• 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.

The perfect load for an OTL

This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
If you could wave your magic scope probe and create the ideal speaker load for an SET amp (one that would render an output transformer unnecessary), What would its specs look like?

I'm trying to imagine the ideal speaker load that could be driven directly by one or a pair of the better triodes. What impedence does a typical output transformer present to the output tubes?

I'm sure this changes depending on the tube, but what is a happy average? Sorry, I'm not very familiar with tube specs.

The reason I'm asking these questions is that I've been working on a loudspeaker motor design that allows for an exceptional amount of magnetic gap length and height, with high flux density. It occurred to me that a very long underhung voice coil winding (up to 1000 ohms) was conceivable while maintaining wide bandwidth and high effeciency.
The perfect load is pretty much always going to be purely resistive. Winding enough turns of wire to get a DC resistance of 1k is going to give you a pretty healthy dose of inductance, not to mention the distributed capacitance.
That's not to say that it's unworkable. Output transformers have inductance and capacitance, too. Just factor it into your plans.
Incidentally, are you looking for a load for an OTL or an SET? They're two different things.

Yeah, your point about distributed capacitance is a good one.

Without giving away too much (I'm hoping for a patent or two here), the voice coil is air-core and non-cylindrical, so I'm hoping inductance will be minimal, even with a winding of up to, say, 4000 feet of 40 awg aluminum (that'd make it about 6.4Kohms and 15-20W power handling). Cost scales linearly with magnet gap area in my design, so there will be lots of quality gap area to float an underhung coil in, even of this size. (All this is still theory, of course. The physics is simple, but I haven't yet built a proof of concept.)

When I say OTL, I'm not thinking Julius Futterman, circlotron, etc. I'm thinking a loudspeaker load for a plain vanilla high output impedence SET (or any other standard topology) that renders an output transformer unnecessary. The goal is to hear what happens when a single output tube holds hands with a single transducer, without a transformer chaperone between them to spoil the romance.

So when a voice coil is more capacitive than inductive, what does that do to response? Would that make a high-pass filter that counteracted mass rolloff?

I need to get on to other things, but I wanted to take a second to drop an idea on you:
Nichrome. (Or other high resistance wire.)
Yes, I know it'd be hard to get as an insulated wire, but the resistance would rack up faster. That would result in less inductance and capacitance, since they're both going to be proportionate to the number of turns you use.
Fewer turns, however, would result in less flux, hence lower efficiency. A balancing act would result.
Actually, I believe nichrome is somewhat ferromagnetic. You might run into hysteresis problems. Anyway, don't limit yourself to copper and aluminum.

Thanks, Grey.

I remember reading somewhere about higher-resistance material for voice coils in heavy duty woofers. I think it had to do with enhanced thermal properties. Nichrome rings a bell, so it might have been discussed. I wonder how dimensionally stable it is under thermal load.

Hystersis would be good to avoid, however. That's a big benefit of the air core voice coil.

In my design, I've been shooting for all the efficiency I could, ergo the long, highly conductive windings in underhung gaps. I'd like to generate SPLs from SETs.

However, I'm also aware that, above a certain point, increasing motor strength tends to roll off bass response. Higher voice coil velocity within a denser permanent flux field returns greater back EMF, pinching the voltage available for drive. Of course, the corresponding reduction of current in the coil can be a big plus, but only if the amp has voltage swing to keep pushing, ala Bob Carver's True Sub.

So, getting back to my question, does anybody know a good impedence that could be easily driven by most of the better output triodes? My tube knowlege is quite limited, but don't tubes have a certain load window, above and below which efficiency rolls off? Do these windows overlap for a few good examples?


I see that Nichrome C is indeed ferromagnetic. Nichrome A, however, is not. Resistance is pretty high--650 ohms/CMF. A 2Kohm voice coil winding of this stuff, at 33awg, would only be 155ft.

Manganin looks like a promising compromise at 290 ohms/CMF. Then there's about four nickle/copper alloys at a range of lower resistances.

