• 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.

My first design. PSE 211 to review.

Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
Hi all,

It is my first time here. I recently got interested in tube amps and decided to study the subject. After a bit I designed my fisrt amplifier. I would like to submit it for scrutiny here to try to learn a bit more.

I will be happy to receive all comments regarding operating points, values of the elements, design concerns, etc.

My goals were to have a tube amp with only 2 stages, single ended, low distortion and more than 10 Watts. I settled for a parallel single ended 211 output stage driven by a 417A.

For those interested and willing to help develop this a bit further, I attach a PDF with the schematic and some details, an another PDF with the list of parts.

Thanks to all,

Radames

PS: Comments are wellcome here or by e-mail.

PSII: I don't really know how to calculate a value for R1.
 

Attachments

  • my_first_amp_pse_211.pdf
    34.7 KB · Views: 639

PRR

Member
Joined 2003
Paid Member
Why do you need R1 at all? If the step-attenuator NEVER loses contact, you don't. This is an imperfect world, and it is wise to have something to hold the grid down when the drive loses contact. The tube spec will usually give a maximum resistance, usually 1Meg for small tubes.

You don't "need" center-taps on heater windings. Use two 100 ohm resistors and ground the center, or a 100 ohm 2-Watt (wire-wound) pot and trim for lowest hum. (If you have to trim off-center, your heater wiring isn't as tight and neat as it could be.)

You are asking 50V peak swings. Old rule of thumb says for ~5% distortion, the peak swing is 20% of the supply voltage. Getting 2% THD for 50V swings on a 277V supply is darn good. At 27mA??? But still your driver distorts more than your monstrous output. Pour some fuel on the fire, raise the plate supply to the 417. I'd look at 400V-500V supply running 200V-250V on the plate, and set plate resistor so it runs just under the plate dissipation rating. I don't have 417 data written on my shirtcuff, there may be some reason that won't work.

It looks like a very elaborate way to get 10 watts. And eating 100+ watts to do it? Still all in all, a fine design.
 
417A operating point

Thanks a lot for your reply PRR.

About the 417A, I chose it due to the fact that it is a triode (I wanted triodes), it has a good mu to allow direct connection from the CD player or DAC to the power stage without requiring another amplification step.

The Max plate voltage is 200V and the power line 4.5W

I could get an improvement in the distortion figures by changing the loadline an increasing the voltage of the supply as you suggested. Best case is something like this:

303 VDC supply
Vp: still 154V
Ip: 27.62 mA
Loadline slope: 5330 Ohm
THD: 0.77% good decrease
2ndHD: 0.74% good decrease
3rdHD: 0.21% quite an increase compared to the 0.047% before
4thHD: 0.08% slight decrease

The increase in odd harmonic distortion worries me a bit, but at these levels, may be I am wrong.

The other thing I find puzzling is that 100 Ohm change in load will cause abrupt changes in the distortion figures. May be I am not calculating stuff wright...

I basically use the same method explained in:
http://members.aol.com/sbench102/po-dis.html

May be I should choose another tube. But I wasn't able to find something that would be able to drive the 211s with a 1 - 2 Vrms grid swing.

Cheers,

Rada
 
Ex-Moderator
Joined 2003
You don't want to use a 100k volume control with a 416C. The Miller capacitance is too high for the output resistance of the control. Try a 25k attenuator. You could use 220k for R1.

Oh, and don't fuss about distortion figures predicted by PSPICE. They're a guess. The real world is likely to be completely different.
 
You'll find that C8, C9, C11 and C12 are unnecessary, along with the inductors between them.

If there's some hum, elevating the 417 heater by 20V or so might help. Just connect a voltage divider from +277V to the heater transformer CT and add a capacitor to AC-ground it. Something like +277V > 100k > CT > 10k + 47uF (25-50V rated electrolytic) > GND will work good.

Most likely wiring will be your biggest cause of hum. Don't worry about the power supply, just make sure the input jack's ground is connected directly to the 417's cathode ground. If any hum voltage were to get in here, it would be carried through as ground loop. PSU ripple, or ground loop in the 211 or PSU is a secondary concern, although filtering should be well done given the triodes' bad PSRR (as mentioned above, a single stage of CLC will be quite alright).

Edit: looking at your parts list, C10/C11 are small. So you went for an LCLC design? Unfortunately that voltage is wishful thinking, you'll get maybe 650V output. You need cap input for that. In that case, pull C10, C11 and the input inductor, now it'll work good. R7 is also extraneous, unless you added it to tune the voltage exactly (not neccessary in an audio amplifier). Same for the lower voltage supply, although 277V is a lot lower than the 350VDC you'll get following the same instructions. Say, what's the rating on 417 anyway? It probably wouldn't hurt to use the extra voltage anyway, though you'll have to rebias it.

Tim
 
PSU hum

Hey Tim thanks for the reply.

I calculated that every LC stage of the PSUs will give me around 58dB of hum attenuation. So together would give more than 100, and that's what I wanted to get.

Using a single LC could let the hum be heard through the speakers if my calculations are correct. But may be that's all math rubble.

I used the parallel value of load (7KOhm) and the parallel value of anode resistance (1885.5Ohm) to calculate the required L*C value to get >100dB humm reduction in the case of the output stage. The load and ra of the input stage was used for the smaller PSU.

For a single LC filter I calculated an L*C product of 64383 (L in H, C in uF). It was imposible to find parts to satisfy that so I turned to 2 LCs.

C6, C7, C10, and C11 are to small to efficiently filter PSU ripple, they are just there as spike eliminators working as an RC filter with the R of the choke.

Again, this may all be just the math and I have no experience whatsoever. I will eliminate the second LC as a first approach to the circuit during breadbording and see what happens.

Thanks for your suggestion.

Cheers!

Rada
 
That's exactly what I want!

Re-designers is exactly what I am looking for! As I come here because my knowledge ran out. Hopefully, many of my mistakes (I'm sure there is many) can be easily spotted by the more seasoned builders and designers.

With respect to regulation, I contemplated the possibility but I stayed clear of it in the interest of simplicity. I also looked at many designs for SET amplis and many of them came out without regulation.

I do not discard the idea, but I will dive into it only if necessary.

Rada
 
Ex-Moderator
Joined 2003
Fair enough. When you don't have much experience, it's better to keep things simple. Regulators can have all sorts of nasty problems that require considerable investigation with a decent oscilloscope. Consider a dual chassis design with power supply on one chassis and amplifiers on the other. Perhaps leave room for a regulator and choose power transformers that could accomodate a regulator at a later stage.
 
Or a nice MOSFET. Or even a valve regulator. Perhaps even shunt.

Aaaaaggggh! Talk about chrome-plating a pig's ear. A nice, simple, stable bipolar Maida (LM317) reg is easy to implement and (even allowing for popping a transistor or two during initial checkout) much cheaper than proper LC sections. Morgan Jones has a nice version in his book; a simpler and (to my mind) as-good-or-better implementation is Joe Curcio's from the ST-70 mod article in Glass Audio.

Why anyone uses valve regulators in this day and age can only be attributed to sheer cussedness. Or (in my case) laziness.
 
Regulator

I downloaded the LM317 datasheet and I am going to study it.

As for Morgan Jones, I got most of the basics from his book and I found it superb. I remember, however, that in the case of his only single ended example (The Scrapbox Challenge), he went for a solid state regulator because he did not have at hand the adequate components for an efficient LC filter.

I will try to implement a solid state regulator based on his designs. Back to studying again!

Cheers,

Rada
 
Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.