Can Someone Explain How the AI-1 Transformer (Dreadnought) Works?

Joe08867

drumminman wrote: »

My guess is if the DC (or other electrical grunge) riding with the signal is not being passed by the AI-1 transformer, but the audio signal is, the improvement would be from the amps grounding scheme not having to deal with this. Even though it may be designed for it, removing it before it gets to the amp seems like it would be a better strategy than dealing with it once it reaches the amp.

That is what i was trying to say with my post. You said it much more better.

FTGV

I'm certainly no SDA expert but heres my take on the DN common ground deal.If I'm reading it correctly according to the A1-1 construction plans(post #6 here http://www.polkaudio.com/forums/showthread.php?55888-POLK-AUDIO-Speaker-Wiring-Schematics-amp-More-Stereo-Dimensional-Array(SDA) ) the dimensional negative is linked through the winding of the transformer to the negative binding post.While there will be some resistance and inductance in the transformers winding it will essentially result in an unwanted(with common ground amp) direct connection between the negative of the dimensional section to ground and not isolatiion from the amps ground.Then since the amp is common ground there will be continuity between the grounds of both channels .To me it looks like the end result is that the dimensional drivers would now be operating in parallel with the stereo array?But then again the pain meds might be messing with me.

heiney9

Are the dim drivers operating in parallel normally, or does the AI-1 being used on a common ground amp change them to be operationg in parallel?

H9

FTGV

heiney9 wrote: »

Are the dim drivers operating in parallel normally, or does the AI-1 being used on a common ground amp change them to be operationg in parallel?

H9

The latter,though the dim drivers are working in parallel at low frequencies via the large inductor.Otherwise the dim drivers should only be producing the R-L L-R signal.

drumminman

FTGV wrote: »

The latter,though the dim drivers are working in parallel at low frequencies via the large inductor.Otherwise the dim drivers should only be producing the R-L L-R signal.

It sounds right to me. Check out this post from MP; he's responding to a question from a member about building an AI-1 interface:

http://www.polkaudio.com/forums/showthread.php?74472-The-AI-1-Dreadnought-Project-Pt.1/page3&highlight=sda+signal

Post #85:

Yes, I did mean 5 amps. The transformer will have to handle nearly the entire amplifier output for any signal that is primarily in one channel or the other. The impedance of the sda drivers in series is around 6.5 ohms. So, a 5 amp rating will deliver up to about 160 watts to the sda drivers before saturation. The inductance determines the frequency at which the transformer will shunt low frequencies to ground allowing the sda driver to produce the same low frequency signal as the stereo driver in that cabinet. 14mh puts that point well below 100 Hz. That means that the sda signal will be preserved at full strength down to 100 Hz which is important for best imaging. Higher DC resistance just attenuates the sda signal and the bass. I think the transformer in the SRT system had a DCR of 0.3 ohms, 14mh inductance and saturation current around 7.5 amps.

So the way I'm reading it, the SDA drivers reproduce the stereo signal below around 100 hz, and the sda signal above that.

heiney9

drumminman wrote: »

So the way I'm reading it, the SDA drivers reproduce the stereo signal below around 100 hz, and the sda signal above that.

Not all SDA's are this way. The first few generations had the dim driver only playing dimensional signals. The SDA's that can use the AI-1 play the way described above. Just wanted to add the point to clarify that not all SDA's do the above.

H9

FTGV

drumminman wrote: »

The inductance determines the frequency at which the transformer will shunt low frequencies to ground allowing the sda driver to produce the same low frequency signal as the stereo driver in that cabinet. 14mh puts that point well below 100 Hz.

OK given that important little tid bit of info it appears a transformer with fairly high inductance is recommended. That would mean the transformers high inductance would be placed in parallel with the internal 16 mh coil resulting in a combined reduction in inductance.This would likely modify (raise) the frequency at which the dimensional drivers reproduce stereo signal, but SDA signal should remain and not working in parallel full range as I surmized previously.
Do you know the rated inductance of the transformer your using?

drumminman

FTGV wrote: »

OK given that important little tid bit of info it appears a transformer with fairly high inductance is recommended. That would mean the transformers high inductance would be placed in parallel with the internal 16 mh coil resulting in a combined reduction in inductance.This would likely modify (raise) the frequency at which the dimensional drivers reproduce stereo signal, but SDA signal should remain and not working in parallel full range as I surmized previously.
Do you know the rated inductance of the transformer your using?

