The T-Rex SET Amplifier DIY project

[paba]

Looks a little busy, and heavy!

you can have anywhere from 4 boxes to 1 box.

idealy would be two monos with each their power inside but I don't think you have the iron for that. (2 boxes)

Your left with
two monos plus single external PS. 3 boxes! My vote.

1 stereo amp plus external PS.. 2 boxes.

That Piltron chassis is huge and means the rack/shelf required will have to fight against WAF. And besides, seeing your previous chassis creations I'm sure you can do better.

cheers
Paul

[Peter Daniel]

Originally posted by paba
Your left with
two monos plus single external PS. 3 boxes! My vote.

That would be also my preferrence. Now, the filament transformer should be in a PS or each separate in amplifier boxes?

[paba]

I would put the filament supply with the mono blocks that way you will have the ability to have power switches on the monos to turn on the filaments before turning on the B+ on your external power supply.

More steps for power on and power off but better tube life.

/Paba

[planet10]

Originally posted by Peter Daniel
This thing is quite big.

And boring... you could stand it on an end and use it for the power supply (with iron inside)....but i know you are bound to have something much more interesting percolating in your brain

dave

[ThorstenL]

Konnichiwa,

Originally posted by Peter Daniel
I will be definitely interested in any suggestions.

Okay.

1) Dump the feedback, I calculated it noe, it amounts to 0.375db. In other words, practically NON AT ALL. Much ado in the accompanying article for no effective operation.

Saves a resistor and capacitor in the layout.

2) Get the 300B Heater Supply fixed up. There are many approaches, but you might as well get the Modules from Guido Tent and be done.

3) The 5687 SRPP may benefit from some tuning of both heater voltage (lower it to around 5.7V) and from some scaling of the resistor. The 5687 SRPP may very well be owed to the Audio Note Kit One where it is in my view and eperience the greatest sonic liability (apart from 300B Heater Supply), but I remember DO being quite fond of the Kit One.... I'd invest some time there.

Now for things that go past the basics and really get chomping.

WE Output Connection:

Connect a 8.2uF Capacitor of the higest possible quality between 300B Cathode connection and the +B connection of the output transformer. Add around 50 Ohm DCR in series with +B line. This gives around 26db better PSRR at 100Hz, a little more at higher frequencies. It actually works not only for PSU noise but also for any signal induced modulation.

In fact I found this so effective that it seems to give most of the benefits of a complex regulated supply for the cost of a single capacitor with NON of the (sonic) drawbacks regulation causes.

Easy thing to do on any self bias series feed SE Amp and always a surefire winner. The capacitor beween cathode and OPT +B connection is Ck/Mu where Ck is cathode bypass capacitor and Mu is output valves voltage gain. Use NPV (Nearest Preferred Value)

Series bootstrap biasing of driver stage:

Connect the cathode resistor of the 300B to the cathode of lower 5687 (or otherwise the driver cathode) instead to ground. Reduce the cathode resistor of the driver until the operating point folds back to what was there before (for T-Rex that would be around 180V on the anode of the lower 5687 looks like 120R would then be ideal for the lower 5687 SRPP Valve) and dump the cathode bypass cap.

This steals a little gain (around 1db) and eliminates the cathode bypass cap while actually adding VLF positive feedback compensating for the LF losses caused by undersized 300B Cathode bypass capacitor.

Originally posted by Peter Daniel
I was planning to build PS and the amp in separate enclosures. Should I go with monos or one chassis for the amp section?

I would do an open plan signal chassis for the actual Amp (including the actual regulator should you keep it) and a big black box thing for the PSU, making sure to allow for long umbical cables. Make sure to place final PSU Bypasses on the Signal Chassis.

Originally posted by Peter Daniel
If the boards will be offered by Audio Oasis, it will be based on my work here and I will be designing the layout. So again, any input is appreciated.

Thinks as usual, local current/bypass loops, short noise loops etc. You will be mostly there.

Originally posted by Peter Daniel
"The T-Rex is now in the public domain for the benefit of the DIY community", and since commercial rights are reserved, I will be not profiting on those boards.

Peter, I may be not DO, but I have something a lot freakier (but simpler) on the drawing board for a good while now.

Care to co-operate on a kit that uses that particular angle, on a for profit basis? No Plitron Transformers except perhaps mains, 2-Chassis design with a PSU aimed at more or less "universal" operation, active load outputstage with parallel feed output from MQ (TFA2004) or S&B (if I can ever get JB to start design on them something - my bad for getting him to make interesting products - now he has no more time for interesting developments) and some rather interesting ideas on the heater supply side and drivers....

The deal (stated here openly) is one set of (prototypes?) for my use and small donation per unit sold to a charity helping children (my usual "licencing deal").

Sayonara

[jleaman]

Look dave... he's useing them transformers i was going to use and you said sucked... humm... care to change your opinion now ?

[rdf]

Originally posted by paba
Your left with
two monos plus single external PS. 3 boxes! My vote.

I always wondered if this was really the best approach. Your source will always have a single ground reference. Splitting the amp chassis creates the potential for a voltage difference between the the two amplifiers chassis, their common power supply and the pre. Isn't it better, assuming optimum internal layout to minimize crosstalk, to use a single chassis and tight ground reference between the left and right channels and externalize a common power supply.

[ThorstenL]

Konnichiwa

Originally posted by Peter Daniel

That would be also my preferrence. Now, the filament transformer should be in a PS or each separate in amplifier boxes?

Originally posted by rdf

I always wondered if this was really the best approach. Your source will always have a single ground reference. Splitting the amp chassis creates the potential for a voltage difference between the the two amplifiers chassis, their common power supply and the pre. Isn't it better, assuming optimum internal layout to minimize crosstalk, to use a single chassis and tight ground reference between the left and right channels and externalize a common power supply.

Exactly. The "common supply, mono signal chassis" is really the worst possible solution, unless the common supply shares only the Box, but the supplies are electrically seperate and NOT ground referenced.

I would vote the opposite as better, 2 Dual Mono Powersupplies with one Stereo Signal Chassis wired dual Mono....

Sayonara

[Peter Daniel]

Originally posted by Kuei Yang Wang
I would vote the opposite as better, 2 Dual Mono Powersupplies with one Stereo Signal Chassis wired dual Mono....

Because of cost restrictions and preferable smaller size I have to share the main transformer and chokes between channels. I can go with separate regulators though.

I can also make dual channel boxes, but lets say with a possibility of tight ground reference between the left and right channels.

[planet10]

Originally posted by jleaman
Look dave... he's useing them transformers i was going to use and you said sucked... humm... care to change your opinion now ?

I never said they sucked -- that is just your black & white interpretation. What i said was that toroids are very intolerant of DC and you have to be careful. I'd prefer C core anyway (after spending an hour or so listening to 2 very good transformer guys -- Menno van der Veen (designer of the Plitrons) & Bill Perkins -- discuss transformers, that was what i got out of the discussion.

If anyone is going to make toroids work as OPTs, it is Menno, his transformers, generally, but not universally, get very good reviews.

dave