As GJ has noted, and typical of valve era practice, the QCII tape output is unbuffered and needs to look into a relatively high impedance if there is to be minimal effect on the signal passing through the control unit. Quad specified a minimum load impedance of 500k for the QCII tape output, with a maximum parallel capacitance (due to the connecting cable) of 200 pF.
In the valve era, provision of high-impedance “line” inputs on tape recorders/tape decks was not at all problematical, and values such as 1M were commonplace. So the probability of mismatch when control units were connected to tape recorders was quite low. That changed with the advent of solid state equipment, where very high signal input impedances were difficult to achieve, particularly with early bipolar devices, and values in the range 33 to 100k were more common.
Some early solid state tape recorders did offer very high impedance inputs by various means, in part one imagines to provide backward compatibility with valve control unit tape recording outputs. The Revox A77 had a 1M auxiliary input, using a bootstrapped bipolar circuit, the Ferrograph Series 7 had a 2M2 line input using a jfet source follower, and I think that some Uher machines had a 1M input obtained by using an input divider with a 1M input arm and essentially tapping the signal down to somewhere around microphone or DIN level. But this kind of facility disappeared after a few years, leaving a matching problem when older valve control units were connected to recent solid-state tape recorders.
One solution would to interpose a simple buffer amplifier between the QCII tape output and the tape recorder input. An opamp with a very high input impedance, such as one from the TL072 family, would seem to be a logical choice. But I imagine that an NE5534 with a bootstrapped input could also be used.
A passive solution might be possible if the tape recorder has a suitable low level input. By way of a worked example, let’s say that a DIN input of 33k impedance and 3 mV sensitivity is available. Then use a 470k over 220k divider at the QCII tape output. The 220k lower leg is in parallel with the 33k Din input impedance, giving an effective 29k for the lower leg. Thus the QCII tape output sees (470 + 29) = 499k, which is just fine, and the DIN input sees a signal level that is 29/499 of that that provided by the QCII, namely 5.8 mV for the nominal 100 mV QCII output, which should be fine for a nominally 3 mV input. Different divider solutions can be developed for different input impedances and sensitivities.
Quote:
Originally Posted by Ted Kendall
If you do have a tape machine on the other end of the cable, beware - I had a similar thing years ago when I plugged a Revox A77 into a Radford SC22. The impedance match wasn't marvellous to start with, but if I switched the Revox off, the load the inert machine presented did funny things to the sound.
|
Interesting! Here’s a thought as to why that might be. Looking at the Revox A77 input amplifier circuit, an incoming “auxiliary” signal sees the 1M input resistor R401. This appears to be shunted by the 22k R404, but the “other” end of this is also at very close to signal voltage by virtue of the bootstrap capacitor C402. Thus to the signal, it looks like a very high impedance that has negligible shunting effect on the 1M input resistor. But that is true only as long as the input amplifier is powered. With the power off, there is no bootstrapping, so the 22k resistor, via R402 (180k) and R403 (100k) in parallel, provides relatively low impedance (86k) shunting of the 1M, so that the incoming signal would see just 79k. The deleterious effect on a valve control unit tape output that expected to see several hundred kohm would likely be very noticeable. And with an unbuffered tape output, those effects would also be apparent on the "through" signal.
Cheers,