My experience has been similar, but I have found fairly consistent performance across transformers and brands. The best success has been with 6AS7 with additional plate to grid feedback to get really low impedance, this works well and measures well. However i have also had good success with 807's in pentode with high level plate to grid feedback and recently with the EL86 with similar feedback.
I always aim for minimal DC imbalance because I want to preserve the most inductance for the job in hand.
However the real solution is active crossover and a dedicated subwoofer amp - which is what I currently use for bass duties.
i am currently building a Hirage Le Monstre to compare to see if I really have been missing something with the valve amps and their lesser bass response.
Shoog
Last edited:
Yeah, they weigh 11 Kg each! They were on Plitron's surplus page about 5 years ago, 110 USD each. I talked to a guy at Plitron who said that they were designed for Marshall for use in a bass guitar amp that employed 8 X KT88. Marshall didn't like the sound, so they sat for about 10 years and were surplussed.
I did not believe that the saturation limit of a transformer could be affected by what feeds it until I started seeing evidence to the contrary. We all know that DC offset in an ungapped OPT will cause premature saturation. I have found that most OPT's work best with some DC imbalance to counteract winding imperfections so that the static magnetic fields from each half primary do fully cancel.
I bought about 200 surplus "guitar amp" quality OPT's about 15 years ago for the purpose of building guitar amps. They were cheap surplus, rated for "80 VA at 80 Hz", 6600 ohm primary, and weigh about 3 Kg each. I took one apart to find no interleaving at all, just one half primary, then the 0-4-8-16 ohm secondary, then the other half primary wound with slightly thicker wire to offset the longer winding length. I made quite a few guitar amps in various power levels up to 150 watts with these transformers and all worked great for guitar (minimum frequency = 82 Hz). I noticed that some of these amps worked reasonably well for my bass guitar (down to 41 Hz) and some distorted badly at a few watts. Why?
I decided to perform some testing. I wired a pair of these OPT's into several of my HiFi amps and did some testing.
The first surprise came when I wired a pair into a 300B P-P amp. These transformers sounded much better than the UTC's that were in the amp.....10 years later they are still in there! Bass was solid and there was no hint of saturation down to 30 Hz at 28 watts (amp clipping point). The upper 3db point was 25 KHz. The 300B is a DHT with very low plate resistance. No feedback was used.
The other side of this came when I wired a pair into one of my SPP amps. This is a "textbook" push pull EL84 amp design with about 6 db of global feedback. The same transformers showed signs of saturation at 10 watts and 40 Hz. The high frequencies were also rolled off. These transformers were not suitable for HiFi use with this amp design. I tested the same transformers again in the same amp with 6CW5 output tubes and the 8 ohm load on the 16 ohm tap to reflect a 3300 ohm load to the tubes. This change allowed over 20 watts at 40 Hz through the same amp with the same transformers. HF response was improved enough that I built the amp with this configuration, but it still isn't HiFi quality, but it works good as a guitar amp.
Pete Millett designed an amp called the "Engineers Amp". It was a mild mannered amp designed for 18 WPC. I have successfully squeezed 525 watts from it in mono mode with one of the monster Plitrons wired across both channels. I experimented on this amp for over a year and published a reliable version that produced 125 WPC. Almost all of my testing was done using the same guitar amp transformers. This was done because I could afford to fry one of them, I had more! Never did fry one.
I have seen those same transformers crank out a 60 watt sine wave at 30 Hz for an hour. No saturation and the only things that got hot were the tubes and the load resistor. Clearly what feeds the transformers can affect how they perform. The source here was a pair of big fat TV sweep tubes in pentode mode with plate to grid (Schade) feedback to reduce the tubes output impedance.
The load impedance on the transformer will also affect how it saturates. My Yamaha speakers have a woofer resonance around 70 Hz. Here the speakers impedance is over 20 ohms. I have scope pictures from a SSE amp (SE triode wired KT88) showing a 35 volt peak to peak undistorted bass guitar signal across those speakers (about 55 Hz). No the KT88 isn't producing 18 watts, and the OPT isn't passing 18 watts, because the speaker isn't 8 ohms, but the amp has enough voltage headroom (440 volt supply) to avoid clipping, and the OPT is lightly loaded at this frequency. Note that the speaker cones are moving near X-max, but the sound isn't very loud from a 7 inch woofer at 55 Hz.
