Neve 1290 gnd wiring, Whatever i do i'm getting hum!!!!

[ruffrecords]

Toroid radiates 50-60Hz. 100-120 Hz is PSU ripple which is completely different. Shielding the toroid should give you less 50Hz component, but you claim that shielding reduced the 100Hz component which is ripple and propagates through wiring, not via inductive/magnetic coupling.
I agree with Magneto - this is very odd.

In practice this is not true. The waveform in a normal toroidal transformer is quite distorted. It is quite normal for the radiated magnetic field radiated by the transformer to include harmonics of the mains frequency.

Cheers

Ian

 

[Ilya]

In practice this is not true. The waveform in a normal toroidal transformer is quite distorted. It is quite normal for the radiated magnetic field radiated by the transformer to include harmonics of the mains frequency.

You are correct. But it's easy to figure out the dominant component and figure out the probable culprit. At least this is my experience.

 

[MagnetoSound]

Perhaps CJ will chime in and tell us that mains toroids are usually wound with the secondary outermost, so that the rectified harmonic is dominant over the primary freq.

 

[ruffrecords]

You are correct. But it's easy to figure out the dominant component and figure out the probable culprit. At least this is my experience.

These days I use the spectrum analyser in REW to look at the entire noise spectrum. I most often see the fundamental, second and third harmonics, often not very different from each other, followed by diminishing levels of higher harmonics. I spent a lot of time with the Classic Solo mic pre, hunting down sources of hum and from that experience I can tell you it is often very hard to tell what the source is irrespective of the dominant frequency, if indeed there is one.

Cheers

Ian

 

[Spencerleehorton]

I am using a Sowter 1:2 txf (don't know what type but i think its a 4095?) that has internal screen and has mumetal case shield, both connections I’ve put to 0v.

now i know the original Neve's use external power i will work with that for my 4 channel version, its so cramped in there i very much doubt that shielding the toroid will help as its only a couple of inches away from the CH4. i'll try shielding first of all but start building an external PSU as i need one also for the other 2 channel i have made which i want to put into a 2U and include the API 553 EQ.

 

[Monte McGuire]

Mu metal is so expensive!!!
[snip]
and then there is the annealing it to bring back the magnetic properties? any ideas how this is done also?

Mu metal annealing / heat treating is done at very high temperatures (>2100 ºF) in a pure hydrogen atmosphere for several hours. I'm amazed that this can be done safely even in a high tech factory - not gonna happen at home.

External supplies, or altering the relative geometry of your power and signal transformers might be easier than relying on high permeability shielding. Also keep in mind that high perm shields can saturate with high flux, so you may want a lower perm shield to soak up some field strength and then a high perm shield to finish the rest.

Or, maybe just a thicker low tech magnetic shield, like a conventional deep drawn can or two or three.

It still seems easier to try to solve this with geometry, or using an external supply with an umbilical cord. Or, how about a switcher so you can move the problem to frequencies high enough to use simple shielding? Switchers don't have to be awful, and there are probably pre-made switchers that will work very well.

 

[Spencerleehorton]

very interesting Monty, i have a formulated a plan for the rest of my Neve's now, i'm happy with this first 2 channel unit for the time being anyways!!!

I'm going to rebuild an external PSU for the 4 channel unit and like you say "umbilical cord" it!!

The other 2 channel unit i'm going to have to rebuild the PSU anyways to accommodate 2 more voltages for the API-553 and the whole unit need to be re fitted into another box as what its in is not working.

Im going to try a few things out on my other input transformers as well, watch this space!!

 

[MagnetoSound]

I am using a Sowter 1:2 txf (don't know what type but i think its a 4095?) that has internal screen and has mumetal case shield, both connections I’ve put to 0v.

No! Connect the shields to the chassis ground. Did you not read Ian's article?

 

[Spencerleehorton]

sorry i meant to chassis gnd, my mistake!!! been some long days and nights!!

I have read Ian's doc and it has made everything much clearer, thank you Ian.

 

[Ilya]

These days I use the spectrum analyser in REW to look at the entire noise spectrum. I most often see the fundamental, second and third harmonics, often not very different from each other, followed by diminishing levels of higher harmonics. I spent a lot of time with the Classic Solo mic pre, hunting down sources of hum and from that experience I can tell you it is often very hard to tell what the source is irrespective of the dominant frequency, if indeed there is one.