I do not. Just looked at Darqueknight's post on building this and couldn't find it I used the same Avel-Lindberg transformer). Found another piece of the puzzle though, which adds to the above post:

From DK's amd Ben's posts #31 & 32, which can be found here:

http://www.polkaudio.com/forums/showthread.php?74472-The-AI-1-Dreadnought-Project-Pt.1/page2&highlight=dreadnought+inductance

I recall that Polk redesigned the SDA circuit and limited it to a specific band of frequencies below the high frequency crossover point. I don't remember what the SDA signal frequency range is. It's in my notes somewhere.

The SDA signal ranges up to 2-2.5khz leaving out the first gen of coarse.

FTGV

Well yes the SDA signal in later models with no dim tweeter are limited to the upper midrange, but thats not related to the transformers inductance.It (transformer inductance)only affects the range of stereo low frequency content produced by the dim drivers.

inspiredsports

thsmith wrote: »

Is it possible people are hearing improvments in SQ with DN because they are using better cables vs what they were using between the speakers previously?

I know the AQ T4 I am using vs OEM P/B cable has made a very positive improvement in SQ.

Just a thought.

Speaking to that, I found the quality of SDA Interconnect cables to make a noticeable difference. Years ago I did a lot of testing of various wire against my stock Polk SDA interconnect. I ended up determing an interconnect made from 12AWG Silver Coated Oxygen-Free Copper Kapton Insulated Wire sounded best in my systems.

An interesting theory that came out of the testing at the time is we know silver is a poorer conductor of low frequencies, but the SDA effect is intentionally not passed at bass frequencies (and a lot of electrical noise may be present in that lower frequency range).

I encounter improved detail and clarity (the "unveiling" effect) mentioned with the use of a Dreadnought when using this wire.

drumminman

FTGV wrote: »

OK given that important little tid bit of info it appears a transformer with fairly high inductance is recommended. That would mean the transformers high inductance would be placed in parallel with the internal 16 mh coil resulting in a combined reduction in inductance.This would likely modify (raise) the frequency at which the dimensional drivers reproduce stereo signal, but SDA signal should remain and not working in parallel full range as I surmized previously.
Do you know the rated inductance of the transformer your using?

Found the measurements for inductance in DK's post: 153.3 mH primary, 168.8 mH secondary

FTGV

drumminman wrote: »

Found the measurements for inductance in DK's post: 153.3 mH primary, 168.8 mH secondary

If thats the case then when paralleled with the 16mh results in roughly 14.5 mh.Therefore little change will occur in the range of stereo signal low frequencies covered by the dimensional drivers.Given that the only difference I can see by using the DN with a common ground amp appears to be the addition of an extra inductor with little net change in response except for possibly a desirable reduction of the DC resistance across that leg.

drumminman

I have about 5 hours on the Dreadnought AI-1, and plan to do some more listening tonight. Can't wait - it's like having a new toy :cool:

FTGV

drumminman wrote: »

Can't wait - it's like having a new toy :cool:

Enjoy.

FTGV

Are there any differences other than the foil?

FTGV

FTGV wrote: »

Are there any differences other than the foil?

Or do you mean the stock Polk AI 1 vs the custom unit?

drumminman

I think he means compared to the original, stock AI-1.

drumminman

Update - Give the Dreadnought some burn in time. I have around 15 hours on mine. I noticed immediate improvement with the first CD's I played after connecting it into the system per my review above.

But last weekend, something did not sound quite right. The bass seemed to have diminished, and the highs sounded a little hazy. Imaging was still good, but things seemed a little subdued somehow. Hard to put into words, but I thought about pulling the DN out of the system to hear if I was imagining the changes.

I decided to leave everything and do a little more burn in. Sat down for a listen last night and W-H-O-A!

Everything snapped into place: cymbals sounded correct again with the slight hissiness/sibilance replaced by a clear metallic sound, transients (the impact of drums, of sticks on cymbals, of various percussion instruments, the plucking of stringed instruments) were a little crisper, bass returned to previous levels and then some, and imaging became even larger than before, both side to side and front to back. Music played: Flim and the BB's "This is A Recording", Blood Sweat and Tears 2nd album, Dave Matthews Band "Crash".

These changes are similar to what I heard with the rebuilt Xovers, but maybe a little more pronounced. The sound quality went from noticeable improvement initially, to "huh, what's off?", to "damn, that sounds good!"

