Small la2a transformer challenge

GroupDIY Audio Forum

Help Support GroupDIY Audio Forum:

This site may earn a commission from merchant affiliate links, including eBay, Amazon, and others.
I cheated a little bit. Remember, I have a 7903 right here on the bench.

I thought that the goofy level readings around 15 hertz on Brand B were something wrong with your tests, but when I put my 7903 on the scope and saw that instability on the low end with a square wave and 600 ohm load, I knew I had one choice down, as low end instability is pretty rare among transformers.
Also, thats a mic input being used as an output, so I figured that the winding structure might not be the best for an output transformer.

So then I was down to a fifty/fifty chance. I figured that Sowter had about 40 more years on UTC as far as technology is concearned, and that their transformers seem to work well as a group, so that led me to believe that the better curves belonged to them. I also looked at their data sheets. They have a big EI framed core with 50 percent silicon, so I figured that their core would be the strongest as far as low frequency vs loading is concerned. The loading curves helped a lot because you were only loading the output iron.

I also know that Sowter tilts their response towards the high end a bit more than UTC from listening tests. This is probably due to the small high Ni core, where as UTC has always been great from 20 hertz on down, and sure enough, if you look at the low end distortion specs, UTC actually beats out Sowter at 10 hertz. This would be the bigger UTC input core taking over.

But I don't know. You could come back and tell me that you made a mistake and that the order is backwards, and I could probably find new reasons to make that plan fit. Why? Becuase transformers are weirder than I am.

If you guys are still set up, I would love to see that distortion vs level chart done at a low freq, like 50 or 20 hz.

And thanks for this cool test!!!
:thumb:
 
the reason this thread was started as both myself and keef were rather surprised at what we observed.

Our instincts were to have expected better performance from the lundahls.

Once we ran the sowters against the UTC's our attention was grabbed.

I think theres a lot of value for the dollar sitting in those sowters. Compared to how difficult it is to find the UTC's coupled with the ever rising prices on ebay.

Look into the coming weeks for some wave files comparing the UTC's, lundahls, sowters AND Jensens :)

im just waiting on getting the Jensens...

it'll be a blind test :wink:
 
Good job CJ!

Not fair though, having a Lundahl to test and all.... :razz:

I am curious as to why the Lundahl doesn't acheive their spec'd max level for 1% TDH though. The 1922 should get to +26 and I sheepishly admit not checking out the specs for the OP tranformers. How is the output transformer strapped? I'm guessing 2:1?

Which would I rather have....hmm I'm guessing Lundahl 1922 input and Sowter output. Maybe the Jensens will change my mind.

Thanks for the test guys...

Kris
 
OK, I've clearly got a long way to go on the whole understanding iron thing :oops: At least I'm not the only one who got spanked by Jedi Master CJ. Thanks again for the great thread. Very enlightening. I really expected the Lundahls to be better than they turned out to be. I guess my small consolation is that C seemed the best to me from a listening with my eyes point of view (stupid as soundguy reminds us) and that's what I have :razz:

Looking forward to the Jensen info and some sound files. Any chance of trying out some CineMags? I'm sure they've got something that would work with the LA-2A (and they are pretty dang cheap, too).

Analog Packrat
 
He-He...Great Cj...;-)

I´m thrilled however `cause the only reason I tried to do this is that I just knew (95%)the sowter had to be No 3..Why ???

because That the only transformer of the 3 that I have ever worked with..*GGG*... so I´m not that far out..LoooL

Kind regards

Peter
 
Hey guys,

I would love to try out all kinds of combos but there is a limit in the cash department :\ just between the lundahls, sowters and jensens were lookiting at almost $600...

in time I'll be more than happy to try out other trannys.. but theres no real practical way in which they can all be done at one time. As it would break the bank. :sad:
 
Now is it fair to talk about the sonic differences? My instruments don't give a damn how music sounds, but I do!
 
Kevin,

I'm trying to keep it to the measurement at the moment, the sonic differences we can do in another thread, with WAV files. Scenaria and I are trying to figure out the logistics of how to ensure that each transformer combo is fairly and equally fed.

