HT Transformers for Vacuum Tube Preamps

GroupDIY Audio Forum

Help Support GroupDIY Audio Forum:

This site may earn a commission from merchant affiliate links, including eBay, Amazon, and others.
Thanks Thor ,
I had noticed some smaller switcher powered tube amps doing the rounds ,

Many of the Vox amps are class A ,so HT current doesnt vary very much ,
In Class A/B amps the HT current varies appreciably from zero to full signal , I cant imagine an SMPS doing anything good for the tone during heavy clipping , Sure some mains noise moduated into the sound with traditional style PSU's , but thats a devil we know well .

Ive noticed in the age where modern stomp box type effects pedals come with small SMPS wall warts many problems with multiple SMPS interacting and causing horrendous noise/ mush all over the signal .
Fancy DSP stomp boxes which use amps of current are the usual suspects in this regard , garbage is mixing and combing via multiple paths and what comes out sounds little better than an Edison wax cylinder recording as far as signal/noise and distortion levels are concerned .

I know one very tallented guitarist whos put up this bullshit for years , I did take a look at his set up at one point and offer my advice to him ,
But hes a hard head and in the end my time was wasted , he stuck with his 35db S/N ratio .
I think he plays more acoustic these days so crowd noise defines the S/N :)

One time on a recording session with a few friends the Bass player had got the latest and greatest SMPS/Class D/ amp head , weighed only a few kg .
It had a DI out which I decided to try into the line in on the Yamaha HDR/mixer so we could use the amp eq compression etc to tape ..
The Genz-Benz threw a rod when I hooked up, all the virtual knobs bells and whistles were out to lunch , sound gone.....
There was a moment when the owner thought I'd killed the F'cking thing :D
He was having a major malfunction and was about go into meltdown until I flipped the power off and on again on the Genz Binz.
Lucky really because he's a big lad .
A standard active DI box and we were back in the groove again and all the fuss was forgotten .
A really memorable session , I must dig it out and cut the whinging whining and argueing between the takes .
 
The tube rectifier and alnico speaker gave the 60's its sound,smooth mellow . SS state rectifiers /ceramic magnet speakers started to get much more popular towards the late 60s and the agressive hard rock sound was pop music for an angry generation .

Theres no doubt a tube rectified amp puts out a bit less on the peaks than SS rectified for a given HT ,
The limiting effect of the rectifier comes into play only at higher volume ,
A Boogie Dual Rectifier has selectable SS or tube rectification ,
you wont hear any difference at low volume .

The question of LC filter of the screen supply of the output tubes , it makes a difference , its a commonly held belief that you just sub in resistor equivalent to the resistance of the coil and thats fine , The choke gives very good filtering to the screen , its moments of peak power this becomes important .
 
One other interesting thing about the Boogie I got to try out ,
Switching from SS to Tube rectification the nature of the noise floor of the amp changed, measured RMS wasnt was much different as LF noise dominates anyway, but tube mode had noticibly less harmonics and backround mush/hiss .
 
There are actually two output xfmrs in them, much smaller I agree.
And with 8 power MOSFETS, is it still a tube amp?
Well, the output transfer function is precisely that of the Tube Amplifier section, once the switching frequencies are filtered out.

The transformers are not in a classic sense output transformers. If we accept an SMPS and are looking for low weight, this may be a suitable approach

When it comes to this level of complexity, why not not go full solid-state?

I would go hybrid. With a Class D output stage switching >> 1MHz and a proper output filter that produces > 100dB carrier (switching frequency) suppression. The output stage can be quite linear if picking the right platform.

A tube frontend with small signal tubes and solid state parts can be used to simulate tube harmonics, overdrive behaviour and output transformer behaviour minus hysteresis. I made a commercial product with this principle (1.4141/1.536MHz switching, giving in effect a simulated 300B SE Amp (scaled to expected lower speaker sensitivity) with 160W/Ch. Single 2C51/5670/396A tube for stereo. It also uses a SMPS.

Thor
 
I cant imagine an SMPS doing anything good for the tone during heavy clipping ,

That's just making sure you design correctly.

Making a SMPS for a strongly varying load (audio in a power amp) is quite different to lighting up LED's.

Sure some mains noise moduated into the sound with traditional style PSU's , but thats a devil we know well .

As said, IT COULD be simulated in a SMPS.

Ive noticed in the age where modern stomp box type effects pedals come with small SMPS wall warts many problems with multiple SMPS interacting and causing horrendous noise/ mush all over the signal .

Most often this is because the SMPS are badly designed with a 2.2nF EMC "Y" Capacitor and there is no earth brought to the SMPS.