If you do this mod with your common ground amp, give it some play time before making any judgements.

DarqueKnight

FTGV wrote: »

I assume using the IC creates a dead short between the negative outputs of each channel.Since the negative output of non common ground amps is not at ground potential but an inverse of the postive signal, shorting them together would be as undesirable as shorting the two positive terminals together.Using the transformer allows signal transfer while isolating the negative ouputs from each other.

That is correct.

heiney9 wrote: »

So in a sense part of the signal is transformer coupled?

Yes, with the AI-1 interface the dimensional signals are transformer coupled.

heiney9 wrote: »

I still need to understand why this results in an improvement in sound for a common ground amp. Is it likely because of a better grounding scheme with the isolation transformer? It seems more cables and a transformer in the signal path isn't what would be desired. I wonder if the improvement is amp dependant? Again talking using the isolation transformer with a common ground amp.

If I strap either my non-common ground Adcom GFA-5500 or my Parasound JC 1 monoblock amps, and then install the AI-1 (Dreadnought) along with the strap, it sounds better than with the strap alone.

Refer to the attached SDA circuit diagram.

In the top circuit, with the normal SDA interconnect, a common ground amplifier or strapped non-common ground amplifier provides a return path for each speaker's SDA signal througn the amplifier's ground plane. There exists a physical circuit connecting the speakers to each other through the amplifier ground plane and through the SDA cable. In addition to the speakers transmitting inverted spatial information with each other through the SDA cable, they are also able to "hear" the noise from both channel's ground paths.

In the bottom circuit, with an AI-1 isolation transformer, a common ground amplifier or strapped non-common ground amplifier still provides a return path for the SDA signal througn the amplifier's ground plane, but only for each speaker seperately. The difference in this case is that there is no physical electrical connection between the SDA circuits of the two speakers, even though the upper lead of each side of the transformer is connected to the same ground path. Each speaker's SDA circuit cannot "hear" the other channel's ground path noise because a complete circuit to the opposite channel's ground path noise does not exist. Inverted spatial signals still flow between speakers through the magic of inductive coupling, but with lower noise (by half) and higher sound quality.

drumminman wrote: »

Found the measurements for inductance in DK's post: 153.3 mH primary, 168.8 mH secondary

Those were the inductance values provided by Avel Lindberg, but they do not agree with what I measured. Apparently, whoever took the A-L measurements forgot to short the other side of the transformer. My measurements are posted here: The AI 1 Dreadnought Project Pt.2 Upgrade to 1000VA Transformer

Below: Low frequency and SDA circuits of an SDA SRS 1.2TL loudspeaker.

FTGV

DarqueKnight wrote: »

In the bottom circuit, with an AI-1 isolation transformer, a common ground amplifier or strapped non-common ground amplifier still provides a return path for the SDA signal througn the amplifier's ground plane, but only for each speaker seperately. The difference in this case is that there is no physical electrical connection between the SDA circuits of the two speakers, even though the upper lead of each side of the transformer is connected to the same ground path. Each speaker's SDA circuit cannot "hear" the other channel's ground path noise because a complete circuit to the opposite channel's ground path noise does not exist.

Hummm In the bottom illustration it appears to me like there will be a ground path from the DIM to the STR of each speaker through the paralelled combination of transformer winding and 16mh inductor.Since the grounds of both channels of a common ground amp are internally tied together(common) the grounds will not be isolated from each other.

DarqueKnight

I recently asked Ken Swauger if he had the original AI-1 specs. He did not. He referred my inquiry to Stu Lumsden (Polk's engineering VP) and Matthew Polk. Matthew Polk responded with specification information and a general overview of how the SDA circuit operates.

"The original isolation transformer was successful at eliminating the issues of SDA with non-common ground amps but at the cost of a significant compromise of the SDA sound stage. We later figured out how to do this right with the transformer incorporated into the SRT controller.

The transformer is a 1:1 ratio and allows the L-R and R-L SDA signals to be derived while isolating the grounds of the two amp channels. The inverted portions of the SDA signals only need a bandwidth from a little below 100Hz to 1.6kHz. It's helpful if the transformer has more bandwidth on the top end since the high-end cutoff is controlled by a low-pass in the cross-over. At the low end, however, it's best if the transformer windings have a very high inductance as measured with the secondary open, around 14mh, and a very low DCR.

