Neve 10468 Input Transformer Compared to modern Substitutes

DISCLAIMER

Only one transformer was tested from each type for the sole purpose of how it performed in a Neve 1272 normal ,1272 modified ,1064,1073,1076,1078 style Mic Pre circuits and for use with the JLM 1272 to 1073 Module.    . No batch testing was performed at all. So use of this info is at your own risk. These test were done to help DIY'er get there Neve clones closer to the real thing with out the need for expensive test gear. All transformers tested have been ones suggested to me as possible substitutes in over a hundred or so emails.

500Hz Square Test Wave and 20Hz 0.5% & 1% THD level Test

Above graph shows one rising edge of the 500Hz Square Wave taken from the transformer secondary winding with a x10 probe into a Tektronix TDS2000 Digital CRO.

The test wave used was a 500Hz Square Wave of 400mV Peak to Peak into a real 1073 Neve Module with the substitute transformers wired directly into the 1073. Generator Output Impedances of 50ohm and 200ohm where both tried on each transformer to simulate modern solid state output Mics, Dynamic & Transformer output condenser microphones. The Grey wave form shown was taken while the Neve 10468 was tested after the 50ohm or 200 ohm generator output. The Square wave was pure at the zero ohm output point of the generator. The CRO was adjusted to show all the transformer outputs at the same height. The true voltage out is shown beside each transformers name.

The 20Hz THD test shows the level at the input to the transformer at which the core saturates at 20Hz causing the THD to increase dramatically. The test shows the input level at 0.5% and 1% THD.

Ideally a we would like to see one of the substitute transformers wave travel over the top of the Neve 10468 transformer wave so the two looked as one. Even two original 10468 transformers will travel mostly on top of each other but in places the two colours would be still seen. Obviously none of the substitutes tested obtained the holy grail but some are very close and some are just plain wrong. Basic components of the rising edge are the rise time of which we have transformers on both sides of the Neve. Then there is the slight overshoot after which the transformer should settle and not ring. Ringing is the biggest no no when it comes to transformers with to much overshoot being the next in the list of unwanted effects. Overshoot can be treated by zobel or cap loading  to bring it with in tolerance. Ringing can be treated like wise but usually will make a larger over shoot and still have smaller but still visible ringing.

We also want a transformer with a 20Hz saturation point close to the original Neve 10468 to ensure the transformer core has a similar sound to the Neve 10468 at high levels of bass input . The Neve 10468 has a 0.5% THD @ +7dBM & 1% @ +9dBM at 20Hz.

Substitutes from worst to best performance as a Neve 10468 substitute in a 1073 or 1272 Hotrod

NOT SUITABLE. The OEP A262A2E which is sold by RS as part no. 210-6419 with screen can 210-6469. And by Farnell part no. 149-155 with screen can 149-157. The transformer was wired as 1:2 with both sets of winding in series. Giving 600:2k4ohms. The idea being that the 6k load on the 2k4ohm side would reflect over 1200 on the 600ohm side just like the 10468 is right and can be done with many transformers but not this one. The ringing was so big I could not fit it on the screen without making all the other graphs to small to easily see. Tried different loading but nothing could control it. This transformer needed a low resistance like 2400ohm across the secondary plus Zobel to fix it which puts it out of the Neve substitute list.

So unusable did not do 20Hz test

 BUT when I swapped the polarity of the secondary so the red was now the 0v ground all ringing stopped and a good square wave was seen. This was tried on all three 9002 and found to work perfectly. Attaching the black screen wire to 0v ground produced only a just visible ringing but nothing like had been seen with polarity reversed. The three 9002 now had a good square wave response with one overshoot followed by an under then settled into a flat line. All three transformers had different overshoot so there is a bit of tolerance between them. To get the overshoot the same on all three 9002’s only the 180pF cap across the secondary had to be adjusted. 1 needed 180pF, 1 needed 82pF and one needed on cap to get the same overshoot height as the 10468.

The Sowter 9002 has a 0.5THD @ + 12dBM & 1% @ +14dBM at 20Hz which is more than the Neve 10468 can take, so the 9145 if it did not have the ringing problem of the 9002 it would be a closer match as its saturation point is supposed to be a few dB lower than the 9002.

The 9002 now sits speed wise between the Lundahl LL1538 and the Marinair 10468. This linked with the similar 20HZ THD point as the Marinair make it a very good substitute for the 10468.

Conclusion all that is needed for most people to improve the 9002 performance is to swap the input and output polarity by swapping the blue & yellow wires and the red & orange wires. As long as the RED wire is the one connected to 0v ground in a 1272 or 1073 clone this transformer works very good. This may also apply to the 9145 if they are wound the same as the 9002 but I have not tested one.

Changing which output winding is grounded has redistributed how the screen capacitance, and coil inductance and inter-winding capacitance work together.

The Sowter coil layout sheet with the wire colour codes shows the yellow & red wire on top which most people would take as the positive potential input and output since there are no dot marks shown that usually indicate positive potential.

I’m sure most Neve clones will have the orange wire connected 0v ground unfortunately.

I have not added the new Sowter transient plot to the current jpg at the top of the page yet.

 

The OEP A262A3E has a 0.5% THD @ +1.5dBM & a 1% @ 3.5dBM at 20Hz which is less than the Neve 10468 by 5dB but still not to bad.

The Lundahl LL1538 has a 0.5% THD @ +8.5dBM & 1%THD @ +11dBM and can handle a dB or two more than the 10468 which is fairly close but slightly more than the 10468. The Lundahl LL1538XL was also tested and has a 0.5% THD  @ +12dBM  & 1% @ +14.5dBM. So the extra expense of the LL1538XL is not needed for the 1073 clone and the LL1538 is a closer match to the 10468 saturation point.

The Neve 31267 has a 0.5THD @ +7dBM & 1% @ +9dBM at 20Hz which is identical to the 10468 which leads me to believe that the 31267 has two winding each equivalent to the 10468 windings in series which makes perfect sense really.

Wiring 31267 as 10468

You can use either Mic IN as the Mic input and leave the other two winding not connected. The Line IN gives you a 1:1 input with 5k input impedance. If you want to use the 2 or 3 inputs at once they need to be resistively feed for them to sum and mix properly. Basically you could have two dynamic mics connected direct to the mic inputs and a synth out with about a 2k resistor in series with its output and they will all mix together. But if you connect a opamp out straight to any input it will kill the level of the other two inputs.

WARNING. Using only one primary and secondary coil on the 31267 disables the mechanical hum bucking due to there being two separate bobbins on the U laminates that sit mechanically on opposite sides and directions of the formed O core.