I find adding an external EMC filter with mains earth (these come as canned modules) on the mains side and a "soft earth" to the outputs (needs a special adapter due to negative centre for pedals) can dramatically lower these problems, even with badly designed SMPS.

The key is to understand the problem and then to apply a fix.

Thor
 
Well, the output transfer function is precisely that of the Tube Amplifier section, once the switching frequencies are filtered out.
Except the xfmr's magnetic effects (LF and HF roll-off, eddy currents, core losses...) that are arguably a significant part of the "tube sound".
I would go hybrid. With a Class D output stage switching >> 1MHz and a proper output filter that produces > 100dB carrier (switching frequency) suppression. The output stage can be quite linear if picking the right platform.
Do we want a linear output stage for a guitar amp?
There is clearly a divider here.
Guitarists want a tube amp for its numerous imperfections.
Audiophiles are supposed be in search of the most transparent reproduction. A tube amp may not be the best candidate for this role, in particular a SE 2A3, which many audiophiles consider the pinnacle of sonic nirvana..
A tube frontend with small signal tubes and solid state parts can be used to simulate tube harmonics, overdrive behaviour and output transformer behaviour minus hysteresis. I made a commercial product with this principle (1.4141/1.536MHz switching, giving in effect a simulated 300B SE Amp (scaled to expected lower speaker sensitivity) with 160W/Ch.
It is clearly an FX box. General principles of audio engineering (low THD, wide BW, low noise, high linearity...) do not pertain to FX boxes.
 
Except the xfmr's magnetic effects (LF and HF roll-off, eddy currents, core losses...) that are arguably a significant part of the "tube sound".

LF & HF roll-off and the relative increases in harmonic distortion towards the two ends of the passband can be modeled using strictly RC circuits and solid state add on's.

Using real inductors as "amplified inductors" in the tube stage loading can model core effects.

Do we want a linear output stage for a guitar amp?

If you model the rest ahead of the power stage - yes.

Audiophiles are supposed be in search of the most transparent reproduction. A tube amp may not be the best candidate for this role, in particular a SE 2A3, which many audiophiles consider the pinnacle of sonic nirvana.

I'm not sure what audiophiles are supposed to do, neither are they, I suspect.

Many like how "FX boxes" as you call them make their music sound, same for many non audiophiles.

General principles of audio engineering (low THD, wide BW, low noise, high linearity...) do not pertain to FX boxes.

Low THD? Is that a valid design goal, knowing the THD of microphones and speakers? Or just someone looking for his lost keys under a streetlight because the light is better there, compared to where he lost them.

Thor
 
Here's a quick idea:

Battery operated? Pick a battery voltage that will run the tube filaments directly, say 12volts and get a low power 12volt to 120 Vac inverter for a car, like this 60 watter from Amazon:

https://www.amazon.com/110V-240V-NA...s=12v+to+120v+converter&qid=1672677227&sr=8-5

Use a voltage doubler directly off the inverter for about 330 volts DC which should power those 6BQ5s nicely and use traditional analogue R-C filters for the pre and driver stages. If you want supply droop use a resistor, which is close to a tube rectifier = 6X4, 5U4, etc.

Lithium or LiPo batts are light and 4 would get you 14+ volts. Weight would come from the speaker and the output xfmer.

Put 6.3 volt filaments in series and use 12AX7s and 12AU7s and you're good to go. The cab - Maybe use one of those small fender amps as a donor with a speaker. Could be a nice setup esp for SoapFoot
 
LF & HF roll-off and the relative increases in harmonic distortion towards the two ends of the passband can be modeled using strictly RC circuits and solid state add on's.

Using real inductors as "amplified inductors" in the tube stage loading can model core effects.
All this adds complexity. Is it worth it? IDK. As I mentioned earlier, I'm not adverse to a full- SS solution. But I like simple designs. The complexity of replacing the OT with 2 smps seems unjustified to me in both cases.
A guitar amp does not need to have a tube amp without the xfmr - I think it's better repaced with a SS amp (perhaps class D) and some signal processing.
A HiFi amp needs to start with the most possible linear amp, which a tube amp is not.
I'm not sure what audiophiles are supposed to do, neither are they, I suspect.
The original concept of HiFi is pretty clear, and relies on wide BW, low noise, low THD and a few other things that can be measured and quantified.
Audiophoolery is another concept, that tries to find technical justifications in subjective aberrations.
Many like how "FX boxes" as you call them make their music sound, same for many non audiophiles.
Sure; I don't deny it, but a designer must have a set of design and performance targets. Just saying "I want it to sound good" is not enough.
Low THD? Is that a valid design goal, knowing the THD of microphones and speakers?
Is it a good reason to ignore THD in the electronics? I believe you know the characteristics of transducer distortion is different enough than that of electronics for a mildly trained listener to identify and differentiate them in many cases.
Or just someone looking for his lost keys under a streetlight because the light is better there, compared to where he lost them.
Based on this, any kind of measurement is wasted time.
 
