transformer build thread

[abraxalito]

I was thinking that noise other than CM noise transmitted on the signal lines might cause degradation in sound & the trafo may have rejected this due to it's bandwidth limitations?

In EE parlance, noise is either 'normal mode' (i.e. differential, meaning the inner conductor carries it in one direction - forward - and the return path is the outer conductor or shield of the IC) or 'common-mode'. In the CM scenario both conductors are passing noise in the same direction and the return path is another conductor somewhere else - normally the mains wiring.
Does that help any? So all noise is either DM or CM or a combination of the two.

There may be LF or HF noise outside of the trafos pass bandwidth that would be rejected.

Ah so you are thinking that the transformer is going to act as a low-pass filter on the DM signal. I suppose it will but it'll be hard to quantify its effect. I've checked the passband of one of my trafos into a light load (such as an amp input) and its well into 100's of kHz. The low-pass effect would be unchanged when the two sides were shorted together, which is how we started out if you recall? With just the trafo in circuit and no isolation the SQ improvement was negligible if I remember right?

Oh, I didn't realise this choke was meant to go in series with the trafo!

I noticed a few years ago that this technique is used in some ethernet isolation transformers - combining a CM choke with a trafo. Seems to me like a trick too good to be left only to the digital boys...

[jkeny]

I was thinking that noise other than CM noise transmitted on the signal lines might cause degradation in sound & the trafo may have rejected this due to it's bandwidth limitations?

In EE parlance, noise is either 'normal mode' (i.e. differential, meaning the inner conductor carries it in one direction - forward - and the return path is the outer conductor or shield of the IC) or 'common-mode'. In the CM scenario both conductors are passing noise in the same direction and the return path is another conductor somewhere else - normally the mains wiring.
Does that help any? So all noise is either DM or CM or a combination of the two.

There may be LF or HF noise outside of the trafos pass bandwidth that would be rejected.

Ah so you are thinking that the transformer is going to act as a low-pass filter on the DM signal. I suppose it will but it'll be hard to quantify its effect. I've checked the passband of one of my trafos into a light load (such as an amp input) and its well into 100's of kHz. The low-pass effect would be unchanged when the two sides were shorted together, which is how we started out if you recall? With just the trafo in circuit and no isolation the SQ improvement was negligible if I remember right?

Yep, I don;t even know if there was an SQ improvement when we initially tried the trafo in Nige's setup - when there was no isolation. Changing it to isolated config certainly did jump the SQ - so yes, I agree this seems to be where the greatest SQ improvement is to be found

Oh, I didn't realise this choke was meant to go in series with the trafo!

I noticed a few years ago that this technique is used in some ethernet isolation transformers - combining a CM choke with a trafo. Seems to me like a trick too good to be left only to the digital boys...

Excellent lateral thinking - I like it :)

[abraxalito]

The Chinese transformers you found have arrived and boy are they tiny... quite a bit more petite than I imagined.

The laminations look to be held together by wax of some sort and the primary inductance is higher than stated, 1.2H at 1V excitation and 50Hz. Its possible they measured the inductance with a very small signal and at a high frequency to get their lower figure.

[jkeny]

The Chinese transformers you found have arrived and boy are they tiny... quite a bit more petite than I imagined.

The laminations look to be held together by wax of some sort and the primary inductance is higher than stated, 1.2H at 1V excitation and 50Hz. Its possible they measured the inductance with a very small signal and at a high frequency to get their lower figure.

Wax is probably better than varnish for possible disassembly?
I wonder what they are being used for as there seems to be bucket loads of them on ebay & Taobao?
Do they characterise as useful in audio applications??

[abraxalito]

I'll hook one up to the signal generator and AC voltmeter today and put it through its paces. Curious to know how good its low-frequency response is.

Along with the transformers I also received my STM32F072 development boards so I have a lot to be getting on with...

