OTL for dummies.

[Spasticteapot]

As I tinker with my el-cheapo guitar amp, I've been thinking about what I build next. A few guys I know need a new amp, and any project where I make money is a good project in my book.

There are oodles of very nice tubes on eBay. However, output transformers - good ones - are rare indeed. New OPTs, courtsey of Hammond, will cost at least triple the cost of the tubes in the amplifier.

So, output transformers are expensive. (No surprise here.)

Why not ditch it?

Using various combinations of tubes, Phillips made some pretty terrific OTL amplifiers during the 40's and 50's. While these amplifiers did use 600 ohm speaker drivers, I figure that I'll be safe if I triple up the output tubes and use a 192 ohm load. This should also give a reasonable amount of output power.

http://people.cs.uu.nl/gerard/RadioC.../SerBalOut.htm

Then, there's also the issue of high-power tubes used in Russian fighter jets. I very much like the idea of using something intended for killing people instead to play Jimmi Hendrix, and the price is right. It appears that many of these amplifiers' efficiencies are limited by output impeadance; in the case of a guitar amp, 32-ohm drivers are not uncommon, and I can use four in series.

A few examples:
http://www.tubebuilder.com/schematic3.html

There's also the guys at Tubecad, who have a similar project with 12AS7s.
http://www.tubecad.com/2004/blog0018.htm

Furthermore, power transformers are also expensive. (Also, water is wet.) To keep down costs, I'd like to instead use a pair of rectifier tubes to form a bridge rectifier. Is there a reason why this is a bad idea?

[SY]

If your last comment meant that you're considering dispensing with a power transformer, don't even think about it- that's deadly dangerous. More like Stone The Crows than Jimi Hendrix.

[kevinkr]

Sounds like an interesting project, I can think of many reasons why it might be ill-advised, but regardless if you are curious you should try it. Read my missive on my ill-fated OTL project on my site if you have a moment, there are some pitfalls you can avoid.

One thing to bear in mind is that most otl designs are not intended to deliver high levels of output current on a continuous basis, given the trend towards heavily overdriven and saturated tones this might be a serious issue in an otl guitar amplifier.

No reason not to use a bunch of rectifier tubes in a bridge if you are so inclined, just remember to choose ones that can handle the load current. It would also be good if they have a long enough time delay so that no special warm up precautions need to be taken to protect the output tube cathodes. (6C33 will arc if presented with 170V on the plates with not fully warmed filaments... )

Certainly in this application a 4 speaker stack would be your friend if wired in series, you could even go with 16 ohm drivers for a whopping 64 ohm load - this is definitely otl friendly territory.

Avoid the temptation to over complicate, an inverted futterman or futterman is easier to get working that most circlotrons.

I would recommend staying away from the 6C33 unless you have the time and energy to test a lot of tubes after a lengthy burn - in. 6080 are very cheap, rugged and have more modest filament requirements, a quartet of them should put you in about the same power range as a pair of 6C33.

[Jeb-D.]

Certainly in this application a 4 speaker stack would be your friend if wired in series

You read my mind Even four 8 ohm drivers in series would be pretty OTL friendly.

[Spasticteapot]

Originally posted by SY
If your last comment meant that you're considering dispensing with a power transformer, don't even think about it- that's deadly dangerous. More like Stone The Crows than Jimi Hendrix.

Actually, I figured I'd just use some nice, big isolation transformers - likely, two isolation transformers to give me 240v with a center tap. The point of using the two rectifiers is to get me double the wattage.

Originally posted by kevinkr
Sounds like an interesting project, I can think of many reasons why it might be ill-advised, but regardless if you are curious you should try it. Read my missive on my ill-fated OTL project on my site if you have a moment, there are some pitfalls you can avoid.

Fair enough. I was'nt aware of the issues of output current - but, then again, this is a guitar amplifier. Distortion is not necessarily a bad thing, nor is odd characteristics.

Originally posted by kevinkr
No reason not to use a bunch of rectifier tubes in a bridge if you are so inclined, just remember to choose ones that can handle the load current. It would also be good if they have a long enough time delay so that no special warm up precautions need to be taken to protect the output tube cathodes. (6C33 will arc if presented with 170V on the plates with not fully warmed filaments... )

So, why not just add a delay? I figure I'll be using a seperate transformer for the heaters, so I can add some sort of timed delay circuit with a relay.

