Rerurn to Romy the Cat's Site

Audio Discussions
Topic: An amplifier for Tweeters

Page 1 of 1 (18 items)

Posted by Romy the Cat on 09-03-2006

Following my recent “new objectives” with tweeters (partially covered in the last posts of the following thread)

I lately made quite a few very interesting and very educational experiments with tweeters. I would like to capitalize on some of my findings and now I think to use my tweeters differently then I use to do it in context of my new 5 channels:

Anyhow, while I was doing my experiments I was driving tweeters with own dedicated amp and… I kind of like it. I was even using my headphone amp:

to drive the tweeters, it was not enough power to drive them all the way up but it was still very nice and “enough” at the low listening levels. (The headphone amp has no gain and made to have beter bass then HF). So, I become wondering that I would be very-very interesting to have a dedicated channel for tweeters. It is not really necessary this channel to be a full blow amp but it should be something very optimized for over 10.000kHz reproduction. Obviously it might be single-ended amp with no feedback. It should have 10-12dB gain. It should be one-tube amps. If were a SET then the transformer should have a few turns and somewhere a fraction of Henry inductance and “fast” core or perhaps an air core:

, or perhaps even OTL?

I was looking what Vincent Brient proposed in his “DIY 3W 6C33B-C SE OTL amplifier for compression driver and tweeter”: 

I do not know about the compression driver, as I do not like what OTLs do with sound harmonically but for tweeters it might be interesting (almost certainly not around the 6C33C)

So, I am wondering. I do have in the Super Melquiades the necessary voltages and space to put one extra single-tube channel.  I am thinking about a single-gain stage on the triode strapped 6E5P, 7721, 6C45/437A, EC8020, RS241 or some other fast RF tube and a toroid air-core transformer (no hysteresis, no saturations, no eddy current losses...). I might be very-very interesting and perfectly accomplishable. I welcome any ideas at this point.

Romy the caT

Posted by cv on 09-04-2006
Hi Romy,

I think you may be struggling to get 10-12db of gain out of a spud amp (the term is a play on "one tuber").... May just about get it with 6C45.... remember you will be stepping down something like 12 to 18:1 on the OPT.

best case: 10db required (gain = 3.15), transformer is 3k:16, n=13.69, so gain of the tube stage needs to be 43. This needs a mu of about 57 from the tube assuming an RI of about 1k... Need to find high mu examples of the 6C45. With a 2k5 load, I get a mu of about 55.

The thing with an air-core is that you need as low a driving impedance as possible. 6C45 at high current may well be ok, but the best tube in that class is the E55L triode connected, which has an Ri of about 600 ohms IIRC. This means that you could get away with 50mH of primary inductance. But it doesn't *quite*  have the gain you need.

I built some air cores to drive a ribbon project which is on long term hold. I haven't listened to them but they measured reasonably well, something like -1dB at 19kHz when driven from 600 ohms. The primary inductance was about 250 mH and the turns ratio was huge as they were designed to be 3k:0.6 ohms.

I don't have the design to hand but can obtain it in a couple of weeks when I head back home; it was something like a 3" diameter cylinder, with 20 or so grooves machined into it. Each groove was a 3 mm wide and a few mm deep, and the primary.secondary sections were alternately wound P-S-P-S etc. There was also an axial trough machined into it to enable one to get to all the terminations easily.

However, I suspect that the bobbin from a standard SE transformer, sans core, may come close to meeting the specs you require, in which case, just purchase a pair without the iron and give it a go. Maybe give Bud a call?

The only issue I can see might be signal radiating from the cores.

Alternatively, for a fast core - a whlle back I had Sowter wind some 100% mumetal outputs, good for 2kHz up (135mA, 3k:8).  They sounded absolutely *magical* on my tweeters. I hope to get them rewound for use with the S2s and a different OP valve at some point... but the point is that a 100% mumetal core in your application could be tiny, and very, very "fast", whatever that means...

It's a tradeoff between:
high interwinding capacitances and no-core (air) or
 magnetic effects and low turns count, low capacitance (mumetal).

I can't guess which would be preferable, especially with the 60kHz cutoff on the tweeter x/o!
If you can persuade a winder to lend you an existing bobbin or have some you can use, that would  be a quick test of the concept.

Brainwave - the 7788/E810F has a higher gain than the 6S45.... should just about do it!

Hope this helps,

Posted by Romy the Cat on 09-04-2006

Thanks, Chris.

