The making of: The Two Towers (a 25 driver Full Range line array)

I think my list would be quite different, although there are a few on there that could be fun. :)
Browsing that list on a mobile phone immediately showed advertisements on books about Lucid Dreaming...
That's taking it a little too far for me :D.
I just thought it was a fun/fitting title for my setup. The stuff of dreams (lol). ;)
(Don't ever take me too serious, I know I don't)
 
Last edited:
This may turn into a lengthy post so be warned in advance ;).

On fluid's thread I have mentioned that I could never get the Goldmund amplifier to do some things the Pioneer did so well. I have worked hard on my processing, figuring that as the Goldmund actually sounded more detailed the fault had to be coming from somewhere else.

After our joint listening session, where koldby brought his amps to try and BYRRT was there with a couple of his favorite toys as well, the dust settled and the Goldmund stayed for a while longer.
After rearranging my balance to get the Goldmund to work, I noticed some balance changes. I've measured the Goldmund and compared it to what I had before and could not quite understand these differences. I moved on and started fresh armed with prior experience.

I noticed things to improve after re-balancing the setup but some things remained different.

With the right settings and my previous amplifier, the Pioneer 757 MKII (assisted by the Pioneer A400 series ambience amp), I could shape the imaging/staging at will. This has always been a delicate process of balancing everything just right. I could literally move the stage around, position every individual sound in the stage. (not literary, mind you) What I mean is I could make it as wide as I could, but also steer the side panned sounds to imaging in front of the speaker position if I felt they should be there (based on prior experience with these songs).
This was also true with center panned sounds. Some mixes have an up front vocal part that is clearly meant to be there. Other songs might place the singer further back. I'm sure all of you with an interest in imaging have noticed things like this.
With mid/side EQ and ambience on, I could place the singer in front of the speaker plane if it was recorded as close up and personal. Floating positions where there is black between the notes as well as the separate instruments/vocalists (as we so often read in the prose of audiophiles).
The mid/side EQ played a large part in that, where on some songs, where the actual words of the vocalist were never really that clear became easy to follow. That was a real strange feeling. I could get close to that with cross talk schemes and phase shufflers too, but they each had their own drawbacks.
Fast forward to the Goldmund amplifier. I could still shape the sides, much like before. But it was way harder to get that central vocalist to move closer. I could not get the "why"? More work for me I assumed and tried a few different ways to process. Having a ball, mind you, it wasn't a punishment to listen. But I simply could not get it to do what the Pioneer exhibited. That close up front singer presented in an eerie scary real way.

I improved the balance over time, having a steady enjoyable sound, but always failing on that last part. I did not think much about it for a while, just figured it was up to the last details and I would get it. I did not get that with the Pioneer in a day either.

One thing that stayed with me in my thoughts though. In our joint listening session, two amps were excellent in getting that up close and personal feel. One that makes you believe you can look straight down the singer's mouth and see the vocal cords move :D. Those were the Goldmund and the Fetzilla.

So how could I not get that back having that same amp that impressed me so much in there. One thing that had been in the back of my mind was that to get the levels to match how I had setup the Pioneer, we used BYRTT's HP-1 headphone amplifier as a pré-amp. Could that have been the real clue? And why would that be true, what reason would there be?

After talking it through with BYRRT a couple of times (we had talked about it before when I noticed that balance difference, after replacing the Pioneer with the Goldmund) I figured it might be worth a simple experiment. In those first talks, we (that's including koldby, BYRRT and myself) came to one conclusion, the DAC I have should be able to get the amplifier up to full swing, supplying it with enough power to do every thing it should.
Quite a while ago, in one of our email exchanges BYRTT mentioned an interesting headphone amp. A relative cheap amplifier that showed promising results. When I recently restarted this conversation about pré-amps with BYRTT he mentioned he had already brought forth a viable option to try.

The headphone amplifier BYRTT mentioned was the JDS Atom. At about $ 99,00 (plus tax and transport) it could be easy and affordable enough to try.
If it works, other, maybe even better solutions would be available.

So I bought that JDS Atom amp. A few weeks ago I put it into the chain to see what it could do. It has a double function of being a headphone amp as well as a pré amp (having both RCA in and outputs) with 4.5x gain. I reduced the level of the main left and right speakers within JRiver and used the gain of the Atom after the DAC to power the long lines toward the Goldmund. This difference wasn't subtle :eek:. I set the levels with my SPL meter to match what I had before. Its hard to describe this difference though. I can feel it in my whole body. I was so relaxed when listening, no strain etc. While the level I was listening to felt very comfortable. (basically always the same level, as guarded by the JRiver Volume Leveling.

