• WARNING: Tube/Valve amplifiers use potentially LETHAL HIGH VOLTAGES.
    Building, troubleshooting and testing of these amplifiers should only be
    performed by someone who is thoroughly familiar with
    the safety precautions around high voltages.

Transcendar

Also disagree about Hammond OPT's being crap. They might not do DC to moonlight as some OPT's specs claim but then I'm not a bat so don't need an OPT that goes up to 5GHZ.

I used the 1650TA's on some 120w monoblocks, the frequency response was flat as it should be and low THD. They handle over 120w too, no saturation at low frequency.

For those of us on a budget who can't afford to drop 500 sheets on an OPT they do the job.

Re Raphelite OPT's, they do look swish, plenty of laminations, some more info and discussion here - Raphaelite Audio Transformers & Sino Tube Amplifier Kits - What's your experience ? including in post #10 a spec sheet. Notice there's no primary inductance figures given or what laminations they use.


Andy.
 
What a brilliant thread.

I started one not too long ago, asking about UK O/T suppliers where I didn't need a second mortgage and all I got were people telling me how important O/Ts are.
While I understand and agree, with that point of view, the info in this thread was what I needed, so thanks very much everyone.
 
I'll disagree on the Hammond comments. I've built quite a few amps by now with Hammond PT's and they are a good product. I used Edcor too a few times and they buzzed just as much in a given circuit.

In fact I've had more problems with Edcor PT buzzing, one to the point of having to mount it in rubber isolators to tame it. The Hammond PT I've used have been quiet and run reasonably cool. They seem fine to me.
 
The original Hammond 1628SE (not the newer 1628SEA) was junk. There were so many parasitic's in that design that the resonant notch, present to some degree in all OPT's, fell INSIDE the audio band. Depending on what external electronics were connected to the OPT, the notch could be at 16 to 18 KHz and up to 10 dB deep resulting in significant rolloff at 10KHz (about 3db).

All OPT's have unwanted parasitic capacitances and inductances. The leakage inductance and the distributed winding capacitance combine to create a notch in the frequency response. A good OPT may have this notch well above 30 KHz.

If you consider the whole picture, the depth of the notch depends on the "Q" of the resonant circuit, which is often determined by the resistance in the circuit, as "R" is the dominant element in the equation for "Q." This includes the Rp of the output tube if no feedback is employed.

Drive that crappy 1628SE with an 811A or other high Mu tube, and don't put feedback around it and your cell phone will have better treble. Drive that same Hammond with a 300B, and you may get satisfactory results.

The external driving circuit has a LARGE effect on how a given OPT will perform. If the driving tube has a low Rp, either by design, or helped with feedback, that resistance shunts the resonant circuit lowering its "Q." This has the effect of reducing the depth of the notch.

All OPT's exhibit loss, some of the signal put into the primary will never make it out of the secondary. Loss is caused by DC resistance in the wire, and magnetic losses in the core. Bigger OPT's have more wire, and more iron, so there will be more loss in a bigger OPT than a small one. The magnetic loss curve is not linear, so that the loss becomes larger at very low signal levels, and very large signals. This is why a big (Hammond 1628 size) OPT will lose low level detail on a 1 watt amp, and a little 125 CSE will saturate if asked to pass 15 watts at 30 Hz. The Hammond 1628SE or to a lesser extent the 1628SEA will blur details in a 2 watt 45 based amp through 106 dB speakers where the average power is about 100 milliwatts where the little Edcor XSE15-8-5K shines, and was my first choice for a triode wired 6V6GT amp at 1.5 watts.

This is why you may get good results with a given OPT in one amp design, and poor results with the same OPT in another.

Back around 2005 Hammond redesigned the 1628SE, and most of that series. The 1628SE became the 1628SEA which was a big improvement, but still far from ideal.

I ordered a pair of 1628SEA's from Antique Electronics Supply, but they shipped me 1628SE's instead. They insisted that I send the OPT's back to them at my expense (about $50) even though it was clearly their mistake. This phone tag argument went on for a couple months and eventually involved Hammond. It resulted in me keeping the 1628SE's and buying a new pair of 1628SEA's at a considerable discount which were shipped directly from Hammond.

I had a pair of each Hammond, a pair of the first 5K Edcor CXSE25-8-5K's, and a box full of Transcendar 3K ohm "300B" OPT's. I used all of these for amp design and testing.

