31267 trafo in 1272 clone problems..

[Family Hoof]

[quote author="CJ"]Current is not an issue with line and mic in x-formers.[/quote]
Yes, I remembered this after posting. Wrong choice of words . In your opinion, CJ, do these Marinair input trafos need to have both coils used to get low noise and good mileage? Judging by what is written on this thread I've already assumed yes. Of course I will be finding out first hand sooner or later. Maybe I'll send you one.

Transformers and mag amps are a fascinating subject to me. I've got tons of homework to do. :grin:

 

[Family Hoof]

31267 update:

Today I tried a St Ives 31267 as a mic in transformer for a 1272 module. This was an actual 1272 module wired like the first two stages of a 1073 in the "-50dB" sensitivity position. My grounding differed in that I tied shields to the chassis instead of B-. B- is tied to chassis at the power supply via a 0.1uF bypassed 10ohm resistor. A little bit quieter than Neve's wiring spec this way. I had a Marinair T.1444 (10468) to compare it to. This was a listening test only, with mics, no sweeps or anything like that. Here are my findings:

1) using only half of the windings, in reverse, for 1:2 (1k2:5k) = sounds exactly like the 10468 in 1:2. I'm still not sure if I actually can hear any difference at all - it's that close.

2) in reverse, series:series (2k4:10k) for 1:2 or parallelarallel (600:2k4) for 1:2 = in either case it still sounds much like the 10468, but now with very subtle differences. I'm not sure how to describe, just slightly less focus in the uper mids and highs, and a small amount of low mids resonance. These are not so much freq reponse variances, but more like different phase shift in places - slight smearing - at least to my ears. In spite of this it did not sound bad at all, totally usable IMHO, just different relative to the 10468.

In all cases I'm maintaining correct polarity, and pin 6 is tied to the trafo case and chassis ground. No problems with noise like Svart reported unless I use an uneven number of windings (i.e. only one secondary with the primaries in parallel) or disconnect pin 6/trafo case from the chassis ground. When I do this a large amount of 60Hz hum is heard.

After all was said and done I checked the JLM page again and it seems his findings were very much the same.


Anyway, this was part of whole N*eve input trafo shootout I've been conducting. Hopefully I'll get a detailed site and some listening examples posted next week.

 

[Rob Flinn]

Hoof,

Can you verify which pins you connected to when you used half the windings ? Was it the same way that I quoted in the other thread ? This way we can get a known working configuration fro all circumstances.

 

[rascalseven]

I'm not the most informed person regarding things electronic, but I've built many channels of Neve class-A preamps, and I've learned, by trial and error, and the occasional instruction from Geoff Tanner, how to achieve a dead-silent, excellent-sounding preamp. BTW I have used almost exclusively Marinair or St.Ives 31267's wired backwards per Joe Malone's instructions (one primary and one secondary, pin 6 to chassis, and no connection for the unused windings).

I am confident that much of the problems people experience with noise and hum in class-A Neve kit come from the grounding and shielding scheme used for the entire unit. Until I understood how to properly execute the grounding I had all kinds of noise/hum issues, but once I started consistently grounding and shielding these units as I describe below, ALL of these issues dissappeared completely.

To understand this post, as well as the many posts Geoff Tanner has made on this subject on various forums (and there are many) it is helpful to understand Neve's terminology:

"B+" = +24v
"B-" = +0v from the PSU (consider this 'audio ground')
"Earth" = chassis metalwork and AC ground pin

It seems many single-ended grounding schemes seem to equate B- and Earth, but they are very different when it comes to things Neve.

First I'll hit on grounding, then shielding. Assuming the PSU is in the chassis with the actual audio circuits do this:

 Run a wire from AC inlet ground pin to chassis. Run a second wire from the same AC inlet ground pin to the +0v (B-) of the power supply (don?t connect the 0v of the power supply to the chassis?. ONLY to the AC inlet ground pin). (so you have two wires attached to the AC inlet ground pin? one to chassis, the other to PSU 0v)
 Attach a second wire to the 0v of the PSU board and connect the other end to the B- (audio ground) of the audio circuit PCB.
 Connect the +24v of the PSU to the appropriate spot on the circuit board
 Attach the core of the input and output transformers to the chassis (pin 6 in the inputs if they?re Marinair or St.Ives).

