Output Transformer questions / General Science and Electronics / Forums

Forums

4hv.org :: Forums :: General Science and Electronics

« Previous topic | Next topic »

Output Transformer questions

1 2 3 4 next last

Move Thread

LAN_403

Ash Small

Fri Dec 06 2013, 02:03PM

Registered Member #3414 Joined: Sun Nov 14 2010, 05:05PM
Location: UK
Posts: 4245

The subject of output transformers seems to be cropping up quite regularly at the moment, and I'm wondering if it's possible to series the primaries of a few transformers to provide the high impedancve required by the output stage of a valve/tube amp, and then parallel the secondaries, either to run miltiple speakers or to feed into one large speaker.

Would this work in principle, or am I missing something?

Shrad

Fri Dec 06 2013, 06:51PM

Registered Member #3215 Joined: Sun Sept 19 2010, 08:42PM
Location:
Posts: 780

the leakage inductance is a key parameter for bass response, so you have to either make an air gap by sawing a part of the core, or perhaps short one secondary to get higher inductance

the best solution, for what I know, is to use a double C core like the ones you find for old power transformers (old tape recorders, 70's and 80's era consumer electronics, etc) as they provide an easy solution for assembly/disassembly and testing, and you can disassemble, swap bobins, and reassemble to quickly test parameters

Ash Small

Fri Dec 06 2013, 07:18PM

Registered Member #3414 Joined: Sun Nov 14 2010, 05:05PM
Location: UK
Posts: 4245

Shrad wrote ...

the leakage inductance is a key parameter for bass response, so you have to either make an air gap by sawing a part of the core, or perhaps short one secondary to get higher inductance

the best solution, for what I know, is to use a double C core like the ones you find for old power transformers (old tape recorders, 70's and 80's era consumer electronics, etc) as they provide an easy solution for assembly/disassembly and testing, and you can disassemble, swap bobins, and reassemble to quickly test parameters

You're saying you 'want' leakage inductance?.....That could make sense. You certainly don't want the core saturating.

I have been looking out for some of those very 'double 'C' cores', but not found any yet. Maybe I should spend some time on Ebay this weekend?

I have been looking around at 'other core materials', apart from silicon steel. One that looks interesting, and which is used where corrosive conditions are encountered, like the magnetics in automotive fuel injection systems, is 430 series ferritic stainless steel, which, in the fully annealed condition, has 25% of the permeability of top grade silicon steel, and twice the resistance, so presumably half the eddy currents. You'd need cores four times the size, though, but you'd have a larger winding window, so less copper losses too.

It just so happens that dishwashers are lined with this very alloy, and it is thin sheet, so could be good for laminating. I have a scrap one out the back, which I may try cutting up and experimenting with. If I can find a scrap washing machine motor or two, maybe I can get the copper for nothing too? (I don't know how well this would work. Any comments would be appreciated.)

Sulaiman

Fri Dec 06 2013, 10:53PM

Registered Member #162 Joined: Mon Feb 13 2006, 10:25AM
Location: United Kingdom
Posts: 3140

A bit of distraction;
if I was going to make a valve amplifier I would seriously consider push-pull ultra-linear output

my school friend's father made an ultra-linear (mono days) that had a very good sound,
better than affordable commercial transistor amplifiers of the time
(but not as good as my STEREO transistor amplifier!)
Rather than make an even harmonic generator why not go for 'Hi-Fi' thermionics?
I have negligible experience with valves so I can't really help in this direction other than
a vague memory that the transformer for push-pull is much easier/cheaper
and both sound and specifications are better.
Even forgetting ultra-linear,
choosing a pair of valves for push-pull and winding a gap-less transformer
should be better, easier and cheaper ... or am I missing something?

Ash Small

Sat Dec 07 2013, 03:31PM

Registered Member #3414 Joined: Sun Nov 14 2010, 05:05PM
Location: UK
Posts: 4245

Sulaiman wrote ...

A bit of distraction;
if I was going to make a valve amplifier I would seriously consider push-pull ultra-linear output

Well, from what little I've read on the subject so far, single-ended class A amps are LESS efficient than push-pull output stage types, but, at lower powers, can provide the same, or better, low distortion levels as push-pull types.

The push-pull type will obviously provide more power because it has two valves, but the 1930's single ended type can provide the same low levels of distortion, is a lot simpler to construct, and seems to be an ideal 'first valve amp project'.

The big attraction of the 'Mighty Atom' design that I'm basing mine on is simplicity, and I've also heard one running. It will do for me

Anyway, back to output transformers, and I've found a place in Birmingham, UK whicvh looks like it supplies everything needed. Not sure about small quantities, though. Maybe we should consider a 'group buy'.

I'm still working my way through it, but they do supply 0.35mm laminations in various grades in all the popular sizes, along with bobbins, etc.

PDF is here:
]asco_components_2007_catalogue.pdf[/file]

I'll report back once I've looked up the various grades they offer.