In another session I decided to give my Home Theatre a try to. So following the same steps I set left + right 9 dB down only to find the center slightly louder and less panning of voices. Then it struck me. I had to reduce the center channel as much as left and right :D. Yeah, now everything was panning along with the visuals. I had it at the same levels as I was always using. Checking within JRiver I noticed no drop in level from Volume Leveling :eek: So I was actually listening much louder than usual, and not by a small amount. The SPL meter confirmed that. While it did feel a bit different, I did not notice it half as much as when the center that was off before. Everything was playing clear and clean.

That actually reminded me of my early days with the arrays. Their ability to turn them up without showing any strain. Not experiencing the true listening levels until everything around you starts rattling and moving along to the music. Which is why I always kept the SPL meter close and used the volume leveling to keep me from enduring damaging loud sessions.

I'll click the submit reply now... but I'm not done yet :)
 
I'm just unloading my thoughts here, hang on :).

Trying to come up with valid reasons behind this story is still a bit shaky. This weekend I've started JRiver in a clean environment and ran some loop back tests on the sound card. It is super clean, whether I test it in Asio or Wasapi it simply does what it's supposed to do. The specs say it can put out 2 volt RMS which should be adequate for loud levels after amplification. These loop backs were run without a pré amp in the circuit.

So nothing wrong there. Nothing wrong with the Goldmund amp either. Up comes one suspect, all the processing I do, with boost as well as cut eats up some of the internal bits, that's for sure. The Pioneer had a pré circuit but also had a very sensitive input. I've clipped the input at times and that has never been pretty. The Goldmund clone only has the amplification stages, the real deal does feature a pré-amp circuit (and even a DAC). So my best guess is that I'm eating up a lot of headroom, even a bit more than my REW graphs show me. That's why reduced levels from the JRiver output and making up for that with amplifier gain gives me more headroom and cleaner sound. That's the first intermediate conclusion I've come to. Despite having much lower boost on the lines after adding the subwoofers, I apparently am still overloading, or straining the internal bitspace available.

So where do I go from here on forward? Fanatic forum browsers (like myself ;)) may have noticed me asking stupid questions on Tom's Universal Buffer thread. I will build up a pré-amp, most probably with a fixed gain and 6 (or 8 to be future proof) channels. Gain has yet to be determined, right now I expect a number around 9 dB.

I've also let loose the notion that I will be getting a new (to me) amplifier for ambience duty. Originally I sourced an old Pioneer A447 (A400 series) to match the 757 MKII I had. This time I want something to compliment the Goldmund clone. I chose one who's result I trust and know, I hope koldby is OK with me mentioning this, I've pried the Fetzilla from his firm grip (lol).
Honestly, I appreciate the assist I get from you guys, koldby was willing to build up the Fetzilla in an enclosure of my choosing, BYRRT enduring many of my brain farts and of course all readers and participants both on this thread and beyond.

Can't state I'm right in my choices, opinion or thoughts but I do feel this is a new way for me moving forward to my ultimate goal. Creating headroom was the thought behind the subwoofers, so I see this as a secondary path to create that headroom. Who knows what's in store next....

Another chapter is starting in the never ending story :D
Oh well, as long as it is all fun, why not?
 
In light with my earlier question about the text on the amp-logo, here's what I had in mind:
attachment.php


Now I know and realize taste is a very personal thing. But this is what I think would look nice in combination with the arrays and subwoofers. I chose not to add handlebars on the ambience amp, as I think it would be a bit much. I like to keep the form follows function philosophy but do not mind a touch of bling (the logo).

Would you understand that I can muck up sizes and shapes like this, however simpe for quite a while to find the balance in it? The second logo is ever so slightly smaller than the one on the bottom amp. The power led is placed slightly different for me to feel it is in balance.

I guess I'm funny like that. :D Caring about little details that most will never notice or care for. Unimportant for their function, yet it has to fit my overall view. Call me crazy and you'd probably be right!

In all other aspects, even on the backside this amp is build up like a little brother to it's bigger companion. Heck, it's almost standard and yet I find great satisfaction in worrying about these little details :).

The million dollar look I talked about earlier, not to be taken seriously, but to me? I am building up something I feel good about, very good in fact and that's what counts ;).

Can anyone relate? Or is it just me, that's this delusional....
 

