The only thing that would give equivalent performance in that circuit is the B2640/D1640 - assuming you can still get those. For new design, simply dont design around them. Use discretes, because there will always be SOMETHING that can be used.
I did design around D2083/B1383, but it was more than 5 years ago and I did a lifetime buy. It was a class G amp for HF driver duty. I’d have to shell out for custom heat sinks and new boards to be able to shoehorn eight more TO-220’s in there, so I need to keep all my extras. I also snatched up 16 pair of B2640/D1640 when I saw the writing on the wall (when they went MOQ with low remaining stock at Digikey). They are a little light on current capacity for bridged 4 ohm, but ok anywhere else.
I did design around D2083/B1383, but it was more than 5 years ago and I did a lifetime buy. It was a class G amp for HF driver duty. I’d have to shell out for custom heat sinks and new boards to be able to shoehorn eight more TO-220’s in there, so I need to keep all my extras. I also snatched up 16 pair of B2640/D1640 when I saw the writing on the wall (when they went MOQ with low remaining stock at Digikey). They are a little light on current capacity for bridged 4 ohm, but ok anywhere else.
If you wanted to ensure the same SOUND as the original one would simply use the highest current audio outputs available (17A, I believe) and a bog standard high fT driver pair. Connect the same base emitter resistors as used in the Sanken darlingtons in the same way (which is NOT the good way). It will have very similar crossover characteristics - as you will end up with the same issues that they create.
Hi Njswede,
All the components are available except the output pair. This amplifier was sold by Denon up to 2020. It is still available in India. I really liked the power and sound quality, besides the circuit looked simple. never really thought the output pair would go obsolete so soon
All the components are available except the output pair. This amplifier was sold by Denon up to 2020. It is still available in India. I really liked the power and sound quality, besides the circuit looked simple. never really thought the output pair would go obsolete so soon
If you’re a semiconductor maker, you make them go obsolete as soon as you can. It’s what you do. If you want to save fifteen percent you switch to Geico. It’s what you do.
A close replacement for 2SD2083/2SB1383 currently available are the KTB2640G and KTD1640G. Will try to source them from https://us.profusion.uk/
Attachments
Sanken just put those originals on MOQ. Need to order full box quantity now (no more onesey twoseys or even 25’s). Not as high a current rating as the D2083, but hFE at 15A is comparable. It just can’t go to 25A (or 40 pulsed). Most amps including this one won’t - the one I did does. If I wasn’t trying to get them into such a tiny case, using heat sinks I already had, I would have used TIP35C and a fast driver, or paralleled those. I did get a couple dozen of the D2640 for future servicing, since all the D2083’s were now spoken for.
The KECs are brand new. “One leader falls, another rises”. Their track record is yet unproven. I’ve started messing with their drivers, and all indications are good enough for my purposes. But those are relatively easy to make. Look at SOA curves for any 25 to 50W sustained-beta drivers. SOA is relatively poor, with second breakdown showing it’s ugly head at relatively low voltages. Doesn’t matter in the applications, as you only start drawing any real current off the drivers at current peaks in the outputs, where the beta is falling. If this happens at all in a modern amplifier the Vce is going to be small. At the zero crossings your outputs are still in the linear portion of the beta curve, and not much (maybe 50 to 100 mA) is demanded off the driver. Power handling can fall and it’s not the end of the world since you might need 5 watts at 50V. Whether KEC can make a 150W output device with 20+ MHz fT, that can handle full power at 30, 50, or 70 volts is yet to be determined. If they can, it is going to be a game changer. EVVO is promising such performance, but their data sheets are all effed up and full of errors. That is not inspiring confidence in what they are shipping. Maybe they are good. But they are going to have to PROVE it. If these second sources are really any good for the highest power devices, you’ll see On exit the audio market faster than you can get your credit card out and do lifetime buys of NJW3281’s. And Sanken will likely get swallowed up, unfortunately.
.
The KECs are brand new. “One leader falls, another rises”. Their track record is yet unproven. I’ve started messing with their drivers, and all indications are good enough for my purposes. But those are relatively easy to make. Look at SOA curves for any 25 to 50W sustained-beta drivers. SOA is relatively poor, with second breakdown showing it’s ugly head at relatively low voltages. Doesn’t matter in the applications, as you only start drawing any real current off the drivers at current peaks in the outputs, where the beta is falling. If this happens at all in a modern amplifier the Vce is going to be small. At the zero crossings your outputs are still in the linear portion of the beta curve, and not much (maybe 50 to 100 mA) is demanded off the driver. Power handling can fall and it’s not the end of the world since you might need 5 watts at 50V. Whether KEC can make a 150W output device with 20+ MHz fT, that can handle full power at 30, 50, or 70 volts is yet to be determined. If they can, it is going to be a game changer. EVVO is promising such performance, but their data sheets are all effed up and full of errors. That is not inspiring confidence in what they are shipping. Maybe they are good. But they are going to have to PROVE it. If these second sources are really any good for the highest power devices, you’ll see On exit the audio market faster than you can get your credit card out and do lifetime buys of NJW3281’s. And Sanken will likely get swallowed up, unfortunately.
.