Piggybacking - reason

[Lino22]

Guys what other reason apart from lowering the current gain is for piggybacking transistors?

[anotherjim]

Lower noise?

[antonis]

https://www.diystompboxes.com/smfforum/index.php?topic=21801.0

[brett]

Hi.  Yes, decreasing gain was the aim.
Other options were to use power devices (I lie BD139s).
But there's a couple of things going on besides hFE (measured at X uA DC).
I think that RG Keen mentions in the technology of the FF that hFE varies enormously, at least in Q1 because it runs at very low current (a 33k resistor on the collector ensures that) and the input voltage is large (often around 1V) wrt the base-emitter voltage (0.25V).

Power Si BJTs have lots of input capacitance, like old Ge BJTs, which some of us thought a good characteristic (reducing gain at RF).  But after many builds I realised that the hefty power BJTs (1W to 10W) often test ok for gain at 1mA or or 10uA but are sadly lacking down below 1uA. I have no direct evidence but suspect that TIPs and MJEs don't sound good in FFs because their gain is too low, or too variable, under normal use.

When RG Keen proposed piggybacking and I built the first of them, we were surprised by how big the emitter resistor was (5k ohms to 10k ohms).  We don't usually think about this small signal resistance being important, but (delta V / delta I) can be very large for a fraction of a micramp of conduction.  e.g. 0.1V/0.1uA = 10k ohms. Small differences in conduction between the passive junction and the active BJT junction make a LOT of extra resistance needed for the non-active junction of the piggyback to work in a similar way to the active one (assuming we want to "waste" about half of Ib by having the junctions function in a similar way).

What's the importance of this?  Possibly...that the high-performance silicon BJT doesn't have the disadvantages of the power BJT.  It's not like operating a train (the high power BJT) operating at walking speed (being fed a tiny amount of current). More like a fast walker carrying their brother or sister on their back.  (My analogies are quite bad, but what the heck).

In the end, I think piggybacking works really well.  Particularly if the hFE isn't reduced too much (eg from 250 to 150).  Devices like 2N3904, 2N2222A, PN100 and many more are good.  Really good.
Better than high power BJTs in the Si-but-soft FFs that I've built. 

The third option, that wors quite well in my opinion, is "driver" BJTs of 1 or 2 W of dissipation.  A BD139 is one example (1.25W dissipation, hFE around 150).  It works OK in Q1 and perfect in Q2 of a FF.  In a world of variable Ge transistors, buying something that works first time, every time for $1 is kinda nice.
Cheers

[Knobby]

One of the things I've thought of, though not investigated, is the use of piggy-backing to provide variable gain transistors, either with pots or several switchable settings. This might allow you use modern high gain transistor to give a pedal that's switchable between a low gain germanium simulation, like the Axis Face, a standard silicon fuzz face, and a high gain nasty fuzz face. Don't know how practicable that would be, anyone tried something similar?