Sigurthr

Sat Dec 07 2013, 07:56PM

Registered Member #4463 Joined: Wed Apr 18 2012, 08:08AM
Location: MI's Upper Peninsula
Posts: 597

Could you elaborate more on the relationship between leakage inductance and bass response please? I recently built a quick and dirty class-A (with AB preamp) single ended valve amp (it is in the projects section) but found that the bass response was too strong for my dinky midrange speaker, so I put in a high pass filter in the signal chain to roll off the lower frequencies. It certainly cuts back on sound reproduction quality, but is better than overdriving my speaker into all kinds of distortion. I was planning on just getting a better speaker for it, but if the issue lies with the leakage inductance of the output transformer network I am using I would certainly like to know before I get a different speaker!

Shrad

Sun Dec 08 2013, 12:56AM

Registered Member #3215 Joined: Sun Sept 19 2010, 08:42PM
Location:
Posts: 780

I guess it is related to the size of the bobbin, and the way it is constructed, as the bigger and heavier the output transformer, the better said sound quality

this is something I was told by a couple people already, and I remember having read things about that, maybe on diyaudio, but I don't know

maybe someone could try out a small NST? I have some small 15mA ones...

anyway, I'll have to experiment about this as soon as I have finished to tidy the room I use for electronics

I have a signal generator in my to-do list, with a small DDS synth and a PIC, to make some sweeps and profile response of such devices over a range...

Proud Mary

Sun Dec 08 2013, 02:10AM

Registered Member #543 Joined: Tue Feb 20 2007, 04:26PM
Location: UK
Posts: 4992

The primary inductance of the output transformer is the key determinant of low frequency response in a thermionic single ended amplifier.

Inductive reactance falls with frequency, which means the AF voltage appearing across the primary must fall as the frequency falls.

10H was often the chosen compromise inductance of primaries for O/P transformers in ordinary domestic audio applications using SE amplifiers, such as TVs, radios, and record players. At 10H we have XL = 188.5kÎ© @ 3 kc/s, 18.85kÎ© @ 300 c/s, 1.885kÎ© @ 30 c/s and a totally useless 188.5Î© at 3 c/s.

To improve bass response in costly hi-fi amplifiers which very few could afford, primary inductances as high as 50H were used in push-pull configurations using specially wound split-load primaries, where the tetrode or pentode screen grids were supplied by auto-transformer action provided by balanced complementary taps about 20% in from the two ends of the primary.

These high inductance primaries were wound with great cunning - multiple split bobbins etc - to keep the internal capacitance as low as possible, since the higher frequencies are always looking at any little bit of capacitance they can as a way to get out of the hard work of being in a transformer.

So the design of a valve audio output transformer must always be a compromise. It doesn't matter how much time and money is spent on designing and building a valve amplifier from the choicest parts, and the most wonderful low-hum super-stable layout, it can never be better than the quality of the output transformer.

Decide how much time, money, and grief you are prepared to spend on the output transformer, and design the rest of the amplfier to the build quality, bandwidth, linearity, of this weakest link. Any other approach is a waste of time that can only result in disappointment.

Rather than try to re-invent the valve output transformer, why not simply copy the winding and core details generally included in the constructional plans of advanced amplifiers built by Mullard and others to showcase their valves?

Sigurthr

Sun Dec 08 2013, 09:04AM

Registered Member #4463 Joined: Wed Apr 18 2012, 08:08AM
Location: MI's Upper Peninsula
Posts: 597

Ahh, thank you for the info, PM!

My output transformer measures in at 5H, I wonder if it would be worth it to use a second one where I series the primaries and parallel the secondaries to get to 10H. This would also better match the output impedance to my speaker, though my speaker can't handle the bass as is so perhaps it isn't worth doing until I get a better speaker. Then again there is a noticeable 60Hz hum on the amp anyway, greater bass response may worsen that. I may just file the info away for later when I get around to making a high quality all tube amp.

Ash Small

Sun Dec 08 2013, 12:42PM

Registered Member #3414 Joined: Sun Nov 14 2010, 05:05PM
Location: UK
Posts: 4245

Sigurthr wrote ...

Then again there is a noticeable 60Hz hum on the amp anyway, greater bass response may worsen that. I may just file the info away for later when I get around to making a high quality all tube amp.

I understand you can eliminate a lot of the 'mains hum' by powering the filaments from a DC supply.

Also, thanks to PM for the 'textbook stuff'.

I basically look at it this way, a larger core will improve bass response at low frequencies, by avoiding saturation, but will increase eddy current losses at the top end.

The only way to reduce eddy currents is thinner laminates, unless you use a smaller core, or a core material with a higher resistance, so you want large core, and thin laminates.

You then need to keep the self capacitance to a minimum.

I'm still not sure about the effects of 'leakage inductance'.

BTW, I've found a 5k Ohm output transformer in my shed, which I'll use as a 'starting point', and for comparison with others that I may try later.

1 2 3 4 next last

Moderator(s): Chris Russell, Noelle, Alex, Tesladownunder, Dave Marshall, Dave Billington, Bjørn, Steve Conner, Wolfram, Kizmo, Mads Barnkob