Results from the low frequency response testing - they're not going to be any good for line level audio. Very noticeable distortion at 50Hz, 1VRMS on my scope. Just about passable at 50Hz with 500mVRMS. So if you want full audio bandwidth out of these you'll be limited to about 200mVRMS, at this amplitude they'll do 20Hz, just. Or if you don't much mind bass distortion, give them a whirl. HF-wise, no issues with frequency response, they'll do well above audio.

I stripped off the tape from the outside so you can see (although rather a fuzzy shot) what the trafo looks like 'under the hood'. Of interest here is the spare space to either side of the winding, this tells me that if I can get the wax out of the core (or alternatively find a source of the raw laminations and bobbins) there's room for improvement by adding more turns and/or using thicker wire.

It turns out not too difficult to remove the laminations - just cover the trafo with freshly boiled water for about 15s or so then extract it and while its still warm the laminations aren't too sticky to each other. I needed to give mine two treatments to get all of them out. Then one weakness of the design became apparent - the laminations are all E with no I. This makes the core cross-section half what it could be. Not sure where I could find I laminations though, perhaps butchering a second transformer could provide me with some...

[jkeny]

Very good, Richard thanks for that
So this is an E trafo instead of an EI trafo?
I don;t understand how this is constructed - E & interleaved E face each other? Makes for a difficult construction.
Do you consider it a cheap source of laminations - are they steel?

[abraxalito]

It has only E laminations so each layer is a single E where the Es are turned around 180deg on each layer. Which fills in the central pillar with steel very nicely but the rest of the magnetic circuit has gaps in it meaning it can't support so much magnetization. I reckon they're steel as nickel (as claimed) would be extremely expensive. But I don't know of a way to test what they're made from.
I've never bought just laminations so don't know if buying the whole trafo is a cost-effective way to obtain them. If you value your labour, definitely not!

[nige2000]

i see line level traffos here at premium prices
https://www.don-audio.com/Audio-transformers



how do we figure out what impedance we require?

example
the impedance of soekris is 625ohm, so impedance od trafo needs to be circa 600ohm?
output impedance of trafo would need to match input of amplifier so if it were feeding a potentiometer for volume control that was 15k
ideally we would need 15k impedance on secondary ?
but is having a 600z input and 15k output transformer not a large step down in voltage aswell?


http://www.n9xh.org/License/PCARA_Gener ... son_08.pdf

what about something like this
https://www.don-audio.com/Edcor-PC10K-600_1
https://www.don-audio.com/Edcor-PC600-600
https://www.don-audio.com/Edcor-WSM600-600

[abraxalito]

how do we figure out what impedance we require?

I still haven't figured that out for myself! In general at line level, higher impedance is better. So if you see 10k go for that. 600ohm would be fine for an output stage happy to drive 600ohm headphones but most line stages will be too wimpy.
In the case of Soekris, that 625 is the output impedance meaning its very wimpy indeed. I guess that's just passive resistance from the R2R ladder? In which case preferably buffer it with a classA buffer (not a classAB opamp) before feeding into a trafo. If you have some BUF03 they'd do a good job but they're expensive nowadays.

the impedance of soekris is 625ohm, so impedance od trafo needs to be circa 600ohm?

In this case definitely not! Too low a load impedance, an R2R DAC needs as close to infinity as possible on its output.

output impedance of trafo would need to match input of amplifier so if it were feeding a potentiometer for volume control that was 15k

No need to match, preferably sources should be low impedance, loads high impedance. Matching would lose you 6dB and that would be the least of your troubles....

but is having a 600z input and 15k output transformer not a large step down in voltage aswell?

Most likely the opposite in fact. 15k means high inductance which means many turns which means higher voltage. So 600ohm in, 15k out is a step-up transformer.

[nige2000]

i see

10k/10k 1:1
15k/15k
15k/600
10k/600
on there

whats most suitable for direct out of ladder input impedance on trafo?
just to get the ball rolling