Hi guys.
I have a question for you.
I have a class A power amplifier built many years ago by an electronic engineer friend of mine that uses a single pair of complementary MJ3001 and MJ2501 darlingtons for a power of 32W per channel, powered by 32V. It worked perfectly and with a nice sound. Unfortunately, one channel no longer works. I would take advantage of this to update some components such as filter capacitors and other things. I need to know which darlingtons I can buy as a replacement since the originals are no longer available and also to use something more modern and superior, since all the work is entrusted to a single pair.
I would have thought of the BDV66D and BDV67D pairs or, even better, 2SD2449 and 2SB1594, but I need your authoritative opinion.
Thanks.
I have a question for you.
I have a class A power amplifier built many years ago by an electronic engineer friend of mine that uses a single pair of complementary MJ3001 and MJ2501 darlingtons for a power of 32W per channel, powered by 32V. It worked perfectly and with a nice sound. Unfortunately, one channel no longer works. I would take advantage of this to update some components such as filter capacitors and other things. I need to know which darlingtons I can buy as a replacement since the originals are no longer available and also to use something more modern and superior, since all the work is entrusted to a single pair.
I would have thought of the BDV66D and BDV67D pairs or, even better, 2SD2449 and 2SB1594, but I need your authoritative opinion.
Thanks.
The performance of your amp is likely not limited by the choice of output transistors. You can still buy MJ3001/2501 and BDX66/67D (BDV66/67 and 2SD2249/1594 are a TOP-3 package not TO-3). The BDX's have a higher Vceo and current rating. Whatever you buy, when it comes to these vintage types, you are not buying the 'real' thing, but something whose parameters meet (or exceed) the minimum requirement of the original. Especially if fT is far exceded, you can easily run into oscillations issues with your amp. I would try the 'originals' from a trusted manufacturer and suppplier.
Likely be easier to find
MJ11015
MJ11016
Currently manufactured by OnSemi and typically stocked in
my region, likely vary in others.
120 volt 200 watt 4 MHz
should exceed 80 volt 150 watt originals.
MJ11015
MJ11016
Currently manufactured by OnSemi and typically stocked in
my region, likely vary in others.
120 volt 200 watt 4 MHz
should exceed 80 volt 150 watt originals.
The MJ11015 - MJ11016 pair is the first one I thought of for its excellent qualities. Unfortunately, during the refurbishment work, I replaced the cooling fins with larger ones and with a different drilling from the TO3, so I cannot take them into consideration except as a last choice.
I have to do my research in the field of plastic Darlingtons.
I have to do my research in the field of plastic Darlingtons.
I used Sanken 2SB1647 / 2SD2560 and they work really nice for me, but my amp is class AB. I also have few NOS MJ2501/3001 in my drawer, don't know what to do with them. If you have original schematics of that amp, pls post.
For plastic Darlington
MJH6284G
MJH6287G
100 volt 160 watt 4 MHz
Decent pad area with TO-247
far far better than junk TIP 142/147
The typically " Sanken" recommendations likely will come in.
Those are typically higher speed from 20 to 40 MHz
Then it would be helpful to see the amplifier schematic
To determine what compensation would be increased.
Otherwise the normal MJH6284/87 are 4 MHz
better matched with better gain.
and junk tips around .5 MHz poorly matched pnp/npn
" Class A" amp or overbiased A/B
would likely have 2 to 4 pairs minimum not Single pair.
Depends on voltage and how much into the A region your biased too
MJH6284G
MJH6287G
100 volt 160 watt 4 MHz
Decent pad area with TO-247
far far better than junk TIP 142/147
The typically " Sanken" recommendations likely will come in.
Those are typically higher speed from 20 to 40 MHz
Then it would be helpful to see the amplifier schematic
To determine what compensation would be increased.
Otherwise the normal MJH6284/87 are 4 MHz
better matched with better gain.
and junk tips around .5 MHz poorly matched pnp/npn
" Class A" amp or overbiased A/B
would likely have 2 to 4 pairs minimum not Single pair.
Depends on voltage and how much into the A region your biased too
Sankens I used are actually even bit faster, 45MHz (P) and 70MHz (N) specified, DC hfe 13,000 (Y grade)and 11,000 (P grade), both very consistent in batch measured by myself
I think that's the same amplifier I made many years ago. It was published in CQ Elettronica in the '80s.
Class A amplifiers are more forgiving of fT than class AB. It doesn’t need to be as high to work well, and stability is not as sensitive.
