100Hz hum in preamp

GroupDIY Audio Forum

Help Support GroupDIY Audio Forum:

This site may earn a commission from merchant affiliate links, including eBay, Amazon, and others.
Hi Scott

Green/yellow is the mains earth, it's connected to the chassis but not the pot - the pot is just obscuring it I think  :)
 
Oh yes forgot to mention , I found with single ended EL84 ,triode connected around 220volts at the anode was the max, early 30's ma current , I think the screen grid seemed to draw a disproportionate amount of current if I pushed the voltage much higher than 220v without a current limiting resistor in series. Going 250 or 270 volts in triode on the EL84 does nothing for your sound , and I tested that out with both transformer and resistor load

What I found also was, once I had 'brute forced' out Ht hum , now Lt hum became the dominant factor ,then induced hum in the output transformer ,grid to ground .

A picture  might give some clues as to improvements to layout ,  at some point in your quest for 'pure resistance noise' in your outputs ,  mains Tx will become the limiting factor , you could consider housing mains tx , rect ,first filter or more seperately ,

Methodically working your way down through the various sources of noise is great fun ,Ive never used anything buy my ears as the judge , but the benefits of REW  are obvious
 
Tubetec said:
Oh yes forgot to mention , I found with single ended EL84 ,triode connected around 220volts at the anode was the max, early 30's ma current , I think the screen grid seemed to draw a disproportionate amount of current if I pushed the voltage much higher than 220v without a current limiting resistor in series. Going 250 or 270 volts in triode on the EL84 does nothing for your sound , and I tested that out with both transformer and resistor load

What I found also was, once I had 'brute forced' out Ht hum , now Lt hum became the dominant factor ,then induced hum in the output transformer ,grid to ground .

A picture  might give some clues as to improvements to layout ,  at some point in your quest for 'pure resistance noise' in your outputs ,  mains Tx will become the limiting factor , you could consider housing mains tx , rect ,first filter or more seperately ,

Methodically working your way down through the various sources of noise is great fun ,Ive never used anything buy my ears as the judge , but the benefits of REW  are obvious

Thanks for the input Tubetec; see Scott's post above for the link to the pictures

So based on PRR's calculations and your observaions, the focus needs to be improving HT filtering and lowering voltages.  I did originally plan for lower voltages, but ended up leaving them higher.  Can I ask, what kind of sound differences running a triode EL84 at lower/higher voltge makes?

Am I looking a recalculating all dropping resistors, or would I be able to get an additional RC filter stage in, let's say just after the choke, before the EL84?

It's a while since I worked everything out and I still have load lines and everything filed away somewhere, can't remember exactly what plate current I calculated for but 30mA sounds familiar.  Last night I bumped the cathode resistor up to 400 ohms (2x 200ohm temporarily while I wait for a 470ohm) as abbey road suggested

Looking at my pictures I've just noticed that my OT is grounded at the same point as the mic input stage... I should shift that up to the base of the first filter cap
 
Your focus should be first on determining where the hum gets in. Several of us have suggested the last stage as the most likely candidate but you need to confirm this by testing before setting of on the redesign trail. As I suggested before, short the output tube grid to ground and measure the hum level. If it is largely unchanged then the output stage is where you need to concentrate.

Even then, you still do not know for sure if the hum gets in through the HT or induced in the output transformer. So your next test should be to disconnect the op transformer and replace it with an equivalent resistor and measure the hum at the output  again. If it goes down somewhat then you have an induction problem. If it is unchanged it is the HT supply.

Divide and conquer.

Cheers

Ian
 
ruffrecords said:
Your focus should be first on determining where the hum gets in. Several of us have suggested the last stage as the most likely candidate but you need to confirm this by testing before setting of on the redesign trail. As I suggested before, short the output tube grid to ground and measure the hum level. If it is largely unchanged then the output stage is where you need to concentrate.