Thx Ian
This is very helpful info...it seems my dwindling collection of old Stancor or Motorola are too hot and need to be saved for future larger projects. I have never built with a torroidal...is this really of great value in a B+ supply ? You are such a helpful cat...thx for sharing
 
Thx Ian
This is very helpful info...it seems my dwindling collection of old Stancor or Motorola are too hot and need to be saved for future larger projects. I have never built with a torroidal...is this really of great value in a B+ supply ? You are such a helpful cat...thx for sharing
There is nothing special about a toroid that lends it to an HT supply. it is just that in general they have a lower external magnetic field which is important if you are using audio transformers in the circuit. They do also tend to have better regulation than the EI type transformer but that is not a crucial factor in an HT supply,

Cheers

Ian
 
Last edited:
Marketing blurb seems to have pulled up the popularity of toroids for tube amps .
Theres no doubt if you have space constraints the lower profile is handy .
Bill did express some reservations about them recently , saying the point where the leads exit can leak magnetically . I wonder about capacitive coupling between primary/secondary and the resiliance of the insulating tape material in the longer term , as well as wires with very different potentials laid on top of each other with only the insulated coating to prevent break down .
In an old style EI transformer the core istelf makes contact with chassis , where a toroid core is insulated ,maybe thats a factor . Im definately interested in hearing more opinions on the subject .
 
Ian,
If I may, what are your thoughts on using an SMPS as a filament supply for a directly heated single ended triode (2A3, 45, 300B, 845, etc)?
These days, having lived to an age I thought to be impossible (68) pretty much all I build are single ended triode hifi amps and tube mic pres, both of which I have a waiting list for.
Being a bit of a left over hippie type there are only two sources of misery in my life: my Xwife and hum. While I’ve become rather skilled at ignoring the X, any sort of hum or noise in my valve amp builds is a source of misery for me. The high sensitivity speakers my clients and I tend to use (100 Db+ doesn’t help).
An LC filter design works very well but results in gear I can’t comfortably lift these days. A SMPS filament supply is a very attractive alternative, particularly with a high current triodes such as an 845.
Should you have the time to respond I’d appreciate and look forward to it.

I think I’ve mentioned before I’m a big follower of your postings.

Peace,
DrRick
 
Bill did express some reservations about them recently , saying the point where the leads exit can leak magnetically .
That is true, but the leakage there is way smaller than the leakage from a EI core, for the same nominal flux.
A significant issue is that most often, toroidals are operating at higher flux than EI or R cores, since it allows publishing a higehr power-to-bulk ratio.
When I had the luxury of having custom wound xfmrs, I always asked operating at lower flux, which generally implied using a bigger core than normallly accepted.
I wonder about capacitive coupling between primary/secondary and the resiliance of the insulating tape material in the longer term , as well as wires with very different potentials laid on top of each other with only the insulated coating to prevent break down .
It's true that the construction of toroidals increases capacitive coupling and that's why adequate filtering and wiring must be applied, nominally the mandatory protective earth and the three capacitors two to earth, one across) in the mains connection.
In terms of insulation, the outer layer of primary and the inner layer of secondary is the critical area. A dedicated insulation layer is generally applied.
 
The chances of a professionally wound xfmer made with today's insulation shorting to other windings or ground are practically nil for either an EI, C core or toroid. UNLESS you overheat it. If memory serves they are wound for 100deg C rise but ran at about 40.

If you want high isolation between pri and sec use one with independent bobbins. If you want low cap pri to sec put a e-static shield between the windings on EI or toroids.

If you are putting a high current filament winding on for DC the charging current will make it more difficult to contain the magnetic field. I know from experience. And a magnetic field can be induced in circuit board traces.

In a full function tube preamp including a hi gain (80db @ 1khz) phono stage with DC filaments I designed in the 80s I used an EI core xfmer. By experiment I found an exact angle whereby no hum was induced in the preamp. Pretty good intuition and luck. The phono pre has 100db of gain at 20 hz due to RIAA net.

This is from MY experience which may help someone.

While I'm not a fan of switchers I guess today's PSUs are run at a lot higher f than in the old days. BUT STILL, it's not supposed to be in the music.

I don't like distracting the worker electrons with partying electrons doing a fast tango in the filament next door to the cathode when they're just supposed to jump off the cathode, say hello to the grid and land quietly on the plate then return home. They get jealous. I can hear them carrying on.

Good luck.
 
Back
Top