• 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.

Need advice regarding SET 45 Amplifier

Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
Ex-Moderator
Joined 2011
It appears to be built from hard-to-get vintage parts, if you already have the 45 tubes, the price seems reasonable, but it is likely to sound quite different than Tubelab's SE offering. Also note that the OPT matches to 16Ω, if you deviating too far from it, that will increases the distortion and/or lowers the output power.
 
Old WE transformers often have very limited bandwidth, being rolled off both at the top and bottom ends. The ones with any real or imagined claim to audiophile performance go for silly money, and are usually sold separately to cashed up asian audiophiles. In this case I suspect they are being used for their name, which obviously helps channel amp zen master YB Lee's chi to any circuit, automatically catapulting its performance into the major league.

The famous WE91A theatre amp used nfb to compensate for its output transformer's rather limited intrinsic bandwidth. And the theatre speakers it drove also had no real top end extension to worry about.

The maker doesn't specify if this design uses negative feedback, doesn't make any bandwidth claims, and even worse the WE transformer model number doesn't appear in any google search results except ads for his own products.

Maybe it does have a very nice midrange, however I would be very wary of buying something like this without being able to make a prior audition.
 
Last edited:
The claimed 2 Watts per channel per 45 tells me something about the designer. Just my opinion, but trying to push more than 1.5 watts out of a 45 is not such a nice thing...

I don't see 2-watts as a problem. The basic datasheet shows a 2-watt bias point as well. Also note that a later datasheet increased the ratings on the 45 (ST style) with a plate dissipation of 10 watts and voltage at 300.

My 45 SET amps use a 5K load, cathode current of 34ma with a plate supply of 360 volts (self-biased). Plate dissipation is 10-watts. The amps are rated at 2-watts output with clipping starting at 2.25-watts.

They've been in constant service for over 8-years without issue and have several thousand hours of use on them. I'm also getting 2-watts from other 45 designs using lower plate voltage as well but in all cases with a 5K load and a driver circuit capable of delivering the high voltage swing required.

Regards, KM
 
....

And how it would compare to a tubelab SE up to 1000$.
......

Please comment today :D !

Tubelab SE has low noise input tubes, DC filament supply, fixed bias mosfet powerdrive, capability for 2A3 and 300B tubes.

They're really not very comparable.

I would expect the Tubelab SE to beat that amp like a full house beats two of a kind - even with very modest output transformers.

Win W5JAG
 
I don't see 2-watts as a problem. The basic datasheet shows a 2-watt bias point as well. Also note that a later datasheet increased the ratings on the 45 (ST style) with a plate dissipation of 10 watts and voltage at 300.

My 45 SET amps use a 5K load, cathode current of 34ma with a plate supply of 360 volts (self-biased). Plate dissipation is 10-watts. The amps are rated at 2-watts output with clipping starting at 2.25-watts.

They've been in constant service for over 8-years without issue and have several thousand hours of use on them. I'm also getting 2-watts from other 45 designs using lower plate voltage as well but in all cases with a 5K load and a driver circuit capable of delivering the high voltage swing required.

Regards, KM

Hi KM. you have probably built more than me, but plate dissipation is not really the question.

When I run the 45 at a bit under 300V the sound is really best to my ears. My 45 amp doesn't push much beyond 1.5 watts and I carefully designed it not to swing fully into clipping. I measured my power using an oscilloscope. Maybe my output transformer has greater primary inductance than yours? ;)

At least we agree that the load has to be at least 5k.

I'm doubtful of this amplifier on the 'bay. With no schematic to go by, and a 6SN7 on each channel, I can only assume the circuit uses two differential blocks, resulting in too much swing and too little drive. Couple that with two moderate OPT's and you get something that might only do mid-range reproduction relatively well. Then to top if off, when you turn all the way up, it will not clip cleanly....

You know how that sounds too I think. Maybe its not bad though and the designer is a genius. I can only speculate.

My suggestion to the original poster is that a nicely built tubelab SE is worth the moderate effort to build instead.
 
Last edited:
Hi Soulmerchant,

No idea on who's built more or less but it probably doesn't matter. My only problem was your statement that designing to get more than 1.5 watts from the 45 relates to poor design, which simply isn't true.