Originally posted by kevinkr
Certainly in this application a 4 speaker stack would be your friend if wired in series, you could even go with 16 ohm drivers for a whopping 64 ohm load - this is definitely otl friendly territory.

Actually, Jensen makes a variety of 32 ohm drivers. I figure that by using a 128 ohm cabinet, I can improve efficiency by a large margin.

Originally posted by kevinkr


Avoid the temptation to over complicate, an inverted futterman or futterman is easier to get working that most circlotrons.

I would recommend staying away from the 6C33 unless you have the time and energy to test a lot of tubes after a lengthy burn - in. 6080 are very cheap, rugged and have more modest filament requirements, a quartet of them should put you in about the same power range as a pair of 6C33.

6080 it is, then! I'm all in favor of a simple, reliable schematic. And I'm really, really in favor of cheap tubes.

[Audio_idiot]

Hi,

I have half dozen of Sovtek 6080/6AS7 collecting dust somewhere in house. Would love to tag along with this thread and hopefully I can learn enough to be dangerous and make good use of them...

Just recalled I also have 2 quads of 6C41C lying somewhere....... Will they do as well?

Cheers

[SY]

Spaz, isolation transformers will work perfectly. I'd avoid tube rectifiers- OTLs are very demanding of low source impedance from the power supply.

Kevin, what was the main issue with 6C33, part-to-part variation?

[LineArrayNut]

these are copies of russian 6H13C themselves copies of 6AS7G, but anyway in my totempole amps they are MUCH more rugged, sound better and are well behaved. $6 USD each @ diy hifi supply

[Jeb-D.]

(6C33 will arc if presented with 170V on the plates with not fully warmed filaments...

I've never seen that and I've built quite a bit with this tube. Though every circuit has had a choke, resistor or relay controlled current inrush limiter in the power supply.

As far as your complaints against the 6c33c most are valid, I've also bought a box of them and found them to be quite mis-matched and they do change after burn-in. Though 6as7's, and pretty much any high Gm/ low Rp tube can also suffer from this. 6c33c's can still be good for some situations. In a circumstance like your 6 output tube Circlotron, the 6c33c could be quite difficult (It's a very impressive looking peice though). When used in fewer pairs and in variations of futterman there not a bad tube at all. Futterman spinoffs are a lot less prone to biasing drift causing DC offset on the output when you leave the center tap of the power transformer floating.

1 pair of 6c33c's into 32+ ohm load would probably produce really good results. It's not too hard to match 1 pair.

[kevinkr]

Originally posted by SY
Spaz, isolation transformers will work perfectly. I'd avoid tube rectifiers- OTLs are very demanding of low source impedance from the power supply.

Kevin, what was the main issue with 6C33, part-to-part variation?

Originally posted by Jeb-D.


I've never seen that and I've built quite a bit with this tube. Though every circuit has had a choke, resistor or relay controlled current inrush limiter in the power supply.

As far as your complaints against the 6c33c most are valid, I've also bought a box of them and found them to be quite mis-matched and they do change after burn-in. Though 6as7's, and pretty much any high Gm/ low Rp tube can also suffer from this. 6c33c's can still be good for some situations. In a circumstance like your 6 output tube Circlotron, the 6c33c could be quite difficult (It's a very impressive looking peice though). When used in fewer pairs and in variations of futterman there not a bad tube at all. Futterman spinoffs are a lot less prone to biasing drift causing DC offset on the output when you leave the center tap of the power transformer floating.

1 pair of 6c33c's into 32+ ohm load would probably produce pretty good results. It's not too hard to match 1 pair.

Hi SY,
I had a lot of issues with section to section mismatch. I burned in several at a time and selected the ones with the best section to section match by running one filament at a time and measuring the plate current at two plate voltages and three different grid voltages - very tedious and time consuming, and of about 32 I purchased I ended up with about six that were pretty closely matched, say within 5%, another six that were within 10% and a number that were so far apart as to be useless. I also had some failures in the test jig.

Hi Jeb-D,
I had no time delay and no current limiting and the original Russian language 6C33 specification is very emphatic about not applying plate voltage until the filaments have fully warmed up. I violated this rule and had repeated arc overs, in the first instance it destroyed a woofer. My fault totally of course for not heeding the warning. I think in your case there may be just enough of a delay to avoid this issue. Others have also mentioned that internal contamination might account for these arc overs and that a long burn in may eliminate this as an issue.