ampage Tube Amps / Music Electronics	For current discussions, please visit Music Electronics Forum.

ampage archive

Vintage threads from the first ten years

Re: EH SLapback echo

:

2/10/2001 4:10 AM

Re: EH SLapback echo

xpect about 60-70 "usable" milliseconds of delay for every 1024 stages....max. You CAN get more (up to 100msec or so), but you're talking bandwidth narrower than a $12 telephone or a $4 AM radio. The SAD-1024 and the MN3007/3207/3307 would all be fine for slapback echo, since you only need 50-70msec of delay to get that sound, and you kind of want less bandwidth in the delay than you want in the straight signal anyways.

300msec seems to be sufficient to generate delays long enough to be considered distinct echoes rather than reverb or slapback. Any 4096-stage BBD (like the MN3005), or cascaded combination of BBD's yielding that number of stages, can yield a decent sound in that delay range.

As has been noted, companding can dramatically improve the quality of the sound, both by reducing clock noise and hiss coming from the BBD segment of the circuit, and by serving as a kind of limiter for anything going into the BBD. Companding tends to make the overall circuit more complex however. Another alternative is to just stick a compressor ahead of your delay unit to keep the signal from distorting in the BBD, and use lots of lowpass filtering, both pre- and post- BBD-chip, to keep that damn clock signal outa there. By "lots", I don't mean an unreasonably low rolloff frequency, but more stages of lowpass filtering to provide a steeper cutoff at the same frequency (e.g., 3 or 4-pole instead of 2-pole).

Tom G had a very simple MN3005-based delay line at his site. I looked for it the other night but couldn't find it again. Maybe someone knows where the Xavax site went to.

2/10/2001 4:49 AM
Floydfan	Mark, Would all the poles of filtering that a CE 1 uses before and after the 3005 work for a delay? If so, you could just look at the CE 1 schem for what you'd need. But then again, I've always thought that a CE 1 had a narrow bandwidth (to my ears anyway) but thats what i love about em so much.

2/10/2001 7:55 PM
Maneco	you should build low pass filters before and after the delay line,with a cut off of half the sampling rate,according to the nyquist theoreme.and yes,the ce1 has a narrow bandwith,due to the slow clock needed

2/11/2001 5:05 PM

Agreed.

The filtering serves several functions. One of them is to make any clock signal that leaks in a VERY tiny, and hopefully inaudible, component of the the output. Another function is to attenuate some of the grittiness that occurs when the BBD is pushed to its limits. Those teensy FET's and caps in the BBD don't do their job perfectly, and when you slow the clock down too much and make the signal recirculate too much, the output gets VERY nasty, and not in a good way. Companding helps to offset that somewhat, but the post-BBD filtering helps too. Finally, the filtering helps to create the tone of the effect; in effect, its "personality".

The latter function, however, can be done in many ways. When you consider that changing the rolloff frequency of 3 or 4 pole filtering before and after the BBD requires simultaneously switching a big handful of caps, it is far easier to stick a unity gain stage before the mixer stage, and stick a couple of caps in the feedback loop of that stage that you can easily switch with a 3-position minitoggle or a 6-position rotary switch, or even a dipswitch (ˆ la SansAmp). Alternatively, one could use a "Rat-style" variable lowpass filter to tame the high end to taste.

Page 1 of 1