I am refurbishing an old vintage tape recorder. Its headphone amplifier is an NJM386. Is there a better alternative?
njm386 with 0.2% distortion was designed for portable am/fm radios.
Not a direct replacement, but many op amps could drive headphones. How about njm4556.
Not a direct replacement, but many op amps could drive headphones. How about njm4556.
The NJM386 is a second-source version of the ancient LM386. If you have 8-pin DIP format the LM386 is still available, easily sourced, and very cheap - and will drop-in.
Must be 50yrs old...
Must be 50yrs old...
Well, I could make an adapter board to convert a modern opamp to the NJM386's pin out. An OPA 1655 would seem to be a viable option, but I can't tell from its datasheet if it would drive 32Ω headphones. The portion of the 165x datasheet devoted to headphone application shows an additional BUF 634A to handle the current.......
A better alternative might require a better power supply. I am thinking you could adapt the existing circuit to use LM1875.
Why bother?
There is zero wrong with the NJM386. It was designed as a headphone amp for starters, and it has lower distortion than your headphones do by far.
Now... if you want higher power you can go discreet or use a composite op amp (buffer within the feedback loop).
Now you said tape recorder. I serviced tape decks from the 1970s on and recording studio machines. If you were aware of how much distortion and noise in on any tape, it would put everything into focus for you. I did Revox, Teac / Tascam and Nakamichi (and other good brands) under warranty. The tape itself is not a low distortion media.
There is zero wrong with the NJM386. It was designed as a headphone amp for starters, and it has lower distortion than your headphones do by far.
Now... if you want higher power you can go discreet or use a composite op amp (buffer within the feedback loop).
Now you said tape recorder. I serviced tape decks from the 1970s on and recording studio machines. If you were aware of how much distortion and noise in on any tape, it would put everything into focus for you. I did Revox, Teac / Tascam and Nakamichi (and other good brands) under warranty. The tape itself is not a low distortion media.
I suggest to wire a RCA jack to the volume control of the tape machine so that it can be connected to a hifi. That will bypass the 386 circuit. If sound is improved than maybe 386 replacement is satisfactory.
The 386 has excessive gain and it's distortion is probably worst than any other audio chip. It probably masks the tape distortion.
Here are suggestions for making the best use of the 386 .. https://electro-dan.co.uk/Electronics/LM386.aspx
The 386 has excessive gain and it's distortion is probably worst than any other audio chip. It probably masks the tape distortion.
Here are suggestions for making the best use of the 386 .. https://electro-dan.co.uk/Electronics/LM386.aspx
Actually, it doesn't. Gain is set by external components. "Excess" gain is then used in feedback, and that lowers distortion and noise.
Now given that this chip has low distortion and tape distortion can easily exceed 1 %, what are you saying? If you survey op amps, you'll find the NJM386 isn't that bad for distortion or noise compared to tape. Are there better? Yup, you'd use those in head preamps or magnetic phono preamps, mics and stuff like that. The early LM386 was worse, that was back in time.
It is designed for low current draw at low voltages with a single supply. Now I guess what should be asked is ... what is the make and model of tape machine? If it is a portable, the correct part is being used.
An NJM4556 is designed as a headphone amp, it's dual and uses a bipolar power supply. If the machine has a bipolar supply, maybe a new PCB can be designed and installed since you're going that far. The x386 is a completely different circuit configuration than other devices. A special part.
So look at what you have to work with, go from there.
Now given that this chip has low distortion and tape distortion can easily exceed 1 %, what are you saying? If you survey op amps, you'll find the NJM386 isn't that bad for distortion or noise compared to tape. Are there better? Yup, you'd use those in head preamps or magnetic phono preamps, mics and stuff like that. The early LM386 was worse, that was back in time.
It is designed for low current draw at low voltages with a single supply. Now I guess what should be asked is ... what is the make and model of tape machine? If it is a portable, the correct part is being used.
An NJM4556 is designed as a headphone amp, it's dual and uses a bipolar power supply. If the machine has a bipolar supply, maybe a new PCB can be designed and installed since you're going that far. The x386 is a completely different circuit configuration than other devices. A special part.
So look at what you have to work with, go from there.
The deck is a Sony Walkman WM-D6C with a single supply (~6v) and very limited space for any mods. There is already a separate line out that is not driven by the 386.
@fubar3: Thanks for the link to that 386 page! There are indeed some good suggestions there that I will include.
This deck is forty years old so I was just thinking there might be a plug-in alternative to the 386 that has come out in those years. Like I said, I might be able to squeeze in an adapter board to convert a modern opamp to the existing 8-pin DIP pinout but probably nothing more than that. The OPA1656 is the best I could find but still no indication of whether it will drive 32Ω headphones. I don't think I'll have the need for 8Ω drive capability.
@fubar3: Thanks for the link to that 386 page! There are indeed some good suggestions there that I will include.