If the PSU is external attach like so:

 Run a wire from AC inlet ground pin to PSU chassis. Run a second wire from the same AC inlet ground pin to the +0v (B-) of the power supply PCB (don?t connect the 0v of the power supply to the chassis?. ONLY to the AC inlet ground pin). (This is the same as first step above)
 Run three separate wires from the PSU to the actual unit: 1) +24v, 2) B- (0v of PSU board), and 3) Earth (chassis connection). Connect the chassis lead to the chassis, the +24v lead to the circuit board (obviously), and the B- (0v) to the audio ground on the circuit PCB (Not to chassis).

In this fashion you will notice that with the PSU disconnected from the main unit there is no electrical connection between the chassis and the audio ground (B-), but when you connect it to the PSU, the electrical connection is made. By having the connection between B- and Earth at the PSU only, you eliminate all kinds of noise. The argument can, of course, be made that electrically B- and Earth connect ultimately anyway, so why bother with all the routing of wires? This I cannot really explain fully, due to my lack of real hard electrical design knowledge, however, as I understand it, because the audio transformer cores are connected to the chassis, those cores can be excited by the current modulation of the power supply (current, not voltage) creating noise in the signal. By connecting the B- and Earth at the shortest distance to actual Earth (the AC inlet ground pin), this noise is eliminated.

I can?t explain it better than that, but I can assure you this REALLY WORKS!!

As for shielding, if you look at Neve schematics (or inside an actual Neve module) you will see LOTS of shielding. Basically, with few exceptions, every wire that carries an audio signal is shielded to B- (audio ground). Use individually shielded wires for each connection to the gain switch of a mic preamp, shield the input XLR?s, and the leads from the secondary of the input transformer and connection to the primary of the output transformer, as well as leads to the output XLR. That?s a lot of shielding, but by doing so you get rid of all the interferance.

Don?t make the mistake (as I did) of using shielded pair for connections that terminated close to one another? if those two wires are carrying different signals, they will interact with one another and create problems. Use shielded pair only for the inputs from and outputs to the XLR?s as well as the send to and from the fader/trim pot. Other than these three connections, stick with individually-shielded wire. All XLR pin 1?s go to B- (audio ground), not the chassis.

I hope this makes sense. By doing this my DIY 3-stage preamp?s (think 1290) are actually quieter than a REAL 1290 that I have, especially at the full 80dB of gain. Every other grounding/shielding combo created problems.

Hope this helps! :thumb:

JC

 

[rascalseven]

Oh yeah,

With regards to running the 31267 reverse-wired in series/series, the HF seems a little rolled off compared to a series/series 10468. Just one primary and one secondary sounds EXACTLY like a 10468 to my ear.

I recommend connecting it like Joe shows on his site. I'm reasonably certain that the other issues will be resolved by the previous post.

Peace,

JC

 

[Family Hoof]

Basically, my findings are the same as what rascal said. I disregard a lot of the stuff on Neve diagrams where B- and earth seem to be the same thing, and instead use shielded cable throughout, always tieing one end to chassis, and making sure 0V (B-) is only connected to mains earth at one point only. I never have any noise problems.

It does not seem to matter which side of windings you use on the 31267, so long as you're consistent --- ONE primary and it's associated secondary. Leave everything else disconnected.
look here

pin 8 = + input (pin 2 XLR)
pin 10 = - input (pin 3 XLR)
pin 3 = + output (to BA283 pin U)
pin 5 = - output (to BA283 pin V)

connect a 180pF polystyrene across pins 3 and 5 in parallel with any sort of resistor, pot, or sensitivity switch with a net resistance of approx 5k.

connect pin 6 to one of the screws on the transformer case, and then screw the transformer to the chassis

connect XLR pin 1 to the chassis

Even if your grounding is good you can still get a ton of hum like Svart was getting when you try to connect to an odd number of windings on the input transformer. For example, if I use pins 8 and 10 as the secondary and then parallel the primary windings for a higher step up.. HUM CITY.