One question that arises is whether we should even compress the signal? If we're talking about transformers as interface devices, the compression might logically be considered less important. If we do compress, how do we ensure that differences in optical setup and drive don't affect the result.

We could probably set up a static compression test using a tone and use a common T4 module, (prior to the program material, used for setup only, and not presented as part of the test) but we assuredly want to be honest and make sure that everyone gets 'dealt the same hand' as it were. These excercises are all but worthless in my way of thinking, if they're not equal-handed.

I don't think that anyone has used the term 'sonic differences' in this thread yet, though I could be wrong and it might have been a slip of the keyboard. Certainly, I'm considering this a measurement excercise, and the next stage is to post wavs (NOT mp3's!!!) which might necessarily have to be short.

One thing is for sure, The Lundahl combination tested measures by far the best of any thus far for clean signal transmission below a certain level, which is not restrictively low, and I've recommended them as transformers of choice for a well-respectedjazz musicion who lives locally, has an API studio and was considering getting a couple of Bloo kits, but unsure which transformers. He's a spec-head, as well as a listener. I think he'll sleep best at night with the Lundahl combination, out of those tested thus far.

Keith
 
For the sake of comparing sonic signatures of the various transformer combinations I vote you sample NON-compressed information. The LA-2A is a nice line amp/colorizer without the opto. I think removing the optos will make it easier to determine the effects of the transformers alone.

The timing of this is perfect for my situation, BTW, as I'm in the process of building two LA-2A clones, and I have UTC's for the I/O. If I like any of the other options better I could launch the UTC's on ebay to pay for them, and possibly pocket the difference!

This is fantastic stuff, guys, thanks. No doubt I'm only one of many here looking forward to hearing samples.

And I agree, Keith, WAV's would be much better than MP3's.

:thumb: :thumb: :thumb:

JC
 
Sounds good, Keith. I vote for compressionless listening, too, to reduce the variables.

I would be happy to like the Lundahls best, of course. :wink: But, in any case, this is a rare opportunity to hear several transformers side-by-side in what you have made to be a careful comparison and I am most anxious for the education implied. Thanks for all of your efforts!
 
There are no "bad" transformers in my opinion. Just "different" transformers. Which is good, because if they were to all sound the same, then we would less options as far as making a song sound different than the rest. Like you hear Soundguy talk about from time to time, making something "sit" in the mix somewhere. It's nice to have a whole toolbox full of stuff in case you are not getting the sound you like.

The laws of physics make it impossible for one transformer to do everything well. There is a choice of balance that has to be made. Ultimately, this choice has to be made by the transformer company. And once made, they tend to stick to the same phylosophy, which keeps Brand A from stepping on Brand B's toes.

Most companies tend to use their own lam shape over and over for say, an input transformer. This sets boundries on the coil struture from the get go.
And I believe they are all getting most of their lams from the same vendor, althought that vendor might offer some custom services in the way of anealing. But the vendor does not make the steel. They source it from another company in the way of large rolls of steel that get stamped and processed at the lam vendor. Some companies, like Lundahl in particular, might actually do some stamping themselves, possibly on some custom dies that have probably been in the family for a while.
I did see some large rolls of steel in pictures of the factory, although some of that steel was mu metal used to make the cases.

If you take a frequency curve done on a piece of log paper, you will see two rolloffs, one at each end. Take this same curve and slide it left or right on the graph paper and you have an example of how Brand A m ight be different Brand B. But there are other differences, such as phase shift, distortion, insertion loss etc, which also come into play. So it's a big balancing act, with plenty of variables to choose from.
 
I'm sorry to revive a so old post but i'm trying to find some information of distortion. let's say it is a zombie  ;)

is the first graph at an specified frequency??? 1kHz? 20Hz
 
Old posts like this should always be revived.  I'll guess 75% of the people here haven't seen it, and some other % are always asking these questions. 
 
AnalogPackrat said:
p.s.  Out of curiosity, did you use the same tube sets in each test?

I had not seen this thread before, but the question above was the first thing I thought of as well.  This never appears to have been answered either.
 

Latest posts

Back
Top