AMS Neve 1073 Sweeps

[AErige]

the same sweeps with no termination

 

[AErige]

Super high res

Btw did you do this measurement with the TF12000 output Tx you bought or with a standard carnhill Tx?

 

[Aaronrash]

Hmm... The little ripples in the high end look a bit strange. What is that?

My measurements were done with the AMS TF12000

I don't really want to post this publicly but I'm just gonna go ahead and say it. My EZ1073's sound nothing like a AMS 1073 and I've wasted 7 months of time as well as a ton of money trying to figure out why. I'm about ready to sell my pair.

I went to great lengths to get these to sound as good as possible. Here's the components used...
Phillips BC Series elctros
NOS Motorola 2N3055
Tantalum caps
Polystyrene caps
NOS BC184C's selected for high gain
AMS TF12000 output TX
Mounted the Mic input TX on the frame and earthed like the originals
Added filtering on the PSU
Mullard C296 caps throughout

Back and forth emails over a series of months with a very high respected tech trying a million different things
The sound just isn't there
The AMS sounds big and punchy with the classic Mr Neve Sheen...
These sound like a pillow blanket is thrown over them.
They are transparent. No thickness or color. What goes in is what comes out
Running mixes through the AMS left a unmistakable Neve character
Running mixes through my EZ's I can't tell any before or after difference and they nearly phase cancel with ITB

Not sure what to do at this point besides blame it on the PCB deisgn

 

[3nity]

Why would you like to blame somebody?
As its been stated before by who knows who,  Carnhill transformers sold by Audio Maintenance are not the same 100% specs as the ones used by AMS.

But if you really want to blame it on the PCB design try the EZ1084 Line EQ... to my ears they are not at all transparent they have a nice colour and thickness.

However im very proud to say i come from the diygroup as all of our projects are top sounding ones!

 

[Aaronrash]

Why would you like to blame somebody?
As its been stated before by who knows who,  Carnhill transformers sold by Audio Maintenance are not the same 100% specs as the ones used by AMS.

But if you really want to blame it on the PCB design try the EZ1084 Line EQ... to my ears they are not at all transparent they have a nice colour and thickness.

However im very proud to say i come from the diygroup as all of our projects are top sounding ones!

Hi,

My EZ1073's have the transformers purchased from directly from AMS... If you would have read my post.

I am blaming it on the PCB because everything else at this point has been replaced and replaced again. Checked and quadripple checked again over a series of 7 months. There are no errors with the components. All the components in my builds are the exact same as the ones in the AMS re-issues, so why do they not sound even close?

It's disappointing when you spend allot of money and time as well as almost a year sourcing parts for the build and then they end up being very far off.

I will post comparison samples if you'd like to hear.

Aaron

 

[Aaronrash]

Ok here's the comparison file for all those that are interested.

First in the file is the AMS 1073, followed by the EZ1073

All the bright/silky mid-high range in the AMS is not coming through in the EZ.

Again these are using the EXACT same transformers

We think it might have something to do with capacitance of the traces on the PCB, but there is no way to find out.

The B-0 ground plane on the EZ is so massive it takes forever for it to take solder, even set to 850F

I personally think this must be affecting the sound in some way.

Here's the file: https://soundcloud.com/aaronrash/ams-vs-custom

Can other EZ1073 owners tell my your experiences? Has anyone else done a proper shootout with their EZ's against a real AMS?

I'm not blaming Colin, he makes nice products and has allot of happy customers. But I went into this thinking they would sound dead on accurate. Perhaps allot of EZ1073 owners have never used a AMS unit and are therefore happy with their builds

 

[ruffrecords]

Not sure what to do at this point besides blame it on the PCB deisgn

Anything that makes such a blatantly obvious difference in sound between two units is unlikely to be something as subtle as PCB layout. It is well known that the amount of distortion a circuit produces is proportional to the level at which it operates and the power it produces. The early stages of all Neve channels operate at around -8dBu, about 300mV, and the distortion they produce is negligible. The stage responsible for most of the 'sound' is the output stage which operates at a nominal +4dBu and can produce up to 400mW of power into a 600 ohm load. In my view that is where you should look for differences. The most obvious attribute of this stage is that its bias is adjustable. How did you set this up?