Even then, you still do not know for sure if the hum gets in through the HT or induced in the output transformer. So your next test should be to disconnect the op transformer and replace it with an equivalent resistor and measure the hum at the output  again. If it goes down somewhat then you have an induction problem. If it is unchanged it is the HT supply.

Divide and conquer.

Cheers

Ian

Good point Ian, I'm getting ahead of myself; I'll try shorting the grid tonight :)

I tried a handful of small 'improvements' to see if anything noticable happened; moved the 6.3v heater ct bck to the EL84 cathode, tightened up the PT bolts, inserted a tough rubber washer between PT and chassis incase there was any subtle vibration, shifted the output XLR ground point, and added a small metal 'shield' pinched from another chassis between the choke and OT.  Very negligible difference but it appeared that the 50Hz peak, altohugh lower than the 100Hz, may have dropped off.  But I'll investigate that once the bigger problem is sorted  :D
 
When I said raising the voltage did nothing , it didnt change anything soundwise really  , appart from seeming to raise hum in general ,and draw more current through the tube .

Wow yeah I just looked at the photos ,you had your work cut out just trying shoehorn all that in Im sure , usually what I found best is mains and signal tx's at 90 degrees to each other ,in other words the orientation of the cores relative to each other . Rotating the various iron cored components relative to each other will very often show marked improvement in induced noise at some position  , distance and a layer of steel help also ,  but do bear in mind sometimes times different hum sources can be additive or subtractive depending how things are arranged , tieing the grid of the EL84 to ground kills all noise from previous stages ,extactly what you want when checking induced hum from mains to output tx for instance . in the building stages it often very usefull to keep the leads on the transformers long ,just so you can experiment with positioning before finally bolting things down to chassis , its  impossible for me to know the best positions even from the look of it as the best arrangement will be completely unique to that project and only found by experimenting , its a bit hairy moving things around inside a live chassis , so be very very carefull if you do , add extra layers of insulation and adhesive tape in the short term to prevent unwanted contacts .instinct would say move the output transformer as far from the mains as you can and even add a metal partition if possible .
Looks like an old Akai/Roberts tape pre amp you used as donor for your project , old tape preamps are a great starting point for mic amp projects ,but  adding input/output transformers will always bring the possibillity of induced hum  . 
As you search out and destroy the dominant source of noise ,the next most dominant source will rear its head and so the hunt begins again and again several times until you hear the 'hush' of pure resistance noise ,with the  mains tx on the signal chassis and ac heaters  you will have to end up accepting some small  level of hum in my experience .
 
Tubetec said:
When I said raising the voltage did nothing , it didnt change anything soundwise really  , appart from seeming to raise hum in general ,and draw more current through the tube .

That's good! Soundwise I'm pretty happy with it, just that hum getting in the way  ;D

It's a tight fit but there is actually a bit more room than I thought in places - I could cerainly swing the OT further away but I'm not sure if it might interact with the mic input TX too.  I have the space to bolt some bigger, fatter filter caps to the back of the board or onto the chassis and still have them in the right spots for each stage.  I might have to drill new holes for the OT but if it cures the noise it won't matter.

I have used a metal partion pinched from an Akai M8 but it's not really big enough...  I do have a few sheets of mesh left over from a microphone project - could that be useful, even if just for testing?

It is indeed an old (very old!) Akai chassis and it's solid and looks good, seemed a shame to waste it.  I'm prepared to accept some level of hum, this was more of a learning experience for me and to be quite honest I'm surprised it sounds as good as it does!  Just in the last couple of days the noise level seems to have fallen a few dB from about -62 to closer to -67, and I'm using very old EL84 valves for testing.  I'm confident that a bit more tweaking will bring it further down.

I'm tempted (depending on what the grid short reveals) to add another RC stage just to see what effect it has - is there an ideal place to put another stage, and can anyone suggest a resistor value?  I have another 47uF left over but how I calculated the dropping values last time has just gone from my mind!  I guess I should be aiming to drop that 300v down to about 220v, if only to help the valve live longer
 
untune said:
I'm tempted (depending on what the grid short reveals) to add another RC stage just to see what effect it has - is there an ideal place to put another stage, and can anyone suggest a resistor value?  I have another 47uF left over but how I calculated the dropping values last time has just gone from my mind!

