Hi Bama Slamma,
That would put you in the neighborhood of about 100 watts continuous. A little conservative, but wise given the SOA plus lower than nominal impedances with possible large phase angles for current vs voltage. That will increase your output dissipation more than most people expect.
Can you get more out by pushing them? Yes, at the expense of a less reliable design. Don't skimp on the heat sinks either.
Also, don't forget that you may run out of driver transistor SOA. Keep these things in mind.
-Chris
That would put you in the neighborhood of about 100 watts continuous. A little conservative, but wise given the SOA plus lower than nominal impedances with possible large phase angles for current vs voltage. That will increase your output dissipation more than most people expect.
Can you get more out by pushing them? Yes, at the expense of a less reliable design. Don't skimp on the heat sinks either.
Also, don't forget that you may run out of driver transistor SOA. Keep these things in mind.
-Chris
Hi Chris,
Baby steps on startup - +/-15 volts. +/-70 volts now.
Heatsink is on the small side (homemade), and would not do for full musical program.
How hot? After watching a bass heavy movie for an hour and a half it's pretty warm, but easy to keep hold of. I had originally planned on a fan but decided against it. Unit's been in use for more than 2 years. No smoke.
Also it's EQ'ed with a Linkwitz transform, so the LFE track can push it.
Baby steps on startup - +/-15 volts. +/-70 volts now.
Heatsink is on the small side (homemade), and would not do for full musical program.
How hot? After watching a bass heavy movie for an hour and a half it's pretty warm, but easy to keep hold of. I had originally planned on a fan but decided against it. Unit's been in use for more than 2 years. No smoke.
Also it's EQ'ed with a Linkwitz transform, so the LFE track can push it.
If you are running 8 ohm loads, you can even get 200 watts per channel from 2 pairs, however as Chris says... any lower impedance or reactive loads could fry them. So for an actively crossed 8 ohm subwoofer application, I would go ahead use +/-70-80 vdc rails on 2 pairs. Any funny loads and make that 4 pairs.
A good design with a powerful PSU should yield a 100wpc amp with 2 pairs, with a skimpy PSU, you ccould do a 150wpc amp . (Because with a skimpy PSU, the rails sag much more and protect the devices).
A good design with a powerful PSU should yield a 100wpc amp with 2 pairs, with a skimpy PSU, you ccould do a 150wpc amp . (Because with a skimpy PSU, the rails sag much more and protect the devices).
That would be a negative on the clipping Chris. My temperature probe during testing shows 58-62 C, so heatsink is doing ok. There's more of it under the board.
Not driven hard normally, but has no problem providing enough juice to drive 2-18" subs to the near destruction of my house.
As you could probably guess from my handle, I'm fond of these BJT's.
Not driven hard normally, but has no problem providing enough juice to drive 2-18" subs to the near destruction of my house.
As you could probably guess from my handle, I'm fond of these BJT's.
Attachments
On the subject of the 21193 and 21194:
I was operating single pairs in a test amplifier of +/-40V into a 1ohm load and heated one of the transistors enough to cause it to de-solder itself.
I soldered the lead back on expecting a dead transistor; the same amplifier is still in use today and that was two years ago.
They are tough transistors.
EDIT: Don't ask what maniacal thing I was doing with that poor amp (of my own design), but it was very fun.
I was operating single pairs in a test amplifier of +/-40V into a 1ohm load and heated one of the transistors enough to cause it to de-solder itself.
I soldered the lead back on expecting a dead transistor; the same amplifier is still in use today and that was two years ago.
They are tough transistors.
EDIT: Don't ask what maniacal thing I was doing with that poor amp (of my own design), but it was very fun.
Bama Slamma said:
Amp is pure class B. No prptection circuit
. It's used on the sub pictured above.Duo said:
Unsoldered? Excellent.
During full power testing, I accidentally shorted the output. The two power supply fuses melted, and the 15 amp breaker on my house electrical panel tripped! None of the transistor were damaged.
Tough as nails.
Hi,
from 35 +35Vac you can get about +-50Vdc.
2 pair of mj21193/4 should get about 100 to 110W into 8ohms and phase can be 60degrees with Tc<=60degC.
you can get near 200W into 4ohms from the same supply and the reactive load can be just as severe. Again this amp will work all day long with the most severe 4ohms loading as long as the devices are kept below 60degC.
This will require a heatsink temperature no higher than about 50 to 55degC.
You could use a 42Vac transformer and get the same 200W but into an 8ohm load, again @ 60degree phase angle.
This time though you must ensure that 6ohm (@ 45degree phase angle) or 4 to 8ohm speakers do not get used PA fashion.
But the 2pair do make a superb domestic power house.
from 35 +35Vac you can get about +-50Vdc.
2 pair of mj21193/4 should get about 100 to 110W into 8ohms and phase can be 60degrees with Tc<=60degC.
you can get near 200W into 4ohms from the same supply and the reactive load can be just as severe. Again this amp will work all day long with the most severe 4ohms loading as long as the devices are kept below 60degC.
This will require a heatsink temperature no higher than about 50 to 55degC.
You could use a 42Vac transformer and get the same 200W but into an 8ohm load, again @ 60degree phase angle.
This time though you must ensure that 6ohm (@ 45degree phase angle) or 4 to 8ohm speakers do not get used PA fashion.
But the 2pair do make a superb domestic power house.