Also, perhaps my post of specifications was a bit misleading, as I stated plate supply voltage, which is not the same as plate voltage in a self-biased stage. The actual cathode to plate voltage is just under 300 volts, with ~59 volts of grid bias via the cathode resistor. Plate dissipation can be a problem if you exceed the ratings too far and can damage the tube.

A larger concern is the cathode current. I prefer to keep the current at 34ma and push the plate voltage up a bit higher. With a good driver design, you can achieve good linearity up to 2 watts. Limiting the driver also limits the overall performance of the amplifier.

Also, I can't comment or speculate on transformer differences, as neither of us has stated any specifications other than 5K. I'm using the Hashimoto H-507S OPTs which are excellent but I wouldn't push them for more than 4 watts or so.

As for the ebay listed amplifier, I wouldn't speculate on the topology either as there's no telling what was done. The OPTs and other components could be limiting factors but again requires speculation. I certainly can't understand your comment on too much swing.

I spent a great deal of time designing and testing the input/driver circuit for the 45 SET amp. It is a direct-coupled 2-stage using a 5814a. It has ~43.5dB of voltage gain and can cleanly swing 150 volts pk-pk. The driver stage is biased at 2.8ma and results in very low measured THD and flat frequency response. Measured performance of the amplifier is 25Hz - 50KHz within 1dB at 1-watt output. THD is around 0.35% at 1-watt. Having an input/driver stage that can provide a clean and linear (output following input) swing that is greater than the output tube requires simply means the input/driver stage is always in a clean and linear operating range.

I guess we have difference design parameters here. I'm looking for wide frequency response, low distortion, low noise and dynamic linearity throughout the operating range. Here's a link you may find useful as it has some good design parameters and shows the calculations necessary to realize the performance, albeit I'm not that fond of his driver topology. All the best.

www.wavelengthaudio.com/bugle.pdf

Regards, KM
 
THD is around 0.35% at 1-watt.

Ok, so you measure similar things as I do at 1 watt. Push it up to 2 watts and look at your THD. My normal listening doesn't often go past 1 Watt, and I like to keep harmonic distortion low. I kept that in mind when I designed.

If you use 5814a/12au7/ECC82 then how do you swing 150v p-p other than to use two differential amplifiers in a row? I suppose it can work, but the output impedance will be higher than if you use a follower to drive the grid. I suppose it is still low enough looking at the specs sheet. I would design the 2nd stage for a bit less gain and more current.

I prefer to Triode strap a pentode like a D3A or C3G or E180F and directly couple it to a 45. You can also use 6SL7 in bootstrap follower circuit also directly coupled to the grid of the 45. These circuits works well for 2a3 as well.

ian.
 
Last edited:
Hi Ian,

We're likely more in agreement than not. At 2-watts the THD is still well below 1% (I don't recall the actual number off hand, likely around 0.75%). The input/driver is configured as two common cathode stages cascaded together via direct-coupling. Coupling to the 45 is via a blocking capacitor. I'm not sure how you interpret differential stages.

I'm old school, I used the plate curves for the 5814a and picked two operating points which can work well via direct-coupling. First (input) stage is 1.2ma and the second (driver) stage is 2.8ma. I did make a couple minor adjustments during testing as I actually cycled over 50 different 5814a, 12AU7, ECC82 tube types in the circuit. The final operating points yielded the lowest overall THD and also has excellent square-wave response out to 10KHz (at the output of the amp).

The plate supply for each stage is separate with the second stage being quite a bit higher (from the same supply point as the 45 output stage). So without any issue the output from the driver easily swings 150 volts pk-pk, which is more than required, which is ~120 volts pk-pk. This input/driver also drives a 2A3 with similar specifications. The slew rate is high enough (by design) which allows flat frequency response to 50KHz.

I did a different design using a 12AT7 where the second stage is configured as a cathode follower. I've also built a few amps with this topology, less gain of course but still easily swings the voltage required for both the 45 and 2A3.

Average listening levels are also below 1-watt, so for practical purposes, I have a 1-watt amplifier with 3dB of headroom ;-)

Regards, KM
 
Va max 275V..no way to squeeze more than 2W; with A2 +10V suddenly doable
Will grid survive..?:eek:
 

Attachments

  • Screen Shot 04-03-16 at 10.35 PM.PNG
    Screen Shot 04-03-16 at 10.35 PM.PNG
    86 KB · Views: 131
Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.