Best way to orientate stereo output transformers

[Tubetec]

I have a single chassis housing a large power transformer , two chokes and a pair of output transformers . Do I want the outputs 90 degrees to each other or is my main concern avoiding feild from the mains transformer/chokes ? Noise was already quite low ,the outputs living in the shaddow of the chokes , I may well end up with worse induced noise when I change the transformers .

 

[Tubetec]

I dont have the space to mount the C-cores lenght ways on the chassis ,
Using an off the shelf metal L bracket allows vertical mount with a smaller foot print .
I can experiment to find the point with minimum induced noise or minimum interaction before commiting to a final layout .

 

[Ciro Marzio]

If the induction in those OPTs is reasonably low you should not have significant crosstalk positioning them as in your picture.
I would be more concerned by noise induced by power chokes and transformers

 

[Tubetec]

I see in Audio Note gear they often mount the c core output transformers at 90 degrees to each other ,
because I dont have much space if I mount both this way their cores have a 45 angle of incidence to the mains transformer and chokes , which probably leads to worst induced noise level .
I can also place them back to back , that way the substantial metal mountings and brackets form a magnetic wall between the two transformer cores , they also end up at 90 degrees to the mains transformer field but at maximum interaction with the chokes as the cores are parrallel to each other .
I'll post a few more pics with the possible layouts for clarity , theres no way of knowing what will happen in practise when the different sources of noise interact until we try it .

Added,
After taking another look this is more or less the only option due to chassis constraints and making it so it fits back into the enclosure , the tube sockets can also remain where they are , anything else would require more reworking .

All of the iron is way overrated in terms of current carrying capabillity so that should help keep radiated noise down , the power transformer was from 30 watt 4xEL84 amp , the chokes are around 20H at 150ma .
Output transformers are gapped for around 65ma but in this case Im running the EL84's with around 30ma /220v at the anode /screen . Its more like the kinds of voltages and currents you find in a tube radio than a hot rod 5W single ended that eats tube for breakfeast dinner and tea . Ive got 20 years use from the tubes I put into this box when I renovated it , it was permanently powered for around 5 years at one point ,never a glitch hiss or thump.
I didnt put this thing up on the scales yet but I dont think its an ounce under 30lbs and just 1.5w per channel for a total power consumption of around 50w . Doubles as a small space heater very nicely in a colder room for instance .

 

[thomasdf]

I usually crank my volume up and move my transformers around until I find the quietest spot (to my hears at least). I also do that with the mains transformers

 

[Tubetec]

I looked at things with fresh eyes today and this arrangement seems to work in terms of physical layout at least ,
I drilled the metal bracket to accomodate the tube socket ,below you can just see the cathode bias pot , the tubes sit alot higher in the cabinet , before I had the bias adjust on the deck it was hard to adjust with the unit installed in the cabinet , now I can just drill a hole in the alu-mesh so I can access it externally .
The iron is arranged in a way that should minimise magnetic coupling ,obviously the mounting plane of both the mains and output transformer cores are the same but Im hoping the chokes provide some sheilding .

I pulled out some of the original caps I installed , they were from a batch I got almost 30 years ago , the 100uF blues were rated at only 250vdc , good for the job as long as the PSU was never switched on unloaded ,which it never was . I think they can safely be retired now in favour of some much smaller 100uf 330v flash camera caps
Previously I had allowed only one stage of decoupling per channel , ie cap, choke, cap, choke,resistor,cap,to two resistors each with its asscociated cap , with one channel was driven into distortion , the other with its grid to ground , the harmonics were audible via the psu in the undriven channel output ,
I plan on changing the earlier smoothing stages to dual path resistive and putting a choke/cap filter feeding the plate supply per channel , that should prevent any unwanted interactions that mush up low level detail on the peaks . Ive used a dual concentric 50k audio pot on this unit all along , mostly it sits in the fully on position for +4dbv line level drive . Im not sure I can do with the loss of individual channel trim on this unit but one of the dact style smd mini pots would do a very nice job for repeatabillity factor /interchannel channel balance.