Attachments

  • Amplifier-assy.jpg
    Amplifier-assy.jpg
    103.4 KB · Views: 258
Sounds looks and performance improve bit for bit as you say never ending story, love it and makes me happy and where is that like botton .. oh there ...:up:

btw how did JDS Labs amp sound with headphones it should be up in performance gear and transperance close to that HP-1 beast even JDS is so low in cost.
 
The specs say it can put out 2 volt RMS which should be adequate for loud levels after amplification.
The 2V will be at 0dB, are you running your output that hot? My peak levels are more like -5dB on the worst songs to give me some leeway.

So nothing wrong there. Nothing wrong with the Goldmund amp either. Up comes one suspect, all the processing I do, with boost as well as cut eats up some of the internal bits, that's for sure.
With a 64 bit internal path there is effectively unlimited headroom with any practical audio signal and as long as the end result doesn't clip overall it should present no issue digitally. I have noticed myself that the overall filter gain used does make quite a difference in that it soaks up more amp power to reach the same SPL. Reducing some of the boosts that weren't necessary made me have to turn the volume down a fair bit :)

That's why reduced levels from the JRiver output and making up for that with amplifier gain gives me more headroom and cleaner sound. That's the first intermediate conclusion I've come to. Despite having much lower boost on the lines after adding the subwoofers, I apparently am still overloading, or straining the internal bitspace available.
This is the bit that doesn't make sense to me as reducing the volume coming out of Jriver will reduce resolution in the DAC, as there are less bits available to represent the signal. This should not be an improvement although unless you lose more than 8 bits due to volume loss it shouldn't matter that much either.

Most amps have around 26dB of voltage gain internally to allow them to reach maximum output with 1V or 2V input. Maybe by feeding the Goldmund with a higher input voltage and requiring it to amplify it less you have found it's sweet spot :)
 
Most amps have around 26dB of voltage gain internally to allow them to reach maximum output with 1V or 2V input. Maybe by feeding the Goldmund with a higher input voltage and requiring it to amplify it less you have found it's sweet spot :)
This is not the case. The Goldmund has 26dB gain and it is fixed, so you have to feed it with the exact same voltage to get the same SPL, no matter if you use a preamp or not.
As I see it, there are a couple of explanations. Either the sound card has a hard time driving the cables and the preamp does that better. Or the soundcard is stressed by sending out a higher level and performs better when delivering a lower output (maybe again driving the cables is the issue) to a preamp. Or it has to do with the digital levels. As the sound can be changed by just changing the levels between the DAC and the preamp/amp, the Goldmund is out of the equation.
 
Indeed drive level remains the same. The amount of gain added by the pré-amp is exactly the number that it is dropped internally in JRiver.
Is the 26dB gain the gain of the end stage? As the original schematic did feature a pré-amp stage. As does the original as far as I can tell. Goldmund took down their original site and lots of info is missing. I can't seem to find much anymore. I do have the manual but that really doesn't have anything useful in it. :(

Browsing through the internet archives I do find the Telos 600:
Power
Nominal continuous power :
350 W RMS (2 - 8 Ohms).
175 W RMS (1 - 16 Ohms).
Maximum instantaneous power : 600 W RMS (8 Ohms).
Maximum voltage swing : 85 V peak.
Maximum current swing : 35 A peak.
Frequency Response
These figures are valid for the amplifier circuit alone,
loaded by 8 Ohms and working at any level between 0
and nominal power.
+/- 0.1 dB, 0 - 300 kHz, +/- 1 dB, 0 - 1000 kHz.
+/- 3 dB, 0 - 3 MHz.
Input Sensitivity
Nominal level : 1.45 V RMS.(Adjstable 0->-10dB).
Voltage Nominal Gain : 28.2 (29dB).
Input impedance : 10 kOhms.
Group Delay
Propagation delay < 100 ns stable with frequency from
DC to 200 kHz.
Distortion
Figures valid for all levels from 0 to 40 V / 8 Ohms
Dynamic : TID < 0.01 % (- 80 dB) unweighted.
Static : THD < 0.01 % (- 80 dB) unweighted.
Circuit Speed
Slew rate : > 400 V/us.
Rise time : < 300 ns.
Noise
Signal-to noise ratio : > 115 dB.
Weighted ASA A : > 130 dBA.
Operating Temperature
Room temperature : -30 to +40 degrees Celsius
(-22 to +104 degrees Fahrenheit).
Internal temperature : +45 to +65 degrees Celsius
(+113 to +149 degrees Fahrenheit)
 
Last edited:
This is not the case. The Goldmund has 26dB gain and it is fixed, so you have to feed it with the exact same voltage to get the same SPL, no matter if you use a preamp or not.
As I see it, there are a couple of explanations. Either the sound card has a hard time driving the cables and the preamp does that better. Or the soundcard is stressed by sending out a higher level and performs better when delivering a lower output (maybe again driving the cables is the issue) to a preamp. Or it has to do with the digital levels. As the sound can be changed by just changing the levels between the DAC and the preamp/amp, the Goldmund is out of the equation.
Yes the second part of that statement is not accurate as I had in my mind that extra input voltage was being provided to the amp, which makes no sense as I had thought that the digital volume was turned down, anyway... :eek:
 
The following will be highly subjective but will try to get some data on it soon.