The output stage is the only one where PCB layout could possibly make a difference to the sound as Douglas Self so admirably demonstrated on his series on transitor power amplifier distortion in Wireless World many years ago.

Cheers

Ian

 

[Aaronrash]

Not sure what to do at this point besides blame it on the PCB deisgn

Anything that makes such a blatantly obvious difference in sound between two units is unlikely to be something as subtle as PCB layout. It is well known that the amount of distortion a circuit produces is proportional to the level at which it operates and the power it produces. The early stages of all Neve channels operate at around -8dBu, about 300mV, and the distortion they produce is negligible. The stage responsible for most of the 'sound' is the output stage which operates at a nominal +4dBu and can produce up to 400mW of power into a 600 ohm load. In my view that is where you should look for differences. The most obvious attribute of this stage is that its bias is adjustable. How did you set this up?

The output stage is the only one where PCB layout could possibly make a difference to the sound as Douglas Self so admirably demonstrated on his series on transitor power amplifier distortion in Wireless World many years ago.

Cheers

Ian

Hi Ian,

The bias has been set up properly. I've also tried several different output transistors.
I'm not sure what else to look for at this point I've run out of options.
It's not unheard of for PCB designs to have errors, but I sort of doubt that too, Colin knows what he's doing.

I've bypassed as much as the PCB as possible by wiring straight off the output  TX secondary to the XLR's as well as doing the same with the input TX.

I'm interested in how much this change in sound from layout differences is possible as Douglas demonstrated...

 

[Aaronrash]

I would like to measure the current my output TX is drawing through the primary.

Is there a specific way to measure this?

 

[AErige]

Let's try to think with logic. Like you said here http://www.gearslutz.com/board/high-end/494329-bae-1084-vs-neve-1084-a.html you find your ez1073 to sound "identical" to your friends BAE's 1073's but not like the AMS.

Yet Bae uses the same boards/wiring than ams/old N so like Ian said it is unlikely due to the PCB layout. So what's left ? you use the ams TF12000 output Tx ...what about the input Tx ? ams Tx too ? (btw its an ST transitor on the ams unit not a moto)

On your soundcloud file https://soundcloud.com/aaronrash/ams-vs-custom even if not at the same level/same phrase, differences are pretty obvious and I can see what you mean by tight bottom end/silky midrange on the ams. To me it almost sounds like slight saturation.

Did you measure THD between the 2 units or when you set the bias up? or you just adjusted for 110-130mA ?

Is there a specific way to measure this?

measure DCR between the primary legs 1-3 (should be around 14 ohms) then measure DC tension and I=VDC/DCR

If you have a voice sample you'd like me to put through my units (line input) to compare with yours, tell me. I don't have my mytek around in my lab but the rme device should be enough.

 

[Aaronrash]

Let's try to think with logic. Like you said here http://www.gearslutz.com/board/high-end/494329-bae-1084-vs-neve-1084-a.html you find your ez1073 to sound "identical" to your friends BAE's 1073's but not like the AMS.

Yet Bae uses the same boards/wiring than ams/old N so like Ian said it is unlikely due to the PCB layout. So what's left ? you use the ams TF12000 output Tx ...what about the input Tx ? ams Tx too ? (btw its an ST transitor on the ams unit not a moto)

On your soundcloud file https://soundcloud.com/aaronrash/ams-vs-custom even if not at the same level/same phrase, differences are pretty obvious and I can see what you mean by tight bottom end/silky midrange on the ams. To me it almost sounds like slight saturation.

Did you measure THD between the 2 units or when you set the bias up? or you just adjusted for 110-130mA ?

Hi AE,

Yes these are good points but BAE doesn't stick to orignal specs like AMS and the EZ1073 do. The only thing different on the EZ is the gain structure is a little different.

Colin stayed pretty true to the schematic and I still have hope I might be able to get them sounding right.