Here is the long thread where Petes amp unfolded, and was morphed into a monster.
http://www.diyaudio.com/forums/tube...-new-p-p-power-amp-design.html?highlight=6jn6
All of the power toroidals I have used employ loose bifilar winding of the two primaries with the secondaries bifilar wound over the top. Its difficult to see how the primaries couldn't be almost perfectly matched in this arrangement. I will be sticking to matched DC bias matching. I usually aim for less than 3mA imbalance using CCS's
Shoog
Hi all,
It's been a while since i last posted here. I've been playing with tube amps for some time now but i seem to have hit a bit of a brick wall in my quest for audio heaven. The problem i have is this...:-
I've been trying to avoid using commercial Output transformers due to their cost and the fact that i like to be individual and keeps things diy. In pursuit of this i have been using mains torriods in push pull with remarkable success. They perform flawlessly from 50Hz to around 100kHz. The efficiency of these off the shelf mains torroids is good along with phase angle. So whats the problem i hear you ask??
The issue is after a few years playing with mains torroids i thought it was about time i built a finished amp. to this end i ordered a commercial audio transformer thinking all would be well, or at least a great improvment on my mains torroids. My mains torroids dont really go much lower than 40Hz on full power unless i use very large ( over 400VA )torroids. Upon trying the commercial audio unit i found that it also will not put out full power below 40Hz. It can only deliver about 1 watt at 20Hz before saturation. An improvment on my mains torroid which can only put out around 100mW at this frequency but still rarther poor for a comercial transformer rated at 40W audio. The commercial units high end responce dies off much sooner than the mains power toroids but its phase angles are better.
What i'm wondering is how much power at low frequencies below 50Hz should i expect? Am i really expecting too much to get full power at 20Hz and below or is this a common limitation of tube amps?? What is your real measured full power lower limits before distortion/saturation??
Mine are as follows:-
Off the shelf 100VA power torriod:-
50Hz/14Wrms 40Hz/15Wrms 30Hz/9Wrms 20Hz/2Wrms 10Hz/100mWrms
Quality audio torroid rated 40W
50Hz/15Wrms 40Hz/15Wrms 30Hz/15Wrms 20Hz/8Wrms 10Hz/1Wrms
Leigh
well its the quality of the transformer, and this insertion loss is there regardless of output device (tube, transistor, mosfet)
something a little bit more audiophile:
http://plitron.com/wp-content/uploads/file/4070SSCR.pdf
Plitrons are expensive. but you pay for quality, not at name.
for a push=pull amp, general prupose quality is built like a power transformer.
the plitrons are wound differently. like a signal transformer. this is why they are expensive.
Last edited:
This is all very interesting. Thanks PPl, just the kind of info i was looking for. I was almost thinking of ditching the transformer altogether and experimenting with ridiculus over the top OTL, but your input has now recindled my faith in the output transformer. What i'm feeling from all this information is that Bigger really will give me more down below, and that i should not rely on the specs so much. Hmm... Maybe this subject really is a 'Black Art' lol 
On a conventionally wound (EI) transformer that was designed for audio output duty, weight is a good measure of output power at low frequencies. It is impossible to make big power down low without sufficient iron.....and copper. It's back to that inductance thing.....physics is physics and it's hard to cheat those laws.
A transformer that is designed for mains duty needs only enough inductance at the mains frequency to limit the idle current. The manufacturer keeps cost under control by using just enough iron and copper to do the job. The number of turns of wire is the biggest factor in determining inductance, and the amount of iron in the core (and the alloy) determines the magnetic saturation point. A really big toroid (1KVA) may have nearly the same primary inductance as a small one. It WILL have more winding capacitance that the small one limiting it's high frequency response. You are trying to stuff a lower frequency through a transformer than it was designed to see, so depending on the tradeoffs used in the design, some will work better than others. Testing in the circuit that you intend to use is the only way to tell how it will work.
I realize now that your "audio" toroid is an Amplimo. These use the same design techniques as the Plitron's that I have. Both are regarded as very good transformers, so there must be something else limiting your success. I have noticed that my Plitrons are somewhat sensitive to DC imbalance and exactly matched currents are not the best point. It is also possible that mismatched tubes are causing a "bias shift" as the power level is increased. If you have several tubes, try swapping them around to see if the saturation point changes. I always have a current meter in the cathode of each output tube when I am working on something new, especially when working at high power levels. I use a $5 digital multimeter wired across a 10 ohm resistor in series with each cathode. Set the current for a match at idle, then crank up the power. You might find that one meter reads higher than the other as the power is increased. If so nudge the bias of the hotter tube downward and see if that improves the saturation issue. Repeat at lower frequencies.
It is not uncommon to find "matched" tubes that aren't so well matched as they are cranked up in power. This isn't a big deal with a large EI transformer, but most toroids don't like it. I have seen 50 mA difference at 400 mA on tubes that are matched at 40 mA.
Note, back in the 60's I made a few guitar amps using ordinary EI power transformers for OPT's. You use the HV winding for the primary, and all the heater windings in series for the secondary. Old TV transformers had several heater windings. This results in a rather high load impedance, but at the time it was all I could get.
