suitable transformer for an audio isolation transformer?

[PRR]

> why couldn't it be a "line" transformer? (with 600 ohm input impedance)?

For the same core (same size/cost), a 10K will accept 4 times higher signal voltage than a 600, for same bass distortion.

Unless you must drive 600 ohms or long lines, the 10K is your better buy.

Yes, if you are flooded with good 600:600 iron, use them.

> the cheapo tiny Radio Shack audio isolation transformers

The naked modem transformers. The freq response is much better than you read: a modem tranny must have TIGHTLY controlled ratio and loss 300-3K, tighter than we demand in general audio. Response is useful at least another couple octaves either way.

However the one RS stocks is the smallest size made. I had trouble with bass overload, far less than 1V at 50Hz came out rotten. That's why there's bigger ones, which cost more, the RS buyer didn't know why, so we get the least which is legal on a phone line, and not ameniable to general audio.

OTOH: over in the Car Sound rack there is a lump with RCA plugs. Pull this open, it has two very good 2K:2K transformers inside a shield, and this will take over a volt at thumper frequencies. Cost is $18 or so, but worthwhile.

> I don't really think of them as great either

The little iron sold by Mouser _IS_ good stuff, for its size. The silly-low price is NOT related to "greatness": the price is low because millions of these parts have been made for 50 years, by competing factories. The specs are not hard to exceed: 1K-10K windings are not prone to the troubles of high impedance windings, these parts are small, treble will come out far better than specs.

> stereo to mono summing transformers

Transformers is parallel do not "sum", they "short". Mix with resistors. (Yes, there are ways to sum with transformers, but this opens new problems.)

[wavley]

> stereo to mono summing transformers

Transformers is parallel do not "sum", they "short". Mix with resistors. (Yes, there are ways to sum with transformers, but this opens new problems.)

Sure looks to me like it's a left and right signal inducing a mono signal.  I've worked on a few hammonds in my day, it was pretty common to mix with transformers.  I know it opens new problems, I'm hoping to exploit these problems for unique sonic characteristics.  If I wanted clean summing then I'd just stay inside the computer and would certainly stay away from transformers, but as it is my computer will send out a ping to adjust for latency, I can send out a few tracks to be summed (I'm welcoming crosstalk at this point) and use a matched set of Neumann v476b's and then mix it back in as parallel processing.

[familyortiz]

The Xicon transformers from Mouser perform extremely well... a bit of peaking at 20-25kHz but the 3db range is down to about 20Hz.
As to driving the transformers passively, I got a clue from the disclaimer from Radial for the all passive Bigshot ABY:

"Keep in mind that the trade off here is tone. When you engage the transformer, even though you are still passive, you are no longer technically 'true bypass'. This means that the tone going through the transformer will be affected unless you place a buffer in front of it. Many pedals incorporate buffers into the signal path. This can be good or bad, depending on the quality of the buffer."

i.e. Save yo money and get, or build, something that works!

[wavley]

I used the Triad trannies to build a Hum Free, they are spec'd at 200 Hz instead of 300, I liked them a bit better than the Xicon's, not a huge difference.  I got a lot of bass out of them at a slow roll off, but I also play Bass VI so they just aren't cutting it the way I would like when I switch from guitar to bass.  There can be a nice fattening sound from transformer core saturation, this is why I'm thinking of using the Edcors in the same circuit because it looks like they are made with better materials and with a little more care without going expensive.  That said, there is something very pleasing about pushing a nice transformer hard, I've got gear with Haufe, Cinemag, Jensen and a few other nice things and they are definitely very musical though not pristine, but musical is what I'm going for.  I don't always need the output to be identical to the input as long as it sounds good.

[PRR]

> it looks like they are made with better materials

Both use standard power transformer iron, which is actually very good stuff. And enameled wire, which is usually blameless. Edcor does use nicer-looking bobbin shapes.

> a little more care

There's not a lot of "care" involved. These are simple machine windings.

> nice fattening sound from transformer core saturation

If you are pushing big bass, and it gets mushy, your option is SIZE. Buy bigger. (Yes, Edcor goes as big as you want.)

If size is limited by case or by stray capacitance, there are magic alloys, but saturation can be abrupt, and cost is very much higher. As a part-way compromise, Edcor does offer several "Nickel core" line-match transformers.

Note that small power transformers work at audio. You want both windings in one limp. There is a style with primaries on one leg and secondaries on the other leg, these have a strong ~~1KHz high-cut. However for med-Z to med-Z, a dual-winding 120V+120V primary may be used like a 2K:2K transformer, a good fit between an opamp and a medium-long (<100') line.

[Mike Burgundy]

Good  find on those transformers! I've been mulling over the Xformer coupled ABY for low B (31Hz), looks like this is a real contender.
Hope you get around to testing them RG - this should be really interesting!

[R.G.]

Good  find on those transformers! I've been mulling over the Xformer coupled ABY for low B (31Hz), looks like this is a real contender.
Hope you get around to testing them RG - this should be really interesting!

I will get to them. I re-found them just a few days ago, so I at least know where they are in the shop - not an insignificant advantage! 

[wavley]

I got my edcor 15k/10k transformers is, I got those because I actually wanted them to pad the volume just a little bit.  I quickly wired a couple up passive just to check them out and running them after my Boomerang+ it actually sounds really good, there is a little high frequency loss but it only caused me to turn my treble knob from 4 to 5, so not a big deal.  It doesn't like being run after my GGG phase 45 and loose all low end when it's on.  So this week I'll build them in to a hum free.  But all and all, has more low end than the actively driven triad transformers.

> a little more care

There's not a lot of "care" involved. These are simple machine windings.

"EDCOR uses the finest M6 line grain oriented steel cores. Each unit is hand stacked to assure at least 92% stack."

Sounds like a little more care than a wage slave factory in china to me, I know the chinese ones are hand stacked too, but are they paying attention to grain orientation?  It may not make much difference, but the extra attention most likely means that the lamination of these cores is better, I've seen a few poorly laminated trannies in my day.

[R.G.]

"EDCOR uses the finest M6 line grain oriented steel cores. Each unit is hand stacked to assure at least 92% stack."

Sounds like a little more care than a wage slave factory in china to me,

It's supposed to. However, it's written in Advertising Language, which is an artificial language invented jointly by an adversarial cooperation between advertisers and the courts. It's confusing because they use the same words you do - English in this case - but certain words don't mean what you think they do.

"Finest" for instance, indicates that the thing being described is a member of a "parity class", a group of things which are all very much the same, so they're *all* the finest. M6 grain oriented steel cores are all made the same way, and have to be to call it M6 grain oriented laminations.

Hand-anything is supposed to sound carefully crafted, the idea being that it's done by a senior technical staff member with a graying beard and a love for getting laminations ....just... right. However, I can assure you that stacking laminations is done by the lowest cost worker they can hire (USA, Japan, China, wherever). It's tiring, mind-deadening work. Any worker that can contrive to do something else will do so.

I recommend the book "Doublespeak", which was enlightening for me. It's quite old now, but you may be able to find a used copy.

Oddly enough, the term "wage slave" was invented in the USA, and applied to the US factory worker, who was bored, tired, and didn't give a hoot about whether what they did on the factory line ever made anyone happy or not. In the 70s and 80s, the oddity that Japanese goods were valued because they worked and American made junk didn't was a massive shock, especially to the rulers of the US industrial world, like GM, Ford, and General Electric. And today, the Chinese are busy offshoring the low-tech labor to cheaper places like Vietnam and starting in Africa themselves.

"Chinese-made junk" is becoming a self-denying phrase like "Japanese-made junk" became a few decades ago. The Chinese sell the world cheap stuff because the foreign (to them) businessmen demand the cheapest possible junk.

And I can promise you that you can get more attention to detail in China than you'll get in the USA, even if you pay a lot less. Stacking laminations is a lot more mind deadening than burger flipping, and you know how much burger flipping jobs are valued here.

And for paying attention to grain orientation, last time I looked, the grain orientation in E-I laminations is along the length of the tongues, so it's impossible to not stack them with the grains oriented the same way. They're punched from the rolled strip in that orientation.

I don't mean to be argumentative; it's just that certain things get repeated over and over and taken for the truth when it's just convenient to say them.

Just out of curiosity, could you measure the low end response on the two setups, the passive Edcors and the actively driven Triads, and post the low frequency -6db points?