Back when Gerry (Transcendar) decided to get out of the transformer business the first time, 2005 maybe, he listed a bunch of test and manufacturing equipment that was used in transformer manufacturing, of all kinds including ferrite based stuff on Ebay. A listing for "surplus OPT's used in a high end 300B amp" caught my eye, so I bought a pair. I liked them, so I got in touch with Gerry and bought them ALL. I also get him to make me the mother of all OPT's, a monster 5000 ohm transformer that would be used on an 833A running on 1500 volts.

I never could make the Hammond 1628SE sound good in a HiFi application, so I sold them. I used most of the Transcendar OPT's making 300B and KT88 amps. My computer tells me that I should have 8 left not counting the 4 test mules, but I can't find them.

Each of the "test OPT's" have been used in one or more amp builds at one time or another. This was during a time when I made lots of amps, and grew bored with them and either sold them or took them apart to make another. Some amps lived for years, some a few weeks. I only have four amps from those days, and only one still works. It has Electra-Print 300B OPT's, from back when they were under $100 each. My SSE uses Transcendars, and technically works, but is rather bent from being dropped during a house move.

I would choose the Transcendar's for the best overall well rounded OPT. It is what's in my personal SSE amp. It slapped the big speakers around with authority with a KT88 in triode on 460 volts at 3000 ohms, but does Norah and Enya quite well on single driver horns with an EL34 on 430 volts at 6000 ohms.

The big Hammonds can push big speakers around the room, but lose about 10% more power in ALL tests compared to the big Edcor's. I have used both in the same amp three times and I prefer the sound of the Edcor's each time, especially in a zero feedback DHT amp. The Hammonds have multiple secondary taps so that they can be used at 2500 ohms. The Edcors do not.

So here it is 16 year later and most of those "test OPT's" are lined up on my bench being used to develop a new kind of SE amp which can squeeze the best possible performance out of any OPT due to it's low dynamic resistance. This picture shows the 2006 vintage Hammond 1628SEA's with the pair of 3K Transcendars between them, with a pair of 1500 ohm Toroidys on top of one Hammond and one Transcendar, and one Edcor CXSE25-8-5Kon the right. Yes, all the Edcor transformers were painted "pimp my ride" bright metallic blue in the first year. That died quickly.

This amp can run into a 1500 ohm load, a 5000 ohm load, and anything in between just by changing the B+. The Toroidy 1500 ohm OPT's will work at 3000 ohms with some performance degradation, as the Transcendars will work at 6000 ohms, and the Hammond at 2500 ohms. I am still making measurements on all of the OPT's and I can squeeze over 40 watts through some of them.

Here, I have finally got a sound and measurements out of the Hammonds that I like, but they still are 10% lower than the Edcors in power. So what's going into the first amp that I will actually build in nearly 10 years?

For now it's a toss up between the Toroidy's at 1500 ohms on 420 volts for 25 watts, or the big Edcor's on 620 volts for about 18 watts. The Toroidy's just seem to sound better than all the others with my DIY test speakers, but my opinion may change when I try some good speakers. Note that NONE of my current speaker systems have any real response below 70 Hz, so is real bass was required the Edcors would be the top choice.
 

Attachments

  • P3990734-x.jpg
    P3990734-x.jpg
    912.3 KB · Views: 385
Last edited:
Banned Sock Puppet
Joined 2020
All good stuff as usual from TL.
To me the big influence on sound is the core size with an optimum around 3.5-6.6k.
Those transcendars I wanted had this sort of intermediate size between the large core and the medium size core which typically Bogen* made and were smack in the optimum 6.5k area.

On my tests the FR is good, because they were reverse engineered from H H Scott, which is an outstanding transformer squeezed into a really small space.

Again, what measures well, and what sounds good are 2 slightly different things.
It's certain that the usual culprits from the USA made excellent transformers with often a unique sound. (Dyn, Scott, Pilot, Marantz, Eico, Bogen...etc etc)

I have lots of MO & MX output transformers* from the 1960s, which I got from breaking them for bits.
Lots of those are seriously underestimated.

Having tested many, can deffo say the LARGE core ones are the ones to have, (but won't say which) unless you want to derate the medium size core 120 and 60.
Nothing today really comes close, and a lot is to be said for "drop thru" construction which few if any now do. (saves space/height and is aesthetically much superior).
 
Banned Sock Puppet
Joined 2020
Theres way better transformers out there
Not an opinion shared by people like D Gillespie who has done wonders on Eico and other classic amps.