Dima I discussed today many imaginable scenarios. Initially we thought drive a single 6C19P from MF channel of super Milq. But it has some negative consequences… Eventually we decided (it is not final decision but the final with 90% certainly) the in order do not compromise anything to do for another full blown channel of Super Milq only using 6C19P as output. So, most likely it will be 6E5P to 6C19P and then will be sitting in own sub-chassis above the Milq chassis (in the empty spot between the large caps and tubes). I have all voltage and perfect PS right inside the Super Milq chassis and the capacity of those PS is good enough to drive another 4434 channels. So, it would be a little work to put it all together. The 6C19P is a miniature version of 6C33C and they are basically the tubes of the same family, so it might be good.

In addition I might filter LF out by coupling cap making the transformer job easer. I hope that in case of a driver operating above 12Kz (I will be back to the first order on my tweeters) the sonic problems that I had with Super Milq MF (when I went line level filer) might not manifest themselves…

For Tweeter I decided to loose a little gain but drop the harmonics, loading the tube more idle (1:15). So, not the only problem is to fine an interesting transformer. It would be an air-core of some fast metal core of 0.1H and gaped for 200mA. It looks like I will be able to get around 4W in this setting that it much more then I need. If you know sores for such a little Trans then let me know. If I have the transformers and it will be really a half day project.

Romy the Cat

Posted by cv on 09-04-2006
If you want to go for a cored device, I would recommend Sowter in the UK for a custom unit on a 100% mumetal core. I'm not sure what the delivery time would be though. 

You would definitely need to filter the LF before it hit the output stage if going for a mumetal core. The other thing is that these cores with large gaps can sing along to the signal, so you may wish to try and cure some resin around it if it becomes an issue.

If you decide to go for an air-core - well, any transformer winder should be able to wind the required bobbins for you. I'd proably go for the air-core myself... 

good luck,

Posted by Romy the Cat on 09-04-2006

Sure, Chris, if I would be able to make the air-trans of acceptable size then I would defiantly go for it. If I pretend that the Super Milq pushed some boundaries (and I do feel it) then to use an air-core transformer for HF channel is exactly what in on the specs of such an amp. I have no doubt that the air-core will do wonderful and I also, since it will be very little values in there I was planning to use my collection of vacuum caps between the sages. Still, I would like the amps with the new channel to look and to “feel” good, ergonomically and ideologically. BTW, a very crazy, almost cartoonist idea juts hit my mind. Because it looks like there is nothing ferromagnetic in electrolytic I wonder why don’t I wind my air-coil around my large caps  Talking about the looks and the ergonomics! I think then I would need to rename the Milq into the Tesla-Melquiades Coil :-)

The caT

Posted by cv on 09-05-2006

Ok, so it would look cool, but I'm not sure inducing an HF signal into your caps and having anything ferrous near the coil is a good idea. You'll get eddy currents even if there are non-ferrous conductors around.

How this affects this end result is another matter of course...


Posted by Romy the Cat on 09-05-2006

Well, I do not care about it anymore. Last night I spoke with Bud and, thanks God, he took this project form my hand under his wings. As the result I will be having whatever I need, implemented by the best method possible. I am very enthusiastic and my tail is trammeling.

BTW, I told Bud about your experiments with air-core and he asked pictures and more technically civilized description. I did not find on my mail server the pictures as it was quite long time ago. Can you upload the images to the site or send them to him directly?

Another BTW, Bug come ups with an idea of increasing the primary and secondary coupling in my air-core transformer by separation the electrons by sex in the different coils: male electrons will be in primary and the females electrons will be in secondary. Presumably those toe groups of electrons be naturally inclined to go closer to each other and that will increase the magnetic union of the coils.  However, I know that I will be working is Seattle but I do not know how about Boston. Since Massachusetts alloy gay marriages it might loose some coupling in my new transformers.

The caT

Posted by cv on 09-05-2006

Allo R,
I'm afraid I don't have access to them or pictures for a while. I will try and get some towards the end of next week.
However, I did just find an old post of mine on one of the other forums ... I think this should give Bud a decent idea of what I had.

The device is basically a vertically sectioned bobbin.
The bobbin is a solid nylon cylinder, outer diameter 70mm, length 78mm.
Into the cylinder were turned 17 grooves. Each 3mm wide, with separation 1.5mm between them.
Can't remember how deep they were, but  I seem to remember 5 or 6mm.