Another possibility is using toroidal isolation transformer with 4 115 volt windings, and winding your own secondary over the mylar tape. Run all 4 115 volt windings in series as the primary. This gives you plenty of inductance!
you should just wind the right number of turns if your amp's numbers don't fit what the power xfmr is already wound with
maybe just as a test of where your issues are - some awkward adjustment can come from using multiple of the same standard power line xfmr - series the pri, parallel the secondaries - the secondaries must be paralleled, must be the exact same spec xmfrs for equal V division across the n series pri V - and you only get n times pri Im, should be able to go 1/n lower in f
and as usual you are assuming all risk of the differences in the xfmr safety isolation ratings and your amp's operating V
maybe just as a test of where your issues are - some awkward adjustment can come from using multiple of the same standard power line xfmr - series the pri, parallel the secondaries - the secondaries must be paralleled, must be the exact same spec xmfrs for equal V division across the n series pri V - and you only get n times pri Im, should be able to go 1/n lower in f
and as usual you are assuming all risk of the differences in the xfmr safety isolation ratings and your amp's operating V
Smaller mains toroids do seem to sound better down low although I have not measured this.
I have tried two 35VA 2x115V primary transformers in a sort of **** eyed series arrangement. Each valve sharing half of each transformer primary to keep the current the same in each half of the composite transformer. Secondaries in parallel. Works well, certainly better than one transformer set up for the same ratio.
You could take it further and put one of the transformers in the cathodes ala Crowhursts twin coupled design
Cheers
Matt.
I have tried two 35VA 2x115V primary transformers in a sort of **** eyed series arrangement. Each valve sharing half of each transformer primary to keep the current the same in each half of the composite transformer. Secondaries in parallel. Works well, certainly better than one transformer set up for the same ratio.
You could take it further and put one of the transformers in the cathodes ala Crowhursts twin coupled design
Cheers
Matt.
For anything greater than about 3K PP using a power transformer - the ratio is such that there are insufficient secondary turns to adequately couple the signal across. it can be just 10 turns for each volt of secondary output rating.
Tubelabs idea of using an isolation transformer to get higher primary inductance and better ratios is a very good one. Winding a secondary isn't that difficult (though its no fun either).
Shoog
Tubelabs idea of using an isolation transformer to get higher primary inductance and better ratios is a very good one. Winding a secondary isn't that difficult (though its no fun either).
Shoog
I dont get you , years ago an Mc3500 would produce 350 watts @20hz and was flat 20-20K (1960's) not possible today .... ?
Yes but an MC3500 has a wonderful output transformer with extra windings for bootstrapping the drivers. Also uses 50% cathode feedback and a whole host of tricks. Still not ruler flat but pretty amazing none the less. Pretty sure it has enough inductance to do full power at 14Hz. A work of art.
We are talking about mains transformers vs Plitron/Amplimo transformers and toroidal ones at that with all their inherent problems.
Cheers
Matt.
We are talking about mains transformers vs Plitron/Amplimo transformers and toroidal ones at that with all their inherent problems.
Cheers
Matt.
It's possible today. In fact I have bread boarded such an amp. You just need 4 to 6 BIG TV sweep tubes per channel, a power supply that produces 650 volts at 1 amp per channel and a pair of these OPT's.....Oh, BTW these OPT's cost $450 USD each plus shipping for 23 Kg from Canada plus UPS brokerage charges. The Plitrons I have are essentially the same transformer, but unpotted. The shipping and brokerage was over $100 5 years ago. That's $1000 for the OPT's. You need another 10 Kg power transformer, or a pair of 400 VA Anteks.
I have measured 500 watts at 1 KHz, over 400 watts at 20 Hz, and almost 300 watts at 20 KHz (limited by excess tube current using only 4 output tubes). Frequency response (-3db points) was 6 Hz to 76 KHz at 100 watts.
Smaller mains toroids do seem to sound better down low although I have not measured this.
I have tried two 35VA 2x115V primary transformers in a sort of **** eyed series arrangement. Each valve sharing half of each transformer primary to keep the current the same in each half of the composite transformer. Secondaries in parallel. Works well, certainly better than one transformer set up for the same ratio.
You could take it further and put one of the transformers in the cathodes ala Crowhursts twin coupled design
Cheers
Matt.
This is very interesting, never considered doing that. Seems too good to be true, in theory, twice the inductance, double the core material and half the potential running through each Xformer... A bit of brute force should be able to over come the increased capacitance. It would be interesting to see how all this effects the low end, should be a massive increase in low freq performance.. Hmmm...
Last edited:
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.