Ideal Speaker Load for OTL SET amps:
There are many issues at stake here.
Two of the most popular tubes today are 2A3 and 300B. Those in need of more power use 211 and 845. Those who wish to be off the beaten track use 10, 45, 50 and so on and on.
I dislike all Directly Heated Triodes. However, the tube types are there for you to look up the plate resistances from the manufacturers data books. Note that the specified plate resistances are for a given set of operating conditions.

In the design and prototyping stages, I prefer to load the output tube with an impedance 4-5 times the plate resistance and gradually reduce this to reach a compromise between power output and distortion.
Popular loadings:
211: 10,000-18,000 ohms
2A3: 3,000-3,500 ohms
300B: 2,500-5,000 ohms
45: 3,500-4,500 ohms
845: 10,000 ohms

Output Transformer or Output Capacitor:
Performance wise, my preference will be a transformer until I am able to find an acceptable (performance wise) capacitor. In this context, I quote from my reply to SE or PP in a previous discussion at this forum:
“To enjoy the maximum benefits that Single Ended amps are capable of, we need speakers without crossovers. This means that the voice coil of the driver should be connected to the output transformer. In the old days, Philips made 600 ohm voice coils to suit their OTL amps. I wish a few of the present manufacturers would consider making full range (well, almost full range should be adequate for 90% of recorded music) drivers with 600 ohm voice coils. Some 5 years ago, I did take up the matter with Lowther and when I replied that I am willing to purchase 10 pairs, there was a brief silence at the other end of the telephone line and the discussion topic changed!”

Pardon my tube ignorance, but let me try to understand this.

Were those old Philips voice coils 600 ohms only because it was impractical to wind a higher load? Would a higher load be beneficial in otl designs, provided provided speaker efficiency and bandwidth can be kept high?

In terms of SET transformer vs. capacitor, I was thinking of skipping both by allowing idle current through the voice coil. I know that would multiply thermal load on the voice coil and cause forward displacement in the magnetic gap. In my design, what I'm counting on is A.) good intrinsic heat sinking within the motor structure and B.) the voice coil would have some built-in setback so the forward displacement would re-center it. Since the coil will be underhung, it will remain in a linear field until Xmax, no matter its position within the gap.

Alternately, wouldn't there be some benfit to shunting idle current with, say, a 1:2 ratio transformer driving this high-impedence speaker?

I do not know as to why Philips stuck to 600 ohm.

In your experiments, have you tried DC through the voice coil? Did you use a Battery or a regulated power supply?

In terms of Power Output, I do not see any benefit in lowering the turns ratio. Turns ratio is one issue and the number of turns required to maintain a given inductance at say 50Hz is another issue.

Apologise for this delayed post. I am battling with PSpice during my Christmas-New Year break.


Thanks, I really appreciate your input.

I haven't built this driver yet. I've just been fiddling with the voice coil wire types and the numbers that go with them. As soon as I get my thinking straight, I'll reduce it to practice.

The reason I ask about the winding on the Philips OTL speaker is because I'm looking for feedback on whether or not a high-effeciency, wide bandwidth loudspeaker of 2Kohms (or pick a number) would be any kind of boon to music lovers.

As I wrote previously, I have a loudspeaker motor design that I believe would lend itself very well to this possibility. So now I'm wondering what kind of impedence to shoot for. 500 ohms is obviously easier to engineer than 2000, but I believe this and more is possible if it would free amplifer designers to explore new, higher ground.

In my limited understanding of tube amps, these possibilities include:

A) Balanced OTL designs (circlotron, et al.) using fewer output tubes at higher voltage, allowing greater efficiency, lower distortion, and lower cost.

B) Lower turns ratio output transformers (if this is a benefit) or, in low power designs, possibly allowing bias current through the voice coil.

C) Allowing high output impedence amps to still have decent damping factors.

I know the trouble for most will be to suspend disbelief that this motor design can avoid some of the compromises that are inevitable in more standard designs attempting high impedence. Leave that up to me.

So just dream with me for a second: Is a high-impedence full-range loudspeaker worth bringing to life? If so, how high impedence?

Hello there.