 

[CJ]

Great research guys! Good to see this x-former stuff investigated in real life applications.

The Marinair Mic In has no phase shift at the bottom end, where as the Line In starts shifting from about 100 hz on down, somethimg I doubt you could hear nor shoulfd care about.

The mic in has a lot less ringing, which might be the reason you hear a little top end difference with the line in used backwards.
I have not checked ringing on a line in used backwrads.
The network will dampen the ringing, but the transformer still uses high end energy from the source. It just does not show up as a damped sinusoidal on the secondary.

 

[rascalseven]

That's interesting, CJ! Do you have a 31267 (line in) to test wired backward with just one primary and one secondary? I'd be interested to know how it compares to the 10468 (mic in.... Neve numbers, not "T-xxxx" used by Marinair).

Your post reminds me of something Geoff Tanner said regarding the EMI/Neve's. According to him the EMI/Neve consoles (which were the result of a collaboration between EMI and Neve engineers.... they were actually unique consoles compared to other Neve boards) used ALL 10468's for the inputs, mic AND line (wired reverse). According to Geoff the EMI engineers preferred the performance of the mic input transformer wired backwards for line inputs compared to the performance of the 31267 used as normal in other Neve designs.

I wonder what the difference is.......???

CJ, if you're interested, I've got some of each I'd be happy to loan so you can investigate further.........

:?: :green:

JC

 

[CJ]

The "T" stuff is the same as the 31267 stuff, just a different can and paint job. Same lams, same turns, same wiresize and same geometry.

I have a lot of test info on those x-formers. I didn't have a test bed for listening, something I need to do in the near future. I called Dan Alexander last week and he says he has a ton of sheetmetal, so my plan is to build an exact Neve channel strip, eq anfd the whole enchillada. I see a lot of Neve projects around here, has anyone actually built the whole strip, or just bits and pieces?

 

[Family Hoof]

CJ,

So are you saying the 31267 from Marinair (T.1442, 1452) and St. Ives (VT22671) have absolutely identical internal construction? My listening tests would point to yes, but I have no other testing methods at my disposal to verify. I know in theory they certainly were supposed to be the same, as Neve gave both manufacturers the same specs and used either version freely.

Also, I am very interested in any and all N*eve transformer test data you have. I've already saved that big "mo metal" thread to my hard drive, but I don't have anything else. Whatever there is I'd be more than happy to host a webpage for it.

If by whole strip you mean preamp + eq, then yes, many people have done this. The inductors are available and several companies already market module clones. Only a few of us here at the lab have done n*eve EQ, however. I am not one of them.

 

[CJ]

I have not "tested" the St. Ives input iron. But I would be happy to! :green:

Ask Tommytones if you can have the data. He has a whole stack of it. I have not seen him in a while. Maybe he is winding Neve transformers.

 

[khstudio]

I'm using the JLM ACDC & following the directions below for my "JC" version of the 1272.

http://www.groupdiy.com/index.php?topic=19648

QUESTION: JC says to bring the 0v of the PS back to the AC GND lug BUT the JLM has a 10ohm resistor which connects the 0v to the case... :?

Should I lift the 10ohm & run from this point back to the GND Lug?
Also, should the 10ohm be removed completely?

I'll be using a 31267 for the input transformer so I'd like to keep things quiet per JC instructions.

http://jlmaudio.com/Neve transformer info.htm

Thanks,
Kevin


[quote author="rascalseven"]I'm not the most informed person regarding things electronic, but I've built many channels of Neve class-A preamps, and I've learned, by trial and error, and the occasional instruction from Geoff Tanner, how to achieve a dead-silent, excellent-sounding preamp. BTW I have used almost exclusively Marinair or St.Ives 31267's wired backwards per Joe Malone's instructions (one primary and one secondary, pin 6 to chassis, and no connection for the unused windings).