I also have had some exchanges with the bosses of Electraprint & Sowter.
They also know what they are on about, and neither would agree with blanket statements inc "advice" about the UTC LS series or Partridge (which I own). :rolleyes:

Some stuff (eg like old Neumann / B & K mics) simply doesn't date.
 
What annoys me about buying/sourcing OPT's is the lack of detail of the kind George provided back in post #25. Most OPT specs on a makers site give the obvious Zpri/sec, frequency response, Vmax and maybe Hipot test results, maybe primary L and physical dimensions, that's it.


Often the potential buyer only has some subjective hearsay to go on, it's worse than buying posh caps. Voodoo, faery fart magic starts creeping in, EG don't buy Glopo OPT's they're all crap, I always buy Superupyourarse OPT's they're great, and they come in nice gold plated diamond encrusted personally signed boxes. Before you know it where in audiophool territory and we wave science goodbye.


While I'm having a rant why are Hammond/Edcor OPT's cheaper than Sowter etc OPT's? They are all made the same way, onna machine, once the design is sorted apart from the more expensive OPT's probably having more windings what justify's the extra 400 sheets? £400 is a lot of money to pay someone to solder the extra windings together.

Winge over, any tips on spotting a good OPT? Andy.
 
Last edited:
It looks like every time someone mentions Hammond, all of a sudden endless critics.

I will agree - their OT are not the ultimate in sound reproduction but for DIY public - they are perfect starting point in evaluation. I would never spend 4-500 on a single OT, maybe not even on whole amp, otherwise just buy new amp from manufacturer if your BOM approaches 2K in parts only, what to speak of time needed to complete the amp.

To support my story, I'm at the moment busy with building EL84 SEPP and using Hammond 125GSE for OT duty. I will gladly report how they sound in month or two, from my previous experience they don't lack anything in musical terms, and I don't care if they don't reach flat 100K freq. response.

For that kind of freq. response I fire up my 300 dollar solid state gear...

I based my experience on many tests, one of them > Conrad Johnson Tube PP amp 50 WATT vs. Audio Innovation First PP 7.5 WATT amp. The CJ sounded lifeless and washed out, even boring with all the tech data supporting huge linearity but AI killed in detail response and beauty of sound reproduction on the same KEF LS3/5a and SMG B model... so low efficiency and low/high impedance didn't matter...

I had many other disappointments with Audio Research, Audio Note, Cary Audio, Jadis... to mention few... so that's why I build myself mostly :)

Take care!
 

Attachments

  • IMG_1651.jpg
    IMG_1651.jpg
    769.8 KB · Views: 282
What annoys me about buying/sourcing OPT's is the lack of detail of the kind George provided back in post #25. Most OPT specs on a makers site give the obvious Zpri/sec, frequency response, Vmax and maybe Hipot test results, maybe primary L and physical dimensions, that's it.

Even if lots of data is given we don't have any clue how it was measured.

I purchased a large quantity of "guitar amp grade" open frame push pull OPT's 30 or so years ago to make guitar amps. The published specs from Schumaker state "6600 ohms to 0-4-8-16, 80 VA, From 80 Hz to 5 KHz." Also included are size, weight and a dimensional drawing including a schematic with wire colors and lengths. On paper these transformers suck.

Before buying 100+ OPT's I got a few and did some experiments to see if they would work in something more demanding than a guitar amp. At the time I had a Scott 272 Laboratory Reference Amplifier (2 X Mullard EL34's per channel) in my home stereo system. I have ben listening to this for some time, so I knew how it shold sound. I simply wired a pair of these into the Scott in place of the much larger Scott OPT's. I was not prepared for the surprise that came. They didn't suck, in fact they were not bad at all. The Scott was returned to normal and I began to investigate deeper.

A teardown reveals a single half primary, then the entire multi tapped secondary, then the other half primary wound on top of each other with no interleaving. The outer half primary is wound with slightly thicker wire to equalize the DCR of the two halves.

The 80 Hz to 5 KHz spec is obviously just the guitar's expected frequency range. These transformers were still in their shipping boxes destined for ADA's warehouse in 1990. At the time Schumaker made battery chargers (I have one from 1984) and OEM transformers. The core of the transformer in my 12 volt 6 amp battery charger is exactly the same as these transformers, so that's where the "80 VA" comes from. Clearly these specs are by design or market need, rather than actual measurements, so what do they really do?

