Has anyone seen this front-end before?

[homemodder]

Ill need some software to convert pdf pages to image files first so Ill need a day or two as Ive been planning to purchase said software for a while anyhow.

[HarryDymond]

Originally Posted by homemodder

Ill need some software to convert pdf pages to image files first so Ill need a day or two as Ive been planning to purchase said software for a while anyhow.

No need to purchase anything! The free version of PDF-XChange viewer will do this for free (assuming you're using Windows; in OS X the bundled Preview can also do this).

[homemodder]

Here are the schematics. I was wrong about the manufacturing date though, got some more info via Pioneer, the first amps sold using this outputstage was in 1982. Note their are six pairs of LAPT output transistors driven by this stage which are not shown.

[HarryDymond]

Thanks homemodder, much appreciated.

I note earlier you said:

Originally Posted by homemodder

The driver is composed of two transistors in parralell to achieve higher bandwith and lower device capacitances at optimal bias point.

Presumably you are talking Q25+Q27? Device capacitances depend upon voltage across the relevant junction rather than current flow, here the relevant junction is c-b. So, placing two identical devices in parallel will give you slightly under double the capacitance, not lower capacitance. (current per device will be halved; Vbe will therefore be lower for a given standing current, very slightly increasing the c-b voltage drop and reducing an individual transistor's Ccb; however, now there's two transistors in parallel and capacitors in parallel add.)

[homemodder]

Maybe I shoud have explained, two devices used in parralell instead of one device for the same SOA. The advantage is that two lower Ic capable devices have much lower capacitances even when used in parallel than a higher Ic capable part and then their is also the advantage of FT. Compare the device capacitances of say of two 2sa1837 in parrallel to 1 onsemi mje1503xx series driver. When using transistors as emitter followers I find the Cbe of much more importance than Ccb.

[HarryDymond]

Originally Posted by homemodder

Maybe I shoud have explained, two devices used in parralell instead of one device for the same SOA.

Ah, ok.

Originally Posted by homemodder

The advantage is that two lower Ic capable devices have much lower capacitances even when used in parallel than a higher Ic capable part and then their is also the advantage of FT.

Sadly many high-voltage, high-Ft, ultra-low capacitance BJTs have been discontinued by their respective manufacturers (mostly Sanyo (now ON) and Fairchild) As far as I can tell this is because their main market was actually CRT televisions.

Originally Posted by homemodder

Compare the device capacitances of say of two 2sa1837 in parrallel to 1 onsemi mje1503xx series driver.

I don't like ON's drivers. I don't understand why, when they can make the excellent 3281/1302 etc. output transistors, they don't also make something more like the 2SC5171/2SA1930 from Toshiba as drivers. Alarmingly, these Toshiba parts have become NRND (Not Recommended For New Designs), which in turn will become "obsolete" soon. Apparently they are working on replacements though, so fingers crossed! (the 2SA1837/2SC4793 are also NRND).

Originally Posted by homemodder

When using transistors as emitter followers I find the Cbe of much more importance than Ccb.

Really? When operating in common-collector, a BJT's Cbe is effectively bootstrapped due to Vbe being more-or-less constant. Cbe shouldn't have much of an effect.

[Bonsai]

"Sadly many high-voltage, high-Ft, ultra-low capacitance BJTs have been discontinued by their respective manufacturers (mostly Sanyo (now ON) and Fairchild) As far as I can tell this is because their main market was actually CRT televisions."

This will be a real pain for high power design. We will be left itch running TIS stages at very low Iq (so we can use small ignal devices), or MOSFET based solutions.

[HarryDymond]

Originally Posted by Bonsai

This will be a real pain for high power design. We will be left itch running TIS stages at very low Iq (so we can use small ignal devices), or MOSFET based solutions.

I know! MOSFETs would be OK but of course require higher voltage rails due to high threshold voltage and low gm.

If I was insanely rich I'd get a manufacturer to resurrect some of these obsolete BJTs.

[HarryDymond]

Originally Posted by dadod

What do you think about this OPS?

Looks interesting. Q3 and Q4 have low Vce drops so you can use much faster transistors such as the 2SC5707/2SA2040 (ten times the Ft of the MJEs). If you download my zip file of transistor models, I have all the data necessary to build models of these transistors but I haven't had the chance to develop the models yet.

[HarryDymond]

Use current-feedback op-amps as high-input-impedance "transistors". This requires a floating supply for the op-amps which I've shown driven from the amplifier output. I'm concerned this may cause stability issues (at HF the power supply pins of the op-amps become additional signal inputs so there's a feedback path from amplifier output to the op-amp outputs). Solving this would require driving the floating supply from another amplifier, but this could be a very simple amplifier. Non-linear distortion performance of this second amp wouldn't really matter, but it would have to have a flat frequency response, like the main amplifier.

I've drawn an output stage with cascoded output transistors, allowing use of the aforementioned 2SC5707/2SA2040.

Here's the schematic. The references to Q10 and Q11 collectors refer back to Edmond's schematic in the first post of this thread.