100Hz hum in preamp

[untune]

Ive seen it done with smaller rectifier tubes ,6x4 ,Ez 80 etc , depends on the transformer I think , the resistors are added if the tubes charachteristics would otherwise be exceeded , but it might allow a slightly bigger first filter cap and keep peak currents down  so the rectifiers tube live long and happy life.

As has been noted this was originally pulled from an old akai recorder and as such had an extra tube that related the the tape/bias circuit I think. With that gone there's less draw on the transformer and a bit of extra heater headroom. Does this help/hinder things?

Ian mentioned tying the 6x4 cathode to a leg of the heater... I didn't try this yet, though

Two 470uF caps have arrived and they're huge, I might be able to fit one next to the choke and one down near the first 6au6. Other option is to lose the choke and get another in there.  I'm going to add another stage right after the choke with a 470uF, keep the rest of the 47uF in place and see how it improves it.  Then I can start looking at putting more in.

So it'll go 6x4 → 22uF → choke → 470uF then 2k2 / 47uF → 5k / 47uF → 33k / 47uF

 

[scott2000]

I have a double 6x4 board on order off ebay  , it places two 100 ohm resistors in series with the cathodes of both tubes , presumably to help the load seen by the tubes up a bit , easy to see how it keeps up the resistance seen by the rectifier , especially with empty caps at switch-on,

Is this similar to what I've read about where you put a resistor on each lead of the transformer secondary going to a 6x4 to limit inrush??

 

[Tubetec]

If you put the resistors in the primary leads or after the tube like in the 6x4 board ,it  amounts to the same as far as I cas see ,but maybe wiser people know better the difference .

If say you had your one side of your rectifier winding  grounded  like a standard Lt winding ,it would work but then the tube has to stand the hundreds of volts difference between its cathode and its heater , by referencing its heater to its cathode theres no big potential difference and so chances of catastrophic failure of heater/cathode insulation is impossible . I did see several comercial recievers which ran the rectifier and other tubes from the same Lt supply ,so cathode left disconnected from heater , its not a good sceme really as extra noise gets in from the influence of the Ht current in the rectifier circuit ,it also seemed to cause EZ81's glass envelope to become clouded with vapourised material ,where with heater/cathode referenced and on its own discrete winding its better ,safer and with less noise .


I can see because the silicon diodes do 50 % of the work in this circuit the two 6x4 are run nice n easy ,well down from the tubes maximum ratings .

I have a seperate 6.3v winding for each  6x4 tube  in this build ,so I'll try putting the windings in series and reference  the centre tap to the point after the 100 ohm resistors , hopefully that might allow some extra noise to cancel compared to the usual arrangement with a single rectifier , I might have a closer look at some of the big old dual rectifier supplies that were common on amps of 100w or more  back in the old days ,I think I saw something similar there before. 

 

[untune]

I've just finished adding in another filter stage as per the last post... So I basically made the first cap after the choke a 470uF and shifted the rest of the stages down, dropping the voltage slightly using a 2k2 resistor and the a 47uF for the EL84.  Good news... the 100Hz has all but gone... The 50Hz hum is dominant now but is low in level, and there are additional harmonics in there (150Hz) but it's a good start.

I could potentially add another 470uF at the bottom of the chain, or I could probably quite easily step the 47uF caps up to 150uF without using any more space.

It's out of the case at the moment so I'm not sure if the 50Hz might drop slightly once it is boxed up and properly shielded again but... progress!

 

[ruffrecords]

An easy way to work out the rough ripple reduction of an RC filter is simply:

2 x PI x f x R x C

So for your 2K2 and 470uF case you get:

2 x 3.14 x 100 x 2200 x 470E-6 = 650 => 56dB

Cheers

Ian

 

[untune]

An easy way to work out the rough ripple reduction of an RC filter is simply:

2 x PI x f x R x C

So for your 2K2 and 470uF case you get:

2 x 3.14 x 100 x 2200 x 470E-6 = 650 => 56dB

Cheers

Ian

That's a great tip Ian thanks, I'm going to make a note of that calculation.

It seems that 50Hz is my new nemesis - will additional filtering affect this, or will I have to focus on grounding and maybe elevating heaters?