The other thing is the internal speaker wiring , Ive it wired to combine the outputs of both channels to mono when the headphone jacks are removed , its a neat way to small speaker/mono compatibility check a stereo signal .
I have a set of reclaimed nokia carphone speakers in sturdy plastic enclosures , there 2ohm as far as I remember , I was planning on mounting them up in place of the single eliptical speaker , to the eye it all fits , just need to get all the parts together and size them up to be sure .

The input level requirements of the EL84 is a good match for modern audio interfaces , in this case its a little under powered for my output transformers , I grab can grab an octal socket and see how it mounts up on the plate for maybe a bit of 6V6 flavour. Anything bigger needs a driver tube which Id prefer to avoid .
I'll post more details on the O/P transformer later .

 

[Tubetec]

These C core transformers I got off ebay from a Chinese maker.
3.5k pri 4,8 ohms sec
They appear to be quite well matched ,the only measurement I took so far that varied to any great degree was primary Rp measurements with the LCR meter ,
Inductance ,resistance figures were in close agreement with the makers specs .
Its gapped for a bigger tube , with upto 65 ma anode current , reducing the gap a bit might get a bit more than the 26H it has now .

I measured the secondary taps with the lcr and they gave 3 and 6 ohm Rp readings at 100hz on the 4 and 8 ohm windings with the primaries unterminated .Likewise primary Rp measurements were made with secondaries unconnected .

An octal socket isnt far away from fitting the pre existing diagonal holes on the L bracket , propped up on a couple of brass pillars , bomb proof,
a side by side shootout , EL84 vs 6V6 is on the cards .The ears combined with REW will give the full picture,

I dug out a bunch of tag strip today to mount up the caps/dropper resistors in the old fashioned way
I have a free supply of 100uF HT caps so lets go to town ,
I found a 10uF440v MKP motor cap I'd like to use after the rectifier , then dual multi section RC filters with eletrolytic caps ,WW resistors with the chokes in series with output transformer primaries apart from a final film cap of around 4uf placed close to the o/p transformer HT connection .

 

[Disco Volante]

Cool project. I gather it's a rebuild of something you did decades ago? What's it been used for?

 

[Tubetec]

Hi Disco ,
I made this unit originally to power headphones , I got tired of the usual suspect op amp based H/P amps distorting on transients and causing distress to myself and the people Im lucky enough to be recording .
As I said previously I kept the internal speaker , its usefull for checking mixes for how they might sound on small open baffle speakers , like in a portable radio or tv for instance .
I also use this on ocasssions to drive reverb tanks which it works well doing, more than enough power for that job .

My latest plan is to use this amp to power my studio monitors , of course its only capable around 1.4watts at high distortion , but even at higher drive levels where distortion is apparent ,its not anywhere near as aggravating as op amp distortion which tends to become really ugly really fast .

Ive read plenty of articles on EL84 based SE triode amps , most seem to skimp on PSU smoothing with only 1 or 2 sections , Ive found much better results can be got with anything upto 5 RC or LC sections , previously I had noise levels of around -90 in this amp ,just about detectable by an ear very close to a speaker , but with headphones a tiny amount of residual hum remains and is audible .

As I said previously I plan on upgrading the psu filtering , of course at some point this will cease to matter when induced noise becomes the dominant factor . I plan on using MKP motor run caps as first filter and last , just before the o/p transformer . As we noted from the amp techs here ,the first filter cap in a HT supply has to withstand the majority of the ripple current , this leads to extra wear ,which over time changes the value of the capacitance , a foil motor cap is rated for much higher voltages and currents than you'd ever see in a tube amp , so it should have a long easy life and keep a very stable voltage off the rectifier , I'll place good quality modern electrolytics next with resistance droppers and a final foil cap after the choke before the o/p transformer.
Again the foil has the long term stability so our LC filter stays tuned and doesnt tend to drift . The electrolytics may vary a bit over time but it shouldnt effect performance to any great degree for a couple of decades at least.

 

[Disco Volante]

Thanks, I'll be following your progress;-) Those OPTs seem very nice, I like that they're well overrated;-)
Currently gathering parts for a small PP amp along the same lines. Have some PCL86's and some tiny PC92's lying around...