I am confident that much of the problems people experience with noise and hum in class-A Neve kit come from the grounding and shielding scheme used for the entire unit. Until I understood how to properly execute the grounding I had all kinds of noise/hum issues, but once I started consistently grounding and shielding these units as I describe below, ALL of these issues dissappeared completely.

To understand this post, as well as the many posts Geoff Tanner has made on this subject on various forums (and there are many) it is helpful to understand Neve's terminology:

"B+" = +24v
"B-" = +0v from the PSU (consider this 'audio ground')
"Earth" = chassis metalwork and AC ground pin

It seems many single-ended grounding schemes seem to equate B- and Earth, but they are very different when it comes to things Neve.

First I'll hit on grounding, then shielding. Assuming the PSU is in the chassis with the actual audio circuits do this:

 Run a wire from AC inlet ground pin to chassis. Run a second wire from the same AC inlet ground pin to the +0v (B-) of the power supply (don?t connect the 0v of the power supply to the chassis?. ONLY to the AC inlet ground pin). (so you have two wires attached to the AC inlet ground pin? one to chassis, the other to PSU 0v)
 Attach a second wire to the 0v of the PSU board and connect the other end to the B- (audio ground) of the audio circuit PCB.
 Connect the +24v of the PSU to the appropriate spot on the circuit board
 Attach the core of the input and output transformers to the chassis (pin 6 in the inputs if they?re Marinair or St.Ives).

If the PSU is external attach like so:

 Run a wire from AC inlet ground pin to PSU chassis. Run a second wire from the same AC inlet ground pin to the +0v (B-) of the power supply PCB (don?t connect the 0v of the power supply to the chassis?. ONLY to the AC inlet ground pin). (This is the same as first step above)
 Run three separate wires from the PSU to the actual unit: 1) +24v, 2) B- (0v of PSU board), and 3) Earth (chassis connection). Connect the chassis lead to the chassis, the +24v lead to the circuit board (obviously), and the B- (0v) to the audio ground on the circuit PCB (Not to chassis).

In this fashion you will notice that with the PSU disconnected from the main unit there is no electrical connection between the chassis and the audio ground (B-), but when you connect it to the PSU, the electrical connection is made. By having the connection between B- and Earth at the PSU only, you eliminate all kinds of noise. The argument can, of course, be made that electrically B- and Earth connect ultimately anyway, so why bother with all the routing of wires? This I cannot really explain fully, due to my lack of real hard electrical design knowledge, however, as I understand it, because the audio transformer cores are connected to the chassis, those cores can be excited by the current modulation of the power supply (current, not voltage) creating noise in the signal. By connecting the B- and Earth at the shortest distance to actual Earth (the AC inlet ground pin), this noise is eliminated.

I can?t explain it better than that, but I can assure you this REALLY WORKS!!

As for shielding, if you look at Neve schematics (or inside an actual Neve module) you will see LOTS of shielding. Basically, with few exceptions, every wire that carries an audio signal is shielded to B- (audio ground). Use individually shielded wires for each connection to the gain switch of a mic preamp, shield the input XLR?s, and the leads from the secondary of the input transformer and connection to the primary of the output transformer, as well as leads to the output XLR. That?s a lot of shielding, but by doing so you get rid of all the interferance.

Don?t make the mistake (as I did) of using shielded pair for connections that terminated close to one another? if those two wires are carrying different signals, they will interact with one another and create problems. Use shielded pair only for the inputs from and outputs to the XLR?s as well as the send to and from the fader/trim pot. Other than these three connections, stick with individually-shielded wire. All XLR pin 1?s go to B- (audio ground), not the chassis.

I hope this makes sense. By doing this my DIY 3-stage preamp?s (think 1290) are actually quieter than a REAL 1290 that I have, especially at the full 80dB of gain. Every other grounding/shielding combo created problems.

Hope this helps! :thumb:

JC[/quote]