Multiple inserts in series/bypass question.

Say I build a routing box containing three send/return insert points in series with the possibility of bypassing each one.

I then add a hardware processor to each insert (A-B-C), where A goes into B goes into C in series.

If I  bypass insert B,  A will now be connected directly to  C.

Question: Should I worry about any impedance variations that may come with a "true" bypass and consider a buffer on each insert, or am I making up issues?

For some reason, true bypass seems less true to me if I am not only hearing the effect of the unit being bypassed, but the effect of A going directly into C instead of B. 

Might be a made up problem.

Gustav

> Should I worry about any impedance variations

Depends what stuff you have connected.

Most modern stuff is medium/high input and very low output impedance. It is *made* to be patched this to that in many combinations. Nothing to worry about.

The gear will be acting as buffer amps. Since all the gear has a high input impedance (and nothing is in parallel) the load will always be high and the source will always be low.

If you are worried about level variations on the order of less than 0.5dB then it's something to consider.

Thanks!

Ill do it without the buffers (and save some relays).

Gustav

Depending on how much level change you can tolerate you might worry a little.

If you plan to use the bypass to punch effects in and out, and using less than professional (balanced 600 ohm drive) gear, you might get noticeable level shifts when punching in/out.

Years ago I made a more complicated midi controlled insert patch bay (AMR MAP8x4) for effects where 4 different hardware effects could be "inserted" into 8 different channel inserts controlled by midi program changes.  I buffered everything since my design was expected to change effects inserts on the fly seamlessly. 

You way also need to worry about DC clicks/thumps when switching hardware in/out, if inputs or outputs have DC leakage.

JR

I see possible problems if no buffering is used, based on "real world fun" using a patchbay.

Some gear, such as many of the Digidesign D/A outputs, don't play nice into unbalanced destinations; the level drops by 6 dB, and often times a lot of current drives from the "-" output into the ground plane.  Hence, in your example, if a Digi 192HD is the source, and you route that signal source as balanced into an unbalanced input gizmo, problems arise.

In a similar vein, some gear (vintage as well as reissues) has a 600 Ohm input impedance which can annoy a signal source with a "wimpy" output stage.

One can argue that all of your Pultecs, 1176's, LA-2 and LA-3 units should have  some sort of impedance transformation device (typically active) so that no source will ever see that 600 Ohm input impedance.  Likewise, your API EQ's in a lunchbox would have some sort of permanent "balun" on the inputs.


Bri