Gearslutz (
-   Geekslutz Forum (
-   -   IBP Analog Phase Alignment Tool - DIY's ? (

hrn 9th January 2004 08:42 PM

IBP Analog Phase Alignment Tool - DIY's ?
Is there any DIY's doing the same do you think?


Lee Cardan 13th July 2011 09:57 AM

7 year bump
I too would like to know if there's a schematic out there for this, extremely useful device!

even if someone could identify how the delay is achieved I'd be greatly appreciative!
A balanced phase shifter with a simple pot instead of LFO comes to mind... hoping it would be simpler than this though, seriously can't be bothered

Jim Williams 13th July 2011 04:37 PM

About as easy as cake. Make up 2 all pass MFB stages. Connect them in series. There will be 2 resistors tied to ground to set the frequency. Replace those with a dual log pot. 100k works good. Now you can tune the phase with a knob.

If you need example circuits, plenty are found in Don Lancaster's classic Opamp Filter Cookbook. It's basically the same as 2 stages of an old MXR Phase 90 pedal using a tuning pot instead of jfets sweeping up and down the resistance of the all pass filter. In the Phase 90, the phase delayed signal is mixed with the dry signal to obtain the nulls.

TAVD 13th July 2011 05:30 PM

Here's a good place to start:
The technology of Phase Shifters and Flangers

Lee Cardan 14th July 2011 04:18 AM

so, lets base it on:

two stages in series, dual gang 100K linear pot (1 gang per stage).
Sounds easy enough, now, question:

How many degrees shift will each stage produce? And one stage is not enough?
1uF NP decoupling cap on input, output, and between stages?
10K for gain setting resistors, and 100nF for the cap OK?

And, if I want to work with balanced signals.... ?

I'd want the signal to be perfectly clean, transparent

JohnRoberts 14th July 2011 03:27 PM

You will get 180' of phase shift, or a complete polarity inversion, beginning with the signal inverted at DC, and rotating through to non-inverted above the RC tuning frequency. Two in series will give you 360' of phase shift..

Are you trying to make an effect? or something to compensate for say the delay between a mic and direct feed from a bass cabinet? For that application you want to tune the RC pole for a relatively LF, and be aware it will only correct for one spot frequency, not all frequencies because it isn't a real delay.


Lee Cardan 14th July 2011 10:51 PM

Thanks for the response john

Not trying to make an effect, as you correctly stated this is for phase aligning of signals. Main instance I can see it being used is Bass cab / DI, and Keys Speaker / DI. I really like to get an idea of what my tracks combined will sound like when tracking, but can never get this happening the way I'd want to.

Pretty much wanting 4x Little Labs IBP Jnr, without spending 2K

JohnRoberts 15th July 2011 04:28 AM

It's about a $2 worth of circuit, but I'm not sure it's worth that....

It's not really correcting the delay alignment error... It fixes one frequency while probably screwing up several others..

A good cheap trick that makes stuff sound different... but IMO little more.

If you are serious, get an inexpensive delay line and do it right.

Also cheaper than a few $k


Lee Cardan 15th July 2011 07:08 AM

I've got some spare bucket brigades lying around, but they'll heavily colour the sound I imagine.. would be cool if the lynx converters could do this themselves!

JohnRoberts 15th July 2011 05:22 PM

Bucket brigade delay is NOT what I had in mind... yes, they would color the sound and be noisy. Far from improving the result.

While not very sexy, a simple digital delay to compensate for a few mSec of delay should not cost a few $K. It is probably free in some digital mixing systems.


Lee Cardan 15th July 2011 10:40 PM

eh, really wanted to keep away from digital

I'm still keen on giving this all pass filter a go. what's the bandwidth of one? and how can it be maximised? and can stages be put in parallel to cover a wider range? :)

daniel c 16th July 2011 02:31 AM

Hi Lee,


what's the bandwidth of one? and how can it be maximised?
An ideal all pass filter by itself doesn't have a limited bandwidth, it has a flat magnitude (amplitude) response at all frequencies in the audio range and beyond (in an ideal circuit).


can stages be put in parallel to cover a wider range?
Interesting idea. You could bandpass the signal (like a 2-3 way crossover) and put an all pass filter in each band with the "tuning frequency" optimised for each band. (the tuning frequency for a single stage/first order APF is the frequency where there is a 90 degree phase shift when the output is compared to the input) This might get a bit messy though, it's a lot of filtering.

I recently had to program an APF in Matlab and I had to do a lot of research on analogue filters, but I am by no means an expert on them.

My advice is to head down to one of the Uni engineering libraries and look in the 700-800 section. Now that's like 3-4 long rows of shelves but if you search the catalogue for "filter" it will give you a long list of titles. (I think most of them are around 725 and 728). All pass filters don't occupy much space in most of the texts but you can collect a lot of information about them. A lot of the references are for radio frequency APF's, but the principle is still the same.

To be honest, from my experience you have to start with a theoretical understanding how a basic filter works (i.e. poles and zeros and where they sit in relation to each other when you plot them) before you can get into designing an all pass filter.

Have fun! kfhkh

EDIT: Section 16.4 will get you started

JohnRoberts 16th July 2011 03:14 AM

For the record that polarity inversion circuit is not a simple all pass filter. It's more like a polarity inversion circuit.

All pass filters are archaic technology used in the pre-digital days to make a short time delay. It only looked like a constant delay within a limited bandpass. So they were difficult to apply effectively. As I recall one application was to correct fractions of a mSec delay issues between loudspeakers drivers, to time align them, for the driver's different depth behind the loudspeaker baffle plane.

To make several mSec of delay (a lot more than needed in a crossover) you could stack up multiple all_pass sections, but this seems awkward.

Maybe look for old analog loudspeaker crossover schematic, I think Rane may have used adjustable APF in a few models back in the '80s for time alignment.. One more thing that modern digital technology is superior at.

I do not think this is a profitable use of time, but it's your time not mine... so enjoy.


Wavebourn 16th July 2011 10:00 AM


Originally Posted by Lee Cardan (Post 6849172)
and can stages be put in parallel to cover a wider range? :)

If you put in parallel stages with different time constants you will get uneven frequency response.