PC -> DAC -> long RCA -> HP-1 -> Goldmund clone:

Highly seductive sound, fast, scary bass that really captivated me

PC -> DAC -> long RCA -> Goldmund clone:

Plenty of power, gain set to same level as before, had to bring down high frequency balance. Bass not as powerful or rather fast as I remember, not as impressive even after redoing all DSP. Playing it any louder sounds off. High frequency less sweet.

PC -> DAC -> JDS Atom -> Long RCS -> Goldmund clone:

Seductive sound is back, bass is impressive again, fast sound, could turn back up the high frequency balance. Can turn it up without any obvious change in balance. Imaging is freed up, like with HP-1 (meaning not stuck in the plane of the speakers)

Fact: I had to rebalance the frequency response once the HP-1 was out of the circuit. Yet it never did sound as sweet (or fast) again.
With the DAC feeding the Goldmund the high frequency did sound more gritty, more tiring over long runs unless I turned it down a bit more than usual. (removed way more of the peaks than I ever had to in DSP to get it sound good)

For that reason I wanted to try the pré-amp. I can feel the difference in my body (lol). Sounds more pure, less gritty and is sweet again.

I cannot understand the balance change with the DAC feeding the Clone directly. All I found in the specs is that the DAC has higher output impedance compared to the HP-1 and JDS Atom. All of them should be able to get the max out of the Goldmund, right? Then why is there a less fast paced, less sweet sound without a decent pré-amp?

This all above is coming from a guy that never figured amps could make these kind of changes :). Yet the Goldmund and the Fetzilla changed my view on it. Could it be the Goldmund really needed that (or at least a good enough) pré-amp to shine? (not getting at the pré-amp that was omitted, I mean the HP-1 or JDS Atom)
That the DAC is simply not up to driving the cable + amp?

One more thing to add, at lower listening levels with the DAC feeding the Goldmund there was way more detail and a rather pleasant sound that the Pioneer never managed. Wen turning it up to the same levels as usual, a ~85 to 87 dB average (on midrange) it seems to loose that advantage.

One other possibility I have to mention; the Asio driver from my Xonar that has some strange quirks. Depending on how I use the driver it powers both headphone and speaker RCA outputs or just one or the other. But this does not seem to account for the direct swap without adjusting that driver on at least 2 times swapping gear.
This will be harder to figure out because I think it is a driver fault.
 
Last edited:
Do you know what the output impedance of the DAC is and the input impedance of the amp?

10K listed in one of the last posts is not unusual for an amplifier (but it is nothing like the input impedance of a good buffer) the output impedance of the DAC would need to be quite high to make signal transfer between them a big issue. The long lines would certainly make the situation worse and explain why a buffer is beneficial.
 
The following will be highly subjective but will try to get some data on it soon.

PC -> DAC -> long RCA -> HP-1 -> Goldmund clone:

Highly seductive sound, fast, scary bass that really captivated me

PC -> DAC -> long RCA -> Goldmund clone:

Plenty of power, gain set to same level as before, had to bring down high frequency balance. Bass not as powerful as I remember, not as impressive even after redoing all DSP. Playing it any louder sounds off.

PC -> DAC -> JDS Atom -> Long RCS -> Goldmund clone:

Seductive sound is back, bass is impressive again, fast sound, could turn back up the high frequency balance. Can turn it up without any obvious change in balance. Imaging is freed up, like with HP-1 (meaning not stuck in the plane of the speakers)

Fact: I had to rebalance the frequency response once the HP-1 was out of the circuit. Yet it never did sound as sweet (or fast) again.
With the DAC feeding the Goldmund the high frequency did sound more gritty, more tiring over long runs unless I turned it down a bit more than usual. (removed way more of the peaks than I ever had to in DSP to get it sound good)

For that reason I wanted to try the pré-amp. I can feel the difference in my body (lol). Sounds more pure, less gritty and is sweet again.