I used these OPT's in several amps including my 300Beast, a zero feedback 300B push pull amp made with 100% junk box parts back in the 1990's. That amp measured good and sounded better. It is one of the four amps that I still have from those days, and the first one built. It has a blown power supply and I stole a few part from it, but I would like to fire it back up and look at it again to figure out why it worked so well. I built several KT88 push pull HiFi amps using these OPT's and all worked well enough to be sold. There were also a few TV sweep tube amps, and 3 amps built from Pete Millett's big red board. All measured differently in frequency response and power output. In most cases the OPT was the limiting factor. The KT88 amps were limited to about 35 WPC due to the poor quality of the 1990's Chinese tubes and the small power transformers from military scrap. At that level they were flat from 20 Hz to 20+ KHz. 60 WPC could be had in the same amp with Sovtek 6550WA's and more B+ voltage, but they would not do that over 20 Hz to 20 KHz.

How much power will they handle? Well I was somewhere north of 160 watts at 1KHz when my dummy load blew leaving the transformer unloaded at which point it burst into flames. In another test a poor layout resulting in an arc from plate to ground on a PC board left one permanently magnetized, again same tube same 160 watts, but these were my mistakes.

In post #55 of the UNSET thread I successfully stuffed 253 watts at clipping using the same 36LW6 tube , and 220 clean watts through this "80 VA" transformer without incident. It didn't even get hot. WHY is this possible?

UNSET is coming?

A power transformer needs to work at one or two frequencies. An OPT needs to work over a wide range of frequencies.

Low frequencies are limited by the inductance of the primary (amount of wire) and saturation of the magnetic core (amount of iron), so for big bass, size matters! As you go lower in frequency lack of primary inductance will drive the tube current up. As you go lower in frequency saturation will cause the tube current to increase and cause distortion. If the tube, power supply, and circuit design can support the higher current without failure or excess distortion the apparent low frequency response for a given OPT will look better, and it's low frequency power handling ability will be extended.

Take one of these Schumaker transformers, wire it into a SPP board (a typical EL84 / 6BQ5 amp with GNFB) and I can get 15 watts down to about 35 Hz before the THD goes over 5%. Take that same OPT and wire it into Pete's Engineers Amp board with my 6HJ5 / 6GU5 mods and I can shove 60 watts through that same OPT at 35 Hz, and 125 watts from 80 Hz to 22 KHz at 5% THD. Attempting to stuff 220 watts through those transformers at 35 Hz would have blown something up for sure. If the amp had enough tubes and power supply to force the needed current without failure, the current would have likely caused wire failure in the OPT itself, because the core would have been fully saturated such that no more power input could pass to the secondary.

High frequencies are limited by stray capacitance and leakage inductance. That same "lots of wire and lots of iron" that help the low frequencies kill the highs. Again tube current goes up as the frequency increases until the resonant point is reached, then the output just drops off. Here too much size works against you.

Many years ago I found mention of the method of measurement on Hammond's (I think, it could have been another transformer company) web site somewhere. I haven't seen it since. The transformer was driven by a "600 ohm signal generator" and a swept frequency measurement made at 1 watt. This will produce a much better result than will be obtained in most amps. Many "600 ohm" signal generators have a much lower output impedance than 600 ohms to obtain low distortion into a 600 ohm load, especially the old HP tube generators like the 200AB. They will drive over 1 watt into an 8 ohm speaker! (2 X 6K6 tubes in push pull for outputs).
 
Last edited:
It looks like every time someone mentions Hammond, all of a sudden endless critics.

I will agree - their OT are not the ultimate in sound reproduction but for DIY public - they are perfect starting point in evaluation. I would never spend 4-500 on a single OT..

I feel the same but I do wish Hammond would make an OPT between their huge SEA series and the small open frame guitar amp models. Something around 15W, like Edcor makes, with end bells etc. would fill the needs of most DIY users. Or that someone/edcor would stock some of the more popular models to get around the 8 week special order thing.
 
It looks like Transcendar is gone for real this time. I had a pair made by them less than a year ago then I tried to get another pair and they said they were too backed up to take orders. Now my last two emails have gone unanswered. Too bad because they were great for the money, now what will I do? Any ideas on better than Hammond but faster than Edcor?

You can always source Ebray for some decent used transformers.
 
Hi Stephe,

The someone that has a stock of Edcor and Hammond are this guys:

Lundahl & Edcor Transformer Distribution

This is where I bought my last couple of OT's...

Sorry, I forgot to add that I am in the US. Shipping would be pretty expensive from the EU. I wish someone over here would stock some of the more common ones like this distributor has!