• 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.

Differential SE Amp

Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
frugal-phile™
Joined 2001
Paid Member
I was speculating today. The picture:

mirrored-50EH5.gif


I used my a variation on my 5-buck amp as a prototype because i have 4 identical channels in the wings with better OPTs. This would be integrated with a pair of bi-poles like my BD-Pipes.

This is intended to be a Frugal-phile(tm) oriented pheonix (risen from the dead) amp. The less is costs the better.

Does this work? Are the speakers in phase? Am i missing anything? If you put an input transformer on the front for single ended in (ie Edcor 10k-10k - is this high enuff R?) do you need the 47k grid leaks? Does the DC in the OPTs disappear?

dave
 
I expect you meant to show both ends of the transformer secondaries joined, i.e. paralleled and in phase. The primaries out of phase. This will drive the speakers ok but then you only need 1 not 2. Both cathodes sharing the one constant current source is something I have been thinking about too. Makes the circuit a power long tailed pair and provided you have both tubes in class A it should be very well balanced even if the grid signals are unbalanced wrt each other.

Thinking further, with a normal single ended output tranny it has an airgap so it can tolerate the dc flowing in it. All the E laminations are on one side and the I laminations on the other with a strip of paper or something in between. Maybe you could try removing the stack of I laminations from each tranny and the paper strips, and face the open side of the E's of each tranny together. If the primary phasing is right, the two core halves will repel each other because of the dc flowing in each half. Clamp the core halves together and the dc magnetic flux will be neutralised. Should give better performance all round.:)

GP.
 
frugal-phile™
Joined 2001
Paid Member
Circlotron said:
I expect you meant to show both ends of the transformer secondaries joined, i.e. paralleled and in phase. The primaries out of phase. This will drive the speakers ok but then you only need 1 not 2.

Like this?

parallel-opts.gif


Maybe you could try removing the stack of I laminations from each tranny and the paper strips, and face the open side of the E's of each tranny together.

I'm not yet ready to try reassembling OPTs :^)

dave
 
I would consider doing either:

1. Use a balanced setup to get double output voltage and thus 4 times the power (minus a bit)
2. Use the two setups to drive the speakers in paralell much like your first post.
3. Use a balanced setup to drive the two output transformers in series (phase to phase) with separate outputs to each speaker element. This would be balanced and also separate the speakers from each other.

With your second picture, you will should consider looking for a way to share current between transformers.

Going to a zero flux variant as suggested above is a good idea. Be ware thought that flux around zero can be a little troublesome which is one of the reasons for having bias flux. You will get many opinions on this.

I have not tried it myself, but www.tubecad.com has a tool that looks very interesting for designing tube amps with transformers.

Petter
 
frugal-phile™
Joined 2001
Paid Member
Petter said:
1. Use a balanced setup to get double output voltage and thus 4 times the power (minus a bit)
2. Use the two setups to drive the speakers in paralell much like your first post.
3. Use a balanced setup to drive the two output transformers in series (phase to phase) with separate outputs to each speaker element. This would be balanced and also separate the speakers from each other.

You will need to hand-hold me a bit here, i'm not quite sure what you mean. The 1st pic has the drivers wired in series, and i thot the OPTs in series as well.

With your second picture, you will should consider looking for a way to share current between transformers.

Share current? And i'm still haven't convinced myself that, as drawn, the OPTs are in phase with one another.

I have not tried it myself, but www.tubecad.com has a tool that looks very interesting for designing tube amps with transformers.

It does look good, but is Windoz only.

dave
 
frugal-phile™
Joined 2001
Paid Member
dice45 said:
this is purely PP differential then.

That is what i was pushing towards. I'm not convinced the secondaries are wired right thou.

Two SE tranies will have a DC tolerance any PP trannie only can dream of .... should be a very stable setup.

The little OPTs usually have a problem with saturation in the LF, so i was hoping this arrangement would help.

What about an input tranny for SE input?

dave
 
Handholding does not seem to be required

quote:
--------------------------------------------------------------------------------
You will need to hand-hold me a bit here, i'm not quite sure what you mean. The 1st pic has the drivers wired in series, and i thot the OPTs in series as well.
--------------------------------------------------------------------------------

The first picture is basically a long-tailed pair which is an electrically balanced connection. However, there are a couple of issues. First, remove the connection between speakers as it is not doing anything other than to ensure the trannies are at the same base potential (as well as confuse you). Secondly, the transformer SYMBOL phases are incorrect as I see it (from memory). The dot sides should be in phase -- ie you don't invert phase "between dots" but that is a mute point. The connections you have made are correct from a phase point of view but you are inverting total phase at speaker.