I can't find the link here but I've attached Ian's HT design paper and it has some good info that may help.....
 

Attachments

  • HTdesign.pdf
    21.7 KB · Views: 12
Awesome, fantastic resources, thanks Scott and Ian  :)

Also, confirmed, shorting the EL84 grid to ground, hum persists at the same level
 
Ive got great pleasure over the years rejigging old bits of valve gear into more usable stuff .

PSUD2 is  indespensible in the war against hum ,might take a small while to get the hang of it ,its always farily bang on accurate for me though , within a handfull of volts of actual circuit readings . If your out of large value caps you might find one in an old desktop pc power supply ,even just for experimentation until you order some new ones . As I said earlier I tend to put the RC sections after the rectifier followed  by the LC's later , but in terms of ht volts at the anode it wont matter much . 

A neat way to drop or raise ht voltage with a tube rectifier is by altering the value of the first filter cap , there is always an upper limit depending on the type of tube used though ,40uF is listed for the 6x4  . Sometimes resistors can be included in the leads from the ht windings or directly after the rectifier cathode ,you could use a combination of any or all these ways to lower hum and get the right voltage at the anode of the valve .

As you see from PRR's simulation with  3 filter stages your getting into good S/N territory , dropping another 20 volts on a fourth section wont make much difference to the output of your stage but it will mean that the contribution HT ripple makes to overall noise output is  very small indeed  , going that little bit further in reducing ht noise is worth it because as you subsequently  remove other hum sources you could find HT hum might once again become the limiting factor .

Once you do get theHT  noise to a satisfactory level with the grid grounded , its time to remove the ground link , now your going to hear a mix of induced hum in the grid wiring and some noise from the 6.3 volt ac coupled in , I cant see if you have screened cable right upto the grid of the EL84 , but thats a very good plan in any case , chopsticking and moving heater wiring can also find some extra db's but I see you've already done a neat job wiring the heaters up and away from signal and with tightly twisted leads .
 
PSUD2 was always a little confusing for me but I do still have it installed... my dad has some old PC PSUs so I might see what I can pinch but I've found that I can get my hands on 470uF caps rated for 400v or 450v.  I'm using 500v axials at the moment but the 470uF are too big and are in cans that I'll have to attach with brackets and solder direct to pins.  There should be room though!

In terms of positioning I can probably fit another stage after the choke, I can potentially fit another between choke and the reservoir cap although I don't know exactly how that affects the pi filter formed by the 22uF, 6H and 47uF.  It could potentially be 22uF, 470uF, 6H, 470uF there.  I've calculated tht a 2K2 5W resistor should drop about 70-80V at the EL84 stage right after the first smoothing cap and lower everything down to what I originally planned for but I might have to move some values round again, lowering the current draw of the EL84 also.

I had considered changing the value of that reservoir cap but I read conflicting info about the 6x4 all the time -some datasheets say 10uF, some say 20uF, others say 40uF.  I don't want to negatively affect the lifespan or potentially damage anything.  I could put another 22uF in parallel with it, or swap it out for a 47uF but I worry that'd be going too high.

PRR's calculations are incredbly useful and I could lose the choke but then, I wonder how that would impact the overall sound.  One of the goals was to keep a bit of character (and recycle as much as possible) rather than going for anything too clean - I've got other pres for that :D

Only the input stages use screened cable - the instrument inputs to V2, the mic XLR and line inputs to V1. Everything else is either twisted or just standard stranded.  There are some long single wires tht run between pots and a switch, they might benefit.  Plus I've neve rbeen completely satisfied with the twisted pair running to the output XLR, as they pass the power quite closely.  But I'll focus on fixing that later :)
 
untune said:
I had considered changing the value of that reservoir cap but I read conflicting info about the 6x4 all the time -some datasheets say 10uF, some say 20uF, others say 40uF.  I don't want to negatively affect the lifespan or potentially damage anything.  I could put another 22uF in parallel with it, or swap it out for a 47uF but I worry that'd be going too high.