Into each groove was wound a winding section. There were 9 primary sections, 8 secondary.
My winder managed to get 220 turns in each primary section. Afraid I can't remember the wire gauge used. Total of 1980 primary turns.
Secondary was something like 3 (!) turns per section, 24 turns total,  for a ratio of about 82:1

The windings were arranged P-S-P-S....-P

182mH for the primary

Driving from a claimed 600 ohm sig gen output, I got a bandwidth (-3db) of about 650Hz to 48kHz.
This, btw, is at a ratio of 82:1 into a 0.5ohm load...

Hope this is of use.... you know, you could probably get a *magnificent* OPT along these lines for the upper S2 as well...
Following your theme above, I'd be inclined to omit the bobbin and have the coils naked...or maybe dipped in hot wax to provide just a little support....


Posted by Romy the Cat on 09-13-2006

OK, I have conceptualized what I need for the tweeter amp. It will be MilqEZ – the same Milq one with 6C19P at output. I will use the Super Melquiades power supplies but the HF channels will host own last caps of the filters and own bias of the both stages. It will also have a vacuum coupling capacitor and air core output transformer. I did some estimate for the space and layout and come up the design of the enclosure… It is not finally painted (there are no holes in there) but the idea is there.

Also, in the enclosure for this HF channel I will accomplish something that I always intend to do. I will put in the elapse time meter to monitor the hours on the tubes. I have no idea way such a useful tool is wildly used.


Posted by Romy the Cat on 09-16-2006

I am putting temporarily the project with tweeter amplifier on hold. The problem is that I can’t get what I need form my tweeter.

Since I removed the T350 tweeter from transition slope and begin to use it as it “meant to be” I do have good HF but it is completely not suitable with what S2 driver does. It is not sound that I would consider acceptable. I still will go for HF SET but I will look more for different HF solutions, of different use of my T350, and therefore the amp might have different output criteria and power requirements…

Romy the Cat

Posted by Paul Butterfield on 09-24-2006
Hello Romy,

Which other output tubes have you considered? I researched ‘suitable’ output tubes to drive tweets a while back for compression drivers [Be domes] — but not phenolic domes. Reading of your results with the EV tweet reminded me of how sweet they sound.

The tubes that made it to the top of my theoretical ‘A’ list were the 10Y, the 45, and the 46.

I have finished collecting the above NOS/UOS (and others) and will start breadboarding this fall, health permitting.

I hope to be able to do a bit of experimenting with some phenolic compression domes. For the first Frankensteinian co-combination, I’ll replace the Be dome on a TAD 2001 with a phenolic epoxied to the TAD surround VC. I have only one used old stock phenolic that can match — with some removal of the junction phenolic/VC surround — so I’ll start with a mono proof of concept first.

Did you mention a potential source for replacement or currently manufactured phenolic domes?

Paul B

Posted by Romy the Cat on 09-24-2006

The EV 350 cones are still available form many sources, they are $25 per cone. There are some other phenolic domes for 1” driver, or smaller. I do not know anyone who does 3”-4” phenolic domes that would make them suitable for MF drivers.

As far as output tubes for tweeter amplifier – I do not know – I’m kind of in tweeter crisis and unlit I fugue out what tweeter I am will be using I do not know what would be power requirement for output state. If I stay with EV350 then my original idea with 6C16P would be ideal. However I might go for some crazy tweeters (ribbon, plasma, electrostatic, Hails, silk cone, different compression driver, inverted compression driver or something like this…) and then I might need more power. As I said now I am in confusion as I recognize that the way in wish I use my EV T350 recently does not work for me. It is not about the tweeter but rather be able to wok with very specific sound that Vitavox S2 does, and it is very hard to find as I learned…. I did not spend any efforts to look for different solution this time. After I nailed down the tweeter and it’s application then I will be able to see what amp would be able to drive it. For now I suspended all efforts to male the tweeter amp and to wind the air transformer…

The caT

Posted by Bud on 09-24-2006

Now that you have removed the slope try out the treated tweeter I provided you with so long ago, the one that sounded exactly like the untreated EV's. We would all like to know how that matches with the S2, in an unlimited application..... might just be the answer here.


Posted by Romy the Cat on 09-26-2006

Bud, perhaps it does makes sense to try the “dotizm” one again. I quite severely review my view on the way in wish I use tweeter. I still am trying to make T350 to work, let see what I end up with. Nowadays my T350 sitting in 3rd order crossover, it does not do exactly what I need but it is very much not final setting. There are two major issues that I deal with: precision of the alignment and presents of something that call “HF sonic aura”

The precision of the alignment is a bitch and the blessing, depends how it is being used. I have written quite a large article about the “precision of alignment” but did not post it yet as I would like to bring my tweeter to the state when it does the necessary for me things and where I would have some moral rights to educate others based upon my experiences. In short: after a tweeter was aligned there is a next level of alignment, sort of “DPOLS for tweeter”, when further improvements become geometric instead of incremental. Those “geometric improvements” in fact are capable to cancel out many nastiness, those that that we usually attribute to “digital” or “recording problems”. I will be talking about it when I will be through with my experiments. Still I was much enthusiastic when I discovered that effect of “DPOLS for tweeters”...

