# Amplifier output impedience question.

Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.

#### woody

Paid Member
Supose I am using this crossover for 8 ohm drivers crossed over at 500hz at 12db/octive. For the moment lets assume the drivers are a perfect resistive load. Is there a easy way to model how the responce would be affected by using it with an amp with an output impedience of 4 ohms or for that matter 10 ohms.

#### Attachments

• calc_cr_12db.gif
1.8 KB · Views: 218

Last edited:

#### AllenB

Paid Member
You could achieve this with a crossover simulator as well.

#### woody

Paid Member
Well could you recomend a crossover simulator that would do this
without needing driver measurements ?

#### AllenB

Paid Member
IIRC VituixCad and XSim both begin with a driver that has a default 8 ohm load and flat response.

#### Douglas Blake

Supose I am using this crossover for 8 ohm drivers crossed over at 500hz at 12db/octive. For the moment lets assume the drivers are a perfect resistive load. Is there a easy way to model how the responce would be affected by using it with an amp with an output impedience of 4 ohms or for that matter 10 ohms.

Unless you are playing with transformer outputs, most modern amplifiers do not use impedance matching. They use a technique called impedance bridging wherein the actual output impedance of the amplifier is much less than 8 ohms, sometimes in tiny fractions of an ohm.

You can calculate this from the amplifier's specifications. Look for "damping factor"... this is the ratio between the amplifier's output and the load impedance. For example a damping factor of 40 with an 8 ohm load implies an output impedance of 8 / 40 == .2 ohms.

This allows you to design crossovers, etc. without worrying about matching the actual impedance of the amplifier. In fact it makes that concern somewhat impractical.

The spec for 16, 8 or 4 ohms is mostly related to the amplifier's current capabilities... not it's actual output impedance. Basically it is telling you that "at full output this amplifier delivers enough current to drive a speaker load of X ohms"

Last edited:

#### woody

Paid Member
AllenB thanks for the subjestion about Xsim. I am thinking about biamping
and untill I get a crossover I wanted to continue using passive xovers. The small
single ended triode amp I may use above 500 hz does indead have an output
impedience of 4.3 ohms. Pluging this into Xsim it looks at first glance like my
500hz passive xover will be acting more like a 400hz xover. I will probably have
to break down and do something I don't like that is measure it.

#### system7

I can tell you that the solution that gets flat impedance is Butterworth. This is a constant power solution.

Putting 500Hz and 6 ohms into a decent 2nd order Butterworth calculator comes up with 37.5uF and 2.7mH for both arms:

2-Way Crossover Calculator / Designer

This ought to give you a flat impedance. You will take your beating on the bass Fs peak. And you might need to do a bit of Zobel impedance equalisation on driver inductance.

I shamelessly stole this idea from the Rogers LS5/9 monitor. Passive LR bafflestep and otherwise a smaller bass coil.

Where the theory departs is in bass loading, here reflex, and not really suitable for SET valve amps, which have an easier time with closed box.

But in general, Butterworth slopes of whatever order tend to flat impedance. Linkwitz-Riley gets the familiar "omega" impedance peak, best equalised with an LCR as in this closed-box example:

I just don't think you can lose the bass peak, so an amp with significant output impedance has less control of the bass.

Last edited:

#### AllenB

Paid Member
it looks at first glance like my 500hz passive xover will be acting more like a 400hz xover.
Could it be as simple as using a properly sized coupling capacitor at the input to your amp to modify this back to 500Hz?

most modern amplifiers do not use impedance matching.
Many of us design our own amps. Often with a finite output impedance.

#### Attachments

• 20191023_195345.jpg
984 KB · Views: 35

#### woody

Paid Member
Allen, yes it may be that easy but I will have to play with Xsim some more. And like you I build amps and at the moment am undecided whether to build a Zen9 with an output impedience of about 1.3 ohms or build a F2 clone with an output impedience of 15 ohms. And I went the F2 rout I might just use a current source crossover network like Pass wrote about. Dang might just be best to build an active xover ! Oh that crossover came from Nelson's article about current amp. crossovers.

#### Attachments

• passxoverpassxover.png
6.1 KB · Views: 111

#### AllenB

Paid Member
Even active, your left and right impedances will have to match well, or you may as well use speaker level conjugates to fix them instead.

#### bentoronto

Paid Member
Unless you are playing with transformer outputs...

Sounds like OP really is talking about a current-source amp! Wow! Maybe?

If a current amp, I suspect all bets are off at to XO and go straight to active XO which removes all those undesirable pieces between the amp and the driver.

If I were experimenting with a current amp, I guess I'd grab some components lying around my bench and see what drive is getting to the driver (although sound to the room is the real criterion).

Interesting.

B.

#### system7

That's series crossover with voltage drive...

I like to use this mental model for simulation. Just putting a resistor in front of the filter tells you how response changes.

Flat impedance is common enough:

Elsinore speaker: Elsinore Speakers DIY

Arpeggio SET speaker: Arpeggio Loudspeaker - diyAudio

I would think you can get Michael Chua's Starling flat enough with a bit of fiddling and butterworth values:

Seas ER18RNX with 27TDFC

Closed box bass loading gets what you get. Adjust box size.

Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.