Those are available. Fairly low current device. Did you look at the KSA1220 datasheet? 160v 1.2a 175mhz
What do you mean by low current? 100mA?
Not enough for VAS?
1381/3503 are the best for the purpose - high fT, low reverse capacitance, high linearity, high voltage.
Not enough for VAS?
1381/3503 are the best for the purpose - high fT, low reverse capacitance, high linearity, high voltage.
So what? Calculate the input impedance of the OPS or simulate the circuit - if you see some current, close to 100mA, at the VAS stage output - this design does not make sense as distortion level will be enormously high.
Normal VAS current is 5-10mA and the more constant it is - the higher quality the amplifier is.
Normal VAS current is 5-10mA and the more constant it is - the higher quality the amplifier is.
Member
Joined 2009
Paid Member
Nothing wrong with that idea. I did exactly what you suggest: http://www.diyaudio.com/forums/solid-state/196973-tgm5-all-bjt-simple-symmetric-amplifier.html
Post #169 shows a TO-92 device face down onto the power device case. The power device is bolted to the heatsink, the pcb is bolted down over the top of the power devices. There's a hole in the pcb to expose the surface of the power device package. The TO-92 device is soldered onto the top of the pcb and bent over so that it can be placed in contact with the power device package.
I used a 2N3904, you can use almost any 'regular' transistor in this application.
Last edited:
I have a few other 2n5551s in the circuit so they're on hand.
I'm still having a bias issues. Took a propane torch to the heatsink. It's stable anyways. Just higher on one channel then the other.
I'm still having a bias issues. Took a propane torch to the heatsink. It's stable anyways. Just higher on one channel then the other.
Yes, no problem with using TO-92 device, like 2n5551, it's just not so convenient to attach it to the heatsink. I often use MJE340 (TO-126) because I have experimented with them a lot, so I know some good configurations, suitable for different OPS arrangements, plus it's easy to screw to the heatsink, plus I've got a ton of them in stock 🙂
It's rather difficult to say why you've got noticeable difference between the channels.
But you can start from measuring the VAS current in both thannels and Vce of the bias spreader. Are they close to each other, or somewhat different?
It's rather difficult to say why you've got noticeable difference between the channels.
But you can start from measuring the VAS current in both thannels and Vce of the bias spreader. Are they close to each other, or somewhat different?
All the devices are clamped to the heatsink using a single bar of aluminum that lays over the outputs. In the middle are the VBE and drivers. A stiff foam pad presses them to the heatsink. A TO92 would work here, mechanically.
Like some power IC's, eg. TDA7293, TDA1514A, various japanese types and even the very old TBA800/810. The TBA810 have the output driver and various internal CCS all bootstrapped to output, in despair to better its raw efficiency. Then the driver fully saturates and the N-P charge storage begins to show...
Personally I prefer to lose some voltage in order to not have a deep saturation (the VAS saturates but at least is a small transistor), unless the output is a MOSFET. That said, one of better sounding transistor amplifiers I made is the bootstrapped VAS...
Last edited:
Since I already had boards made, the 2n5551 won't work in place of the BD139 due to pinout. I need a TO92 with BCE pinout.
A plenty Japanese TO-92 transistors have a ECB/BCE pin-out (ie. with collector at center).
2SC1815 pdf, 2SC1815 description, 2SC1815 datasheets, 2SC1815 view ::: ALLDATASHEET :::
One example, being ECB.
Almost all 2SA/SB/SC/SD japanese TO-92 (no-RF ones) are ECB. So is true, BCE is very rare (hmmm. now I understand that you need a true BCE pin-out, not a reversed ECB).
Of course, for BCE use is only a matter to invert position, if don't have mechanical complaints.
I believe some 2Nxxxx being BCE, if memory serves now...
One example, being ECB.
Almost all 2SA/SB/SC/SD japanese TO-92 (no-RF ones) are ECB. So is true, BCE is very rare (hmmm. now I understand that you need a true BCE pin-out, not a reversed ECB).
Of course, for BCE use is only a matter to invert position, if don't have mechanical complaints.
I believe some 2Nxxxx being BCE, if memory serves now...
Last edited:
It's only to mount a TO92 where the BD139 is. To have the TO92 flat side against the heatsink, it would need to be BCE because a BD139 is ECB.
I suppose I can use the round side and just put a thick thermal pad behind it.
Some time ago I assembled one germanium amplifier, and used a transistor for Vbe multiplier, and not a thermistor like original design (and make some other modifications). The Ge transistor are round-shaped, and sometimes have a internal tin-wisker "leakage" to case. I insulated the transistor body with a small plastic spaghetti, and put it against the heatsink with heavy use of copper thermal grease (one cooper/brown-coloured), pressed against the heatsink with a single stranded Cu wire 😱 (I made a couple of holes for that). One can use a cable tie.🙂
Even with the notorious know thermal dependence from Ge transistors, the quiescent current is rock stable.😎
So, I imagine a good result from a silicon transistor with grease and some gentle pressure against heatsink.
OOPS:
So, I imagine a good result from a **round** silicon transistor with grease and some gentle pressure against heatsink.
- Status
- Not open for further replies.