[sdiy] TR909 VCA circuits

[Trevor Page]

Hi list,

First, a quick update on the 909 clone project : the second board will be
finished this weekend (honest!) and off to the PCB house. PCBs available
'soon'. ish. Honest. :-)

In the meantime, I'm trying to work out why the Snare drum circuit includes
two 'special' transistors, while the whole machine otherwise uses general
purpose transistors throughout. One for the transistor boffins?

Here we go. All drum circuits in the TR909 have one or more VCA circuits,
and this VCA function is normally implemented (pretty crudely) using a
single NPN transistor. Audio in via resistor to the collector, envelope
control current to the base, and emitter is connected directly to the
inverting input of an opamp (= virtual ground).

I'm presuming here that these transistors are being operated in a region
where Ic is no longer anywhere near constant, but is very dependent upon
Vce. This is the region around where Vce is saturated, and the transistor
now looks something like a crude current -controlled resistor. (Equivalent
Ron swings about with varying Ibase). Is this right?

General purpose 2SC2603 NPN devices were used throughout the original
design. Those used in the VCA circuits are marked 'S' on the schematic
suggesting that they have been selected for some characteristic. I presume
that those transistors used to perform the VCA function have therefore been
selected for their value of Vce(sat) (and hfe). Would this sound reasonable?

Now, the snare circuit uses a 2SD1469 in each of two VCA circuits which
shape the noise part of the drum sound. I'm trying to work out why they
chose to use these devices here, but used the gen purpose devices for the
VCAs elsewhere.

The 2SD1469s apparently have a much, much lower Vce(sat) than the general
purpose transistors. So, would I be right, or would I be talking total
nonsense to suggest that these devices can therefore act as much better
current-controlled resistors, since they can operate at much lower values of
Vce where Ic starts to become very (linearly??) dependent upon Vc? In other
words, this characteristic makes the device act like a better VCA? The
datasheet for this transistor (got it from www.partminer.com <<< a VERY good
site for datasheets) actually shows how this part makes a good I-controlled
resistor, and shows typical curves for the equivalent Ron vs. base current.

If so, then maybe these transistors result in a much nicer shaped noise
envelope, which was particuarly important for the snare sound? Substituting
a more general transistor (say BC549, which I've substituted for all the
other transistors in my clone with good results) results in an envelope
shape that is quite different. I find that I cannot alter the gain anywhere
in the circuit to compensate; it doesn't sound right.

If I wish to use a more modern substitute for those two transistors, I'm
guessing that some medium-power transistor of some kind would do the job,
since sa low Vce(sat) / Ron is a desirable characteristic for a power
device. I tried a couple of FMMT489's (1A medium pwr transistor), and with
these the snare sounded pretty much just how it should be. These devices
feature a very low Rce(sat) (175milliohms at 1A). (But are smt).

Would anyone have any comments on this?

I know I could just specify 1469's in the partslist and forget about them,
but I'd like to understand this from an educational point of view - and be
able to recommend a (modern) substitute if I have to.

If anyone would like the datasheets I have for any of these devices, or the
snare schem, email me and I'll gladly help out.

Regards,

Trev




--
trevor at resonance.fsnet.co.uk