OT centertaps

GroupDIY Audio Forum

Help Support GroupDIY Audio Forum:

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

5v333

Well-known member
Joined
Jun 30, 2013
Messages
631
Location
Gothenburg
Hi!

My make up gain circuit for my NYD EQ is more or less finnished but still something quirkie about the sound. hard to explain, like it has somekind of harshness/messyness but on scope it looks quite good and thd is very low on my computer when doing a round with sinewaves.

the OT is  9+9 : 1+1+1+1  and i have connected half of the sec as feedback to the first cathode like in the pultec.

yesterday i experimented with hooking the centertaps of the sec windings together with 4.7k resistor (without resistor it would osc like crazy) and even grounded the output ct.
this seemd to make the sound more open, breathy and sweet.
i cant fig out whats happening and how/if i should persue something here.
the OT feeds my DAC. i thought maybe the OT wants to see a transformer balanced load and felt better with a ground referenced CT for the job..?

sorry for the uncertainty but i cant get a good thought out of this yet.

any ideas??


 

Attachments

  • Tubeamp.png
    Tubeamp.png
    60.2 KB · Views: 54
you have a dual trace scope?  put pri and F/B waves on scope at same time,  look for phase shift as you sweep from 10 k hz on up to 100 k hz,

probably osc caused by f/b wind above audio range,  (ultrasonic osc) which affects output sound but invisible  maybe see iy will scope set to 100 k hz horz  sweep,

don't no if Lundahl designed that for F/B, you might drop him an email,
 

Attachments

  • 1689.pdf
    41 KB · Views: 8
Cj!
i just probed one of the pri ends and one of the FB ends

phase shifts:

22º = 50Khz
30º = 100Khz
90º = 175Khz
180º =  238khz

i asked lundahl about the transformer and he replied : it has been done before.

got to go now.
 
Ok thanks. good to know.

still, CJ mentions oscilation in the ultra range. could be oscilations elsewhere.
i have put alot of investigation in ground layout when it comes to rf oscilations. they move around depending on how the ground is layd.


about the output.
maybe im measuring the wrong way...

when probing the output i have put the ground clip of the probe on the centertap. effectivly grounding it. then i probe for ex one of the grids of the first tubes as the ref.
perhaps this shows how the circuit would behave if it in fact had a grounded center tap on the output...

i have tried to find a matching phase response on both halves of the amp.
flat response with no overshooting/ringing.
and also with the least of rf crap in the response.

i also have my ADC hooked up all the time to show how the amp reacts to external load.
 
i cant even find R23  ;D is it even there?

the phase splitting transfomer is not the best for its job, but it puts out a fairly high bandwidth and very low lofreq distortions. lundahl has some transfomers that states having great symetry for phase splitting, this is not one of them... the two phases vary quite alot. R30 is tuned so that booth phases hit 180º at about the same freq.
i think the phase symetry varies with source impedance, so again  not the best transformer for the job. the R´s and C´s on booth sides was the best i could find to make everything look somewhat good. it looks a bit mad right?

R40 and R41 has been there for a long time from when i was chasing some resonance. they never solved the original problem but they killed alittle bit of RF so they just has been left in.


hmmm. i just tried probing my OT with one end grounded with the clip.
now i see some overshooting that i havent seen before.
connecting the the FB and OUT centertaps seem to damp it...
 
abbey road d enfer said:
R26; my mistake.

if i remember... R26 was added to dampen some small ringing which i guess was due to the transformer windings in series with the caps. that value is also about the same as the DCR of each of the windings. it has four sec windings but they are paralleld.


 
CJ said:
do you really need  feedback, if it sounds good without...

thing is. it has never sounded quite right. but i start to see the end i think.
i have been chasing distortions for over a year and even if the sound improved every time i solved something, the big faults was still there somewhere.
when i managed the transformer feedback i knew that i wouldnt wanna go back.
bass is rather tight. if i could do even more feedback i prob would want that.
i think that with a 60hz sinewave at 10dbu output, i have about -77db thd.

without the transfomer feedback i think that was more like -50db.

the connection of the centertaps with a resistor seems like a good idea.i have 1k2 there know.
i think i might have to open up the loop again and redo some things.
 

Attachments

  • 60hz10dbu.png
    60hz10dbu.png
    981.4 KB · Views: 5
here is a pic of the output with two probes, no ground clip attached, none of the centertaps connected to ground or together and with no external load.

notice the diffuse state of the waveform.
 

Attachments

  • 20190214_120544.jpg
    20190214_120544.jpg
    1.9 MB · Views: 5
then i hook the output centertap to ground.


...more to come
 

Attachments

  • 20190214_120755.jpg
    20190214_120755.jpg
    1.9 MB · Views: 2
5v333 said:
This is when i connect my ADC via the XLR connector.
It shows a large impedance unbalance between the two legs. It is probably due to the xfmr's construction, but also don't neglect the possibility that your ADC's input stage introduces some unbalance too.
 
Abbey
i just had i look in the manual of my ADDA and both analogue inputs and outputs are transformerless balanced floating. could this explain the behaivior of my output?

grounding the output centertap definitely improves the sound.

but if this is only required when feeding OTL floating inputs, maybe it should be possible to change the circuit in some way...
 
5v333 said:
then i hook the output centertap to ground.

Ok, that's disturbing. In the few tests I've done with and without center tapped ground I did not see any meaningful difference, actually, none.  I'll have to explore this a bit. 
 
5v333 said:
Abbey
i just had i look in the manual of my ADDA and both analogue inputs and outputs are transformerless balanced floating. could this explain the behaivior of my output?
Probably not. I believe the xfmr has a dominant HF unbalance due to stray capacitance. Very hard to compensate.
 
:)

im going to disconnect the centertap ground connection and also the XLR and then hook up two 10K resistors to each output end, then tie the resistors to the internal ground of the amp to see how it reacts to a non floating load. (is that even the way you say it?)

im also thinking about how the windings are hooked up.
i asked lundahl which windings would be best for feedback and also if it was a good idea to invert the middle/feedback winding against the others to avoid positive feedback among eachother. they thought it was a good idea. but now i wonder if this might be incorrect...

heres a pic of how its connected.

 

Attachments

  • Feedback LL1689.pdf
    7.8 KB · Views: 8

Latest posts

Back
Top