The Neve DPA is on a PCB but allot of people say they sound very different than their re-issues. I personally haven't heard the DPA yet.

I've also tried the ST by the way. Actually some re-issues have Mospec in them too witch I've also tried as well as vintage 70's era Moto 2n3055 and even the Mexico Moto.

No THD measurements have been taken and I actually dont have the AMS unit anymore so I'm stuck where I'm at for the time being.

From what my tech has told me, the BAE units sound very very close to the AMS re-issues after their common errors are fixed. Something he does allot.

Oh yea... what is the little spike in the high end of your units?

If you're willing to run a vocal sample through you're unit that would be cool and interesting to see how ours differ. I'll send you the wav and you can see if you can get it to match the AMS. My guess would be to drive it kinda hot. PM me your email

Aaron

 

 

[ruffrecords]

I just listened to the samples several times. It seems to me the AMS has too much top end. It is almost as if part of the EQ is still in circuit. The EZ sounds much more like I remember Neve modules sounding. I would suspect something is going on in the AMS. I think it would be instructive to carry out frequency response tests on both.

Cheers

Ian

 

[radardoug]

It's very likely that the difference has to do with the in/out transformers and their termination. Not likely that the pcb is the problem.
You should do some carefull measurements of frequency response to see the differences. The input and output Zobels make a big difference to frequency response.
Also I am sure some of the Neve sound is due to the bootstrapped output stage.

 

[MatthisD]

I would also suspect a problem with the AMS based on those vocal samples. The first one does not sound unprocessed, it is missing low frequency content. Have you checked the response of that preamp?

 

[Aaronrash]

Yes I have checked the frequency response.

To me, the 10XX modules have always had a "processed" sound and definitely far from transparent

My friend has a studio full of them as well as 1084s. I owned 1084s from AMS for several years I know the sound.

Put it this way, the AMS add something, the eZ1073 does not.

My before after mixes passed through the EZ phase cancel out with the before.

There is no difference when going mic in between my super transparent focusrite interface pre and the EZ they sound the same.

I am trying to get the mid forward bite and sheen the AMS has.

Still, two units that graph the same can sound completely different tone wise

Both transformers are the same.

The impedance  are both set to 1200 on both with 55db of gain.

We really need some AMS sweeps

 

[ruffrecords]

Yes I have checked the frequency response.

I find it hard to believe the two frequency responses are identical from the samples you posted.

Cheers

Ian

 

[Aaronrash]

The only difference with my EZ is that from 3k upwards there's a .05 droop until it rolls off somewhere about 50K.

A AP would be helpful. And especially AMS sweeps. They were set up the exact same. EQs left disengaged.

Something interesting, a guy on here that makes ELA M 251 mics on here demonstrated the difference between one wired PTP and then one in a PCB and the difference was massive. Almost like how my EZ sounds to the AMS.

The users name on here is OPR. It's in the mic section, I suggest everyone to check it out as well.

 

[bruno2000]

Yes I have checked the frequency response.

To me, the 10XX modules have always had a "processed" sound and definitely far from transparent

My friend has a studio full of them as well as 1084s. I owned 1084s from AMS for several years I know the sound.

Put it this way, the AMS add something, the eZ1073 does not.

My before after mixes passed through the EZ phase cancel out with the before.

There is no difference when going mic in between my super transparent focusrite interface pre and the EZ they sound the same.

I am trying to get the mid forward bite and sheen the AMS has.

Still, two units that graph the same can sound completely different tone wise

Both transformers are the same.

The impedance  are both set to 1200 on both with 55db of gain.

We really need some AMS sweeps

55 dB of gain?  Thought you were feeding them with a line level signal,
Best,
Bruno2000

 

[3nity]

Seriously to me the EZ1073 sample sounds better to my ears.

 

[Aaronrash]

Nope this was a comparison of the mic pre. But the situation is the same when passing mixes through.. The AMS sound different. That was with a stereo pair so I'm pretty sure there is nothing wrong with the AMS units.