Does anyone have any experience or data to prefer ON Semi versus STM BD139/BD140? The application is a Class A headphone amplifier. I assume either are equally suitable.
For the gain grades (BD139, BD139-10, BD139-16) is there a best choice/trade off of current gain linearity versus higher gain? Which is the best choice in a little Class A headphone amplifier output? Since it is Class A just get the highest gain?
I am leaning towards the ON Semiconductor since the STM datasheet doesn't even include the current gain graph but that doesn't necessarily mean it is better or worse.
For the gain grades (BD139, BD139-10, BD139-16) is there a best choice/trade off of current gain linearity versus higher gain? Which is the best choice in a little Class A headphone amplifier output? Since it is Class A just get the highest gain?
I am leaning towards the ON Semiconductor since the STM datasheet doesn't even include the current gain graph but that doesn't necessarily mean it is better or worse.
You don’t need wattage, but you want class A. It is the right size device.
Gain tends to be correlated (inversely) with actual breakdown voltage. You’re not going to run them anywhere near 80 volts for headphones, so there’s no penalty in using the higher gain grade.
I have electronic chokes that use BU508 transistors, or something like that, wildly too powerful. Those are used in CRT TV SMPS circuits.
Asked the maker, he said I had to take harmonics into account...
No issues since 1996, going well. 40 Watt fluorescent lamps.
Using a device well below spec is nice, but at low volumes some amps sound dull, that was the reason for my comment.
Thanks for the clarification.
Asked the maker, he said I had to take harmonics into account...
No issues since 1996, going well. 40 Watt fluorescent lamps.
Using a device well below spec is nice, but at low volumes some amps sound dull, that was the reason for my comment.
Thanks for the clarification.
Are these plug n play compatible with the BD139 or would there need to be circuit changes?
Thanks for the suggestion. I have them and have used them to replace drivers in various AB power amplifiers, but only to replace already high speed drivers.
What HPA have you used them in?
I was actually considering them but I want to be careful to not create an oscillator. I tried an OPA1611A in a Graham Slee Solo HPA clone and that was the end of the OPA1611A. Somehow it killed the OPA1611A in a couple of seconds
I put an NE5534 back in and the Solo HPA clone was still functioning. I don't want to do that again. At least it was the OPA1611A and not my headphones.
A typical HPA uses an op amp driving a CCS-loaded emitter follower, inside the feedback loop. If the op amp is compensated for the closed loop gain, you shouldn’t have any trouble with it using medium speed BD devices (which as as fast as modern outputs like the C5200) or high-speed driver-of-the-week. Doubtful you’d hear a difference either. I’ve used D44/D45, when it’s what I *had*. If a different type (within reason, not an old 2N3055) causes oscillation there are other problems, probably with the layout/implementation. If you have a healthy supply of the Toshibas use them. If you prefer to hang on to your supply of the Toshibas to use FOR power amp output stage drivers, use the BDs. Rationing parts may be the rule rather than the exception these days. If an oscillating op-amp based circuit of any kind causes a dead op amp there are definitely issues that need to be resolved. You’re supposed to be able to *make* oscillators out of op amps.
Thanks.
I was rationing parts but then I bought a batch of defund receivers (for little dollars) that came with nice extruded heatsinks and 7 channels of real Toshiba and Sanken transistors. 2SC2240, 2SC5171, 2SA1930. So now I am looking at a bunch of new projects.
I can even use genuine 2SC5171 and 2SA1930 which I previously carefully guarded.
The Graham Slee Solo headphone amplifier that killed the OPA1611A (and an OPA228) sounds pretty nice with the NE5534 or RC5534. It is a simple circuit and I am so far mystified as to what went wrong. I have attached the schematic however it came with A42/A92 (KEC). I might need to try to use LTSpice to figure it out but I am not sure I want to go that far.
It sounds (quite a bit) better than I expected. I though the A42/A92 was an odd choice. Any thoughts?
I was rationing parts but then I bought a batch of defund receivers (for little dollars) that came with nice extruded heatsinks and 7 channels of real Toshiba and Sanken transistors. 2SC2240, 2SC5171, 2SA1930. So now I am looking at a bunch of new projects.
I can even use genuine 2SC5171 and 2SA1930 which I previously carefully guarded.
The Graham Slee Solo headphone amplifier that killed the OPA1611A (and an OPA228
) sounds pretty nice with the NE5534 or RC5534. It is a simple circuit and I am so far mystified as to what went wrong. I have attached the schematic however it came with A42/A92 (KEC). I might need to try to use LTSpice to figure it out but I am not sure I want to go that far.It sounds (quite a bit) better than I expected. I though the A42/A92 was an odd choice. Any thoughts?
Attachments
Last edited:
Here is another application for the TTA004B.
The diyAudio First Watt M2x
The diyAudio First Watt M2x
Patrick
The diyAudio First Watt M2x
The diyAudio First Watt M2x
Patrick
The Graham Slee Solo headphone amplifier that killed the OPA1611A (and an OPA228
It sounds (quite a bit) better than I expected. I though the A42/A92 was an odd choice. Any thoughts?
That’s a total mystery why the op amp would fail. No possibility of anything getting driven beyond the rail voltage even by accident. Some op amps ARE fragile and require inverse parallel 1N4148’s across the differential inputs, but that’s the extent of what’s needed unless there is a possibility of getting more than 15 volts applied to something from the outside. MOST op amps can handle higher input differential voltage, either by design or they have the diodes inside.
Why A42/92 is also a mystery - high voltage transistors always have lower gain than their normal voltage counterparts. 2pF Cob isn’t needed in a follower, either. Personally, I’d run the outputs in class A (necessitating TO-126 or higher devices) because a couple watts isn’t going to cause global warming. The similar ones I’ve built run a single ended emitter follower with a CCS at 100 or so mA.
Yes, I was a little annoyed when it took out one of my few OPA1611A and a little more annoyed when it took out the expensive OPA228. I still don't understand why that happened. Perhaps the OPA228 is too similar to the OPA1611A and perhaps I should not have grabbed that one next off the bench.
I had not really intended to use that particular HPA (Graham Slee Solo Clone) but I was doing comparative listening tests so I was dusting off old projects just to compare.
I was thinking of modifying the Graham Slee Solo Clone bias. In series with the 1N4148 diodes I was thinking of either adding a Schottky or a small resistor so that I can get a Class A bias. I don't see why a Schottky would not work (adding a little to the 1N4148 voltage). The SAP15P/SAP15N use a combination of PN junction and Schottky diodes for biasing.
I could try the TTC/TTA004. I think the real Graham Slee Solo used BC546/BC556.
I wonder if some of my clone/generic 2SC2383/2SA1013 would be suitable there.
I had not really intended to use that particular HPA (Graham Slee Solo Clone) but I was doing comparative listening tests so I was dusting off old projects just to compare.
I was thinking of modifying the Graham Slee Solo Clone bias. In series with the 1N4148 diodes I was thinking of either adding a Schottky or a small resistor so that I can get a Class A bias. I don't see why a Schottky would not work (adding a little to the 1N4148 voltage). The SAP15P/SAP15N use a combination of PN junction and Schottky diodes for biasing.
I could try the TTC/TTA004. I think the real Graham Slee Solo used BC546/BC556.
I wonder if some of my clone/generic 2SC2383/2SA1013 would be suitable there.
Last edited: