What Difference Does An Output Transformer Make?
Do Transformers Matter?
Spoiler Alert, They Do. (to varying degrees)
By Steve Graham
Bored? Take an OK amp, buy two new transformers
Yes, I was bored. Two years and counting of near-isolation in a global pandemic can do that. A lot of YouTube rabbit holes gone down too, more on that later. What you see above is the prototype EL84 Simple Push Pull (SPP) amplifier from my six-part DIY series published in 2020. DIY EL84 Amp Update!
I’d been moderately impressed with the sound of the prototype, using transformers reclaimed from a five or six decades-old Heathkit. But the amp shown in the DIY series – using all-new Hammond “iron” – bested the prototype. Erik, the kind patron who commissioned the amp, will weigh in with his impression of its performance later in this report. So, the plan was to try and capture some of the magic Erik’s amp possessed compared to the slightly muzzy sound of my prototype. Swapping output transformers would be a piece of cake electrically but more of a challenge mechanically. A root through my scrap metal bits and pieces yielded some usable 1/8” aluminum plate. A trip to buddy Dan’s tool shop and couple of hours thrashing about on his knockoff Taiwanese Bridgeport milling machine, and it was trannie swap time. It turned out I had to remove the amp circuit board but that only took about 10 minutes. Another 20 minutes or so with a Dremel tool and I was ready to start reassembly.
Adapter plates and their new output transformers were bolted down and the circuit board reinstalled. All wires were connected back up to the board. The Hammond transformer wire colour codes were different than the old trannies but a quick surf through the WOS.ca website and I was at the instructions I needed. Part-4-Attachment-2.pdf
*There are Hammond haters in the DIY community. Hammond transformers don’t pretend to be something they’re not. They are not exotic nor are they overly expensive. If you want ultimate performance there are exotic trannies available from Japan and Europe, at a price. One little-known fact is that artisanal Japanese amp and speaker builder Shindo uses Hammond transformers in their least expensive power amp the Montille CV391, US $6,000 in 2019.
The solder lugs on the speaker terminals were removed to prevent overheating and melting the terminal’s plastic. I’d snipped the leads from the old transformer’s secondaries. Removing the remnants of the old wires in the lugs necessitated a bit of heat from the soldering iron and the usual expletive-assisted encouragement. The secondary wires from the new trannies were attached and soldered to the lugs then the retaining nuts tightened down and staked with a dab of nail polish.
Tubes installed, speakers connected, source plugged in, power on and voila, music again. You wouldn’t think it, I sure wouldn’t have thunk it, but transformers need break-in. I discovered this big-time when I swapped Lundahl transformers into the Elekit TU-8340VK back in 2017. Elekit TU-8340VK Follow-up: Reviewed with Lundahl transformers and Triangle speakers Was that really five years ago? Anyway, the Lundahls sounded absolutely dreadful when first installed. Harsh, spitty, real ear-burners. Fifty hours of running later and they started to sound really good, and with more time, even better. So, without even listening to it, other than to ensure it was working properly, I connected dummy load resistors to the SPP amp’s outputs and put 50 or 60 hours of streamed music through it on and off over the following week or so.
I set the amp up in my main listening room fed by my NAD 50.2 digital source into a PS Audio Direct Stream DAC. My version of the SPP has a volume control so the DAC was connected directly to the amp (no line stage). On the output side the amp drove Triangle BR02 speakers on DIY stands. Review: Triangle Borea BR02 Bookshelf Speakers AC power was, as always in my main listening room, supplied by a PS Audio AC regenerator.
I know what many of you are thinking. How can an informed comparison be made with a ten-day gap, auditory memory being notoriously fickle? Well, before the swap I listened carefully to some of my favourite, intimately-known audition tracks:
After the swap, every track sounded better. Voices were more articulate. The instrumentation underpinning Ms. McKennitt’s tracks were more clearly defined. Ditto the Knopfler and Knopfler/Harris tracks. It was easier to distinguish the multiple guitar parts, and the vocals were more immediate-sounding. The Crosby, Stills and Nash vocals were more, to use that somewhat antique audio descriptor, palpably present.
The Vaughan-Williams piece, even though suffering the indignity of being LP-sourced-then-ground-into-digital-hamburger before being turned back into an analogue steak – insert vegan alternative if you wish – was noticeably improved. The violin sound was clearer and the hall sound of the recording venue was easier to distinguish.
All in all, through the mid and high frequencies the new Hammond transformers were the clear winner. At the bass end of the spectrum, I’d say differences were less apparent. Overall, I’d say the improvements wrought by the Hammond transformers were noticeable rather than earth shaking. Was it worth the time and expense for an amplifier that will go back into my workshop system? Hard to say but I mostly listen to that system in the nearfield so more resolution should be easily apparent.
Thinking back to my comparison review of the Ortofon Quintet Red versus the Quintet Bronze, the level of improvement offered by the Bronze was in many was similar to the general improvement heard with Hammond transformers in the SPP amp. Note that the Quintet Bronze costs approximately 500 dollars Canadian (~$400 US) more than the Quintet Red. To put in that perspective, the approximately $300 CDN I spent on the Hammonds seems like a fair deal. Ortofon Quintet Red Vs. Quintet Bronze Comparison Review
One more data point:
Not to open the objectivist/subjectivist and measurements versus listening but I will, a bit. If you fall on the opposite side of the divide from me that’s fine. I won’t change your mind and you won’t change mine. No nasty comments please, it’s only audio after all.
In the absence of meaningful measurements, the method I most rely on for judging reproduced sound quality is to compare the component under evaluation to a known standard. The standard doesn’t have to be perfect – however one might define that – just its strengths and weaknesses acknowledged.
I was finishing up my listening evaluation and was wishing I had another amp to compare the newly Hammond-shod SPP. I was in my main listening room when the obvious hit me. My Audio Research Reference 3 line stage and Reference 110 power amp, their bright shiny face plates gazing at me as if to say, “Pick me, pick me”. Well why the heck not. This beguiling pair, though now superseded by newer models, still offer very high levels of musical enjoyment. When new they sold for around ten grand US each. I didn’t drop that much on them. The REF 3 I bought second hand for about fourty cents on the dollar when someone traded it to buy the then latest Reference 5. The REF 110 was one of those, be careful what you wish for, scenarios. Someone bought the dealer demo, used it for a few months, then their family situation suddenly changed and needed to realize some quick cash. I came along at the right time and paid about fifty cents on the dollar compared to buying new. All this happened over ten years ago and I haven’t regretted it for a moment. These are my “foundational” components and they will see me out.
Anyway, I connected up my DAC to the REF 3 and the REF 110 to the Triangle speakers, let the amps warm up for half an hour or so then played my evaluation tracks again. I resisted the urge to crank it up. Higher resolving, more powerful amps tend to encourage such urges. Well…. Audio Research don’t call their components, “High Definition” for nuthin’. There was more of everything on offer with the ARC pair. This was especially noticeable with the Vaughan Williams track. More details of the recorded space were audible and the violin sound was purer and more detailed. The recorded space was laid out in a more three-dimensional manner with more explicit side to side and front to back imaging.
One area where the SPP yielded little to the ARC behemoths was in pace and rhythm (most apparent on the Knopfler and CSN tracks). Good pace and rhythm are one of my must-haves. If lacking, it feels to me as if the equipment can’t get out of its own way and let the music progress at its intended speed. Poor pace and rhythm and it’s, “next component please” as far as I’m concerned.
Yes, the ARC gear had more of most everything, as we would expect. But here’s the thing; the little 15 watt per channel SPP got the essence of the music right. It’s engaging, easy on the ears and has no obvious shortcomings when paired with sympathetic speakers. In other words, it’s thoroughly enjoyable.
To those of you not remotely interested in DIY – assuming you’ve made it this far – what can the take-away be? Shop carefully when buying a tube amp. Have a look under the lid. (Make an excuse if you have to. Say something like wanting to see how hard it would be to replace tubes.) Listening must be the ultimate arbiter of a purchase decision but in two similarly powerful amps the one with the larger transformers** might tip the balance in its favour. Unfortunately, the one with the larger transformers will likely cost more.
In the case of the upgraded Elekit mentioned earlier in this report, the $500 US transformer upgrade really, and you must forgive me for this, transformed the amp. The performance leap was that significant.
**This assumes we are comparing the same type of amplifier topology. A push-pull amp of a given power will typically have much smaller transformers than a single-ended amp of the same power. The electrical characteristics of SE amps requires them to have physically larger transformers to achieve bass response equaling push-pull amps. That’s why a 300B SET amp, which might be rated at 8 to 10 watts, will have huge, honking output transformers compared to an EL84 amp rated at 10 to 15 watts. It’s one of the reasons why SET amps cost more per watt. The other is six EL84s can be had for roughly the same price of one 300B.
Erik, our commissioning patron for the SPP build series, chimes in:
March 19th 2020 I was barred from working at my office. That period of ‘two weeks’ stretched to over two years – I am still working from home. When I set up an office in my garage this little beauty became my daily driver.
Over the past two years I have fed this Chapman Audio, A/D/S, Dynaudio, and JBL speakers. From tiny bookshelves to a power-hungry pair of 3-ways with 4 big 8” cones to move. This amp does admirably. For times I need more bass I integrate a REL sub. On occasions I want more volume than this amp is capable of I will swap. These swaps have included Threshold 400A, a Krell KSA-100, and a fully redone ADCOM 555-II. This amp has a sweeter sweet spot than any of those amps. The only thing it cannot do is match the volume and low-end control of a big solid-state amp. The signal source has been a fully redone Krell PAM-3, by way of a Emotiva DAC, and a (gasp) cheap Bluetooth receiver mostly streaming Spotify.
When I first received the amp I figured I would start buying quads and roll some tubes…. but why? This thing is perfect as it is. When the tubes eventually get worn, I will find a matched quad of equivalent or higher quality and rock the amp again for 100s more hours. I am addicted to audio gear and will continue to find reasons to unplug this amp and try out other gear, but there is not a single weakness in the mid or upper end and therefore will be using this as my reference amp for years to come.
Rant no more
There’s enough ranting, raving and other nastiness in the world, so I thought I’d close with an audio anecdote.
Many years ago, I worked with a fellow whose wife had grown up in Toronto in the nineteen-fifties and early nineteen-sixties. Her father was a huge jazz fan and would frequent clubs featuring live music. It was a not uncommon occurrence for her to get up in the morning and go into the living room of the family home to find several black men sleeping on the couch or the floor. It seems her dad would offer touring musicians a place to sleep when the club-provided accommodation was less than appealing, if any had even been offered. Just think of the jazz greats she might have tripped over on her way to breakfast.
What brought this recollection to mind was one of the YouTube rabbit holes I’ve gone down in the last two years. See the link below to Rick Beato’s channel, where he features Oscar Peterson in an incredible live performance from 1974. Rick is a musician, producer and musical educator – but I think most of all a music lover. His passion for music is infectious. Rick’s channel has almost 3 million subscribers and the Peterson video alone has almost 2 million views. Even if you hate jazz, you owe it to yourself to experience Peterson’s other-worldly musicianship. There are also more than one hundred instalments of Rick’s “What makes this song great?” series where he dissects songs from mostly the classic rock era.
Thanks for visiting Wallofsound.ca, and please be kind to each other.
Send a Donation
If you've enjoyed this content, consider supporting Wall Of Sound with a donation. It only takes a moment and will be greatly appreciated.
Steve, Great to have you back in the contribution game. It was getting a bit lonely. Your EL84 experiment reminded me of the Leak Stereo 20 I had many years ago. It finally died a noble death as the heat it produced just wore out the bakelite tube sockets. But it was a hum dinger of an amp. I ran it until it gave up and was not worth repair (in that era). The stupidest move I ever made was not keeping those lovely hand wound trannys. They got junked with the chassis. I sometimes cry about this as I went on a mission to find another EL84 design with that magic and never found it. Even the mighty AN OTO didn’t do the trick. Live and learn. Cheers, David Neice
You’re killing me. (Sigh)