What you have is balanced input driving split loads. The normal way to use balanced outputs is to take them between the phases, not separately for each phase. Consider what happens when you modulate the power supply. Eventually you will get some cancellation, but that is mechanical between speaker elements instead of electrical before which is usually one of the goal of balanced signals. You are correct, however that you should be relatively immune from common mode on the input.

***edit*** I got second thoughts. I guess you are using electrical summing but consider what happens when speaker elements and transformers are not identical between phases ... ***end edit***


quote:
--------------------------------------------------------------------------------
Share current? And i'm still haven't convinced myself that, as drawn, the OPTs are in phase with one another.
--------------------------------------------------------------------------------

The second photo uses electrical summing. The problem with connecting transformers in parallell is that you usually have limited control over how current is shared between windings unless there is considerable resistance in the windings (which there probably is here). Otherwise you need share resistors (say .2-.5 ohm in series with each winding preferably on the secondary) but this will also affect your bass box. However your approach is absolutely valid. You have the same output voltage driving half the "resistance" so you get double the power.

I would consider going for a series connection on the secondary with series configuration of drivers. This gives you double voltage into double "resistance" which gives you twice the power. Thus you don't have to worry about sharing current (which may not be a problem anyhow).

If the amps are very poorly constructed, consider letting two of them drive just one element in parallel + beef up the power supply somewhat. You won't get more power (well you will get something) but you will get a lot more control. If they are really cheap, consider using more than two per channel.

As for using unbalanced inputs you are right. A transformer or simple phase splitter will work for you. You have essentially designed a phase splitter into your circuit. If you connect one of the inputs throught the other's source impedance to signal ground you have it. However this is usually done in a low power stage. If you are looking for common mode rejection per se, consider a transformer on the input which will solve common mode problems relatively well without having to revert to balanced circuitry "inside". Balanced is great, but if your source is single ended, you don't get all the benefits.

Petter
 
diyAudio Senior Member
Joined 2002
Hello Dave,

On the OPT side of the design:
For what it's worth,am I correct in thinking that at the end of the day you're going to end up with a centertapped SE on the primary side and a "regular" or possibly centertapped secundary ?
This would theoretically combine the best of both worlds but would this be economically viable ?:angel:

I'm curious to see what this is going to sound like.Looks promising though.Keep on tinkering...

One caveat: Dc resistance on both SE trannies should as symmetrical as possible for stability.
If executed with care on the centertapped SE OPT this should come naturally.

Greetz,
 
frugal-phile™
Joined 2001
Paid Member
fdegrove said:
For what it's worth,am I correct in thinking that at the end of the day you're going to end up with a centertapped SE on the primary side and a "regular" or possibly centertapped secundary ?
This would theoretically combine the best of both worlds but would this be economically viable ?

Frank,

The intention is to use a pair of SE OPTs connected together at the B+ end to get a PP OPT. A chat with Bill Perkins at PEARL got things straight enuff in my head to make it clear how the OPTs need to be wired (ie what Circlotron said)

Since i started this threa, Decware has started shipping such an amp -- The Torii.

I have iron for at least 4 prototypes -- 1 50EH5 & 3 EL84s. The primamry pieces of the chassis for the 1st one are cut (a plexi version of the Fi X-amp with Garry Oak accents). I'm going to use some trafo phase-splitters i have gathered up.

dave
 
Dave,

This concept was discussed some time ago on RAT.
The conclusion, as I remember it was that if you get 2 SE amps and bridge them, you'll end up with the harmonic signature of a P-P amp. Also, if you bridge connect 2 P-P amps, you end up with (almost, but not quite) the harmonic signature of an SE amp.

By bridging, I mean feeding each with opposite phases and connecting the outputs in series.

BTW This is a real torii:
 

Attachments

  • torii.jpg
    torii.jpg
    9.3 KB · Views: 437
frugal-phile™
Joined 2001
Paid Member
Mach_Y said:
sounds like a great project.... any thoughts on going to el34's?

Maybe down the road... i have a lot of iron to go throu and experience to gain before i go out and buy iron for a project like that.

When i get all the amps cleared off these shelves and everywhere else i have them tucked away then, if the prototypes work out, i may tacle something larger.

dave
 
Re: TORII

fdegrove said:
Hi,



Does that mean gate(s) or something like that?



The principle should work with any output tube.

Cheers,;)



The Torii is a Vietnamese symbol denoting warriors. It is the symbol given to the Rakkasan's of the 101st Airborne Division (Air Assault) for combat operations during the Vietnam war by the Vietnamese people. They are the 3rd Brigade of the 101st. The red Torii can be seen at their headquarters, on their vehicles, and on their uniforms.
 
Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.