From what I hear, these BJT's are pretty tough. Using a 40-0-40V transformer would give me +/-56V rails (unloaded). I am using full range, passive crossover speakers rated 8 ohms nominal. As it's a full range amp I want to make, I guess I need to add some bias so it's AB rather than pure B. 200wpc channel would make a bad-a@#$ home theatre system.
Hi,
40Vac will give a bit more than +-56Vdc, expect 58 to 59Vdc unloaded and 57.5 to 58.5Vdc with a low bias current (Ib=100mA).
These voltages only just work with 63Vdc capacitors. If you go higher with the transformer you will need 80Vdc capacitors, 100Vdc if you can afford them.
+-58Vdc PSU into 2pair of MJ21193/4 will produce a maximum of 170W if you optimise all the components for maximum power.
It may sound a little better (higher Re=0r47 requires Ib=50mA per transistor) if you aim for 150W to 160W.
This arrangement will drive 6ohm quite a little better (full 60degree phase angle) than suggested in my last post. And 4ohm phase to 40degrees.
40Vac will give a bit more than +-56Vdc, expect 58 to 59Vdc unloaded and 57.5 to 58.5Vdc with a low bias current (Ib=100mA).
These voltages only just work with 63Vdc capacitors. If you go higher with the transformer you will need 80Vdc capacitors, 100Vdc if you can afford them.
+-58Vdc PSU into 2pair of MJ21193/4 will produce a maximum of 170W if you optimise all the components for maximum power.
It may sound a little better (higher Re=0r47 requires Ib=50mA per transistor) if you aim for 150W to 160W.
This arrangement will drive 6ohm quite a little better (full 60degree phase angle) than suggested in my last post. And 4ohm phase to 40degrees.
Hi John,
They are very good parts for sure! If you want a real treat for lower power, look at the MJW0281A and MJW0302A. Those are the best parts I have seen in a long time. They would probably make excellent drivers for a bunch of the MJ2119x series.
Hi Andrew, Joe,
I was thinking along those lines. I will say that often a home environment is harder on amps than a pro environment. At least the owners manual isn't sitting over the vents.
-Chris
Yep. From the first time I saw your avatar.
They are very good parts for sure! If you want a real treat for lower power, look at the MJW0281A and MJW0302A. Those are the best parts I have seen in a long time. They would probably make excellent drivers for a bunch of the MJ2119x series.
Hi Andrew, Joe,
I was thinking along those lines. I will say that often a home environment is harder on amps than a pro environment. At least the owners manual isn't sitting over the vents.
-Chris
Thanks for all the advice. This is one of those projects I'll attempt when I have some extra cash. Maybe get some Yorkville YS-115's (8 ohms) for the finished amp. Specs I'm going for: 1vrms in for full output, protection circuit included so the thing doesn't blow up if shorted or operated with too low a load impedance, full-range response, class AB. I already have a kit amp I built that uses a single pair of Toshiba 2SC5200/2SA1943 on each channel. The supply for it is +/-45 volts and it does 100w into 8 ohms. Having never designed an amp from scratch before, some pointers would be greatly appreciated.
anatech said:
Hey Chris,
Have you used these as drivers? Or are they better used in the output. They are good looking, very linear, well matched gain, and massive SOA.
I have 4 pairs of MJL4281A/MJL4302A collecting dust. Any ideas on possible configurations with the MJW0281A and MJW0302A as drivers?
I've been thinking of a Bryston 3B clone for a while. How about with these devices?
Hi John,
I have built some prototype output stages with these and found them to be excellent. They would prove to be very good drivers as well in a higher power class. I'm thinking the bigger Adcom units also.
These would be deadly overkill for a 3B clone. Also a waste in my view. I'm not a fan of complimentary differential pairs. Consider something more reliable and better sounding. Don't forget these had very, very poor bias compensation circuitry on top of everything else.
I did build a SymAsym (5.3) with these as outputs. I required far lower bias current to get low distortion with no visible notches (using a THD analyzer to get rid of the fundamental frequency). About 22 mA if you can believe that. That really convinced me of the value of the MJW0281A/MJW0302A pair.
-Chris
I have built some prototype output stages with these and found them to be excellent. They would prove to be very good drivers as well in a higher power class. I'm thinking the bigger Adcom units also.
These would be deadly overkill for a 3B clone. Also a waste in my view. I'm not a fan of complimentary differential pairs. Consider something more reliable and better sounding. Don't forget these had very, very poor bias compensation circuitry on top of everything else.
I did build a SymAsym (5.3) with these as outputs. I required far lower bias current to get low distortion with no visible notches (using a THD analyzer to get rid of the fundamental frequency). About 22 mA if you can believe that. That really convinced me of the value of the MJW0281A/MJW0302A pair.
-Chris
I've got quite some expereince with the 21194 and 21193 devices- they are rugged as hell. I however decided to move to 1281/1302's because of Ft considerations. The 21194/21193 have an Ft of about 4MHz whereas the 1281/1302 are good for about 30Mhz, so I can run the amp with a bit more loop gain and therefore lower distortion overall. My amp is heavily class AB biased.
I have to add that the 21193/21194 SOA is very good - I've been using 5 pairs (so 10 in total) for 250W, but will be going to 8 pairs of the 1281/1302 to go up a bit more in power and to make up for the lower overall Pd (250w/device vs 200W per device).
I have to add that the 21193/21194 SOA is very good - I've been using 5 pairs (so 10 in total) for 250W, but will be going to 8 pairs of the 1281/1302 to go up a bit more in power and to make up for the lower overall Pd (250w/device vs 200W per device).
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.