The pattern may probably differ with different topologies, gear and parts. I'm attaching the spectrum of both cases with actual Neve type amp stages. You can clearly see the 50Hz and its harmonics in the bottom plot (close proximity of mains toroid) vs 100Hz and its harmonics in the top plot (PSU ripple). These graphs were taken from the actual Neve type output stage. Although not exactly 1290, but the topology is very similar.

I agree, it's not correct to say that there's an dominant fundamental here (although in both cases the fundamental is a bit louder), but actual fundamentals are different for each case thus giving out a clue as to what the problem might be.

 

[ruffrecords]

The pattern may probably differ with different topologies, gear and parts. I'm attaching the spectrum of both cases with actual Neve type amp stages. You can clearly see the 50Hz and its harmonics in the bottom plot (close proximity of mains toroid) vs 100Hz and its harmonics in the top plot (PSU ripple). These graphs were taken from the actual Neve type output stage. Although not exactly 1290, but the topology is very similar.

I agree, it's not correct to say that there's an dominant fundamental here (although in both cases the fundamental is a bit louder), but actual fundamentals are different for each case thus giving out a clue as to what the problem might be.

I think semiconductor gear is easier to sort out in this respect because there is only one power supply. In the Classic Solo, as well as the HT supply, there were also AC heaters as well a possible coupling from the mains toroid. This meant the waveforms were rather more complex than the ones you showed and hence much harder to interpret.

Cheers

Ian

 

[Spencerleehorton]

Wanted to ask which way is best to have a toroid? Flat or vertical?

 

[tmuikku]

Wanted to ask which way is best to have a toroid? Flat or vertical?

depends, changing magnetic field though a loop (area) will induce current to the loop. Check google image search for magnetic field of a toroid, altough in a chassis the field is not so perfect Imagine loops in your electronic circuit pcb and wiring (including your io transformers!) and try to position your toroid (and io transformers) in a way that imagined magnetic flow lines won't pass any (at least relatively large) loops.

now you'll have some idea how to orient the transformer, then just experiment.

also minimize all loop areas by twisting wires etc borrow this one from your nearest library: Noise Reduction Techniques in Electronic Systems by Henry W Ott

 

[ruffrecords]

Although toroids have a much lower external magnetic field than regular transformers, it is not zero. You can achieve a useful reduction by enclosing it in a steel box. Lat time I had to do this I managed to find a suitable size box on ebay:

https://www.ebay.co.uk/sch/i.html?_odkw=100mm+x+100mm+steel+boxes&_osacat=0&_from=R40&_trksid=p2045573.m570.l1313.TR0.TRC0.H0.X100mm+x+100mm+steel+enclosure.TRS0&_nkw=100mm+x+100mm+steel+enclosure&_sacat=0

I made a small hole in the side for the wires to exit.
Cheers

Ian

 

[Spencerleehorton]

Ah ok, me using M6 lams looks ugly but seems to work fine.
I have about 3-4kg of them so will continue to use them!!

Just waiting on some new toroidals then I’ll sort the other Neve channels.

Thanks again for all your help and knowledge everyone.

Regards

Spence.

 

[Spencerleehorton]

on my four channel version i've taken the power transformer out of the box and i'm still getting big hum, must be ripple on the rail?
Ive checked the GND scheme and is correct now.

Could anyone explain how to check for ripple please, this seems to be something i really need to figure out!

 

[Ilya]

Try powering a single channel and see if you get any hum. Then add the second one and check again. Repeat for all channels and see what you get.

 

[Spencerleehorton]

Ok sounds like a good plan.

 

[Spencerleehorton]

tested PSU pcbs and both are shot to bits!!!

0.5 on the scope of ripple on one pcb, the other one needs to be binned!!

gonna build a couple more and get them perfect this time!!!

 

[ruffrecords]

Try powering a single channel and see if you get any hum. Then add the second one and check again. Repeat for all channels and see what you get.

You have only got a 24V and 48V supplies so a regular scope should be able to handle that. Probe the supply positive, switch the scope to ac coupling and turn up the sensitivity.

Cheers

Ian