I cannot understand the balance change with the DAC feeding the Clone directly. All I found in the specs is that the DAC has higher output impedance compared to the HP-1 and JDS Atom. All of them should be able to get the max out of the Goldmund, right? Then why is there a less fast paced, less sweet sound without a decent pré-amp?

This all above is coming from a guy that never figured amps could make these kind of changes :). Yet the Goldmund and the Fetzilla changed my view on it. Could it be the Goldmund really needed that (or at least a good enough) pré-amp to shine? (not getting at the pré-amp that was omitted, I mean the HP-1 or JDS Atom)
That the DAC is simply not up to driving the cable + amp?

One other possibility I have to mention; the Asio driver from my Xonar that has some strange quirks. Depending on how I use the driver it powers both headphone and speaker RCA outputs or just one or the other. But this does not seem to account for the direct swap without adjusting that driver on at least 2 times swapping gear.
This will be harder to figure out because I think it is a driver fault.
It is pretty obvious to me that the problem is that the DAC has serious problem driving the long cable at HIGH volumens, as the problem is gone when the output level is reduced, either by lowering the listening SPL level or inserting a preamp to compensate fro the missing level and restoring the listening SPL level
 
Last edited:
Trying to understand more about the Goldmund philosophy I was reading a thread on this forum about a Goldmund Memesis.

My impression is that Goldmund took a knows amplifier topology (similar to Hitachi) and optimised it performance by simplifying it while keeping it stable in performance. They continued to do so over a long period up till this day.

A few of the more known names on the amp section of this forum compared it to an F1 car. A highly developed combination of parts to make it do what it does best. However could that make it susceptible to any change in the circuit feeding it?

Which could make sense to me in that it needs certain characteristics of the pré-amp circuit to keep it performing stable? It is after all a very wide frequency spectrum amplifier that differs in many regards from other amplifiers.
 
Do you know what the output impedance of the DAC is and the input impedance of the amp?

10K listed in one of the last posts is not unusual for an amplifier (but it is nothing like the input impedance of a good buffer) the output impedance of the DAC would need to be quite high to make signal transfer between them a big issue. The long lines would certainly make the situation worse and explain why a buffer is beneficial.
One thing you are missing is that the HP-1 was inserted AFTER the long cable in the first scenario. So the DAC performs well when it is driving the cable at lower levels.
 
It is pretty obvious to me that the problem is that the DAC has serious problem driving the long cable at HIGH volumens, as the problem is gone when the output level is reduced, either by lowering the listening SPL level or inserting a preamp to compensate fro the missing level and restoring the listening SPL level

You're right, that does make sense to me. It is the conclusion I should have drawn. With proper amplification power it sounds sweet at all levels.

One thing you are missing is that the HP-1 was inserted AFTER the long cable in the first scenario. So the DAC performs well when it is driving the cable at lower levels.

True, which is why I tried to paint the complete picture above. The culprit seems to be the DAC not performing as it should at higher levels. I think this is the most logical explanation. During the course of this project I put in that DAC to get rid of a hum problem caused by the PC's power source. Also during that period I changed how I used the Pioneer. As it had a volume control I had used it somewhere at 10 o'clock position right up till the point I started using the DAC. Not long after that the position of the dial changed to 3 o'clock (max would be 5 o'clock). Listening levels were kept the same, so internal volume level was reduced.
 
Last edited:
As an experiment I will try to go back to the Xonar Essence RCA output without a pré-amp and see how that sounds to these ears.
There might be hum, something I tried to cure with a transformer before I got that DAC running through optical. However lots of things have changed, another PC, different RCA cables etc and another amplifier, so maybe there won't be any hum to worry about.
Measurements of the Xonar in loop back all look as expected, clean performance all the way.
A long time ago I used Wasapi instead of Asio, this weekend I tested both. They still look identical after I tweaked the Asio buffer a long time ago (documented in this thread somewhere).

By the way, for the record, the output impedance of the M1 DAC is 47 ohm.
The JDS Atom amp's output impedance is 0.1 ohm.
Input impedance of the atom is 10 KOhm, the input impedance of the Goldmund is probably higher than that.
 
What "long RCA's" are you using?

For anything over 3 meters I use Canare star quad cable. It makes a noticeable difference. Anything shorter than 3 meters is less influenced by cable. Unless it is REALLY bad cable.

But either way a good buffer will improve a long cable run.

Edit:
I have been using it since the 90's when I found it while living in Japan however here is a more credible option.

YouTube
 
Last edited: