Neve transformer info / inp TX damping

[henryl]

Hi JLM/Forum,

Hope this question is not out of line here. Not currently related to a particular JLM product (though I'm browsing with interest as I've recently started looking into preamps/building/modding!).

I found this JLM page http://jlmaudio.com/Neve%20transformer%20info.htm in the process of experimenting with a Golden Age Projects PRE-73. Having replaced some of the transformers with Carnhill models (9045 mic inp and 1148 o/p) I didn't trust that just replacing them would be a good idea without also considering the zobel/damping/loading components so I started looking with a 'scope and tweaking the input TX loading/damping components but found in the end that I wasn't entirely sure what constitutes a good theoretical result.

I was originally working from some general advice to try and get rid of overshoot and ringing but saw that the JLM traces show some for the Neve 10468.

Ref the above JLM tx-comparison traces:

- what gain/sensitivity setting was used?
- what is the horizontal/time grid scale?

The reason for the first question is that I discovered only after completing a run of tweaks on one gain setting that the overshoot/ringing was then different if I switched to another setting. On some settings there was no overshoot or ringing at all, maybe even a little over-damped, whereas on others some appeared. I'm guessing a Neve would be the same unless the GA implementation of the switched attenuator somehow caused this, which seems unlikely (though possible).

I was using a 0.1V test signal (which is fixed in my test rig) rather than 0.4V as per JLM, so the gain setting was probably not the same and I found differences in overshoot/ringing within 5dB and 10dB steps from my original test setting.

Thanks / Henry

[henryl]

OK so I've now measured, calculated, delved.. and it looks like the 1073 sensitivity selector keeps the Zin, as seen by the mic input tx, fairly tightly around 4k8 so the published JLM traces should probably be reasonably consistent across different sensitivity settings - within the same gain range at least. There seems to be a significantly lower Zin of 3k7 in the higher gain / lower sensitivity range [-80 to -55] when the tx just fires straight into the BA284.

Given the level of test signal used in the JLM case (0.4V) then it seems unlikely to have been done in the low sensitivity range - but it'd still be nice to confirm that.

It turns out The Golden Age Pre-73 sensitivity selector is not identical to the Neve in component values, uses a slightly different gain-controlling configuration, and does not keep the Zin so consistent in the [-50 to -20] range, though it's not massively different overall to the Neve.
I've been working on this over at https://groupdiy.com/threads/neve-ish-m ... ice.84275/ if anyone's interested.

cheers

[henryl]

On third thoughts it would be very interesting to know at what sensitivity selector position the JLM traces were taken. ??

From experiments with the GA version there is a clear jump in overshoot/ringing behaviour when the selector is set at -40, -45 and below.

I see that at that point the resistance of the top leg of the sensitivity attenuator drops significantly from 2k to 660R in the GA and 2k2 to 0 in the Neve.

[Joe Malone]

On third thoughts it would be very interesting to know at what sensitivity selector position the JLM traces were taken. ??

From experiments with the GA version, there is a clear jump in overshoot/ringing behavior when the selector is set at -40, -45, and below.

I see that at that point the resistance of the top leg of the sensitivity attenuator drops significantly from 2k to 660R in the GA and 2k2 to 0 in the Neve.

From memory (as this was over 20 years ago) I did this in a 1272 module so had a 5k6 load resistor on the secondary. There are 2 different versions of the gain switching circuit in 1073 and both vary the input secondary load somewhat to help get the correct dB steps with the old standard E12 resistor values. The Tektronix digital scope text overlay shows the measured voltages and 5uS per division time setting of the scope. So you can see the voltage doubling for 1:2 ratio and triple for the 1:3 ratio.

All transformers into 1073 new LATEST 5us.gif

[henryl]

Many thanks Joe. Only 20yrs?!
That's very interesting. I've been working from a 1073 schematic but didn't realise there were two different sensitivity selector versions (I've been referencing EK20033). I'm going to have to catch up with that and the 1272 differences. 5k6 certainly does dampen things...

Based on my tweaking of a 1073-ish clone I've found a significant jump in the visible mic input TX ringing when going from the -35 to the -40 position, which coincides with a significant step change in the attenuator which loses the top resistor of the divider completely (and starts adjusting the gain control resistors at -45 etc.). Hence my wondering at what selector position your traces were taken.
(Based on someones modelling work it seems the Zin seen by the TX nonetheless remains commendably consistent across all settings. )

Sorry my bad re missing the 5uS per div info - I don't know what I was thinking it was referring to, but the penny got stuck.

All the best/ Henry