[wavley]

Just out of curiosity, could you measure the low end response on the two setups, the passive Edcors and the actively driven Triads, and post the low frequency -6db points?

I guess I should have chosen my words a little more carefully, I felt a little attacked by PBR for not knowing what I was talking about, when I've been doing this for a number of years and know full well what I'm getting into with these things.  By no means am I a "chinese junk" person, I am quite fond of a lot of the audio products coming out of china right now and find it to be well made when you buy the right things.  I do think there is a bit of a difference in care taken between transformers made when ordered (it took almost three weeks to ship mine, of course they may just be overwhelmed with work) and hundreds or thousands a day like the triads or xicons.  I think it was more of a point of mass produced vs. made to order.  But you're right, I've never been in a transformer factory, but I have been a wage slave electronics assembler    I paid attention to detail, but I can't say that for everybody there, and for that reason I would never trust the instruments in a Cessna, personally I made ultrasound radiometers, so I had nothing to do with the other stuff.  Guess I'm a little spoiled by my NSF job, where just about everybody cares about what they do an quality is the only priority (not that we get paid well, but I do have time to check this page while conductive epoxy cures)

I pirated parts off my old hum free, so I didn't AB them, but there was a noticeable (not huge, but enough) bass roll off when using my bass vi that wasn't there with the edcors.

Luckily, I'm going to leave one wired up passively, for now, and I'm going to build a hum free with one triad and one edcor because one of the amps I'm splitting to is just a little Kalamazoo that I modded and it doesn't have a lot of low end anyway, I was going to socket the coupling cap for that anyway so I could play with it.  Of course I'm in the middle of some house repairs, I'm planning on getting to it this week, but I may not.

[wavley]

Quick and dirty measurement shows the triad flat down to about 63Hz, then rolling off.  Edcor pretty flat.

Sound wise this shouldn't make much of a difference, but it's just more stuffy on the low end, not as thumpy and clear.

[PRR]

> I felt a little attacked by PBR for not knowing what I was talking about

That was not my intent. I am sorry I gave that impression.

> I'm in the middle of some house repairs

Must be the season. Tearing out two walls, putting in two walls, new tub, vanity, old toilet, and today's chore was converting the WHOLE (94%) water system from haywire cobbled corrupt cracked copper to new tubing and manifolds.

> grain orientation in E-I laminations is along the length of the tongues, so it's impossible to not stack them with the grains oriented

Been messing with a LOT of wood lately. It's the same way. There's a grain, lengthways and cross-way. Wood is several times stronger one way than the other. Joist or shelf really wants the grain running from support to support.

Older transformer iron had about the same property any direction. But they found a way of heating and rolling that slightly reduces cross-grain but improves with-grain property. Like 500 cross and 1000 in direction of grain.

But is that useful? If you build a box, you can't have the grain run all 4 ways, unless you have 4 joints, and joints (butt, dovetail, or lapped) are always weakspots. Transformers are stamped E-I with lap-joint, but you still have some flux going in the un-preferred direction. E-I proportion can minimize but not eliminate cross-grain flux.

Grain-oriented has just enough advantage in larger power transformers that oriented is "the" standard iron for anything bigger than a thumbnail, and not over-priced. (Plenty of exotic cores in the overpriced audiophile world.)

[wavley]

I realized that I measured the xicon tranny the other night, the triad went flat down to 40Hz, they both still sound stuffy on the low end.  The edcor sounds a lot better on the low end, but the trade off is that it's twice the size.

> I felt a little attacked by PBR for not knowing what I was talking about

That was not my intent. I am sorry I gave that impression.

It's cool, that's the problem with the written word, sometimes if we're in a bad mood, we think everybody is in a bad mood.

> I'm in the middle of some house repairs

Must be the season. Tearing out two walls, putting in two walls, new tub, vanity, old toilet, and today's chore was converting the WHOLE (94%) water system from haywire cobbled corrupt cracked copper to new tubing and manifolds.

Ha, well at least I'm not alone, tearing down walls, jacking the basement, tearing out the floor upstairs to sister joists, new water heater, putting in a new kitchen...

