VERY flexible because many average people like to parallel a bunch of speakers on their amps, not knowing any better. many times getting less than 4 or 2 ohms per channel. I'm planning 1 ohm stable, so he can wire whatever combo of speakers he likes. Would one pair of MJL4281/4302 handle these rails at 1 ohm?
Last edited:
2R and 1R , no no no. 8R only , maybe 80W @ 50V rails. OEM's get away with 4R w/ one pair by undersizing the power supply , collapsing the rails before SOA is exceeded.
2 pair NJW0281/0302 @ mouser are the same price , fit in the same space and
will do 120/200W 8/4R w/ 58V rails (all day).
OS
2 pair NJW0281/0302 @ mouser are the same price , fit in the same space and
will do 120/200W 8/4R w/ 58V rails (all day).
OS
OK so lower rails at lower ohms. 8 ohms only, HAHAHAHA I've seen TO264 devices on 75V rails at 8 ohm.
Not concerned with 8 ohms at all. Anything can push 8 ohms.
I can push 4 ohms easy with a pair of TO-220 devices at the right voltage.
Last edited:
into such low impedance it will run like a classA amp, and BIG heatsinks needed
can't do it with 'average cheap heatsink', using your own words
even if classAB, you will have to build like it were a classA
but built like classA, you can do it with about 25-30V rails
1ohm load might be possible with very effective forced cooling
amp design needs to be very stable and reliable
2ohm is possible, but 1ohm is really pushing the limit
playing loud on a really hot day, and it may be gone
can't do it with 'average cheap heatsink', using your own words
even if classAB, you will have to build like it were a classA
but built like classA, you can do it with about 25-30V rails
1ohm load might be possible with very effective forced cooling
amp design needs to be very stable and reliable
2ohm is possible, but 1ohm is really pushing the limit
playing loud on a really hot day, and it may be gone
MJL4281A/4302A
Collector current
-continuous: 15A
- peak: 30A
So I see no problem with supply only +/-18 VDC
and 1 Ohm load regarding the transistors.
But in a high current low voltage amplifier there are other difficulties.
This includes rail inductances.
It is just not like a 'normal' amplifier for 4/8 Ohm.
Collector current
-continuous: 15A
- peak: 30A
So I see no problem with supply only +/-18 VDC
and 1 Ohm load regarding the transistors.
But in a high current low voltage amplifier there are other difficulties.
This includes rail inductances.
It is just not like a 'normal' amplifier for 4/8 Ohm.
If you want low voltage / high current nothing beats the 15003/15004.
Just make sure you get the original ONsemi from a reliable source.
It's only a matter of 6 more holes on the H/S but there is nothing like that for power handling.
Of course the design of the amplfier/PCB/power supply/heatsink/wiring(!) must be adequate if you do not want to cheat on output power.
Just make sure you get the original ONsemi from a reliable source.
It's only a matter of 6 more holes on the H/S but there is nothing like that for power handling.
Of course the design of the amplfier/PCB/power supply/heatsink/wiring(!) must be adequate if you do not want to cheat on output power.
Hmmm. Looking at figure 1 & 2 in the datasheet with Vce=5V the current gain plummets after a few amps, as does the GBWP. Since lesser Vceo is not a problem here I presume, perhaps a higher conductance device would be more appropriate?
I've always wondered how much low Z sub-woofer power could be made with a pair of these super slow things.
They have substantial SOA.
Why would your transformer sag - design it for the purpose you intend and it would not sag. You can design a transformer with very good regulation, it depends on the turns /volt the size of the core and the size of the wire. If you design the transformer for 18V (rms) at 20 amp (rms) then you have no problem.
The gain on those devices drops like a rock past 10A (peak), so figure 7A RMS. That's 50W at 1Ω.
The gain curves are also specified at DC, it will not be that high for audio frequencies.
One pair will need a hefty driver, 0.5A or so, and a front end to match. You might want to run a triple EF output stage.
Carver used the 1302/3281 pair in many amps, running 15A (peak) into 4Ω with one pair. He also used another identical pair for the drivers, and the front end could supply 50mA or so.
The gain curves are also specified at DC, it will not be that high for audio frequencies.
One pair will need a hefty driver, 0.5A or so, and a front end to match. You might want to run a triple EF output stage.
Carver used the 1302/3281 pair in many amps, running 15A (peak) into 4Ω with one pair. He also used another identical pair for the drivers, and the front end could supply 50mA or so.
I've always wondered how much low Z sub-woofer power could be made with a pair of these super slow things.
Their lower-current cousins (2N5886) work very well for 12V inverters pushing 300 watts output. The current gain at 25A is higher than that of the modern sustained-beta high fT types. And I've pushed 2-ohms load from +/-28V rails from these as well - even though the SOA curves say you're not supposed to.
I may still have a few of them left to play with.
Would it not be unrealistic to design for 15-20V rails. The criteria:
But , what if they just hook one pair ? Disappointment might follow (no headroom). 2 chipamps bridged might be better. My daughter's logitech sub uses 25v rails with dual chipamps. Woofer measures 3R.
A real amp really designed for 40V/2R would still be cheap , if you are drive "shorts" you need a "arc welder" (like cordell's triple w/3 pairs OP - below). Optimum would be 2sc3503/1381 - njw0281/0302 - 3 pair njw21193/4 (pre/driver/OP). Anything short of this would be a "hack job". I have the parts for this (will build soon) , costs $32usd .... not bad.
For 18V rails at least a standard triple with 2 pair OP (or 1 pair MJ21193/4 - to-3 metal can).
OS
But , what if they just hook one pair ? Disappointment might follow (no headroom). 2 chipamps bridged might be better. My daughter's logitech sub uses 25v rails with dual chipamps. Woofer measures 3R.
A real amp really designed for 40V/2R would still be cheap , if you are drive "shorts" you need a "arc welder" (like cordell's triple w/3 pairs OP - below). Optimum would be 2sc3503/1381 - njw0281/0302 - 3 pair njw21193/4 (pre/driver/OP). Anything short of this would be a "hack job". I have the parts for this (will build soon) , costs $32usd .... not bad.
For 18V rails
at least a standard triple with 2 pair OP (or 1 pair MJ21193/4 - to-3 metal can). OS
Attachments
I believe the 2N5686 is really just two 2N5886 dies on the same heat slug.
I have some 2N5886/84 somewhere around here too.
Last edited:
For 18V rails
OS
my monos use 30V rails. and single pair of MJL21193/94
A friend owns a pair of Kappa9
are they 5ways or something ?
exstremely low impedance load
known as notorious amp killers
well, my friend had never heard his speakers sound that nice before
not exstremely loud, but still plenty of low end power
I feared to take a pair of burned amps home with me
instead I was rewarded
I now know I own a pair of pretty rare amps
I don't know, but is there a kind of reliability in single output pairs
I mean, with multiple pairs, what happens then
how accurate and reliable is matching when it gets really tough
one pulls more than the other, etc ... and puff
what can I say
these amps 'allowed' me to build better speakers
they simply responds to everything
they can be rough as hell, or sound sweet as sugar
they simply do what is asked
simple as that
Very interesting OP stage . (my book is still hostage of the italian Post..
)Is it some kind of diamond buffer? I have seen something similar around with bootstrapped sorces, recently. I won't call it a standard triple anyway (i.e. a la Leach), since I do not see 3 summing Vbe per rail.
So it should have high current gain with better thermal stability, I guess.
If you build it, keep us informed....
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.