I think you have something revolutionary going on if you manage to make 2kohm loudspeakers that are reliable. I'm not quite sure about those impedance ratings, but 2k sounds much better than 600 ohm to me. Anyway, there are a lot of tubes that should work easily with 2k. I suggest that you contact other tube fanatics ;) also. Try for instance with news://rec.audio.tubes or some other rec.audio newsgroup. If this design is possible to manufacture, you got a (at least a small) market for your product, I think!
Thanks for the link, Johan.

Yeah, I've noticed that this forum is light on the tube, heavy on the solid state. That's fine, I haven't sold my soul to either.

I haven't created a proof of concept of my driver yet, but I do believe very high impedences are possible. I'll be especially interested to see how low the self-inductance of the voice coil will be compared to normal cylindrical steel-core voice coils. There isn't really a way to simulate it, but my first attempts seem to indicate self-inductance will be lowered by a factor of four, at the very least.

By the way, can anyone tell me how typical output transformers pass anything close to the spectrum of human hearing with such great inductance values? Having displayed my tube ignorance, I'll now display my output transformer ignorance.

I beg y'alls pardon,

I said inductance when I meant distributed capacitance.


Does anybody know the approx. capacitance of good output transformer primaries? I understand that HF rolloff is determined by the LC filter created by leakage inductance and distributed capacitance. Which is usually the most irksome/most limiting factor?


Certainly, I would encourage the development of a high impedance speaker. I would buy the first pair that you are happy with.

You have put me on a "Dream-About" mode. In the 'Down-under' part of my sphere, we are all hanging upside down like bats and so there is more blood rushing into one's head. Give me a week or so to think this through. Are you located in Queensland, the land of Banana Benders?


DC through voice coil will hum…as far as the majority of power supplies out there are concerned. I am talking about Single Ended Amps here. In all amplifiers, we are attempting to deliver the energy stored in a power supply to the load and preferably in exact variance with the signal input.

There is no universal load that I can come up with for your motor. In terms of load applied to tubes, there are 3 broad categories.
Group 1: Low rp tubes (load <2000 ohms) such as 6AS7, 6336, 6C33C, 7242…etc
Group2: Medium rp tubes (load 2000-5000 ohms) such as 2A3, 300B, 45 etc.
Group3: High rp tubes (load >5000 ohms) 211, 845, most indirectly heated tetrodes and pentodes…etc.

Your motor may call for tubes with high current capability such as those in Group1. I have not modelled these devices with high impedance loads and I know that these are not in the same category as transistors in terms of current capabilities.

Perhaps it may be more prudent for us to imagine the primary of a transformer as a field coil to energise a temporary magnet and the secondary as a freely suspended voice coil. The field coil is then connected to an output tube swinging say 80V-180V and 50mA-250mA. Will this energy be sufficient to induce a proportionate movement in the voice coil? Or should we drive both the field coil and the voice coil? The term voice coil is used here merely to address this winding separate from the field coil.
Let us have your thoughts.


You've got me thinking in a whole new direction now, let's see...

If you modulate a field coil instead of the VC, the VC would still have to set up a field of its own. Maybe the VC could be a few lowish-impedence turns of wire and run stable low-voltage DC through it, say from a large 6V battery. That would set up a field for the field coil to work on. Back EMF from the VC would only see the battery...might be a plus?

The field coil could be made robust to easily handle idle current from the amp and you could wind any load you want since you're free of VC mass/size/inductance constraints. If you run it single-ended, you'll avoid hysteresis distortion from reversing the field on the core.

The secondary winding (in this case, the VC) could be comparatively short and light. Incidenly, are there worthwhile benefits to the low VC capacitance/inductance that would result? How much of a negative factor is distributed capacitance in the secondary winding of a typical OT?

I guess my main worry is this:

If you effectively have a primary (field coil) with no secondary to couple to, doesn't that mean that the primary inductance is actually equivalent to leakage inductance and therefor greatly bandwidth limiting?


In audio output transformers the capacitance of the secondary is reflected to the primary as the square of the turns ratio. For example;
Primary to Secondary turns ratio 10:1
Capacitance of secondary winding18pF
Reflected primary capacitance 1800pF

Output transformers representing impedances between 3000-5000 ohms usually have reflected capacitance of the order of 3000pF.

I hope this answers some of your questions.

This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.