In the mid-range where SDA operates, the SDA drivers on one side operate in series with the SDA drivers on the other side with the signals passing through one set of SDA drivers then through the transformer and then through the set of SDA drivers on the other side. However, at low frequencies the positive part of the signal passes through the SDA drivers and is shunted directly to ground by an inductor in parallel with the transformer. The result is that the SDA drivers operate in parallel with the stereo drivers at low frequencies but operate in series with the other side to produce the SDA difference signal in the mid-range. If the DCR is too high, as it was in the original isolation transformer, it will reduce the level of the SDA signal and alter the bandwidth. If the inductance isn't high enough, also the case in the original, it will raise the low-end cutoff of the SDA signal further reducing the SDA soundstage and allow the SDA drivers to produce too much upper bass thereby muddying the sound.

The transformer in the SRT controller had a DCR around .3 ohms and an inductance around 14mh as best I recall. Current carrying capability before saturation is also an issue and was not adequate with the original. I think it was only about 3 amps which would be about 60 watts to the SDA drivers and not enough. I don't recall exactly what it was in the SRT but seem to recall something like 10 amps which would deliver about 600 watts to the SDA drivers.

As a practical matter we found the only way to get a low enough DCR was with a 1:1 toroidal transformer. These are available off the shelf but expensive. Because of the efficiency of the toroid the cores must be much larger to avoid saturation.

Hope that helps.

-msp

DarqueKnight

FTGV wrote: »

Hummm In the bottom illustration it appears to me like there will be a ground path from the DIM to the STR of each speaker through the paralelled combination of transformer winding and 16mh inductor.Since the grounds of both channels of a common ground amp are internally tied together(common) the grounds will not be isolated from each other.

In the case of a common ground amp, what we want to isolate is the SDA signal.

FTGV

DarqueKnight wrote: »

. However, at low frequencies the positive part of the signal passes through the SDA drivers and is shunted directly to ground by an inductor in parallel with the transformer. The result is that the SDA drivers operate in parallel with the stereo drivers at low frequencies but operate in series with the other side to produce the SDA difference signal in the mid-range. If the DCR is too high, as it was in the original isolation transformer, it will reduce the level of the SDA signal and alter the bandwidth. If the inductance isn't high enough, also the case in the original, it will raise the low-end cutoff of the SDA signal further reducing the SDA soundstage and allow the SDA drivers to produce too much upper bass thereby muddying the sound.

OK this confirms the points I made in the 1000VA transformer thread regarding parallel inductunce changes modifying the dimensional drivers bandwidth.My uneducated guess is this and the reduction in DC resistance is the main reason for the sonic differences people are hearing when the transformers are used with common ground amplifiers.

FTGV

DarqueKnight wrote: »

In the case of a common ground amp, what we want to isolate is the SDA signal.

Isolate from what?Since only non common ground amplifiers require isolation of their negative terminals.

DarqueKnight

FTGV wrote: »

Isolate from what?Since only non common ground amplifiers require isolation of their negative terminals.

I am going to take one last crack at this:

DarqueKnight wrote: »

In the bottom circuit, with an AI-1 isolation transformer, a common ground amplifier or strapped non-common ground amplifier still provides a return path for the SDA signal througn the amplifier's ground plane, but only for each speaker seperately. The difference in this case is that there is no physical electrical connection between the SDA circuits of the two speakers, even though the upper lead of each side of the transformer is connected to the same ground path. Each speaker's SDA circuit cannot "hear" the other channel's ground path noise because a complete circuit to the opposite channel's ground path noise does not exist. Inverted spatial signals still flow between speakers through the magic of inductive coupling, but with lower noise (by half) and higher sound quality.

FTGV

DarqueKnight wrote: »

I am going to take one last crack at this:

I guess then we will have to agree to disagree.but since I'm the inquisitive type might I ask if this "ground noise" is something identified by Polk as a sonic shortcoming when using just the interconnect and did they officially endorse the use of the transformer with common ground amplifiers?

FTGV

Ok I will ask again because I believe it's a fair question given the recent interest in using transformers with common ground amplifiers and may be of interest to those considering building one.It was implied that the biggest improvement wrought by using this configuration was the reduction in "ground path noise".Is there any references from Polk indicating this ground noise was an inherent compromise when using the simple SDA interconnect?Are they on record endorsing the use of a transformer with common ground amplifier for that or any other reason?