Here's an interesting thread that may help

https://groupdiy.com/index.php?topic=16479.0

untune said:
PRR's calculations are incredbly useful and I could lose the choke but then, I wonder how that would impact the overall sound. 

Yeah, having stuff built already makes me want to question things...lol

It's usually never as bad  or take  as long as I think it's going to  when I actually try something.....

Also, another useful article on grounding I often look at is the Aiken amps page

http://www.aikenamps.com/index.php/grounding

and Ian's grounding paper I've attached...(I really need to save the thread links :-[)

Of course it's just some more info to add to your arsenal if you don't already have it...... never a one size fits all thing yet for me..
 

Attachments

  • grounding101v2.pdf
    339 KB · Views: 6
> some datasheets say 10uF, some say 20uF, others say 40uF. 

Do you know the difference between Rating and Typical Characteristic?

It's never about the uFd. It is about the Ohms.
 
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,  probably a good plan for a class A  preamp where current draw isnt massively different dependant on signal or not and we can afford to loose some volts in any case .
Just dug out a large wall mounted charger case from scrap to house my newest psu project , plenty of slots and air flow , and plenty of room so things arent crowded either .
 
Scott; Yeah it's a bit of a faff to have to rebuild things but I'd rather be doing that here where everything is just soldered point to point, it doesn't take much effort to reroute a few wires :) I was just curious if removing the choke would have a perceived effect on the sound (noise aside :p) or if that would only happen in say, a guitar amp where you're really pushing things hard.  Lots of talk of RC filters 'stiffening' the supply but I'm not sure exactly what that's meant to mean ;D I'll ditch the choke if absolutely necessary but I'll need to do some recalculation.

Thanks for the additional grounding info!  As a side note I tried moving the ouput transformer and reorienting it on the chassis in relation to the PSU/choke and there was no perceived difference. Perhaps 1dB less, if that.

PRR; do you mean the plate resistance?

Tubetec; Interesting  - is that comon practice?  I've seen resistors in series with reservoir caps but it doesn't seem common

I've ordered some WW resistors to experiment with the v drop, and some bigger caps to smooth the supply.  If I can lose 70v ish right after the first filter cap, the HT voltages at each stage should be more reasonable

EDIT: Small aside... I'm using a step down transformer to take my UK mains voltage down to the 117v expected by the transformer in the preamp.  Is there any chance that this could also contribute to the noise?
 
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 .

Below is the schem of what Im using , a pair of 6x4's with a pair of silicon rectifiers to form a bridge ,  I have 2x6.3vac  1A  secondaries, its could permit series heaters at 12.6 ,im not sure which is best ,I also see each 6x4 handles the full ac voltage across the transformer  at each of its anodes ,would it be better if  each tube carried either the upper or lower part of the waveform and so have reduced ac voltage to handle ?
 
Could  a centretapped heater supply with the tap referenced to the cathodes after the resistance  work better here?

 

Attachments

  • Dual 6x4 board.PNG
    Dual 6x4 board.PNG
    236.4 KB · Views: 9
Tubetec said:
I also see each 6x4 handles the full ac voltage across the transformer  at each of its anodes
That's one of the advantages of the full bridge rectifier, compared with the center-tapped secondary arrangement, where the anodes see twice the secondary voltage.

,would it be better if  each tube carried either the upper or lower part of the waveform and so have reduced ac voltage to handle ?
How would you do that? One option is to have a custom transformer feeding separate rectifiers that would be placed in series. Moot in consideration of the fact the 6x4 can withstand 1250Vac.
 
Could  a centretapped heater supply with the tap referenced to the cathodes after the resistance  work better here?
No need for a center-tapped heater, but you need a connection between the heaters and the cathodes.
 
Back
Top