> grain orientation in E-I laminations is along the length of the tongues, so it's impossible to not stack them with the grains oriented

Been messing with a LOT of wood lately. It's the same way. There's a grain, lengthways and cross-way. Wood is several times stronger one way than the other. Joist or shelf really wants the grain running from support to support.

Older transformer iron had about the same property any direction. But they found a way of heating and rolling that slightly reduces cross-grain but improves with-grain property. Like 500 cross and 1000 in direction of grain.

But is that useful? If you build a box, you can't have the grain run all 4 ways, unless you have 4 joints, and joints (butt, dovetail, or lapped) are always weakspots. Transformers are stamped E-I with lap-joint, but you still have some flux going in the un-preferred direction. E-I proportion can minimize but not eliminate cross-grain flux.

Grain-oriented has just enough advantage in larger power transformers that oriented is "the" standard iron for anything bigger than a thumbnail, and not over-priced. (Plenty of exotic cores in the overpriced audiophile world.)

You know, I have a lot of nice equipment with nice transformers, guess I never bothered to look into the actual manufacturing details, just the junk they taught me in school and my ears.

[PRR]

> the triad went flat down to 40Hz

I'm assuming the Edcor is larger (more iron and copper) than the Triad and Xicon.

Try higher level, 1V 3V or 10V, and 'scope the output. At 10V it would be ample to just listen to the output. Even the larger iron may be in distress at 10V 400Hz, but the little iron may not handle 10V below 1KHz-2KHz.

I gather you know the general tradeoffs, but for others:

The bass/amplitude tradeoff is essentially linear. So for the made-up numbers above, the larger can do 1V at 40Hz, the smaller may not manage 1V at 100Hz without gross mangle.

Taking 0.5V at 82Hz as guitar cord envelope, the larger core is at a quarter of its mangle limit, the smaller core is very near its mangle point.

> more stuffy on the low end, not as thumpy and clear

Below gross overload, iron THD goes down, but never goes away. The less-stressed transformer will have lower THD, which in your test seems to mean "more stuffy on the low end, not as thumpy and clear". (Talking about sound = dancing about architecture.)

Some interesting tests here:
http://www.cliftonlaboratories.com/non-linear_transformer_behavior.htm
Lot to read, and some of it inconsistent.


Apology. I missed this bit of marketing-speak:

"Each unit is hand stacked to assure at least 92% stack."

This is "not the user's problem", or shouldn't be. We might like the winding 100% full of iron, but shellac, burrs, and tolerances mean that won't happen. Larger iron is often assumed 94%-95%, I guess in smaller parts you can't do that good. The reference to "hand stacked" seems odd. Perhaps MASS-produced parts are machine stacked and set well shy of the worst-case interference to reduce jam-up. Obviously a machine can be made smarter. I suspect Edcor is a low-tech place, and at least their low-volume parts they may find it cheaper to hire local housewives (the traditional stackers) than to invest in fancy machinery.

If the electrical response is met by re-designing core to the actual production stacking factor, a 92% stacking factor leads to about 109% more copper resistance and height than an ideal 100%-stacked core. Taking 95% assumption, 3% worse if actually produced at 92%. This is mostly insignificant.

Yes, I have seen some no-name cheap iron which seemed to be stacked 75%. Maybe very-variable lams and over-concerned for production jams? Maybe they stacked to 90+%, but iron improved, or customer complained about price, so they left some iron out to just-hit spec on bass or price?

It's a mature boring product, they have to boast about any little thing they can.

[wavley]

Below gross overload, iron THD goes down, but never goes away. The less-stressed transformer will have lower THD, which in your test seems to mean "more stuffy on the low end, not as thumpy and clear". (Talking about sound = dancing about architecture.)

I'm not sure this is an apt analogy, these are pretty common descriptive terms, not to mention all the threads we've had clarifying these terms, like the "Glassy Highs" thread    quick measurements showed me frequency response, my ears told me what sounds good.  I think further measurements are kind of pointless, we can tell that the edcor is at least twice the size of the triad so it's a given that it will have better low frequency response, so we're just proving what we already know.  I do after all like to play my guitar every now and then and this thing is only to solve a problem, it's not even fun or makes a cool noise, hopefully I'll box everything up, plug things into it, and never think about it again!