 

[Tubetec]

Id say sheild up the wires to and from the pot and around the grid of the El84 ,your almost certain to be picking up a large amount of 50 hz there , with the high impedences involved and where you join the screen to ground on the cable will also make a difference .Only ground the sheild at one end closest to where the El84 grid and cathode resistor meet if possible and dont allow that screen to touch down again for instance in the vicinity of the pot case which will be grounded , that would end up forming a loop encompassing the screen and the chassis , which would definately pick up loads of noise .

It could be beneficial to move your input transformer into the far corner of the area by the input jacks and furthest from the mains tx , you'd shorten your grid wires and you would most likely get better screening from magnetic ,dont drill new mounting positions until you have first found the place and orientation with minimum noise . the extra space in the other compartment  could be used to get better mounting positions  of choke and output transformer    and what ever other capacitors you want to add  later if the need arises .                                                                                                         

It would be interesting to see the new spectrum of the noise compared to the graph you showed before  you started                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                 

 

[scott2000]

That's a great tip Ian thanks, I'm going to make a note of that calculation.

this is cool too

https://www.ampbooks.com/mobile/amplifier-calculators/RC-ripple-filter/calculator/

Great work on your progress! I'm not clear enough to say anything about the elevation of the heaters. I'm definitely interested in hearing about the best options here too. Like best place to do it etc.....

Great thread with some solid advice here.......

 

[untune]

Shielding some of the signal wires between switches and pots will be the next port of call, and double checking the grounding scheme for anything I missed first time round.  Input TX can't move far from where it is, as the XLR on the rear goes into it, then it continues to a switch on the front panel and then to the first valve.  All these wires are shielded already.  I have wires running from coupling caps to the pots that run across the whole board that will probably benefit from this come to think of it.

It would be interesting to see the new spectrum of the noise compared to the graph you showed before  you started

I've attached an overlay of the first noise spectrum (green) from about a week ago with the one after adding the extra filter stage (red).  Although there's a big 100Hz drop, 50Hz has gone up and harmonics are much higher. 

 

[untune]

Just in the interests of testing, I might clip bigger caps parallel to the smoothing caps already in the circuit before making any permanent changes.

For example, I might parallel another 47uF, or a 470uF, with the 47uF at the first stage to see what effect it has.

I've also pondered putting a second 22uF in parallel with the 22uF reservoir cap on the 6x4.  Any risk of damage here?

 

[ruffrecords]

I notice from the last spectrum you posted the 50Hz hum has harmonics at 150Hz, 250Hz etc. These odd harmonics are typical of magnetic induction.

Cheers

Ian

 

[untune]

I notice from the last spectrum you posted the 50Hz hum has harmonics at 150Hz, 250Hz etc. These odd harmonics are typical of magnetic induction.

Cheers

Ian

Thankyou Ian, I wondered if that might be the case.  Will this induction only occur between PT > OT, or choke > OT or can it also be induced elsewhere? I removed a small shield between the choke and OT to get the bigger cap in, and I wonder if that might be responsible for the slight 50Hz boost

 

[scott2000]

I'm not sure if I'm following correctly but PRR mentioned in his first couple of posts about the success to be expected from where we are at now?

 

[Tubetec]

I did notice the odd harmonic nature of the spikes ,
interesting how the transformer can throw that out .

Maybe now  could be a good time to try and find a less inductive spot or mounting position  for your output tx relative to the mains ,
normally 90 degrees to each other is best with as much distance as possible ,your at about 45 degrees and only halfway across the chassis which is very unlikely to be the optimum spot  . Adjusting the position of the transformer while listening or viewing the spectrum is the best way to find minimum induction point . A metal screen could certainly make a difference in what your seeing at 50hz, anything steel and grounded  between your mains and your OP tx is probably a good thing ,but locating the best possible position on the chassis now ,means any screens you add later should improve matters another degree of magnitude ,  instead of fighting noise your  inducing by poor layout in the first place ,its very much a round in circles thing but you can incrementally make improvements .

Your choke will be picking up stuff too of course ,  usually I'll use the choke towards the load end of the supply with RC upfront , that reduces the ac current in  the choke , and hopefully the radiated noise as well . Id look at dividing up the tx compartment with a couple of metal sheilds ,one  to the right of the  mains tx , space for smoothing caps and choke then one more behind which you place the output tx  ,up in the right hand corner .

sheet steel , a metal snips and a bench vise will make it a breeze to make the partitions

 

[ruffrecords]

I did notice the odd harmonic nature of the spikes ,
interesting how the transformer can throw that out .