The “HF sonic aura” is another thing that I’m trying to deal with. The Vitavox S2 in my horn has extremely precise and “none-foggy” tone. It has one “issues” but the “HF fogginess” is not one of them. The T350 does have it “HF fogginess” or sort of aura-noise that envelops HF presentation. It is not so bad itself (99% of tweeter do it) but it is completely not acceptable in contrast with Vitavox S2’s high frequency. The T350 just contaminates and dilutes the S2 sound and it is doe not sound kosher. I was trying to lower down the EV-T350 calibrating it not by it’s output but only by the amount of it’s “HF sonic aura” and I end up with something in vicinity of sub 0.1uf against 16R… or I ended up with same use of the tweeter on the transition slope at over 60Khz

Anyhow, I will see how it goes further. I would wish I have enough time to render all of this. Unfortunately it is very painful and time consuming ceremony as at the level of precision when “DPOLS for tweeters” take place the measurements are not working anymore (at least at the level in which I’m able to conduct them) and I have to use just hearing.

Romy the Cat

Posted by Romy the Cat on 09-28-2006
Well, it looks I’m near finding a more or less acceptable setting for the T350 tweeter. Ironically and completely accidentally it is not far form I use to be. It is still on the slope at 20Hz (since the T350 is pretty dead at 16kHz) but now it is  on the third order slope. It use to be at 65kHz slope of the first odder and I was arriving at the Vitavox S2 upper knee at approximately minus 8dB. Now I am running from slightly above 20kz and I arrive to the same crossover point at minus 9dB. Since T350 can not do at it’s lower knee what S2 does (I doubt that any tweeter would – actually it is scare to hear what a well tempered S2 does at 10kHz ) then the their order I find is beneficial, and practically given the sharp S2’s decay. The time alignment is still a complicated bitch but I am near the target already, very close. What I sumized that the way in which I mounted the tweeter is very good but not good enough to set the necessary level of precision for time alignment. I will not redo the mount bracket as it would require a very sever revision of everything and I do not have a desire to dive in there. Still, I would wish I have something like this.

The caT

Posted by Wojtek on 09-29-2006
I plan to attach tweeter in a place of the lense to bellows (I hope thats correct term) used in macro -photograpfhy .They look sturdy enough and should provide 10cm of fine ,easy adjustment and I can buy them cheap ($10) on any photo show. Regards, Wojtek

Posted by Romy the Cat on 09-29-2006

 Wojtek wrote:
I plan to attach tweeter in a place of the lense to bellows (I hope thats correct term) used in macro -photograpfhy .They look sturdy enough and should provide 10cm of fine ,easy adjustment and I can buy them cheap ($10) on any photo show. Regards, Wojtek

Good idea. If your tweeter is light enough then it might work, but you would need to find a way to attach it very strong to the Bellows. There are a lot of Precision Positioning Tables out there that also might be used. The key in here is not to have “Precision” but rather predictability and the comfort to mount a tweeter on it. People who use box speakers is another story but for the horn folks who has no reference baffles to have a calibrated Positioning Table is a good idea. Then the question how to make the Positioning Table to be fixed firm in relation to MF driver… anyhow, Wojtek, when you will be thought with your Bellows then can you post the images how it ended up?

Now I am back to the amplifier for tweeter subject….

The Cat

Posted by Romy the Cat on 09-19-2007

Referring to my very recent acceptance of the 6E6P as a tetrode to drive the tweeter:

I decided to today to stick my beloved 7721/D3A in place of the 6E6P – thanks God they have the same pins layout. I dropped bias to 1.5V and fired up the amp driving 20mA though the D3A’s plate. The 6E6P is has transconductance 35mV/A +/- 5FRT. I measured 5 D3A and all of then has exactly 36mV/A – virtually the same. So, I lighted the amp on and … it worked. I did direct some differences but I have 6dB per octave now in tweeter and when I will have 12dB, and particularly in the slope, I doubt that any difference will be observable. I wish I have a 9 pin tube like this with 50mV/A…

The Cat

Page 1 of 1 (18 items)