This deck is forty years old so I was just thinking there might be a plug-in alternative to the 386 that has come out in those years. Like I said, I might be able to squeeze in an adapter board to convert a modern opamp to the existing 8-pin DIP pinout but probably nothing more than that. The OPA1656 is the best I could find but still no indication of whether it will drive 32Ω headphones. I don't think I'll have the need for 8Ω drive capability.
Go here: https://www.ti.com/audio-ic/amplifiers/headphone-amplifiers/products.html One of the filters is "load (min)", there are a few suitable for <32 ohms
The T.I. link is excellent---thanks! But it appears that all those devices are STEREO, whilst mine are mono---one chip for left, one chip for right. So that would mean a lot more PCB butchering.....hmmmm......
I am familiar with the Sony Walkman WM-D6C from a service standpoint. Don't bother. What is in there is balanced in quality with the machine, and space is extremely limited.
If you are really bent on making things better, make a separate box with a headphone amp and power supply (batteries or whatever) and plug it into the line out. Besides, it has value in original condition, don't wreck it for those who love Sony Walkmans. It will probably sound better purely because you can have higher supply voltages and therefore higher power. We need headroom for peaks.
If you are really bent on making things better, make a separate box with a headphone amp and power supply (batteries or whatever) and plug it into the line out. Besides, it has value in original condition, don't wreck it for those who love Sony Walkmans. It will probably sound better purely because you can have higher supply voltages and therefore higher power. We need headroom for peaks.
@anatech: THANKS for your advice; I think I will just leave well enough alone with the 386; although I did note in one posting that there were several versions of the 386---I wonder if the later ones are better.
Are there any other mods you would recommend to improve performance? I see some coupling caps that seem on the low value side. Might be able to improve the LF response. One component I see that definitely needs upgrading are those awful NJM4558 opamps. An NE5532 or OPA1642 would be a major improvement, IMO.
Are there any other mods you would recommend to improve performance? I see some coupling caps that seem on the low value side. Might be able to improve the LF response. One component I see that definitely needs upgrading are those awful NJM4558 opamps. An NE5532 or OPA1642 would be a major improvement, IMO.
Hi dotneck335,
They were careful to keep power consumption down. The op amps reflect this. I don't think the coupling caps will improve LF performance, and it is limited in headphones anyway - especially the ones popular with that application. So you can use lower noise op amps in the head amp maybe. I can't remember if the head amp was in a big chip or not, so if the head amp is an op amp you can lower noise. If it uses transistors before the op amp in the head amp, changing it won't affect much.
Distortion from the rest of the electronics (Dolby) totally swamps any improvements in distortion. You have to consider this from a system viewpoint or you're only wasting money.
Some products were designed as a system and were balanced from a performance viewpoint. Improving one section might bring marginal improvements, but nothing like you might expect. When considering op amps, look at the current draw as well as minimum voltages. You might consider a rail to rail, single supply op amp.
They were careful to keep power consumption down. The op amps reflect this. I don't think the coupling caps will improve LF performance, and it is limited in headphones anyway - especially the ones popular with that application. So you can use lower noise op amps in the head amp maybe. I can't remember if the head amp was in a big chip or not, so if the head amp is an op amp you can lower noise. If it uses transistors before the op amp in the head amp, changing it won't affect much.
Distortion from the rest of the electronics (Dolby) totally swamps any improvements in distortion. You have to consider this from a system viewpoint or you're only wasting money.
Some products were designed as a system and were balanced from a performance viewpoint. Improving one section might bring marginal improvements, but nothing like you might expect. When considering op amps, look at the current draw as well as minimum voltages. You might consider a rail to rail, single supply op amp.
Yes, the current draw is vital in these devices. The OPA1642 draws half as much current as the 4558, and its rail-to-rail performance could be a real advantage given that the WM-D6C only provides a ~10.8 volt power source. Every bit of headroom always helps. And they're only about $2 each, so replacing the two devices in the unit won't hurt so much.
Well, replace and report - please. I am interested to see if you see any difference.
If possible, measure before and after performance. Op amps available today are so far beyond what we had back then it isn't funny.
If possible, measure before and after performance. Op amps available today are so far beyond what we had back then it isn't funny.
Many years ago there was the headwize site where Mr Chu Moy presented a portable headphone amplifier which worked well with sensible headphones. It was just a plastic box with an opamp and battery. People liked it and created variants such as two batteries in an Altoid minty tin.
* https://headwizememorial.wordpress.com/2018/03/08/a-pocket-headphone-amplifier/
* https://www.minidisc.org/headbanger.html (nope!)
".... NJM386 should be the same as LM386."
Hmmmm....I looked this up. It appears that the LM386 is also available in -3 and -4 packages, which can deliver more power than the NJM386. Methinks I will try one of THOSE due to the higher power rating and the fact that I trust Texas Instruments' products a hell of a lot more than New Japan Radio.
Hmmmm....I looked this up. It appears that the LM386 is also available in -3 and -4 packages, which can deliver more power than the NJM386. Methinks I will try one of THOSE due to the higher power rating and the fact that I trust Texas Instruments' products a hell of a lot more than New Japan Radio.