Transformers saturate first at low frequencies. 50Hz is pretty low and mains transformer manufacturers tend to design their transformers so they operate pretty close to saturation to keep size and cost down. Semiconductor rectifiers and large reservoir caps cause large spikes of charging current which can easily get very close to saturation. As a transformer gets close to saturation it produces odd harmonic distortion which is reflected in the magnetic field it emits.

These days I try to operate mains transformers at close to 50% of their rated power.

Cheers

Ian

 

[untune]

Bit more progress, calling it quits for the evening because I'm exhausted! I sat the pre on it's face, unbolted the output tranny, and tried shifting it around to see if the hum was affected.  Moving it around on the chassis plane did absolutely nothing, but I noticed as I started to rotate it away from the chassis, the 50Hz and harmonics started to drop.  I was limited by the wires in how far I could move it, so I could only rotate it maybe 30 degrees off axis, but my recordings suggest an 8dB drop.  So I've spent most of the night making a bracket out of some old aluminium angle, mounted the TX to that, and will affix it at 90 degrees to the chassis with a bit more scope to rotate it if I need to.

I can get the huge polyprop coupling cap above it and out of the way, opening up some space on the chassis for one, perhaps two more large caps for additional filtering.  I doubt it'll ever be perfect but it's certainly very useable!

A quick thankyou once again to everyone who has contributed help, advice, links, calculations so far - it really is appreciated

I'll post a futher update when I have made the modifications

Cheers!
Lee

 

[Tubetec]

Thats a good result 8db  ,
making your screen from steel would be preferable as Aluminium hasnt much abillity to stop magnetic interference .
Keeping long leads on the transformer until the final installation position is found helps , finding the right orientation of the transformer lowers it a good amount then its down to distance after that , I found moving a choke on a particular design of mine to the other side of a metal partition worked ok maybe about 15 db reduction  , but in the end I located the choke feet away on the signal chassis and all induced 50 hum vanished completely .

I was thinking for screening partitions maybe two layers of sheet steel mounted together with nylon bolts 10-15 mm apart  ,then placed in the path of the offending stray magnetic radiation , in fact three nylon and one steel  bolt might be best so a loop isnt formed but one plate is securely grounded through the other and screwed down to deck. Would probably be twice as effective as a single layer , another nice alternative or possibly in addition to partitions is transformer screening cans , I found one ebay supplier out of China who has three smaller sized cans round the 10 dollars each mark and free shipping , perfect for small chokes, input and output transformers .

I might do up a small example of the screening plate in a while seeing as the roads are frozen and work is 30 odd miles away.

 

[untune]

Thanks Tubetec

Yeah just the bracket is aluminium, just needed sometihng to mount at a 90 degree angle - I'm hoping that extra rotation is going to drop it even more, as it will put the OT, PT and choke at completely different orientations to each other.  I hink your idea of creating steel sheilds is good - it works for console/channel strip designs to keep transformers isolated so I think there's a good chance it will improve things more.

If I move the choke further down the chain it puts it smack bang next to my mic input transformer so that will be a last resort I reckon.

Quick question aside while it's in my mind - where do people consider the optimum point to ground the output transformer secondary?

It would be great to see an example of the screening plate!  I've walked in today but most of my team haven't showed up, snow has hit pretty hard.  Where are you based?

 

[Tubetec]

Heres a pic of the sheild I made ,

One piece is bolted with four bolts and spaced with extra nuts before the other section is placed on , now if you use the double end between your transformer and the source of your noise theres two layers of steel for the magnetic to get through . 

 

[untune]

Looks good! I can get hold of some raw steel sheets and maybe cut something to suit, I've cleared a lot more space on the chassis now.

I reorientated my output transformer on the bracket and the hum did fall by about 8dB - I was hoping for more than that, but it's still pretty good! A steel shield/partition might bump it down a tiny bit more.  While I was boxing it up I changed the EL84 cathode resistor back to 150R from 400R and the hum climbed back up by 2-3dB again so I'll swap it for a 470R to drop the current.  I didn't want to bias it too cold, but I don't really want more noise either.

As it stands the hum is practically inaudible at normal recording levels so it's very usable, I just want to get it as low as I can get it.  A few ground wires might still need to move and I need to follow Ian's advice about grounding the mic in/line out XLRs at the chassis.

I have plenty of room now to swap the 47uF filter cap on the EL84 for another 470uF which should drop things again - although I don't know if the 50Hz will fall with additional filtering.