The Sankens do have higher SOA than the OnSemi. The OnSemi parts are fine at these voltages, though. Those Sanken darlingtons (and even their Korea clones) are getting harder and harder to GET. Often not in stock, with large MOQ’s. KEC’s clones are supposedly available through Profusion, but the last time I checked they were not available to order over the web. I’m working from my lifetime buy from Digikey, and I literally got the last of the PNPs they had. They were down to a handful when I bought tubes of each, and now are non-stocked. The OnSemi MJH6284/7 are generally available. 4 MHz is plenty for raw frequency response. Class A does not have a switch-off issue so fT in the output transistors is almost a moot point.
It seems to me that Sanken (or even KEC) may be trying to keep these devices out of the hands of non-professionals.
The Sankens do have higher SOA than the OnSemi. The OnSemi parts are fine at these voltages, though. Those Sanken darlingtons (and even their Korea clones) are getting harder and harder to GET. Often not in stock, with large MOQ’s. KEC’s clones are supposedly available through Profusion, but the last time I checked they were not available to order over the web. I’m working from my lifetime buy from Digikey, and I literally got the last of the PNPs they had. They were down to a handful when I bought tubes of each, and now are non-stocked. The OnSemi MJH6284/7 are generally available. 4 MHz is plenty for raw frequency response. Class A does not have a switch-off issue so fT in the output transistors is almost a moot point.
It seems to me that Sanken (or even KEC) may be trying to keep these devices out of the hands of non-professionals.
I think I will use the MJH6284G and MJH6287G. I would prefer to have matching pairs. Can you tell me who provides this service?
For small quantities, that’s on the user. I wouldn’t sweat over matching for class A operation anyway.
The KTB2640G-Y and KTD1640G-Y are available and normally in stock at Profusion. They have proven to be great drop in replacements for Sanken 2SB1647 and 2SD2560 in production amplifiers. If replacing Motorola types, some compensation rework may be needed. This depends on the design , and sometimes lowering cap values instead of increasing them works. Simulation can help here, and testing with a dim bulb set up while observing a square wave output can really get it dialed in. Only those with experience and a complete test bench should attempt to replace low speed Darlingtons with fast ones. BTW , the MJH series is designed for switching , and the linear region SOA is quite limited.
The 6284/6287 were originally 2N types. For a while, On offered both. They recently dropped the TO-3. Those particular types were better than most switching darlingtons when it came to SOA. Full power at 30V and 100W at 50. Some would barely tolerate 18V supply operation in class A.
The published SOA for the 2640/1640 almost seems too good to be true. Better than the equivalent non-darlingtons, and even Sanken originals. Do we finally have a fast transistor pair that can safely run (beyond the warranty period) at +/-55V in singles? Driving dual-woofer loudspeakers?
The published SOA for the 2640/1640 almost seems too good to be true. Better than the equivalent non-darlingtons, and even Sanken originals. Do we finally have a fast transistor pair that can safely run (beyond the warranty period) at +/-55V in singles? Driving dual-woofer loudspeakers?
So, do you think I can be reasonably comfortable using 6284/6287 without making any modifications, since I have no other equipment than a tester?
It wasn't easy, but with patience I got the final part of the class A amplifier I'm talking about in the thread, hoping it can be of help in determining which of the replacements you recommended can work without changes.
That looks like a run of the mill class AB output stage. They can be biased into class A - the voltage across the emitter resistor will tell you. You have one channel still working, right? Might want to measure that and see if it was really biased into class A or not.
It’s probably a very basic amplifier circuit in front of that. And if it’s really running AB, you ARE better off (much better off) with the faster types of darlingtons.
To give it some perspective, a proper class A bias current would be around an amp and a half of idle current. That’s damn hot. Milliamp levels is AB operation.
It’s probably a very basic amplifier circuit in front of that. And if it’s really running AB, you ARE better off (much better off) with the faster types of darlingtons.
To give it some perspective, a proper class A bias current would be around an amp and a half of idle current. That’s damn hot. Milliamp levels is AB operation.
Then those MJH628x will be fine, as far as power dissipation, SOA, and speed. But watch bias drift with (any) ordinary darlingtons. A regular Vbe multiplier (which this has) may or may not be able to compensate properly. You really need a sense transistor with a very low Vbe so a high multiplication factor happens by default (tending toward overcompensation rather than under compensation). The BD13x are not exactly known for that - neither are most TO-126 driver types. They tend to run almost 0.7. The original may have run away on you causing the dead channel.
The more common (and stable) way of biasing push pull class A is to configure the entire output stage as a CCS, where the sum of both emitter currents is set by local feedback. When one goes up the other goes down and the difference current flows to the speaker. Tends to be more thermally stable. There may be drift, but always down in current over temperature.
The more common (and stable) way of biasing push pull class A is to configure the entire output stage as a CCS, where the sum of both emitter currents is set by local feedback. When one goes up the other goes down and the difference current flows to the speaker. Tends to be more thermally stable. There may be drift, but always down in current over temperature.